Melting History and Magmatic Evolution of Basalts and Picrites from the Accreted Wrangellia Oceanic Plateau, Vancouver Island, Canada ANDREW R. GREENE 1 *, JAMES S. SCOATES 1 , DOMINIQUE WEIS 1 , GRAHAM T. NIXON 2 AND BRUNO KIEFFER 1 1 PACIFIC CENTRE FOR ISOTOPIC AND GEOCHEMICAL RESEARCH, DEPARTMENT OF EARTH AND OCEAN SCIENCES, UNIVERSITY OF BRITISH COLUMBIA, VANCOUVER, BC, V6T1Z4, CANADA 2 BRITISH COLUMBIA GEOLOGICAL SURVEY, P.O. BOX 9320, STN PROV GOVT, VICTORIA, BC, V8W9N3, CANADA RECEIVED MARCH 18, 2008; ACCEPTEDJANUARY 23, 2009 The accreted Wrangellia oceanic plateau in the Pacific Northwest of North America provides important insights into the volcanic archi- tecture of major oceanic plateaus, as well as the nature of their mantle source, conditions of melting and subsequent magmatic evolu- tion. The 20 000 km 2 Karmutsen Formation flood basalts (Vancouver Island) were emplaced at c. 225^230 Ma onto Middle Triassic marine sediments and Late Devonian to Early Permian island-arc volcanic and sedimentary sequences, and are overlain by LateTriassic platformal carbonates.The basalts form an emergent sequence consisting of a basal sediment^sill complex (600^900 m thick), pillowed and massive submarine flows ( 4 25 km), pillow breccia and hyaloclastite ( 5 15km), and massive subaerial flows ( 525km). Although the Karmutsen Formation is predominantly composed of tholeiitic basalt, the submarine part of the stratigraphy on northern Vancouver Island also contains picritic basalts.These high-MgO (9^20wt %) lavas are depleted in light rare earth ele- ments (LREE; La/Yb CN ¼ 05 02), whereas the tholeiitic lavas (6^8wt % MgO) are LREE-enriched (La/Yb CN ¼ 22 03). Both lava groups have overlapping initial e Hf (þ 103 21) and e Nd (þ 77 13), indicating a common, plume-type Pacific mantle source with geochemical characteristics similar to the source of basalts from the Ontong Java and Caribbean plateaus. Major-ele- ment modeling results indicate that the picrites formed by extensive melting (23^27%) of anomalously hot mantle (15008C), which is consistent with a mantle plume initiation model for formation of the Karmutsen flood basalts on Vancouver Island. Trace element constraints indicate that the picrites require melting of a depleted spinel lherzolite source, whereas the tholeiitic basalts involved melt- ing of garnet and spinel lherzolite.The tholeiitic basalts underwent significant fractional crystallization ( 450%) and the fractionated residues may be represented by high-velocity rocks beneathVancouver Island identified from seismic reflection studies. KEY WORDS: Karmutsen Formation; mantle plume; large igneous prov- ince; picrite; tholeiitic basalt INTRODUCTION Oceanic plateaus are transient large igneous provinces that together with continental flood basalts represent the pro- ducts of the largest magmatic events preserved on Earth from the Phanerozoic era. There are some 10 oceanic pla- teaus that have formed in the last 145 Myr, ranging in size from 100 000 to 2 10 6 km 2 , and collectively they cover an area of about 10 10 6 km 2 (3%) of the ocean basins, mostly in the Pacific and Indian oceans (e.g. Eldholm & Coffin, 2000). Oceanic plateaus form submarine or subaer- ial flood basalt sequences that can be 6 km thick, with total crustal thicknesses (extrusive plus intrusive rocks) of 8^33km (e.g. Coffin & Eldholm,1994). Several key aspects of most large igneous provinces are that they form as a *Corresponding author. E-mail: [email protected]ß The Author 2009. Published by Oxford University Press. All rights reserved. For Permissions, please e-mail: journals.permissions@ oxfordjournals.org JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 PAGES 1^39 2009 doi:10.1093/petrology/egp008 Journal of Petrology Advance Access published March 9, 2009
39
Embed
Melting History and Magmatic Evolution of Basalts and ... · Basalts and Picrites from the AccretedWrangellia Oceanic Plateau,Vancouver Island, Canada ANDREW R. GREENE1*, ... [email protected]
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Melting History and Magmatic Evolution ofBasalts and Picrites from the AccretedWrangelliaOceanic PlateauVancouver Island Canada
ANDREW R GREENE1 JAMES S SCOATES1 DOMINIQUE WEIS1GRAHAM T NIXON2 AND BRUNO KIEFFER1
1PACIFIC CENTRE FOR ISOTOPIC AND GEOCHEMICAL RESEARCH DEPARTMENT OF EARTH AND OCEAN SCIENCES
UNIVERSITY OF BRITISH COLUMBIA VANCOUVER BC V6T1Z4 CANADA2BRITISH COLUMBIA GEOLOGICAL SURVEY PO BOX 9320 STN PROV GOVT VICTORIA BC V8W9N3 CANADA
RECEIVED MARCH 18 2008 ACCEPTEDJANUARY 23 2009
The accretedWrangellia oceanic plateau in the Pacific Northwest of
North America provides important insights into the volcanic archi-
tecture of major oceanic plateaus as well as the nature of their
mantle source conditions of melting and subsequent magmatic evolu-
tion The 20 000 km2 Karmutsen Formation flood basalts
(Vancouver Island) were emplaced at c 225^230 Ma onto Middle
Triassic marine sediments and Late Devonian to Early Permian
island-arc volcanic and sedimentary sequences and are overlain by
LateTriassic platformal carbonatesThe basalts form an emergent
sequence consisting of a basal sediment^sill complex (600^900 m
thick) pillowed and massive submarine flows (425 km) pillow
breccia and hyaloclastite (515 km) and massive subaerial flows
(525 km) Although the Karmutsen Formation is predominantly
composed of tholeiitic basalt the submarine part of the stratigraphy
on northern Vancouver Island also contains picritic basalts These
high-MgO (9^20 wt ) lavas are depleted in light rare earth ele-
ments (LREE LaYbCNfrac14 05 02) whereas the tholeiitic lavas
(6^8 wt MgO) are LREE-enriched (LaYbCNfrac14 22 03)
Both lava groups have overlapping initial eHf(thorn103 21) and
eNd(thorn7713) indicating a common plume-type Pacific mantle
source with geochemical characteristics similar to the source of
basalts from the Ontong Java and Caribbean plateaus Major-ele-
ment modeling results indicate that the picrites formed by extensive
melting (23^27) of anomalously hot mantle (15008C) which isconsistent with a mantle plume initiation model for formation of the
Karmutsen flood basalts on Vancouver Island Trace element
constraints indicate that the picrites require melting of a depleted
spinel lherzolite source whereas the tholeiitic basalts involved melt-
ing of garnet and spinel lherzoliteThe tholeiitic basalts underwent
significant fractional crystallization (450) and the fractionated
residues may be represented by high-velocity rocks beneathVancouver
Island identified from seismic reflection studies
KEYWORDS Karmutsen Formation mantle plume large igneous prov-
ince picrite tholeiitic basalt
I NTRODUCTIONOceanic plateaus are transient large igneous provinces thattogether with continental flood basalts represent the pro-ducts of the largest magmatic events preserved on Earthfrom the Phanerozoic era There are some 10 oceanic pla-teaus that have formed in the last 145 Myr ranging in sizefrom 100 000 to 2106 km2 and collectively they cover anarea of about 10106 km2 (3) of the ocean basinsmostly in the Pacific and Indian oceans (eg Eldholm ampCoffin 2000) Oceanic plateaus form submarine or subaer-ial flood basalt sequences that can be 6 km thick withtotal crustal thicknesses (extrusive plus intrusive rocks) of8^33 km (eg Coffin amp Eldholm1994) Several key aspectsof most large igneous provinces are that they form as a
Corresponding author E-mail agreeneeosubcca
The Author 2009 Published by Oxford University Press Allrights reserved For Permissions please e-mail journalspermissionsoxfordjournalsorg
JOURNALOFPETROLOGY VOLUME 00 NUMBER 0 PAGES1^39 2009 doi101093petrologyegp008
Journal of Petrology Advance Access published March 9 2009
result of unusually high melt production rates are predo-minantly basaltic in composition and their formation isnot directly attributable to seafloor spreading processes(eg Saunders 2005) The high melt production rates arebest explained by high mantle source temperatures in rap-idly upwelling mantle (ie mantle plumes) direct evidenceof high mantle temperatures is the eruption of primitivehigh-MgO lavas (Kerr amp Mahoney 2007) Studies of oce-anic plateaus can be used to constrain aspects of mantleplume-related magmatism including the composition ofthe mantle source involved the temperature and depth ofmelting the volume of magma and melt production ratesand the mechanisms of emplacement of voluminous lavasequences From a geochemical perspective the study ofoceanic plateaus is important because the chemistry of thelavas has generally been unaffected by continental crustalcontaminationAt present one of the great challenges in studying oce-
anic plateaus in the ocean basins is the difficulty in drillingto depths of more than a few hundred meters into thebasaltic basement rocks (eg Kerguelen Ontong Java)Accreted sections of oceanic plateaus (eg CaribbeanOntong Java) provide an opportunity to closely examinetheir volcanic stratigraphyWrangellia flood basalts in thePacific Northwest of North America are part of a majorlarge igneous province that erupted in a marine setting inthe Late Triassic (c 225^231 Ma) and accreted to westernNorth America in the Late Jurassic or Early Cretaceous(Richards et al 1991) Although their original areal distri-bution was probably considerably larger much of the orig-inal stratigraphic thickness (6 km on Vancouver Island35 km in Alaska) is intact Current exposures of theflood basalts are preserved in a narrow belt over 2300 kmin length resulting from transform fault motions afteraccretion which extends from southern British Columbiathrough the Yukon and into Alaska Richards et al (1991)proposed a plume initiation model for the Wrangelliaflood basalts based on the short duration and estimatedhigh eruption rate of the volcanism evidence of upliftprior to volcanism and lack of evidence of associated rift-ing (ie few dikes and abundant sills)TheWrangellia flood basalts are perhaps the most exten-
sive accreted remnants of an oceanic plateau in the worldwhere parts of the entire volcanic stratigraphy areexposed however they have been the focus of few inte-grated field geochemical and radiogenic isotopic studiesThe internal architecture of oceanic plateaus can provideconstraints on the sequence of melting events associatedwith mantle plumes that occur beneath the oceans in ananalogous way to the use of continental flood basalts toelucidate the sequential development of mantle plumesand rifting within continental lithosphere (eg Peate 1997Storey et al 1997) The internal stratigraphy of continentalflood basalts is far more complex than previously realized
(eg Jerram amp Widdowson 2005) and because of theirrelative inaccessibility the structure of flood basaltsequences in oceanic plateaus is not well known In thisstudy we examine the field relationships petrographymajor and trace elements and Sr^Nd^Hf^Pb isotopiccompositions of Wrangellia flood basalts from variousareas of Vancouver Island to assess the nature of themantle source conditions of melting and magmatic his-tory of the basalts in the context of the construction ofthis major oceanic plateau This study is part of a compre-hensive research project on the nature origin and evolu-tion of the Triassic Wrangellia flood basalt province inBritish Columbia Yukon and Alaska (Greene et al 20082009)
GEOLOGICAL SETT INGWrangellia on Vancouver IslandWrangellia flood basalts form the core of the Wrangelliaterrane orWrangellia one of the largest outboard terranesaccreted to western North America (Jones et al 1977)Middle to LateTriassic flood basalts extend in a discontin-uous belt from Vancouver and Queen Charlotte Islands(Karmutsen Formation) through SE Alaska and SWYukon and into the Wrangell Mountains Alaska Rangeand Talkeetna Mountains in east and central Alaska(Nikolai Formation Fig 1)Wrangellia covers 80 of Vancouver Island which is
460 km long by 130 km wide (Fig 1) Wrangellia is theuppermost sheet of a stack of SW-vergent thrust sheetsthat form the crust of Vancouver Island and has a cumula-tive thickness of410 km (Monger amp Journeay 1994) Pre-Karmutsen units of Wrangellia on Vancouver Islandinclude Devonian arc sequences of the Sicker Group andMississippian to Early Permian siliciclastic and carbonaterocks of the Buttle Lake Group (Muller 1980 Sutherland-Brown et al1986)These Paleozoic sequences have variableestimated thicknesses (2^5 km Muller 1980 Massey1995a) and are exposed only on central and southernVancouver Island (Fig 1) The Karmutsen Formation isoverlain by the shallow-water Quatsino limestone (30^750m) and shallow- to deeper-water Parson BayFormation (600m) which is intercalated with and over-lain by Bonanza arc volcanics (169^202 Ma Nixon et al2006b) Bonanza arc plutonic rocks (167^197 Ma) alsointrude the Karmutsen Formation and are some of theyoungest units of Wrangellia that formed prior to accretionwith North America (Nixon et al 2006b) Following accre-tionWrangellia units were intruded by the predominantlyCretaceous Coast Plutonic Complex (Wheeler amp McFeely1991)The Karmutsen Formation (19142 km2 calculation
based on digital geological compilations) is composed ofbasal sediment^sill complexes a lower unit of pillowed
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
2
and massive submarine flows a middle unit of mostlypillow breccia and hyaloclastite and an upper unit of pre-dominantly massive subaerial flows (Carlisle amp Suzuki1974) The pillow basalts directly overlie Middle Triassicand late Paleozoic marine sedimentary sequences that areintruded by Karmutsen mafic sills The boundary betweenpillow breccia^hyaloclastite and massive lava flows repre-sents the transition from a submarine to a subaerial erup-tive environment The uppermost flows of the Karmutsenare intercalated with and overlain by shallow-water lime-stone and local occurrences of pillowed flows and hyalo-clastite deposits occur within the upper subaerial memberThe Karmutsen Formation and Wrangellia on Vancouver
Island have been the focus of detailed mapping and strati-graphic studies by Carlisle (Carlisle 1963 1972 Carlisle ampSuzuki 1974) and Muller (Muller et al 1974 Muller 1977)Recent descriptions of the Karmutsen Formation on north-ern Vancouver Island have been made during regionalmapping studies (150 000 scale) by Nixon and coworkers(Nixon et al 2006b 2008 Nixon amp Orr 2007)
Age of the Karmutsen FormationThe age and duration of Karmutsen volcanism are con-strained by fossils in the underlying and overlying sedi-mentary units and by three U^Pb isotopic age
Fig 1 Simplified map of Vancouver Island showing the distribution of the Karmutsen Formation and underlying Paleozoic formations (afterMassey et al 2005a 2005b) Areas in white are mostly younger units The main areas of field study are indicated with boxes or circles withcapital letters (see legend) Ocean is light blue The inset shows the extent of theWrangellia flood basalts (green) in British Columbia Yukonand Alaska
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
3
determinations on intrusive rocks that are considered to berelated to the Karmutsen volcanics Daonella-bearing shale100^200m below the base of the pillow basalts on SchoenMountain and Halobia-rich shale interlayered with flows inthe upper part of the Karmutsen indicate eruption of mostof the flood basalts between the Middle Ladinian (c 232^235 Ma Middle Triassic) and latest Carnian or possiblyearliest Norian (c 216^227 Ma Late Triassic Carlisle ampSuzuki 1974) The only published U^Pb age is based on asingle concordant analysis of a multi-grain baddeleyitefraction from a gabbro on southernVancouver Island thatyielded a 206Pb238U age of 227326 Ma (Parrish ampMcNicoll 1992) Two unpublished 206Pb238U baddeleyiteages also from a gabbro on southern Vancouver Islandare 226805 Ma (five multi-grain fractions) and228425 Ma (two multi-grain fractions Sluggett 2003)
VOLCANIC STRAT IGRAPHY ANDPETROGRAPHYFieldwork undertaken onVancouver Island as part of thisstudy in 2004^2006 explored the volcanic stratigraphy ofthe Karmutsen flood basalts in three main areas theKarmutsen Range (between Alice and Nimpkish lakes)the area around Schoen Lake Provincial Park and atMount Arrowsmith (Greene et al 2005 2006) areasaround Holberg Inlet in northernmost Vancouver Islandand Buttle Lake in central Vancouver Island were alsoinvestigated (Fig 1) The character and thickness of theflood basalt sequences vary locally although the tripartitesuccession of the Karmutsen Formation appears to be pres-ent throughout Vancouver Island The stratigraphic thick-nesses for the pillow volcaniclastic and subaerial flowunits in the type area around Buttle Lake are estimated tobe 2600m 1100m and 2900m respectively (Surdam1967) estimated thicknesses on northernVancouver Islandare 43000m 400^1500m and 41500m respectively(Nixon et al 2008) in the vicinity of Mount Arrowsmithand nearby areas on southern Vancouver Island thick-nesses are 1100m 950m and 1200m respectively (Yorathet al 1999 Fig 1)Picritic pillow lavas west of the Karmutsen Range on
northernVancouver Island occur in a roughly triangular-shaped area (30 km across) bounded by KeoghMaynard and Sara lakes (Figs 2 and 3 Greene et al2006 Nixon et al 2008) Excellent exposures of picriticpillow lavas occur in roadcuts along the north shore ofKeogh Lake the type locality (Greene et al 2006) TheKeogh Lake picrites form pillowed and massive flow units(515m thick Fig 3) with pillows and tubes of varieddimensions (typically51m wide) Numerous thermal con-traction features in the pillows are filled with quartz^carbonate such as drain-back ledges and tortoise-shelljointing (Greene et al 2006) The picritic pillow basalts
are not readily distinguishable in the field from basaltexcept by their density non-magnetic character and com-paratively minor amounts of interpillow quartz^carbonateFieldwork and mapping indicates that the picrites occurnear the transition between the uppermost pillow lavasand the lowermost part of the volcaniclastic unit (Nixonet al 2008)In and around Schoen Lake Provincial Park where
the base of the Karmutsen Formation is exposed is asediment^sill complex consisting of Middle Triassic andPennsylvanian to Permian limestones and fine-grained sili-ciclastic sedimentary rocks intruded by mafic sills andoverlain by pillow basalt (Figs 2 and 4 Carlisle 1972) Thesediment^sill complex is 600^900m thick whichincludes 150^200m of pre-intrusive sedimentary rocks(Carlisle 1972) TheTriassic sedimentary rocks range fromthinly bedded siliceous and calcareous shales to bandedcherts and finely laminated Daonella-bearing shales(Carlisle 1972) Basal sills and pre-Karmutsen sedimentsare also exposed around Buttle Lake (Fig 1)Exposures with considerable vertical relief around
Mount Arrowsmith (see Electronic Appendix 1 availablefor downloading at httppetrologyoxfordjournalsorg)and Buttle Lake preserve thick successions of submarinepillow basalts and pillow breccias with rare intercalationsof sediment and subaerial flows Massive submarine flowsin the pillow unit are locally recognizable by their irregu-lar hackly columnar jointing The massive subaerial flowsform monotonous sequences marked mainly by amygdaloi-dal horizons brecciated flow tops are rarely observedA locality with well-preserved pahoehoe flow features iswell exposed north of Holberg Inlet (Fig 3 Nixon amp Orr2007 Nixon et al 2008) There is no evidence of significantdetrital material from a continental source in the sedi-ments associated with the flood basalts anywhere onVancouver Island The uppermost Karmutsen flows areinterbedded with thin (commonly 0^4m thick) lenses oflimestone and rarely siliciclastic sedimentary rocks(Nixon et al 2006b) Plagioclase-phyric lavas characterizethe upper part of the subaerial sequence and locallyinclude distinctive plagioclase-megacrystic (1^2 cm crys-tals) trachytoid-textured flows (Nixon et al 2006b)A total of 129 samples were collected from the
Karmutsen Formation on Vancouver Island from which63 samples were selected for geochemical analysis basedon the relative degree of alteration and geographical distri-bution of the samples Fifty-six of these samples weredivided into four main groups based on petrography(Table 1) and geochemistry tholeiitic basalt picrite high-MgO basalt and coarse-grained mafic rocks In additiontwo outlying samples and a mineralized sill are distin-guished in Table 1 The tholeiitic basalts are dominantlyglomeroporphyritic and less commonly aphanitic withan intersertal to intergranular groundmass (Table 1)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
4
Fig 2 Geological map and stratigraphy of the Schoen Lake Provincial Park and Karmutsen Range areas (locations shown in Fig 1 northernVancouver Island) (a) Stratigraphic column depicts units exposed in the Schoen Lake area derived from Carlisle (1972) and fieldwork (b)Generalized geology for the Schoen Lake area with sample locations Map derived from Massey et al (2005a) The exposures in the SchoenLake area are the lower volcanic stratigraphy and base of the Karmutsen Formation (c) Stratigraphic column for geology in the Alice^Nimpkish Lake area derived from Nixon amp Orr (2007) (d) Geological map from mapping of Nixon et al (2006a 2008) Sample sites andlithologies are denoted in the legend The Keogh Lake picrites are exposed near Keogh Sara and Maynard lakes and areas to the east ofMaynard Lake (Nixon et al 2008)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
5
Fig 3 Photographs of picritic and tholeiitic pillowed and massive basalt flows from the Karmutsen Range (Alice and Nimpkish Lake area)and Holberg Inlet (HI) areas northernVancouver Island (locations shown in Fig 1) (a) Massive submarine flow (marked with arrows) drapedover high-MgO pillow basalt of similar composition (b) Picritic pillow basalts in cross-section near Maynard Lake (Note vesicular uppermargin to pillows) (c) Close-up photograph of undulating contact of massive submarine flow draped over high-MgO pillow basalt (markedwith arrows) from photograph (a) (area of photograph indicated with a white box) (d) Photograph of margin of pahoehoe lobe (sledgehammerfor scale) (e) Cross-section of a large picritic pillow lobe with infilling of quartz^carbonate in cooling^contraction cracks (sledgehammer forscale) (f) High-MgO pillow breccia stratigraphically between submarine and subaerial flows (arrow points to lens cap 7 cm diameter forscale)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
6
Plagioclase forms most of the phenocrysts and glomero-crysts and the groundmass is fine-grained comprising pla-gioclase microlites clinopyroxene granules small grains ofFe^Ti oxide devitrified glass and secondary minerals
there is no fresh glass preserved The picritic and high-MgO basaltic lavas are characterized by spherulitic plagi-oclase and clinopyroxene with abundant euhedral olivinephenocrysts (Fig 5 Table 1) Olivine is completely replaced
Fig 4 Photographs showing field relations from the Schoen Lake Provincial Park area northernVancouver Island (a) Sediment^sill complexat the base of the Karmutsen Formation on the north side of Mt Schoen 100m below the Daonella locality (see Fig 2 for location) (b)Interbedded mafic sills and deformed finely banded chert and shale with calcareous horizons from location between Mt Adam and MtSchoen
Fig 5 Photomicrographs of picritic pillow basalts Karmutsen Formation northernVancouver Island Scale bars represent 1mm (a) Picritewith abundant euhedral olivine pseudomorphs from Keogh Lake type locality (Fig 2) in cross-polarized light (XPL) (sample 4722A4 19^198wt MgO four analyses) Samples contain dense clusters (24^42 vol ) of olivine pseudomorphs (52mm) in a groundmass of curvedand branching sheaves of acicular plagioclase and intergrown with clinopyroxene and altered glass In many cases plagioclase nucleated on theedges of the olivine phenocrysts (b) Sheaves of intergrown plagioclase and clinopyroxene in aphyric picrite pillow lava from west of MaynardLake (Fig 2) in XPL (sample 4723A2108 wt MgO)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
7
Table 1 Summary of petrographic characteristics and phenocryst proportions of Karmutsen basalts on Vancouver Island
BC
Sample Area Flow Group Texture vol Ol Plag Cpx Ox Alteration Notes
4718A1 MA PIL THOL intersertal 20 5 3 few plag glcr52mm cpx51 mm
4718A2 MA PIL THOL intersertal glomero 15 1 plag glcr52mm very fresh
4718A5 MA FLO THOL intersertal glomero 10 2 mottled few plag glcr54 mm
4723A4 KR PIL PIC intergranular intersertal 0 3 swtl plag51mm no ol phenos
4723A13 KR PIL PIC spherulitic 24 3 swtl plag51 mm
(continued)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
8
by an assemblage of talc^tremolite^serpentine^opaqueoxides there is no fresh olivine present in the Karmutsenlavas onVancouver Island Plagioclase and clinopyroxenephenocrysts are absent in the high-MgO lavas Thecoarse-grained mafic rocks are characterized by sub-ophi-tic textures with an average grain size typically 41mmand are generally not glomeroporphyritic these rocks aremainly from the interiors of massive flows although somemay represent sillsSample preparation and analytical methods for major-
and trace-element and Sr^Nd^Hf^Pb isotopic composi-tions of the Karmutsen volcanic rocks are given in theAppendix
WHOLE -ROCK CHEMISTRYMajor- and trace-element compositionsThe most abundant type of volcanic rock in theKarmutsen Formation is tholeiitic basalt with a restrictedrange of major- and trace-element compositions The tho-leiitic basalts have similar compositions to the coarse-grained mafic rocks and both groups are distinct fromthe picrites and high-MgO basalts [Fig 6 picrites412wt MgO (LeBas 2000) high-MgO basalts8^12wt MgO] The tholeiitic basalts have lower MgO(57^77wt MgO) and higher TiO2 (14^22wt
TiO2) than the picrites (130^198wt MgO05^07wt TiO2) and high-MgO basalts (91^116wt MgO 05^08wt TiO2 Fig 6 Table 2) Almost allcompositions plot within the tholeiitic field in a total alka-lis vs silica plot although there has been substantial K lossin most samples from the Karmutsen Formation (seebelow) and the tholeiitic basalts generally have higherSiO2 Na2OthornK2O CaO and FeOT (total iron expressedas FeO) than the picrites The tholeiitic basalts also havenoticeably lower loss on ignition (LOI mean1713wt ) than the picrites (mean 54 08wt )and high-MgO basalts (mean 3815wt ) (Fig 6)reflecting the presence of abundant altered olivinephenocrysts in the latter groups Ni concentrations are sig-nificantly higher for the picrites (339^755 ppm) and high-MgO basalts (122^551ppm) than the tholeiitic basalts(58^125 ppm except for one anomalous sample) andcoarse-grained rocks (70^213 ppm) (Fig 6 Table 2) Ananomalous tholeiitic pillowed flow (two samples outlierin Tables 1 and 2) from near the picrite type locality atKeogh Lake and a mineralized sill (disseminated sulfide)from the basal sediment^sill complex at Schoen Lake aredistinguished by higher TiO2 and FeOT than the maingroup of tholeiitic basalts The tholeiitic basalts are similarin major-element composition to previously publishedresults for samples from the Karmutsen Formation how-ever the Keogh Lake picrites which encompass the
Table 1 Continued
Sample Area Flow Group Texture vol Ol Plag Cpx Ox Alteration Notes
4722A5 KR FLO OUTLIER intersertal 3 mottled very fg v small pl needles
5615A11 KR PIL OUTLIER intersertal 3 mottled very fg v small pl needles
5617A4 SL SIL MIN intersertal 5 3 fg
Sample number last digit of year month day initial sample station (except 93G171) MA Mount Arrowsmith SLSchoen Lake KR Karmutsen Range PIL pillow BRE breccia FLO flow SIL sill GAB gabbro THOL tholeiiticbasalt PIC picrite HI-MG high-MgO basalt CGcoarse-grained MIN mineralized sill glomero glomeroporphyriticOlivine modes for picrites are based on area calculations from scans (see lsquoOlivine accumulation in high-MgO and picriticlavasrsquo section and Fig 11) Visual alteration index is based primarily on degree of plagioclase alteration and presence ofsecondary minerals (1 least altered 3 most altered) Plagioclase phenocrysts are commonly altered to albite pumpellyiteand chlorite olivine is altered to talc tremolite and clinochlore (determined using the Rietveld method of X-ray powderdiffraction) clinopyroxene is unaltered FendashTi oxide is commonly replaced by sphene glcr glomerocrysts fg fine-grained cg coarse-grained oik oikocryst chad chadacryst enc enclosing swtl swallow-tail ol olivinepseudomorphs plag plagioclase cpx clinopyroxene ox oxides (includes ilmenite thorn titanomagnetite)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
9
00
05
10
15
20
25
30
35
40
0 5 10 15 20 25
10
11
12
13
14
15
16
17
18
19
20
0 5 10 15 20 25
6
7
8
9
10
11
12
13
14
15
0 5 10 15 20 25
6
7
8
9
10
11
12
13
14
15
16
0 5 10 15 20 25
0
1
2
3
4
5
6
40 45 50 55 60
0
200
400
600
800
0 5 10 15 20 25
(wt)(wt)
(wt)
(wt)
+K2ONa2O TiO2
FeO (T)
Ni (ppm)
Al2O3
CaO
alkalic
tholeiitic
Karmutsen Form
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
(compiled data)
Pillowed flowMineralized sill
SiO2 (wt) MgO (wt)
MgO (wt)MgO (wt)
MgO (wt) MgO (wt)
(a) (b)
(g)
(e)
(f)
(c)
LOI (wt )
(d)
MgO (wt )0
1
2
3
4
5
6
7
0 5 10 15 20 25
(wt)
Fig 6 Whole-rock major-element Ni and LOI variation diagrams for the Karmutsen Formation New samples from this study (Table 2) areshown by symbols with black outlines (see legend) Previously published analyses are shown by small gray symbols without black outlines Theboundary of the alkaline and tholeiitic fields is that of MacDonald amp Katsura (1964) Total iron expressed as FeO LOI loss on ignition oxidesare plotted on an anhydrous normalized basis References for the 322 compiled analyses for the Karmutsen Formation are Surdam (1967)Kuniyoshi (1972) Muller et al (1974) Barker et al (1989) Lassiter et al (1995) Massey (1995a1995b)Yorath et al (1999) and G Nixon (unpublisheddata) It should be noted that the compiled dataset has not been filtered many of the samples with high SiO2 (452wt ) are probably notKarmutsen flood basalts but younger Bonanza basalts and andesites Three pillowed flow samples and a mineralized sill are distinguishedseparately because of their anomalous chemistry Coarse-grained samples are indicated separately Three tholeiitic basalt samples (4724A54718A5 4718A6) with416wt Al2O3 and4270 ppm Sr contain 10^25 modal of large plagioclase phenocrysts (7^8mm) or glomerocrysts(44mm)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
10
Table 2 Major element (wt oxide) and trace element (ppm) abundances in whole-rock samples of Karmutsen basalts
1Ni concentrations for these samples were determined by XRFTHOL tholeiitic basalt PIC picrite HI-MG high MgO basalt CG coarse-grained (sill or gabbro) MIN SIL mineralizedsill OUTLIER anomalous pillowed flow in plots MA Mount Arrowsmith SL Schoen Lake KR Karmutsen Range QIQuadra Island Sample locations are given using the Universal Transverse Mercator (UTM) coordinate system (NAD83zones 9 and 10) Analyses were performed at Activation Laboratory (ActLabs) Fe2O3
is total iron expressed as Fe2O3LOI loss on ignition All major elements Sr V and Y were measured by ICP quadrupole OES on solutions of fusedsamples Cu Ni Pb and Zn were measured by total dilution ICP Cs Ga Ge Hf Nb Rb Ta Th U Zr and REE weremeasured by magnetic-sector ICP on solutions of fused samples Co Cr and Sc were measured by INAA Blanks arebelow detection limit See Electronic Appendices 2 and 3 for complete XRF data and PCIGR trace element datarespectively Major elements for sample 93G17 are normalized anhydrous Sample 5615A7 was analyzed by XRFSamples from Quadra Island are not shown in figures
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
17
picritic and high-MgO basalt pillow lavas extend to20wt MgO (Fig 6 and references listed in thecaption)The tholeiitic basalts from the Karmutsen Range dis-
play a tight range of parallel light rare earth element(LREE)-enriched REE patterns (LaYbCNfrac14 21^24mean 2202) whereas the picrites and high-MgObasalts are characterized by sub-parallel LREE-depleted
patterns (LaYbCNfrac14 03^07 mean 06 02) with lowerREE abundances than the tholeiitic basalts (Fig 7)Tholeiitic basalts from the Schoen Lake and MountArrowsmith areas have similar REE patterns tosamples from the Karmutsen Range (Schoen Lake LaYbCNfrac14 21^26 mean 2303 Mount Arrowsmith LaYbCNfrac1419^25 mean 2204) and the coarse-grainedmafic rocks have similar REE patterns to the tholeiitic
Depleted MORB average(Salters amp Stracke 2004)
Schoen LakeSchoen Lake
Mount Arrowsmith
Karmutsen RangeS
ampl
eC
hond
rite
Sam
ple
Cho
ndrit
e
Sam
ple
Prim
itive
Man
tleS
ampl
eP
rimiti
ve M
antle
Sam
ple
Prim
itive
Man
tle
Mount Arrowsmith
Karmutsen Range
(a) (b)
(c) (d)
(e) (f )
Sam
ple
Cho
ndrit
e
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained mafic rock
Pillowed flowMineralized sill
1
10
100
1
10
100
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
01
1
10
100
1
10
100
1
10
100
Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
Fig 7 Whole-rock REE and other incompatible-element concentrations for the Karmutsen Formation (a) (c) and (e) are chondrite-normal-ized REE patterns for samples from each of the three main field areas onVancouver Island (b) (d) and (f) are primitive mantle-normalizedtrace-element patterns for samples from each of the three main field areas All normalization values are from McDonough amp Sun (1995) Theclear distinction between LREE-enriched tholeiitic basalts and LREE-depleted picrites the effects of alteration on the LILE (especially loss ofK and Rb) and the different range for the vertical scale in (b) (extends down to 01) should be noted
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
18
basalts (LaYbCNfrac14 21^29 mean 2508) Two LREE-depleted coarse-grained mafic rocks from the SchoenLake area have similar REE patterns (LaYbCNfrac14 07) tothe picrites from the Keogh Lake area indicating that thisdistinctive suite of rocks is exposed as far south asSchoen Lake that is 75 km away (Fig 7) The anomalouspillowed flow from the Karmutsen Range has lower LaYbCN (13^18) than the main group of tholeiitic basaltsfrom the Karmutsen Range (Fig 7) The mineralized sillfrom the Schoen Lake area is distinctly LREE-enriched(LaYbCNfrac14 33) with the highest REE abundances(LaCNfrac14 940) of all Karmutsen samples this may reflectcontamination by adjacent sediment also evident in thin-section where sulfide blebs are cored by quartz (Greeneet al 2006)The primitive mantle-normalized trace-element pat-
terns (Fig 7) and trace-element variations and ratios(Fig 8) highlight the differences in trace-element
concentrations between the four main groups ofKarmutsen flood basalts onVancouver Island The tholeii-tic basalts from all three areas have relatively smooth par-allel trace-element patterns some samples have smallpositive Sr anomalies relative to Nd and Sm and manysamples have prominent negative K anomalies relative toU and Nb reflecting K loss during alteration (Fig 7) Thelarge ion lithophile elements (LILE) in the tholeiiticbasalts (mainly Rb and K not Cs) are depleted relativeto the high field strength elements (HFSE) and LREELILE segments especially for the Schoen Lake area(Fig 7d) are remarkably parallel The picrites andhigh-MgO basalts form a tight band of parallel trace-element patterns and are depleted in HFSE and LREEwith positive Sr anomalies and relatively low Rb Thecoarse-grained mafic rocks from the Karmutsen Rangehave identical trace-element patterns to the tholeiiticbasalts Samples from the Schoen Lake area have similar
000
025
050
075
100
125
0 1 2 3 4 5
00
04
08
12
16
0 5 10 15 20 25
MgO (wt)
0
5
10
15
0 50 100 150
Hf (ppm)
Th (ppm)
NbLaNb (ppm)
Zr (ppm)
(a)
(c)
(b)
(d)
Th (ppm)
00
02
04
06
08
10
12
00 01 02 03 04 05
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
NbLa=1
4723A4 5615A12
U (ppm)
Fig 8 Whole-rock trace-element concentrations and ratios for the Karmutsen Formation [except (b) which is vs wt MgO] (a) Nb vs Zr(b) NbLa vs MgO (c) Th vs Hf (d) Th vs U There is a clear distinction between the tholeiitic basalts and picrites in Nb and Zr both inconcentration and the slope of each trend
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
19
trace-element patterns to the Karmutsen Range except forthe two LREE-depleted coarse-grained mafic rocks andthe mineralized sill (Fig 7d)
Sr^Nd^Hf^Pb isotopic compositionsA total of 19 samples from the four main groups ofthe Karmutsen Formation were selected for radiogenicisotopic geochemistry on the basis of major- and trace-element variation stratigraphic position and geographicallocation The samples have overlapping age-correctedHf and Nd isotope ratios and distinct ranges of Sr isotopiccompositions (Fig 9) The tholeiitic basalts picriteshigh-MgO basalt and coarse-grained mafic rocks have
initial eHffrac14thorn87 to thorn126 and eNdfrac14thorn66 to thorn88 cor-rected for in situ radioactive decay since 230 Ma (Fig 9Tables 3 and 4) All the Karmutsen samples form a well-defined linear array in a Lu^Hf isochron diagram corre-sponding to an age of 24115 Ma (Fig 9) This is withinerror of the accepted age of the Karmutsen Formationand indicates that the Hf isotope systematics behaved as aclosed system since c 230 Ma The anomalous pillowedflow has similar initial eHf (thorn98) and slightly lower initialeNd (thorn62) compared with the other Karmutsen samplesThe picrites and high-MgO basalt have higher initial87Sr86Sr (070398^070518) than the tholeiitic basalts(initial 87Sr86Srfrac14 070306^070381) and the coarse-grained mafic rocks (initial 87Sr86Srfrac14 070265^070428)
05128
05129
05130
05131
05132
015 017 019 021 023 025
02829
02830
02831
02832
02833
02834
000 002 004 006 008 010
4
6
8
10
0702 0703 0704 0705 07066
8
10
12
14
5 6 7 8 9 10
12
0
1
2
3
4
5
6
7
0702 0703 0704 0705 0706
(d)
LOI (wt )
87Sr 86Sr
4723A2
Nd
143Nd144Nd
147Sm144Nd
87Sr 86Sr (230 Ma)
(230 Ma)
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
(a)
(c)
176Hf177Hf
Age=241 plusmn 15 Ma=+990 plusmn 064
176Lu177Hf
Hf (230 Ma)
Nd (230 Ma)
(e)
Hf (i)
(b)
Fig 9 Whole-rock Sr Nd and Hf isotopic compositions for the Karmutsen Formation (a) 143Nd144Nd vs 147Sm144Nd (b) 176Hf177Hf vs176Lu177Hf The slope of the best-fit line for all samples corresponds to an age of 24115 Ma (c) Initial eNd vs 87Sr86Sr Age-correction to230 Ma (d) LOI vs 87Sr86Sr (e) Initial eHf vs eNd Average 2 error bars are shown in a corner of each panel Complete chemical duplicatesshown inTables 3 and 4 (samples 4720A7 and 4722A4) are circled in each plot
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
overlap the ranges of the picrites and tholeiitic basalts(Fig 9 Table 3)The measured Pb isotopic compositions of the tholeiitic
basalts are more radiogenic than those of the picrites andthe most magnesian Keogh Lake picrites have the leastradiogenic Pb isotopic compositions The range ofinitial Pb isotope ratios for the picrites is206Pb204Pbfrac1417868^18580 207Pb204Pbfrac1415547^15562and 208Pb204Pbfrac14 37873^38257 and the range for thetholeiitic basalts is 206Pb204Pbfrac1418782^19098207Pb204Pbfrac1415570^15584 and 208Pb204Pbfrac14 37312^38587 (Fig 10 Table 5) The coarse-grained mafic rocksoverlap the range of initial Pb isotopic compositions forthe tholeiitic basalts with 206Pb204Pbfrac1418652^19155207Pb204Pbfrac1415568^15588 and 208Pb204Pbfrac14 38153^38735 One high-MgO basalt has the highest initial206Pb204Pb (19242) and 207Pb204Pb (15590) and thelowest initial 208Pb204Pb (37676) The Pb isotopic ratiosfor Karmutsen samples define broadly linearrelationships in Pb isotope plots The age-corrected Pb
isotopic compositions of several picrites have been affectedby U and Pb (andor Th) mobility during alteration(Fig 10)
ALTERAT IONThe Karmutsen basalts have retained most of their origi-nal igneous structures and textures however secondaryalteration and low-grade metamorphism have producedzeolitic- and prehnite^pumpellyite-bearing mineral assem-blages (Table 1 Cho et al 1987) Alteration and metamor-phism have primarily affected the distribution of the LILEand the Sr isotopic systematics The Keogh Lake picriteshave been affected more by alteration than the tholeiiticbasalts and have higher LOI (up to 55wt ) variableLILE and higher measured Sr Nd and Hf and lowermeasured Pb isotope ratios than the tholeiitic basalts Srand Pb isotope compositions for picrites and tholeiiticbasalts show a relationship with LOI (Figs 9 and 10)whereas Nd and Hf isotopic compositions do not correlate
Table 3 Sr and Nd isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Rb Sr 87Sr86Sr 2m87Rb86Sr 87Sr86Srt Sm Nd 143Nd144Nd 2m eNd
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses carried out at the PCIGR thecomplete trace element analyses are given in Electronic Appendix 3
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
21
with LOI (not shown) The relatively high apparent initialSr isotopic compositions for the high-MgO lavas (up to07052) probably resulted from post-magmatic loss of Rbor an increase in 87Sr86Sr through addition of seawater Sr(eg Hauff et al 2003) High-MgO lavas from theCaribbean plateau have similarly high 87Sr86Sr comparedwith basalts (Recurren villon et al 2002)The correction for in situdecay on initial Pb isotopic ratios has been affected bymobilization of U and Th (and Pb) in the whole-rockssince their formation A thorough acid leaching duringsample preparation was used that has been shown to effec-tively remove alteration phases (Weis et al 2006 NobreSilva et al in preparation) Whole-rock SmNd ratiosand age-correction of Nd isotopes may be affected bythe leaching procedure for submarine rocks older than50 Ma (Thompson et al 2008 Fig 9) there ishowever very little variation in Nd isotopic compositionsof the picrites or tholeiitic basalts and this system does notappear to be disturbed by alteration The HFSE abun-dances for tholeiitic and picritic basalts exhibit clear
linear correlations in binary diagrams (Fig 8) whereasplots of LILE vs HFSE are highly scattered (not shown)because of the mobility of some of the alkali and alkalineearth LILE (Cs Rb Ba K) and Sr for most samplesduring alterationThe degree of alteration in the Karmutsen samples does
not appear to be related to eruption environment or depthof burial Pillow rims are compositionally different frompillow cores (Surdam1967 Kuniyoshi1972) and aquagenetuffs are chemically different from isolated pillows withinpillow breccias however submarine and subaerial basaltsexhibit a similar degree of alterationThere is no clear cor-relation between the submarine and subaerial basalts andsome commonly used chemical alteration indices (eg BaRb vs K2OP2O5 Huang amp Frey 2005) There is also nodefinitive correlation between the petrographic alterationindex (Table 1) and chemical alteration indices although10 of 13 tholeiitic basalts with the highest K and LILEabundances have the highest petrographic alterationindex of three (seeTable 1)
Table 4 Hf isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Lu Hf 176Hf177Hf 2m eHf176Lu177Hf 176Hf177Hft eHf(t)
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
OLIV INE ACCUMULAT ION INH IGH-MgO AND PICR IT IC LAVASGeochemical trends and petrographic characteristics indi-cate that accumulation of olivine played an important rolein the formation of the high-MgO basalts and picrites fromthe Keogh Lake area on northern Vancouver Island TheKeogh Lake samples show a strong linear correlation inplots of Al2O3 TiO2 ScYb and Ni vs MgO and many ofthe picrites have abundant clusters of olivine pseudomorphs(Table1 Fig11)There is a strong linear correlationbetween
modal per cent olivine and whole-rock magnesium contentmagnesium contents range between 10 and 19wt MgOand the proportion of olivine phenocrysts in these samplesvaries from 0 to 42 vol (Fig 11)With a few exceptionsboth the size range and median size of the olivine pheno-crysts in most samples are comparable The clear correla-tion between the proportion of olivine phenocrysts andwhole-rock magnesium contents indicates that accumula-tion of olivine was directly responsible for the high MgOcontents (410wt ) in most of the picritic lavas
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
207Pb204Pb
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
208Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
(a) (b)
(c) (e)
4723A24723A2
4723A2
4723A2
4722A4
4723A4
4722A4
4723A4
4723A134723A3
4723A3
4723A13
1556
1558
1560
1562
1564
1566
1568
186 190 194 198 202 206 2101550
1552
1554
1556
1558
1560
178 180 182 184 186 188 190 192 194
380
384
388
392
396
186 190 194 198 202 206 210374
378
382
386
390
178 180 182 184 186 188 190 192 194
206Pb204Pb(d)
LOI (wt )
0
1
2
3
4
5
6
7
186 190 194 198 202 206 210
4723A2
Fig 10 Pb isotopic compositions of leached whole-rock samples measured by MC-ICP-MS for the Karmutsen Formation Error bars are smal-ler than symbols (a) Measured 207Pb204Pb vs 206Pb204Pb (b) Initial 207Pb204Pb vs 206Pb204Pb Age-correction to 230 Ma (c) Measured208Pb204Pb vs 206Pb204Pb (d) LOI vs 206Pb204Pb (e) Initial 208Pb204Pb vs 206Pb204Pb Complete chemical duplicates shown in Table 5(samples 4720A7 and 4722A4) are circled in each plot The dashed lines in (b) and (e) show the differences in age-corrections for two picrites(4723A4 4722A4) using the measured UTh and Pb concentrations for each sample and the age-corrections when concentrations are used fromthe two picrites (4723A3 4723A13) that appear to be least affected by alteration
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
23
DISCUSS IONThe compositional and spatial development of the eruptedlava sequences of the Wrangellia oceanic plateau onVancouver Island provide information about the evolutionof a mantle plume upwelling beneath oceanic lithospherein an analogous way to the interpretation of continentalflood basalts The Caribbean and Ontong Java plateausare the two primary examples of accreted parts of oceanicplateaus that have been studied to date and comparisonscan be made with the Wrangellia oceanic plateauHowever the Wrangellia oceanic plateau is a distinctiveoccurrence of an oceanic plateau because it formed on thecrust of a Devonian to Mississippian intra-oceanic islandarc overlain by Mississippian to Permian limestones andpelagic sediments The nature and thickness of oceaniclithospheric mantle are considered to play a fundamentalrole in the decompression and evolution of a mantleplume and thus the compositions of the erupted lavasequences (Greene et al 2008) The geochemical and
stratigraphic relationships observed in the KarmutsenFormation onVancouver Island and the focus of the subse-quent discussion sections provide constraints on (1) thetemperature and degree of melting required to generatethe primary picritic magmas (2) the composition of themantle source (3) the depth of melting and residual min-eralogy in the source region (4) the low-pressure evolutionof the Karmutsen tholeiitic basalts (5) the growth of theWrangellia oceanic plateau
Melting conditions and major-elementcomposition of the primary magmasThe primary magmas to most flood basalt provinces arebelieved to be picritic (eg Cox 1980) and they have beenfound in many continental and oceanic flood basaltsequences worldwide (eg Siberia Karoo Paranacurren ^Etendeka Caribbean Deccan Saunders 2005) Near-pri-mary picritic lavas with low total alkali abundances
Table 5 Pb isotopic compositions of Karmutsen basaltsVancouver Island BC
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
require high-degree partial melting of the mantle source(eg Herzberg amp OrsquoHara 2002) The Keogh Lake picritesfrom northernVancouver Island are the best candidates foridentifying the least modified partial melts of the mantleplume source despite having accumulated olivine pheno-crysts and can be used to estimate conditions of meltingand the composition of the primary magmas for theKarmutsen FormationPrimary magma compositions mantle potential tem-
peratures and source melt fractions for the picritic lavas ofthe Karmutsen Formation were calculated from primitivewhole-rock compositions using PRIMELT1XLS software(Herzberg et al 2007) and are presented in Fig 12 andTable 6 A detailed discussion of the computationalmethod has been given by Herzberg amp OrsquoHara (2002)and Herzberg et al (2007) key features of the approachare outlined belowPrimary magma composition is calculated for a primi-
tive lava by incremental addition of olivine and compari-son with primary melts of mantle peridotite Lavas thathad experienced plagioclase andor clinopyroxene fractio-nation are excluded from this analysis All calculated pri-mary magma compositions are assumed to be derived byaccumulated fractional melting of fertile peridotite KR-4003 Uncertainties in fertile peridotite composition donot propagate to significant variations in melt fractionand mantle potential temperature melting of depletedperidotite propagates to calculated melt fractions that aretoo high but with a negligible error in mantle potential
temperature PRIMELT1XLS calculates the olivine liqui-dus temperature at 1atm which should be similar to theactual eruption temperature of the primary magmaand is accurate to 308C at the 2 level of confidenceMantle potential temperature is model dependent witha precision that is similar to the accuracy of the olivineliquidus temperature (C Herzberg personal communica-tion 2008)With regard to the evidence for accumulated olivine in
the Keogh Lake picrites it should not matter whether themodeled lava composition is aphyric or laden with accu-mulated olivine phenocrysts PRIMELT1 computes theprimary magma composition by adding olivine to a lavathat has experienced olivine fractionation in this way itinverts the effects of fractional crystallization The onlyrequirement is that the olivine phenocrysts are not acci-dental or exotic to the lava compositions Support for thisprocedure can be seen in the calculated olivine liquid-line-of-descent shown by the curved arrays with 5 olivineaddition increments (Fig 12c) Fractional crystallization ofolivine from model primary magmas having 85^95wt FeO and 153^175wt MgO will yield arrays of deriva-tive liquids and randomly sorted olivines that in most butnot all cases connect the picrites and high-MgO basalts Anewer version of the PRIMELT software (Herzberg ampAsimow 2008) can perform olivine addition to and sub-traction from lavas with highly variably olivine phenocrystcontents and yields results that converge with those shownin Fig 12
Fig 11 Relationship between abundance of olivine phenocrysts and whole-rock MgO contents for the Keogh Lake picrites (a) Tracings ofolivine grains (gray) in six high-MgO samples made from high resolution (2400 dpi) scans of petrographic thin-sections Scale bars represent10mm sample numbers are indicated at the upper left (b) Modal olivine vs whole-rock MgO Continuous line is the best-fit line for 10 high-MgO samplesThe areal proportion of olivine was calculated from raster images of olivine grains using ImageJ image analysis software whichprovides an acceptable estimation of the modal abundance (eg Chayes 1954)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
25
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
Gray lines = final melting pressure
L + Ol + Opx
07
08
09
Pressure (GPa)
10
3
4
5
6
1
SOLIDUS
01
04
0506
L + Ol
2
03
02
PeridotiteKR-4003
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
Thick black lines = initial melting pressure
Dashed lines = melt fraction
Melt Fraction
2
3 4 5 6
L+Ol+Opx
SolidusSpinel
SolidusGarnet Peridotite
7
L+Ol
0 10 20 30 4040
45
50
55
60
MgO (wt)
SiO2
(wt)
PeridotiteKR-4003
Melt Fraction
0 10 20 305
10
15
CaO(wt)
MgO (wt)
3
4 5 6 7
2
46
2
Peridotite Partial Melts
Thick black line = initial melting pressureGray lines = final melting pressure
Peridotite
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
02
Pressure (GPa)
0302
04
Pressure (GPa)
Picrite
High-MgO basalt
Tholeiitic basalt
(c)
(d)
Karmutsen flood basaltsOlivine addition modelOlivine addition (5 increment)Primary magma
PicriteHigh-MgO basaltTholeiitic basalt
(a)
(b)
800 850 900
910
920
930
940
950
Karmutsen flood basalts
Olivine addition modelOlivine addition (5 increment)Primary magma
Gray lines = Fo contents of olivine
Fig 12 Estimated primary magma compositions for three Keogh Lake picrites and high-MgO basalts (samples 4723A4 4723A135616A7) using the forward and inverse modeling technique of Herzberg et al (2007) Karmutsen compositions and modeling results areoverlain on diagrams provided by C Herzberg (a) Whole-rock FeO vs MgO for Karmutsen samples from this study Total iron estimatedto be FeO is 090 Gray lines show olivine compositions that would precipitate from liquid of a given MgO^FeO composition Black lineswith crosses show results of olivine addition (inverse model) using PRIMELT1 (Herzberg et al 2007) Gray field highlights olivinecompositions with Fo contents of 91^92 predicted to precipitate (b) FeO vs MgO showing Karmutsen lava compositions and results ofa forward model for accumulated fractional melting of fertile peridotite KR-4003 (Kettle River Peridotite Walter 1998) (c) SiO2 vsMgO for Karmtusen lavas and model results (d) CaO vs MgO showing Karmtusen lava compositions and model results To brieflysummarize the technique [see Herzberg et al (2007) for complete description] potential parental magma compositions for the high-MgOlava series were selected (highest MgO and appropriate CaO) and using PRIMELT1 software olivine was incrementally added tothe selected compositions to show an array of potential primary magma compositions (inverse model) Then using PRIMELT1 theresults from the inverse model were compared with a range of accumulated fractional melts for fertile peridotite derived fromparameterization of the experimental results for KR-4003 from Walter (1998) forward model Herzberg amp OrsquoHara (2002) A melt fractionwas sought that was unique to both the inverse and forward models (Herzberg et al 2007) A unique solution was found when there was a com-mon melt fraction for both models in FeO^MgO and CaO^MgO^Al2O3^SiO2 (CMAS) projection space This modeling assumes thatolivine was the only phase crystallizing and ignores chromite precipitation and possible augite fractionation in the mantle (Herzbergamp OrsquoHara 2002) Results are best for a residue of spinel lherzolite (not pyroxenite) The presence of accumulated olivine in samples ofthe Keogh Lake picrites used as starting compositions does not significantly affect the results because we are modeling addition of olivine(see text for further description) The tholeiitic basalts cannot be used for modeling because they are all saturated in plagthorn cpxthornolThick black line for initial melting pressure at 4 GPa is not shown in (c) for clarity Gray area in (a) indicates parental magmas thatwill crystallize olivine with Fo 91^92 at the surface Dashed lines in (c) indicate melt fraction Solidi for spinel and garnet peridotite arelabelled in (c)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
The PRIMELT1 results indicate that if the Karmutsenpicritic lavas were derived from accumulated fractionalmelting of fertile peridotite the primary magmas wouldhave contained 15^17wt MgO and 10wt CaO(Fig 12 Table 6) formed from 23^27 partial meltingand would have crystallized olivine with a Fo content of 91 (Fig 12 Table 6) Calculated mantle temperatures forthe Karmutsen picrites (14908C) indicate melting ofanomalously hot mantle (100^2008C above ambient) com-pared with ambient mantle that produces mid-ocean ridgebasalt (MORB 1280^14008C Herzberg et al 2007)which initiated between 3 and 4 GPa Several picritesappear to be near-primary melts and required very littleaddition of olivine (eg sample 93G171 with measured158wt MgO) These temperature and melting esti-mates of the Karmutsen picrites are consistent withdecompression of hot mantle peridotite in an actively con-vecting plume head (ie plume initiation model) Threesamples (picrite 5614A1 and high-MgO basalts 5614A3and 5614A5) are lower in iron than the other Karmutsenlavas with FeO in the 65^80wt range (Fig 12c) andhave MgO and FeO contents that are similar to primitiveMORB from the Sequeiros fracture zone of the EastPacific Rise (EPR Perfit et al 1996) These samples
however differ from EPR MORB in having higher SiO2
and lower Al2O3 and they are restricted geographicallyto one area (Sara Lake Fig 2)We therefore have evidence for primary magmas with
MgO contents that range from a possible low of 110wt to as high as 174wt with inferred mantle potential tem-peratures from 1340 to 15208C This is similar to the rangedisplayed by lavas from the Ontong Java Plateau deter-mined by Herzberg (2004) who found that primarymagma compositions assuming a peridotite sourcewould contain 17wt MgO and 10wt CaO(Table 6) and that these magmas would have formed by27 melting at a mantle potential temperature of15008C (first melting occurring at 36 GPa) Fitton ampGodard (2004) estimated 30 partial melting for theOntong Java lavas based on the Zr contents of primarymagmas of Kroenke-type basalt calculated by incrementaladdition of equilibrium olivine However if a considerableamount of eclogite was involved in the formation of theOntong Java plateau magmas the excess temperatures esti-mated by Herzberg et al (2007) may not be required(Korenaga 2005) The variability in mantle potential tem-perature for the Keogh Lake picrites is also similar to thewide range of temperatures that have been calculated for
Table 6 Estimated primary magma compositions for Karmutsen basalts and other oceanic plateaus or islands
Sample 4723A4 4723A13 5616A7 93G171 Average OJP Mauna Keay Gorgonaz
Ontong Java primary magma composition for accumulated fraction melting (AFM) from Herzberg (2004)yMauna Kea primary magma composition is average of four samples of Herzberg (2006 table 1)zGorgona primary magma composition for AFM for 1F fertile source from Herzberg amp OrsquoHara (2002 table 4)Total iron estimated to be FeO is 090
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
27
lavas from single volcanoes in the Galapagos Hawaii andelsewhere by Herzberg amp Asimow (2008) Those workersinterpreted the range to reflect the transport of magmasfrom both the cool periphery and the hotter axis of aplume A thermally heterogeneous plume will yield pri-mary magmas with different MgO contents and the rangeof primary magma compositions and their inferred mantlepotential temperatures for Karmutsen lavas may reflectmelting from different parts of the plume
Source of the Karmutsen Formation lavasThe Sr^Nd^Hf^Pb isotopic compositions of theKarmutsen volcanic rocks on Vancouver Island indicatethat they formed from plume-type mantle containing adepleted component that is compositionally similar to butdistinct from other oceanic plateaus that formed in thePacific Ocean (eg Ontong Java and Caribbean Fig 13)Trace element and isotopic constraints can be used to testwhether the depleted component of the KarmutsenFormation was intrinsic to the mantle plume and distinctfrom the source of MORB or the result of mixing or invol-vement of ambient depleted MORB mantle with plume-type mantle The Karmutsen tholeiitic basalts have initialeNd and
87Sr86Sr ratios that fall within the range of basaltsfrom the Caribbean Plateau (eg Kerr et al 1996 1997Hauff et al 2000 Kerr 2003) and a range of Hawaiian iso-tope data (eg Frey et al 2005) and lie within and abovethe range for the Ontong Java Plateau (Tejada et al 2004Fig 13a) Karmutsen tholeiitic basalts with the highest ini-tial eNd and lowest initial 87Sr86Sr lie within and justbelow the field for age-corrected northern EPR MORB at230 Ma (eg Niu et al1999 Regelous et al1999) Initial Hfand Nd isotopic compositions for the Karmutsen samplesare noticeably offset to the low side of the ocean islandbasalt (OIB) array (Vervoort et al 1999) and lie belowand slightly overlap the low eHf end of fields for Hawaiiand the Caribbean Plateau (Fig 13c) the Karmutsen sam-ples mostly fall between and slightly overlap a field forage-corrected Pacific MORB at 230 Ma and Ontong Java(Mahoney et al 1992 1994 Nowell et al 1998 Chauvel ampBlichert-Toft 2001) Karmutsen lava Pb isotope ratios over-lap the broad range for the Caribbean Plateau and thelinear trends for the Karmutsen samples intersect thefields for EPR MORB Hawaii and Ontong Java in208Pb^206Pb space but do not intersect these fields in207Pb^206Pb space (Fig 13d and e) The more radiogenic207Pb204Pb for a given 206Pb204Pb of the Karmutsenbasalts requires high UPb ratios at an ancient time when235U was abundant similar to the Caribbean Plateau andenriched in comparison with EPR MORB Hawaii andOntong JavaThe low eHf^low eNd component of the Karmutsen
basalts that places them below the OIB mantle array andpartly within the field for age-corrected Pacific MORBcan form from low ratios of LuHf and high SmNd over
time (eg Salters amp White 1998) MORB will develop aHf^Nd isotopic signature over time that falls below themantle array Low LuHf can also lead to low 176Hf177Hffrom remelting of residues produced by melting in theabsence of garnet (eg Salters amp Hart 1991 Salters ampWhite 1998) The Hf^Nd isotopic compositions of theKarmutsen Formation indicate the possible involvementof depleted MORB mantle however similar to Iceland(Fitton et al 2003) Nb^Zr^Y variations provide a way todistinguish between a depleted component within theplume and a MORB-like component The Karmutsenbasalts and picrites fall within the Iceland array and donot overlap the MORB field in a logarithmic plot of NbYvs ZrY (Fig 13b) using the fields established by Fittonet al (1997) Similar to the depleted component within theCaribbean Plateau (Thompson et al 2003) the trend ofHf^Nd isotopic compositions towards Pacific MORB-likecompositions is not followed by MORB values in Nb^Zr^Y systematics and this suggests that the depleted compo-nent in the source of the Karmutsen basalts is distinctfrom the source of MORB and probably an intrinsic partof the plume The relatively radiogenic Pb isotopic ratiosand REE patterns distinct from MORB also support thisinterpretationTrace-element characteristics suggest the possible invol-
vement of sub-arc lithospheric mantle in the generation ofthe Karmutsen picrites Lassiter et al (1995) suggested thatmixing of a plume-type source with eNdthorn 6 to thorn 7 witharc material with low NbTh could reproduce the varia-tions observed in the Karmutsen basalts however theyconcluded that the absence of low NbLa ratios in theirsample of tholeiitic basalts restricts the amount of arc litho-sphere involved The study of Lassiter et al (1995) wasbased on major and trace element and Sr Nd and Pb iso-topic data for a suite of 29 samples from Buttle Lake in theStrathcona Provincial Park on central Vancouver Island(Fig 1) Whereas there are no clear HFSE depletions inthe Karmutsen tholeiitic basalts the low NbLa (and TaLa) of the Karmutsen picritic lavas in this study indicatethe possible involvement of Paleozoic arc lithosphere onVancouver Island (Fig 8)
REE modeling dynamic melting andsource mineralogyThe combined isotopic and trace-element variations of thepicritic and tholeiitic Karmutsen lavas indicate that differ-ences in trace elements may have originated from differentmelting histories For example the LuHf ratios of theKarmutsen picritic lavas (mean 008) are considerablyhigher than those in the tholeiitic basalts (mean 002)however the small range of overlapping initial eHf valuesfor the two lava suites suggests that these differences devel-oped during melting within the plume (Fig 9b) Below wemodel the effects of source mineralogy and depth of
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
28
melting on differences in trace-element concentrations inthe tholeiitic basalts and picritesDynamic melting models simulate progressive decom-
pression melting where some of the melt fraction isretained in the residue and only when the degree of partialmelting is greater than the critical mass porosity and the
source becomes permeable is excess melt extracted fromthe residue (Zou1998) For the Karmutsen lavas evolutionof trace element concentrations was simulated using theincongruent dynamic melting model developed by Zou ampReid (2001) Melting of the mantle at pressures greater than1 GPa involves incongruent melting (eg Longhi 2002)
OIB array
PacificMORB(230 Ma)
(0 Ma)
EPR(230 Ma)
-4
-2
0
2
4
8
10
12
14
16
18
20
22
-4 -2 0 2 4 6 8 10 12 14
minus2
0
2
4
6
8
10
12
14
0702 0703 0704 0705 0706
IndianMORB
87Sr 86Sr (initial)
Hf (initial)
(a) (c)
Nd (initial)
Karmutsen tholeiitic basalt
HawaiiOntong JavaCaribbean Plateau
Karmutsen high-MgO lava
high-MgO lavasKarmutsen
1540
1545
1550
1555
1560
1565
175 180 185 190 195 200375
380
385
390
395
175 180 185 190 195 200
(d) (e)
EPR
EPR
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb
Karmutsen Formation (initial)
HawaiiOntong JavaCaribbean Plateau
East Pacific Rise
Karmutsen Formation (measured)
Juan de FucaGorda
ε Nd
(initial)
Karmutsen Formation (different age-correction for 3 picrites)
(0 Ma)
NbY
001
01
1
1 10
ZrY
OIB
NMORB
(b)
Iceland array
6
Fig 13 Comparison of age-corrected (230 Ma) Sr^Nd^Hf^Pb isotopic compositions for Karmutsen flood basalts onVancouver Island withage-corrected OIB and MORB and Nb^Zr^Y variation (a) Initial eNd vs 87Sr86Sr (b) NbYand ZrY variation (after Fitton et al 1997)(c) Initial eHf vs eNd (d) Measured and initial 207Pb204Pb vs 206Pb204Pb (e) Measured and initial 208Pb204Pb vs 206Pb204Pb References fordata sources are listed in Electronic Appendix 4 Most of the compiled data were extracted from the GEOROC database (httpgeorocmpch-mainzgwdgdegeoroc) OIB array line in (c) is fromVervoort et al (1999) EPR East Pacific Rise Several high-MgO samples affected bysecondary alteration were not plotted for clarity in Pb isotope plots Estimation of fields for age-corrected Pacific MORB and EPR at 230 Mawere made using parent^daughter concentrations (in ppm) from Salters amp Stracke (2004) of Lufrac14 0063 Hffrac14 0202 Smfrac14 027 Ndfrac14 071Rbfrac14 0086 and Srfrac14 836 In the NbYvs ZrYplot the normal (N)-MORB and OIB fields not including Icelandic OIB are from data com-pilations of Fitton et al (1997 2003) The OIB field is data from 780 basalts (MgO45wt ) from major ocean islands given by Fitton et al(2003) The N-MORB field is based on data from the Reykjanes Ridge EPR and Southwest Indian Ridge from Fitton et al (1997) The twoparallel gray lines in (b) (Iceland array) are the limits of data for Icelandic rocks
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
29
with olivine being produced during the reactioncpxthornopxthorn spfrac14meltthornol for spinel peridotites The mod-eling used coefficients for incongruent melting reactionsbased on experiments on lherzolite melting (eg Kinzleramp Grove 1992 Longhi 2002) and was separated into (1)melting of spinel lherzolite (2) two combined intervals ofmelting for a garnet lherzolite (gt lherz) and spinel lher-zolite (sp lherz) source (3) melting of garnet lherzoliteThe model solutions are non-unique and a range indegree of melting partition coefficients melt reactioncoefficients and proportion of mixed source componentsis possible to achieve acceptable solutions (see Fig 14 cap-tion for description of modeling parameters anduncertainties)The LREE-depleted high-MgO lavas require melting of
a depleted spinel lherzolite source (Fig 14a) The modelingresults indicate a high degree of melting (22^25) similarto the results from PRIMELT1 based on major-elementvariations (discussed above) of a LREE-depleted sourceThe high-MgO lavas lack a residual garnet signature Theenriched tholeiitic basalts involved melting of both garnetand spinel lherzolite and represent aggregate melts pro-duced from continuous melting throughout the depth ofthe melting column (Fig 14b) Trace-element concentra-tions in the tholeiitic and picritic lavas are not consistentwith low-degree melting of garnet lherzolite alone(Fig 14c) The modeling results indicate that the aggregatemelts that formed the tholeiitic basalts would haveinvolved lesser proportions of enriched small-degreemelts (1^6 melting) generated at depth and greateramounts of depleted high-degree melts (12^25) gener-ated at lower pressure (a ratio of garnet lherzolite tospinel lherzolite melt of 14 provides the best fits seeFig 14b and description in figure caption) The totaldegree of melting for the tholeiitic basalts was high(23^27) similar to the degree of partial melting in thespinel lherzolite facies for the primary melts that eruptedas the Keogh Lake picrites of a source that was also prob-ably depleted in LREE
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
(c)
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
Melting of garnet lherzolite
High-MgO lavas
Melting of spinel lherzolite
source (07DM+03PM)
source (07DM+03PM)
6 melting
30 melting
(b)
(a)
30 melting
Sam
ple
Cho
ndrit
eS
ampl
eC
hond
rite 20 melting
1 melting
Melting of garnet lherzoliteand spinel lherzolite
x gt lherz + 4x sp lherz
5 meltingx gt lherz + 4x sp lherz
Sam
ple
Cho
ndrit
e
Tholeiitic lavas
Tholeiitic lavas
High-MgO lavas
source (07DM+03PM)
Fig 14 Trace-element modeling results for incongruent dynamicmantle melting for picritic and tholeiitic lavas from the KarmutsenFormation Three steps of modeling are shown (a) melting of spinellherzolite compared with high-MgO lavas (b) melting of garnet lher-zolite and spinel lherzolite compared with tholeiitic lavas (c) meltingof garnet lherzolite compared with tholeiitic and picritic lavasShaded fields in each panel for tholeiitic basalts and picrites arelabeled Patterns with symbols in each panel are modeling results(patterns are 5 melting increments in (a) and (b) and 1 incre-ments in (c) PM primitive mantle DM depleted MORB mantle gtlherz garnet lherzolite sp lherz spinel lherzolite In (b) the ratios ofper cent melting for garnet and spinel lherzolite are indicated (x gtlherzthorn 4x sp lherz means that for 5 melting 1 melt in garnetfacies and 4 melt in spinel facies where x is per cent melting in theinterval of modeling) The ratio of the respective melt proportions inthe aggregate melt (x gt lherzthorn 4x sp lherz) was kept constant andthis ratio of melt contribution provided the best fit to the data Arange of proportion of source components (07DMthorn 03PM to
09DMthorn 01PM) can reproduce the variation in the high-MgOlavas Melting modeling uses the formulation of Zou amp Reid (2001)an example calculation is shown in their Appendix PM fromMcDonough amp Sun (1995) and DM from Salters amp Stracke (2004)Melt reaction coefficients of spinel lherzolite from Kinzler amp Grove(1992) and garnet lherzolite fromWalter (1998) Partition coefficientsfrom Salters amp Stracke (2004) and Shaw (2000) were kept constantSource mineralogy for spinel lherzolite (018cpx027opx052ol003sp) from Kinzler (1997) and for garnet lherzolite(034cpx008opx053ol005gt) from Salters amp Stracke (2004) Arange of source compositions for melting of spinel lherzolite(07DMthorn 03PM to 03DMthorn 07PM) reproduce REE patternssimilar to those of the high-MgO lavas with best fits for 22ndash25melting and 07DMthorn 03PM source composition Concentrationsof tholeiitic basalts cannot be reproduced from an entirelyprimitive or depleted source
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
30
The uncertainty in the composition of the source in thespinel lherzolite melting model for the high-MgO lavasprecludes determining whether the source was moredepleted in incompatible trace elements than the source ofthe tholeiitic lavas (see Fig 14 caption for description) It ispossible that early high-pressure low-degree melting left aresidue within parts of the plume (possibly the hotter inte-rior portion) that was depleted in trace elements (egElliott et al 1991) further decompression and high-degreemelting of such depleted regions could then have generatedthe depleted high-MgO Karmutsen lavas Shallow high-degree melting would preferentially sample depletedregions whereas deeper low-degree melting may notsample more refractory depleted regions (eg CaribbeanEscuder-Viruete et al 2007) The picrites lie at the highend of the range of Nd isotopic compositions and havelow NbZr similar to depleted Icelandic basalts (Fittonet al 2003) therefore the picrites may represent the partsof the mantle plume that were the most depleted prior tothe initiation of melting As Fitton et al (2003) suggestedthese depleted melts may only rarely be erupted withoutcombining with more voluminous less depleted melts andthey may be detected only as undifferentiated small-volume flows like the Keogh Lake picrites within theKarmutsen Formation on Vancouver IslandDecompression melting within the mantle plume initiatedwithin the garnet stability field (35^25 GPa) at highmantle potential temperatures (Tp414508C) and pro-ceeded beneath oceanic arc lithosphere within the spinelstability field (25^09 GPa) where more extensivedegrees of melting could occur
Magmatic evolution of Karmutsentholeiitic basaltsPrimary magmas in large igneous provinces leave exten-sive coarse-grained cumulate residues within or below thecrust these mafic and ultramafic plutonic sequences repre-sent a significant proportion of the original magma thatpartially crystallized at depth (eg Farnetani et al 1996)For example seismic and petrological studies of OntongJava (eg Farnetani et al 1996) combined with the use ofMELTS (Ghiorso amp Sack 1995) indicate that the volcanicsequence may represent only 40 of the total magmareaching the Moho Neal et al (1997) estimated that30^45 fractional crystallization took place for theOntong Java magmas and produced cumulate thicknessesof 7^14 km corresponding to 11^22 km of flood basaltsdepending on the presence of pre-existing oceanic crustand the total crustal thickness (25^35 km) Interpretationsof seismic velocity measurements suggest the presence ofpyroxene and gabbroic cumulates 9^16 km thick beneaththe Ontong Java Plateau (eg Hussong et al 1979)Wrangellia is characterized by high crustal velocities in
seismic refraction lines which is indicative of mafic
plutonic rocks extending to depth beneath VancouverIsland (Clowes et al 1995)Wrangellia crust is 25^30 kmthick beneath Vancouver Island and is underlain by astrongly reflective zone of high velocity and density thathas been interpreted as a major shear zone where lowerWrangellia lithosphere was detached (Clowes et al 1995)The vast majority of the Karmutsen flows are evolved tho-leiitic lavas (eg low MgO high FeOMgO) indicating animportant role for partial crystallization of melts at lowpressure and a significant portion of the crust beneathVancouver Island has seismic properties that are consistentwith crystalline residues from partially crystallizedKarmutsen magmas To test the proportion of the primarymagma that fractionated within the crust MELTS(Ghiorso amp Sack 1995) was used to simulate fractionalcrystallization using several estimated primary magmacompositions from the PRIMELT1 modeling results Themajor-element composition of the primary magmas couldnot be estimated for the tholeiitic basalts because of exten-sive plagthorn cpxthornol fractionation and thus the MELTSmodeling of the picrites is used as a proxy for the evolutionof major elements in the volcanic sequence A pressure of 1kbar was used with variable water contents (anhydrousand 02wt H2O) calculated at the quartz^fayalite^magnetite (QFM) oxygen buffer Experimental resultsfrom Ontong Java indicate that most crystallizationoccurs in magma chambers shallower than 6 km deep(52 kbar Sano amp Yamashita 2004) however some crys-tallization may take place at greater depths (3^5 kbarFarnetani et al 1996)A selection of the MELTS results are shown in Fig 15
Olivine and spinel crystallize at high temperature until15^20wt of the liquid mass has fractionated and theresidual magma contains 9^10wt MgO (Fig 15f) At1235^12258C plagioclase begins crystallizing and between1235 and 11908C olivine ceases crystallizing and clinopyr-oxene saturates (Fig 15f) As expected the addition of clin-opyroxene and plagioclase to the crystallization sequencecauses substantial changes in the composition of the resid-ual liquid FeO and TiO2 increase and Al2O3 and CaOdecrease while the decrease in MgO lessens considerably(Fig 15) The near-vertical trends for the Karmutsen tho-leiitic basalts especially for CaO do not match the calcu-lated liquid-line-of-descent very well which misses manyof the high-CaO high-Al2O3 low-FeO basalts A positivecorrelation between Al2O3CaO and SrNd indicates thataccumulation of plagioclase played a major role in the evo-lution of the tholeiitic basalts (Fig 15g) Increasing thecrystallization pressure slightly and the water content ofthe melts does not systematically improve the match ofthe MELTS models to the observed dataThe MELTS results indicate that a significant propor-
tion of the crystallization takes place before plagioclase(15^20 crystallization) and clinopyroxene (35^45
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
31
crystallization) join the crystallization sequence Theseresults suggest that445 crystallization of the primarymagmas occurred (estimated from the residual liquid percent in Fig 15f) and if the original flood basalt stratigra-phy on Vancouver Island was 6 km thick at least anequivalent amount of complementary plutonic rocksshould have crystallized The Karmutsen basalts and theirplutonic residues thus represent a significant addition ofcrust (perhaps 412 km thickness) previously thickenedduring the formation of the underlying Paleozoic arc
Growth of the Wrangellia oceanic plateauon Vancouver IslandThe extrusive part of the Wrangellia oceanic plateau onVancouver Island was constructed in a three-layersequence of mostly tholeiitic basalts with a restrictedrange of composition (Fig 16) The plateau overliesDevonian to Mississippian arc volcanic sequences andmarine sediments of Mississippian to Permian ages thatare intruded by gabbroic rocks related to the flood basaltsThe volcanic stratigraphy of theWrangellia plateau formed
00
05
10
15
20
25
30
35
40
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
16
0 2 4 6 8 10 12 14 16 18 20
10
11
12
13
14
15
16
17
18
19
20
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
0 2 4 6 8 10 12 14 16 18 20
MgO (wt)
0 10 20 30 40 50
percent of total mass
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
) Mineral proportions
Sp (e)
Ol
CpxPlag
PicriteHigh-MgO basaltTholeiitic basalt
TiO2 (wt)
MgO (wt)
Al2O3 (wt) CaO (wt)
FeO (wt)
MgO(a) (b)
(c) (d)
MgO (wt)MgO (wt)
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
)
Residual liquid ()
1220
1190
Ol + Sp
Plag+Ol+Sp Plag+Cpx+Sp
02 wt H2O
no H2O
Residual liquid and crystallizing phases
(f)
0102030405060708090100
plagioclaseaccumulation
plagioclaseaccumulation
plagioclase
accumulatio
n
10
12
14
16
0 10 20 30 40
Al2O3CaO
SrNd
tholeiitic basalts
(g)
Fig 15 Forward fractional crystallization modeling results for major elements from MELTS (Ghiorso amp Sack 1995) compared with picriticand tholeiitic lavas from the Karmutsen Formation (a^d) Major-element compositions of residual liquids compared with Karmutsen lavacompositions For clarity the results are shown only for one starting composition (open circle) using the composition of the estimated primarymagma for the high-MgO lava series from the modeling technique of Herzberg et al (2007) for sample 4723A13 as shown in Fig 12 The esti-mated primary magmas for samples 4723A3 and 93G171 yield similar results A pressure of 1 kbar was used with no H2O and 02wt H2Ocalculated at QFM black line is anhydrous and grey line is 02wt H2OThe end-point for the anhydrous result is at 78 fractionation and11308C and the end-point for the result with 02wt H2O is 75 fractionation and11108C (e) Mineral proportions of crystallized phases fromone MELTrun (sample 4723A13) with no H2O (f) Per cent residual liquid vs temperature labelled with intervals of crystallizing phases (g)Al2O3CaO vs SrNd for the tholeiitic basalts
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
32
in a single eruptive phase (c 231^225 Ma) as an emergentbasalt sequence during a deeper-water shallow-water andsubaerial stage and subsided below sea level relativelysoon after its formation (Fig 16) The deeper-water stagecontains pillowed and massive submarine flows (3000mthick) that built up the submarine volcanic edifice from theseafloor The lowest pillow basalts were emplaced onunconsolidated and lithified fine-grained siliciclastic andcarbonaceous sediments containing Daonella beds(c 235^232 Ma) that are intruded by abundantKarmutsen mafic sills (Fig 16a) The different pillowedand massive submarine flows resulted primarily from dif-ferent effusive rates and local topography Some of themassive submarine flows commonly recognized by radiat-ing columnar jointing represent master tubes for the deliv-ery of lava to distal parts of submarine flow fields Thepillowed flows which commonly have uneven topography
in modern sequences may have served as levees to themassive submarine flows rather than allowing unob-structed spreading and inflation typical of the massive sub-aerial flows that form on shallow slopes in continentalflood basalt provinces (eg Self et al 1997) The rapid suc-cession of flows is evident by the near absence of sedimentbetween submarine flows Minor volumes of picritic pillowlavas occur in the upper part of the pillowed flow sequencein areas of the plateau exposed on northern VancouverIsland (Fig 16 Nixon et al 2008)The growth of the plateau in shallow water led to
emplacement of an increasing proportion of volcaniclastics(pillow breccias and hyaloclastites) that accumulated up to1500m thick in some areas (Fig 16b Nixon et al 2008)Horizons of pillowed flows occur locally within the volca-niclastic stratigraphy Sedimentary structures (eg gradedbedding fluidization structures) and petrographic textures
Late Devonian-Mississippian arc
Daonella beds
Middle Triassic
deeper water
Late Devonian-Mississippian arc
pillowed flows massive submarine
flows
pillow brecciahyaloclastiteshallow water
~4 km
~6 km
Late Triassic limestoneinterflow limestonelenses
Karmutsen Formation
mafic intrusions sediment-sill complex
Karmutsen Formation
(a) (b)
(c) (d)
subaerial flows
Late Devonian-Mississippian arc Late Devonian-Mississippian arc
picritic pillow basalts
Miss to Permian marine sediments
sills
Fig 16 Growth of the Wrangellia oceanic plateau onVancouver Island (a) Deep-water eruptive phase of pillowed and massive submarineflows with almost no intervening sediments onto MiddleTriassic shale containing Daonella beds (c 235^232 Ma) The basement of the plateauis composed of Devonian to Mississippian arc sequences overlain by Mississippian to Permian carbonates and siliciclastic sediments(b) Shallow-water flow emplacement with increasing proportion of volcaniclastics resulted primarily from cooling^contraction granulationPicritic pillow basalts occur primarily near the transition between pillowed submarine flows and volcaniclastic units Pillowed and massiveflows occur locally within the volcaniclastic unit (c) Emergent stage of growth consists of overlapping subaerial flow fields similar to continen-tal flood basalts with pahoehoe flow structures Multi-tiered columnar jointing occurrs locally in response to ponding of flows (eg Bondre et al2004) Interflow carbonate deposits and plagioclase megacrystic flows developed locally during the waning stage of volcanism (d) Post-volcanicsubsidence and deposition of Late Carnian to Norian (c 212^223 Ma) platform carbonates on top of the plateau Units are labeled in the dia-grams Thicknesses of the underlying units are not to scale
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
33
within the volcaniclastic rocks indicate formation primar-ily via cooling^contraction granulation magma^water^steam interaction autobrecciation resedimentation andmass-wasting rather than pyroclastic eruption (Nixonet al 2008)A broad subaerial platform was constructed by emplace-
ment of subaerial lava flows during an emergent subaerialstage (Fig 16c) The subaerial unit predominantly com-prises well-layered sheet flows that formed in a similarway to continental flood basalts as hundred(s) of pahoehoeflow fields erupted on shallow slopes (Self et al 1997) how-ever volcaniclastic and pillowed flow units are locally pre-sent mostly in the upper parts of the stratigraphyLocalized ponding of flows led to the formation of multi-tiered columnar jointing in some subaerial flows (egBondre et al 2004) As volcanism waned local interflowcarbonate deposits developed and plagioclase megacrysticflows were erupted Subsidence of the plateau is recordedby the interflow lenses and deposits of platform carbonates(40^1500m thick) on top of the Karmutsen basalts thatcontain Late Carnian to Early Norian fossils (Fig 16d c216^227 Ma Nixon et al 2008)The evolution of the internal architecture of the
Wrangellia oceanic plateau was primarily related to theconditions of emplacement (ie water depth and eruptionenvironment) There does not appear to have been a sys-tematic coincident evolution or shift of lava compositionsonVancouver Island as has been recognized in most con-tinental flood basalts and parts of the Wrangellia oceanicplateau in Alaska and Yukon (Greene et al 2008 2009)In general with the exception of the picritic pillow lavasin the submarine stratigraphy and local variations betweenhigh-MgO and tholeiitic lavas there are no significantcompositional differences between lavas erupted duringthe different stages of growth of the plateau (ie submarineand subaerial flows) nor are substantial compositional dif-ferences generally found between massive submarine flowsin contact with pillowed flows tholeiitic basalt flows nearthe base of the Karmutsen Formation are compositionallysimilar to flows sampled near the top of the volcanicsequence Where it was possible to sample continuous sec-tions of volcanic stratigraphy greater than several hun-dreds of meters (eg Keogh^Maynard Lake area west ofthe Karmutsen Range Fig 2) there is no clear relationshipbetween stratigraphic position and compositionalvariation
CONCLUSIONSThe Karmutsen Formation covers extensive areas ofVancouver Island (20 000 km2) and forms part of a largeoceanic plateau erupted over a geologically short interval(c 225^231 Ma) The tripartite volcanic stratigraphyexposed on Vancouver Island is at least 6 km thick andcomprises a succession of submarine flows volcaniclastic
deposits and massive subaerial flowsThese volcanologicaldifferences are primarily related to the eruption environ-ment (deep-water shallow-water or subaerial) The rapidgrowth of the plateau prevented significant accumulationsof intervening sediment except in the uppermost stratigra-phy where isolated limestone lenses commonly associatedwith pillow basalt and hyaloclastite deposits preserve arecord of the final stages of subsidence of the plateauThe Wrangellia plateau on Vancouver Island was con-
structed dominantly of tholeiitic basalt with restrictedmajor- and trace-element and isotopic compositional var-iations Minor volumes of picritic pillow basalts eruptedlate in the submarine phase The Karmutsen basalts wereemplaced onto older Paleozoic arc volcanic sequences andlow NbLa for the picrites may indicate a contributionfrom the sub-arc lithospheric mantle Modeling of the pet-rogenesis of the Karmutsen picrites indicates melting ofanomalously hot mantle (15008C) and extensive degreesof partial melting (23^27) and is consistent with aplume initiation model Combined Sr^Nd^Hf^Pb isotopicsystematics indicate that the picritic and tholeiitic lavaswere derived from a depleted plume-type mantle sourcedistinct from the source of MORB that has compositionalsimilarities to the source of the lavas forming theCaribbean Plateau Picrites were generated by melting ofspinel lherzolite whereas the tholeiitic basalts formed asaggregate melts of both garnet and spinel lherzolite Thetholeiitic basalts underwent extensive low-pressure fractio-nation (52^3 kbar) and seismic studies indicate that someof theWrangellia crust beneathVancouver Island may cor-respond to the plutonic residues of the Karmutsen floodbasalts
ACKNOWLEDGEMENTSWe are grateful to Nick Arndt for helping us get this pro-ject started and Nick Massey for insights into VancouverIsland geology We greatly appreciate the assistance withPRIMELT modeling from Claude Herzberg We thankMikkel Schau for his insight and enthusiasm during field-work Jane Barling assisted with analyses by MC-ICP-MSThoughtful and thorough reviews of the manuscript byJohn Mahoney Fred Frey Wendy Bohrson DougalJerram and Teal Riley are greatly appreciated Fundingwas generously provided by the Rocks to Riches Programadministered by the BC amp Yukon Chamber of Mines in2004 by the BC Geological Survey and by NSERCDiscovery Grants to JSS and DW ARG was supportedby a University Graduate Fellowship at UBC
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
34
REFERENCESBarker F Brown A S Budahn J R amp Plafker G (1989) Back-arc
with frontal-arc component origin of Triassic Karmutsen basaltBritish Columbia Canada Chemical Geology 75 81^102
Blichert-Toft JWeis D Maerschalk C Agranier A amp Albare de F(2003) Hawaiian hot spot dynamics as inferred from the Hf and Pbisotope evolution of Mauna Kea volcano Geochemistry GeophysicsGeosystems 4(2) 1^27 doi1010292002GC000340
Bondre N R Duraiswami R A amp Dole G (2004) Morphologyand emplacement of flows from the Deccan Volcanic ProvinceIndia Bulletin of Volcanology 66 29^45
Carlisle D (1963) Pillow breccias and their aquagene tuffs QuadraIsland British Columbia Journal of Geology 71 48^71
Carlisle D (1972) Late Paleozoic to Mid-Triassic sedimentary-volcanic sequence on Northeastern Vancouver Island Geological
Society of Canada Paper Report of Activities 72-1 Part B 24^30Carlisle D amp SuzukiT (1974) Emergent basalt and submergent car-
bonate^clastic sequences including the Upper Triassic Dilleri andWelleri zones on Vancouver Island Canadian Journal of Earth
Sciences 11 254^279Chauvel C amp Blichert-Toft J (2001) A hafnium isotope and trace
element perspective on melting of the depleted mantle Earth and
Planetary Science Letters 190 137^151Chayes F (1954)The theory of thin-section analysis Journal of Geology
62 92^101Cho M Liou J G amp Maruyama S (1987) Transition from the zeo-
lite to prehnite^pumpellyite facies in the Karmutsen metabasitesVancouver Island British Columbia Journal of Petrology 27(2)467^494
Clowes R M Zelt C A Amor J R amp Ellis R M (1995)Lithospheric structure in the southern Canadian Cordillera froma network of seismic refraction lines Canadian Journal of Earth
Sciences 32(10) 1485^1513Coffin M F amp Eldholm O (1994) Large igneous provinces Crustal
structure dimensions and external consequences Reviews of
Geophysics 32(1) 1^36Cox K G (1980) A model for flood basalt volcanism Journal of
Petrology 21 629^650Eldholm O amp Coffin M F (2000) Large igneous provinces
and plate tectonics In Richards M A Gordon R G amp vander Hilst R D (eds) The History and Dynamics of Global
Plate Motions American Geophysical Union Geophysical Monograph 121309^326
Elliott T R Hawkesworth C J amp Grolaquo nvold K (1991) Dynamicmelting of the Iceland plume Nature 351 201^206
Escuder-Viruete J Pecurren rez-Estaucurren n A Contreras F Joubert MWeis D Ullrich T D amp Spadea P (2007) Plume mantle sourceheterogeneity through time Insights from the Duarte ComplexHispaniola northeastern Caribbean Journal of Geophysical Research112(B04203) doi1010292006JB004323
Farnetani C G Richards M A amp Ghiorso M S (1996)Petrological models of magma evolution and deep crustal structurebeneath hotspots and flood basalt provinces Earth and Planetary
Science Letters 143 81^94Fitton J G amp Godard M (2004) Origin and evolution of magmas
on the Ontong Java Plateau In Fitton J G Mahoney J JWallace P J amp Saunders A D (eds) Origin and Evolution of the
Ontong Java Plateau Geological Society London Special Publications 229151^178
Fitton J G Saunders A D Norry M J Hardarson B S ampTaylor R N (1997) Thermal and chemical structure of theIceland plume Earth and Planetary Science Letters 153 197^208
Fitton J G Saunders A D Kempton P D amp Hardarson B S(2003) Does depleted mantle form an intrinsic part of the Icelandplume Geochemistry Geophysics Geosystems 4(3) 1032 doi1010292002GC000424
Frey F A Huang S Blichert-Toft J Regelous M amp Boyet M(2005) Origin of depleted components in basalt related to theHawaiian hotspot Evidence from isotopic and incompatible ele-ment ratios Geochemistry Geophysics Geosystems 6(1) 23 doi1010292004GC000757
Galer S J G amp Abouchami W (1998) Practical application of leadtriple spiking for correction of instrumental mass discriminationMineralogical Magazine 62A 491^492
Ghiorso M S amp Sack R O (1995) Chemical mass transfer in mag-matic processes IV A revised and internally consistent thermody-namic model for the interpolation and extrapolation of liquid^solidequilibria in magmatic systems at elevated temperatures and pres-sures Contributions to Mineralogy and Petrology 119 197^212
Greene A R Scoates J S amp Weis D (2005)WrangelliaTerrane onVancouver Island Distribution of flood basalts with implicationsfor potential Ni^Cu^PGE mineralization In Grant B (ed)Geological Fieldwork 2004 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2005-1 209^220Greene A R Scoates J S Nixon G T amp Weis D (2006) Picritic
lavas and basal sills in the Karmutsen flood basalt provinceWrangellia northern Vancouver Island In Grant B (ed)Geological Fieldwork 2005 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2006-1 39^52Greene A R Scoates J S amp Weis D (2008) Wrangellia flood
basalts in Alaska A record of plume^lithosphere interaction in aLate Triassic accreted oceanic plateau Geochemistry Geophysics
Geosystems 9(Q12004) doi1010292008GC002092Greene A R Scoates J S Weis D amp Israel S (2009)
Geochemistry of triassic flood basalts from the Yukon (Canada)segment of the accreted Wrangellia oceanic plateau Lithosdoi101016jlithos200811010
Hauff F Hoernle K Tilton G Graham D W amp Kerr A C(2000) Large volume recycling of oceanic lithosphere over shorttime scales geochemical constraints from the Caribbean LargeIgneous Province Earth and Planetary Science Letters 174 247^263
Hauff F Hoernle K amp Schmidt A (2003) Sr^Nd^Pb compositionof Mesozoic Pacific oceanic crust (Site 1149 and 801 ODP Leg185)Implications for alteration of ocean crust and the input into theIzu^Bonin^Mariana subduction system Geochemistry Geophysics
Geosystems 4(8) doi1010292002GC000421Herzberg C (2004) Partial melting below the Ontong Java Plateau
In Fitton J G Mahoney J J Wallace P J amp Saunders A D(eds) Origin and Evolution of the Ontong Java Plateau Geological SocietyLondon Special Publications 229 179^183
Herzberg C (2006) Petrology and thermal structure of the Hawaiianplume from Mauna Kea volcano Nature 444(7119) 605
Herzberg C amp Asimow P D (2008) Petrology of some oceanicisland basalts PRIMELT2XLS software for primary magma cal-culation Geochemistry Geophysics Geosystems 9(Q09001) doi1010292008GC002057
Herzberg C amp OrsquoHara M J (2002) Plume-associated ultramaficmagmas of Phanerozoic age Journal of Petrology 43(10) 1857^1883
Herzberg C Asimow P D Arndt N Niu Y Lesher C MFitton J G Cheadle M J amp Saunders A D (2007)Temperatures in ambient mantle and plumes Constraints frombasalts picrites and komatiites Geochemistry Geophysics Geosystems8(Q02006) doi1010292006GC001390
Huang S amp Frey F A (2005) Trace element abundances of MaunaKea basalt from phase 2 of the Hawaii Scientific Drilling Project
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
35
Petrogenetic implications of correlations with major element con-tent and isotopic ratios Geochemistry Geophysics Geosystems 4(6)doi1010292002GC000322
Hussong D M Wipperman L K amp Kroenke L W (1979) Thecrustal structure of the Ontong Java and Manihiki OceanicPlateaus Journal of Geophysical Research 84(B11) 6003^6010
Jerram D A amp Widdowson M (2005) The anatomy of ContinentalFlood Basalt Provinces geological constraints on the processes andproducts of flood volcanism Lithos 79(3^4) 385^405
Jones D L Silberling N J amp Hillhouse J (1977)Wrangellia a dis-placed terrane in northwestern North America CanadianJournal ofEarth Sciences 14(11) 2565^2577
Kerr A C (2003) Oceanic Plateaus In Rudnick R (ed) The Crust
Treatise on GeochemistryVol 3 Oxford Elsevier Science pp 537^565Kerr A C amp Mahoney J J (2007) Oceanic plateaus Problematic
plumes potential paradigms Chemical Geology 241 332^353Kerr A CTarney J Marriner G F Klaver GT Saunders A D
amp Thirlwall M F (1996) The geochemistry and petrogenesis ofthe Late-Cretaceous picrites and basalts of Curacao NetherlandsAntilles a remnant of an oceanic plateau Contributions to
Mineralogy and Petrology 124(1) 29^43Kerr A C Marriner G F Tarney J Nivia A Saunders A D
Thirlwall M F amp Sinton A C (1997) Cretaceous basaltic ter-ranes in Western Colombia Elemental chronological and Sr-Ndisotopic constraints on petrogenesis Journal of Petrology 38(6)677^702
Kinzler R J (1997) Melting of mantle peridotite at pressuresapproaching the spinel to garnet transition Application to mid-ocean ridge basalt petrogenesis Journal of Geophysical Research
102(B1) 853^874Kinzler R J amp Grove T L (1992) Primary magmas of mid-ocean
ridge basalts 1 Experiments and methods Journal of Geophysical
Research 97(B5) 6885^6906Korenaga J (2005) Why did not the Ontong Java Plateau form sub-
aerially Earth and Planetary Science Letters 234 385^399Kuniyoshi S (1972) Petrology of the Karmutsen (Volcanic) Group
northeastern Vancouver Island British Columbia PhD disserta-tion University of California Los Angeles 242 pp
Lassiter J C DePaolo D J amp Mahoney J J (1995) Geochemistry ofthe Wrangellia flood basalt province Implications for the role ofcontinental and oceanic lithosphere in flood basalt genesis Journalof Petrology 36(4) 983^1009
LeBas M J (2000) IUGS reclassification of the high-Mg and picriticvolcanic rocks Journal of Petrology 41(10) 1467^1470
Longhi J (2002) Some phase equilibrium systematics of lherzolitemelting I Geochemistry Geophysics Geosystems 3(3) doi1010292001GC000204
MacDonald G A amp Katsura T (1964) Chemical composition ofHawaiian lavas Journal of Petrology 5 82^133
Mahoney J J (1987) An isotopic survey of Pacific oceanic plateausimplications for their nature and origin In Keating B HFryer P Batiza R amp Boehlert GW (eds) Seamounts Islands andAtolls American Geophysical Union Geophysical Monograph 43 207^220
Mahoney J LeRoex A P Peng Z Fisher R L amp Natland J H(1992) Southwestern limits of Indian Ocean Ridge mantle and theorigin of low 206Pb204Pb mid-ocean ridge basalt isotope systema-tics of the central Southwest Indian Ridge (178^508E) Journal ofGeophysical Research 97(B13) 19771^19790
Mahoney J J Sinton J M Kurz M D MacDougall J DSpencer K J amp Lugmair GW (1994) Isotope and trace elementcharacteristics of a super-fast spreading ridge East Pacific rise13^238S Earth and Planetary Science Letters 121 173^193
Massey N W D (1995a) Geology and mineral resources of theAlberni^Nanaimo Lakes sheet Vancouver Island 92F1W 92F2Eand part of 92F7E British Columbia Ministry of Energy Mines and
Petroleum Resources Paper 1992-2 132 ppMassey N W D (1995b) Geology and mineral resources of the
Cowichan Lake sheet Vancouver Island 92C16 British Columbia
Ministry of Energy Mines and Petroleum Resources Paper 1992-3 112 ppMassey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005a) Digital Geology Map of British Columbia Tile NM9 MidCoast BC British Columbia Ministry of Energy and Mines Geofile
2005-2Massey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005b) Digital Geology Map of British Columbia Tile NM10Southwest BC British Columbia Ministry of Energy and Mines Geofile
2005-3McDonough W F amp Sun S (1995) The composition of the Earth
Chemical Geology 120 223^253Monger JW H amp Journeay J M (1994) Basement geology and tec-
tonic evolution of theVancouver region In Monger JW H (ed)Geology and Geological Hazards of the Vancouver Region Southwestern
British Columbia Geological Survey of Canada Bulletin 481 3^25Muller J E (1977) Geology of Vancouver Island Geological Survey of
Canada Open File Map 463Muller J E (1980)The Paleozoic Sicker Group of Vancouver Island British
Columbia Geological Survey of Canada Paper 79-30 22 ppMuller J E Northcote K E amp Carlisle D (1974) Geology and
mineral deposits of Alert Bay^Cape Scott map area VancouverIsland British Columbia Geological Survey of Canada Paper 74-8 77
Neal C R Mahoney J J Kroenke L W Duncan R A ampPetterson M G (1997) The Ontong^Java Plateau InMahoney J J amp Coffin M F (eds) Large Igneous Provinces
Continental Oceanic and Planetary FloodVolcanism American Geophysical
Union Geophysical Monograph 100 183^216Niu Y Collerson K D Batiza R Wendt J I amp Regelous M
(1999) Origin of enriched-typed mid-ocean ridge basalt at ridgesfar from mantle plumes The East Pacific Rise at 11820rsquoN Journalof Geophysical Research 104 7067^7088
Nixon G T amp Orr A J (2007) Recent revisions to the EarlyMesozoic stratigraphy of Northern Vancouver Island (NTS 102I092L) and metallogenic implicatinos British Columbia InGrant B (ed) Geological Fieldwork 2006 British Columbia Ministry of
Energy Mines and Petroleum Resources Paper 2007-1 163^177Nixon GT Hammack J L KoyanagiV M Payie G J Haggart
JW Orchard M JTozerT Friedman R M Archibald D APalfy J amp Cordey F (2006a) Geology of the Quatsino^PortMcNeill area northernVancouver Island British Columbia Ministry
of Energy Mines and Petroleum Resources Geoscience Map 2006-2 scale150 000
Nixon G T Kelman M C Stevenson D Stokes L A ampJohnston K A (2006b) Preliminary geology of the Nimpkishmap area (NTS 092L07) northern Vancouver Island BritishColumbia In Grant B amp Newell J M (eds) Geological Fieldwork2005 British Columbia Ministry of Energy Mines and Petroleum Resources
Paper 2006-1 135^152Nixon G T Laroque J Pals A Styan J Greene A R amp
Scoates J S (2008) High-Mg lavas in the Karmutsen flood basaltsnorthernVancouver Island (NTS 092L) Stratigraphic setting andmetallogenic significance In Grant B (ed) Geological Fieldwork
2007 British Columbia Ministry of Energy Mines and Petroleum
Resources Paper 2008-1 175^190Nowell G M Kempton P D Noble S R Fitton J G
Saunders A Mahoney J J amp Taylor R N (1998) High precisionHf isotope measurements of MORB and OIB by thermal ionisation
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
36
mass spectrometry insights into the depleted mantle Chemical
Geology 149 211^233Parrish R R amp McNicoll V J (1992) U^Pb age determinations
from the southern Vancouver Island area British Columbia InRadiogenic Age and Isotopic Studies Report 5 Geological Survey of
Canada Paper 91-2 79^86Peate DW (1997) The Parana^Etendeka Province In Coffin M F
amp Mahoney J (eds) Large Igneous Provinces Continental Oceanic andPlanetary Flood Volcanism American Geophysical Union Geophysical
Monograph 100 217^245Perfit M R Fornari D J Ridley W I Kirk P D Casey J
Kastens K A Reynolds J R Edwards M Desonie DShuster R amp Paradis S (1996) Recent volcanism in the Siqueirostransform fault picritic basalts and implications for MORBmagma genesis Earth and Planetary Science Letters 141(1^4) 91^108
Pretorius W Weis D Williams G Hanano D Kieffer B ampScoates J S (2006) Complete trace elemental characterization ofgranitoid (USGSG-2GSP-2) reference materials by high resolutioninductively coupled plasma-mass spectrometry Geostandards and
Geoanalytical Research 30(1) 39^54Regelous M Niu Y Wendt J I Batiza R Greig A amp
Collerson K D (1999) Variations in the geochemistry of magma-tism on the East Pacific Rise at 10830rsquoN since 800 ka Earth and
Planetary Science Letters 168 45^63Recurren villon S Chauvel C Arndt N T Pik R Martineau F
Fourcade S amp Marty B (2002) Heterogeneity of the Caribbeanplateau mantle source Sr O and He isotopic compositions of oli-vine and clinopyroxene from Gorgona Island Earth and Planetary
Science Letters 205(1^2) 91Rhodes J M (1996) Geochemical stratigraphy of lava flows samples
by the Hawaii Scientific Drilling Project Journal of Geophysical
Research 101(B5) 11729^11746Rhodes J M amp Vollinger M J (2004) Composition of basaltic lavas
sampled by phase-2 of the Hawaii Scientific Drilling ProjectGeochemical stratigraphy and magma types Geochemistry
Geophysics Geosystems 5(3) doi1010292002GC000434Richards M A Jones D L Duncan R A amp DePaolo D J (1991)
A mantle plume initiation model for the Wrangellia flood basaltand other oceanic plateaus Science 254 263^267
Salters V J M amp Hart S R (1991) The mantle sources of oceanridges islands and arcs the Hf-isotope connection Earth and
Planetary Science Letters 104 364^380Salters V J M amp Stracke A (2004) Composition of the depleted
SaltersV J amp WhiteW (1998) Hf isotope constraints on mantle evo-lution Chemical Geology 145 447^460
SanoT amp Yamashita S (2004) Experimental petrology of basementlavas from Ocean Drilling Program Leg 192 implications for dif-ferentiation processes in Ontong Java Plateau magmas InFitton J G Mahoney J JWallace P J amp Saunders A D (eds)Origin and Evolution of the Ontong Java Plateau Geological Society
London Special Publications 229 185^218Saunders A D (2005) Large igneous provinces Origin and environ-
mental consequences Elements 1 259^263Self S ThordarsonT amp Keszthely L (1997) Emplacement of conti-
nental flood basalt lava flows In Mahoney J J amp Coffin M F(eds) Large Igneous Provinces Continental Oceanic and Planetary Flood
Volcanism American Geophysical Union Geophysical Monograph 100381^410
Shaw D M (2000) Continuous (dynamic) melting theory revisitedCanadian Mineralogist 38(5) 1041^1063
Sluggett C L (2003) Uranium^lead age and geochemical constraintson Paleozoic and Early Mesozoic magmatism in WrangelliaTerrane Saltspring Island British Columbia BSc thesisUniversity of British ColumbiaVancouver 84 pp
Storey M Mahoney J J amp Saunders A D (1997) Cretaceousbasalts in Madagascar and the transition between plume and con-tinental lithospheric mantle sources In Mahoney J J ampCoffin M F (eds) Large Igneous Provinces Continental Oceanic and
Planetary Flood Volcanism Aamerican Geophysical Union Geophysical
Monograph 100 95^122Surdam R C (1967) Low-grade metamorphism of the Karmutsen
Group PhD dissertation University of California Los Angeles288 pp
Sutherland-Brown A Yorath C J Anderson R G amp Dom K(1986) Geological maps of southern Vancouver IslandLITHOPROBE I 92C10 11 14 16 92F1 2 7 8 Geological Survey ofCanada Open File 1272
Tejada M L G Mahoney J J Castillo P R Ingle S P Sheth HC amp Weis D (2004) Pin-pricking the elephant evidence on theorigin of the Ontong Java Plateau from Pb^Sr^Hf^Nd isotopiccharacteristics of ODP Leg 192 basalts In Fitton J GMahoney J J Wallace P J amp Saunders A D (eds) Origin and
Evolution of the Ontong Java Plateau Geological Society London Special
Publications 229 133^150Thompson P M E Kempton P D amp Kerr A C (2008) Evaluation
of the effects of alteration and leaching on Sm^Nd and Lu^Hf sys-tematics in submarine mafic rocks Lithos 104(1^4) 164^176
Thompson P M E Kempton P D White R V Kerr A CTarney J Saunders A D Fitton J G amp McBirney A (2003)Hf-Nd isotope constraints on the origin of the CretaceousCaribbean plateau and its relationship to the Galapagos plumeEarth and Planetary Science Letters 217(1^2) 59^75
Vervoort J D Patchett P Blichert-Toft J amp Albare de F (1999)Relationships between Lu^Hf and Sm^Nd isotopic systems in theglobal sedimentary system Earth and Planetary Science Letters 16879^99
Walter M J (1998) Melting of garnet peridotite and the origin ofkomatiite and depleted lithosphere Journal of Petrology 39(1) 29^60
Weis D Kieffer B Maerschalk C Barling J de Jong JWilliams G A Hanano D Mattielli N Scoates J SGoolaerts A Friedman R A amp Mahoney J B (2006) High-pre-cision isotopic characterization of USGS reference materials byTIMS and MC-ICP-MS Geochemistry Geophysics Geosystems
7(Q08006) doi1010292006GC001283Weis D Kieffer B Hanano D Silva I N Barling J
Pretorius W Maerschalk C amp Mattielli N (2007) Hf isotopecompositions of US Geological Survey reference materialsGeochemistry Geophysics Geosystems 8(Q06006) doi1010292006GC001473
Wheeler J O amp McFeely P (1991) Tectonic assemblage map of theCanadian Cordillera and adjacent part of the United States ofAmerica Geological Survey of Canada Map 1712A
WhiteW M Albare de F amp Tecurren louk P (2000) High-precision analy-sis of Pb isotope ratios by multi-collector ICP-MS Chemical Geology167 257^270
Yorath C J Sutherland Brown A amp Massey N W D (1999)LITHOPROBE southern Vancouver Island British Columbia Geological
Survey of Canada Bulletin 498 145 ppZou H (1998) Trace element fractionation during modal and
non-model dynamic melting and open-system melting Amathematical treatment Geochimica et Cosmochimica Acta 62(11)1937^1945
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
37
Zou H amp Reid M R (2001) Quantitative modeling of trace elementfractionation during incongruent dynamic melting Geochimica et
Cosmochimica Acta 65(1) 153^162
APPENDIX SAMPLEPREPARAT ION ANDANALYT ICAL METHODSOnly the freshest rocks in the field were sampled and onlythe least altered samples were selected for chemical andisotopic analysis based on thorough petrographic inspec-tion Fifty-nine of the 129 collected samples were crushed(400 g) into small pieces 52mm in diameter in aRocklabs hydraulic piston crusher between WC plates tominimize contamination The coarse-crush was mixedand 100 g was powdered in a planetary mill using agatejars and balls cleaned with quartz sand between samples
ActLabs analytical methodsThe major- and trace-element compositions of the whole-rock powders were determined at Activation LaboratoriesLtd (Actlabs) in Ancaster Ontario Analytical techniquesand detection limits are also available from Actlabs (httpwwwactlabscommethsub_code4erehtm) The analyticalmethod for each of the elements analyzed is indicated inTable 2 For the major elements a 02 g sample was mixedwith a lithium metaborate^lithium tetraborate mixtureand fused in a graphite crucible The molten mixture waspoured into a 5 HNO3 solution and shaken until dis-solved (30min) The samples were analyzed for majoroxides and selected trace elements by inductively coupledplasma optical emission spectrometry (ICP-OES) on acombination simultaneous^sequential Thermo Jarrell-AshEnviro II system Internal calibration was achieved usinga variety of international reference materials (eg W-2BIR-1 DNC-1) and independent control samplesAdditional trace elements were analyzed by both instru-mental neutron activation analysis (INAA) and induc-tively couple plasma mass spectrometry (ICP-MS) ForINAA15^25 g of sample was weighed into small polyeth-ylene vials and irradiated with a control international ref-erence material CANMET WMS-1 and NiCr flux wires ata thermal neutron flux of 71012 neutronscm2 per s in theMcMaster Nuclear Reactor Following a 7 day waitingperiod the samples were measured on an Ortec high-purity Ge detector linked to a Canberra Series 95 multi-channel analyzer Activities for each element weredecay- and weight-corrected and compared with a detectorcalibration developed from multiple international certifiedreference materials For ICP-MS 025 g of sample wasdigested in HF followed by HNO3^HClO4 heated andtaken to dryness The samples were brought back into solu-tion with HCl Samples were analyzed using a PerkinElmer Optima 3000 ICP In-lab standards or certified ref-erence materials (eg W-2 BIR-1 DNC-1) were used for
quality control A total of 15 blind duplicates were ana-lyzed from the same sample powders to assessreproducibility
University of Massachusetts analyticalmethodsFourteen sample duplicate powders taken from the same100 g sample aliquot (high-MgO basalts and picrites)were also analyzed at the Ronald B Gilmore X-RayFluorescence (XRF) Laboratory at the University ofMassachusetts Major elements were measured on a fusedLa-bearing lithium borate glass disc using a SiemensMRS-400 spectrometer Trace element concentrations(Rb Sr Ba Ce Nb Zr Y Pb Zn Ga Ni Cr V) weremeasured on a separate powder pellet using a PhilipsPW2400 sequential spectrometer using a Rh tube LOIand ferrous iron measurements were made as describedby Rhodes amp Vollinger (2004) Precision and accuracy esti-mates for the data have been described by Rhodes (1996)and Rhodes amp Vollinger (2004) Results for each sampleare the average of two separate analyses and are shown inElectronic Appendix 2 for a comparison of analyses ofduplicate powders using XRFand ICP-MS
PCIGR analytical methodsA subset of 19 samples was selected for high-precisiontrace-element analysis and Sr Nd Pb and Hf isotopicanalysis at the Pacific Centre for Isotopic andGeochemical Research (PCIGR) at the University ofBritish Columbia (UBC) Samples were selected from the59 samples analyzed for whole-rock chemistry at ActLabsbased on major- and trace-element chemistry alteration(low LOI and petrographic alteration index) and samplelocation Samples were prepared for trace-element analysisat the PCIGR by the technique described by Pretorius et al(2006) on unleached rock powders Sample powders(100mg) were weighed in 7ml screw-top Savillex bea-kers and dissolved in 1ml 14N HNO3 and 5ml 48 HFon a hotplate for 48 h at 1308C with periodic ultrasonica-tion Samples were dried and redissolved in 6ml 6N HClon a hotplate for 24 h and then dried and redissolved in1ml concentrated HNO3 for 24 h before final dryingTrace element abundances were measured by high-resolution inductively coupled plasma-mass spectrometry(HR-ICP-MS) with aThermo Finnigan Element2 systemfollowing the procedures described by Pretorius et al(2006) within 24 h of redissolution HFSE and LILE weremeasured in medium-resolution mode at 2000 dilutionusing a PFATeflon spray chamber washed with aqua regiafor 3min between samples REE were measured in high-resolution mode and U and Pb in low-resolution modeat 2000 dilution using a glass spray chamber washedwith 2 HNO3 between samples Total procedural blanksand reference materials (BCR-2 BHVO-2) were analyzedwith the batch of samples Indium was used as an internal
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
38
standard in all samples and standard solutionsBackground and standard solutions were analyzed afterevery five samples to detect memory effects and massdrift Results for PCIGR trace-element analyses areshown in Electronic Appendix 3 for the full set of analyzedtrace elementsSample digestion for purification of Sr Nd Hf and Pb
for column chemistry began by weighing each samplepowder (400^500mg) prior to leaching All samples wereleached with 6N HCl and placed in an ultrasonic bath for15min Samples were rinsed twice with 18 M-cm H2Obetween each leaching step (15 total) until the supernatantliquid was clear [following the method of Mahoney (1987)]Samples were then dried on a hotplate for 24 h andweighed again Sample solutions were then prepared bydissolving 100^250mg of the leached powder dissolvedin 1ml 14N HNO3 and 10ml 48 HF on a hotplate for48 h at 1308C with periodic ultrasonication Samples weredried and redissolved in 6ml 6N HCl on a hotplate for24 h and then dried Pb was separated using anion exhangecolumns and the discard was used for Sr REE and Hfseparation Nd was separated from the other REE and Hfrequired two additional purification steps Detailed proce-dures for column chemistry for separating Sr Nd and Pbat the PCIGR have been described byWeis et al (2006) andHf purification has been described byWeis et al (2007) Srand Nd isotope ratios were measured by thermal ioniza-tion mass spectrometry (TIMS) on a Thermo FinniganTriton system in static mode with relay matrix rotation ona single Ta and double Re^Ta filament respectively Fourto five filaments per barrel of 21 were occupied by stan-dards (NBS 987 for Sr and LaJolla for Nd) for each barrelwhere samples were run Sample Sr and Nd isotopic com-positions were corrected for mass fractionation using86Sr88Srfrac14 01194 and 146Nd144Ndfrac14 07219 Each samplewas then normalized using the barrel average of the refer-ence material relative to the values of143Nd144Ndfrac14 0511858 and 87Sr86Srfrac14 0710248 (Weiset al 2006) During the course of the Vancouver Islandanalyses the La Jolla Nd standard gave an average valueof 05118566 (nfrac14 7) and NBS987 standard gave an aver-age of 07102408 (nfrac1411) (2 error is reported as times106) 147Sm144Nd errors are 15 or 0006 USGeological Survey (USGS) reference material BHVO-2was processed with the samples and yielded Sr and Nd iso-topic ratios of 07034607 and 05129786 respectivelyThese are in agreement with the published values of070347920 and 051298411 respectively (Weis et al2006)
Pb and Hf isotopic compositions were analyzed by mul-tiple collector-inductively coupled plasma-mass spectrome-try (MC-ICP-MS) on a static multi-collection on a NuPlasma (Nu Instruments) system The detailed analyticalprocedure for Pb isotopic analyses on the Nu Plasma atthe PCIGR has been described by Weis et al (2006) Theconfiguration for Pb analyses allows for collection of PbTl and Hg togetherTl and Hg are used to monitor instru-mental mass discrimination and isobaric overlap repec-tively All sample solutions were analyzed withapproximately the same PbTl ratio (4) as the referencematerial NIST SRM 981 To accomplish this a small ali-quot of each sample solution from the Pb columns was ana-lyzed on the Element2 to determine the precise amount ofPb available for analysis on the Nu Plasma The SRM 981standard was run after every two samples on the NuPlasma During the time samples were run analyses ofthe SRM 981 Pb reference material gave values of206Pb204Pbfrac1416940322 207Pb204Pbfrac1415495823and 208Pb204Pbfrac14 36713164 (nfrac14 61 2 error is reportedas times 104) these values are in excellent agreement withreported TIMS triple-spike values of Galer amp Abouchami(1998) Fractionation-corrected Pb isotopic ratios were fur-ther corrected by the sample-standard bracketing methodor the ln^ln correction method described by White et al(2000) and Blichert-Toft et al (2003) Leached powders ofUSGS reference material BHVO-2 yielded Pb isotopicratios of 206Pb204Pbfrac14186454 8 207Pb204Pbfrac14 1549105 and 208Pb204Pbfrac14 38222514 and BCR-2yielded 206Pb204Pbfrac141880466 207Pb204Pbfrac14 1562518 and 208Pb204Pbfrac14 3883496 (2 error isreported as times 104) These values are in agreement withthose for leached residues of BHVO-2 and BCR-2 fromWeis et al (2006)Hf isotopic compositions were analyzed following the
procedures detailed byWeis et al (2007) The configurationfor Hf analyses monitored Lu mass 175 andYb mass 172 toallow for interference correction to masses 174 and 176 Hfisotopic ratios were normalized internally for mass fractio-nation to a 179Hf177Hf ratio of 07325 using an exponentialcorrection Standards were run after every two samplesand sample results were normalized to the ratio of the in-run daily average and a 176Hf177Hf ratio for JMC-475 of0282160 During the course of analyses the Hf standardJMC-475 gave an average value 02821679 (nfrac14130)USGS reference materials BCR-2 and BHVO-2 were pro-cessed with the samples and yielded Hf isotopic ratios of02828675 and 02831005 respectively Publishedvalues for BCR-2 and BHVO-2 are 02828717 and0283104 8 respectively (Weis et al 2007)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
39
result of unusually high melt production rates are predo-minantly basaltic in composition and their formation isnot directly attributable to seafloor spreading processes(eg Saunders 2005) The high melt production rates arebest explained by high mantle source temperatures in rap-idly upwelling mantle (ie mantle plumes) direct evidenceof high mantle temperatures is the eruption of primitivehigh-MgO lavas (Kerr amp Mahoney 2007) Studies of oce-anic plateaus can be used to constrain aspects of mantleplume-related magmatism including the composition ofthe mantle source involved the temperature and depth ofmelting the volume of magma and melt production ratesand the mechanisms of emplacement of voluminous lavasequences From a geochemical perspective the study ofoceanic plateaus is important because the chemistry of thelavas has generally been unaffected by continental crustalcontaminationAt present one of the great challenges in studying oce-
anic plateaus in the ocean basins is the difficulty in drillingto depths of more than a few hundred meters into thebasaltic basement rocks (eg Kerguelen Ontong Java)Accreted sections of oceanic plateaus (eg CaribbeanOntong Java) provide an opportunity to closely examinetheir volcanic stratigraphyWrangellia flood basalts in thePacific Northwest of North America are part of a majorlarge igneous province that erupted in a marine setting inthe Late Triassic (c 225^231 Ma) and accreted to westernNorth America in the Late Jurassic or Early Cretaceous(Richards et al 1991) Although their original areal distri-bution was probably considerably larger much of the orig-inal stratigraphic thickness (6 km on Vancouver Island35 km in Alaska) is intact Current exposures of theflood basalts are preserved in a narrow belt over 2300 kmin length resulting from transform fault motions afteraccretion which extends from southern British Columbiathrough the Yukon and into Alaska Richards et al (1991)proposed a plume initiation model for the Wrangelliaflood basalts based on the short duration and estimatedhigh eruption rate of the volcanism evidence of upliftprior to volcanism and lack of evidence of associated rift-ing (ie few dikes and abundant sills)TheWrangellia flood basalts are perhaps the most exten-
sive accreted remnants of an oceanic plateau in the worldwhere parts of the entire volcanic stratigraphy areexposed however they have been the focus of few inte-grated field geochemical and radiogenic isotopic studiesThe internal architecture of oceanic plateaus can provideconstraints on the sequence of melting events associatedwith mantle plumes that occur beneath the oceans in ananalogous way to the use of continental flood basalts toelucidate the sequential development of mantle plumesand rifting within continental lithosphere (eg Peate 1997Storey et al 1997) The internal stratigraphy of continentalflood basalts is far more complex than previously realized
(eg Jerram amp Widdowson 2005) and because of theirrelative inaccessibility the structure of flood basaltsequences in oceanic plateaus is not well known In thisstudy we examine the field relationships petrographymajor and trace elements and Sr^Nd^Hf^Pb isotopiccompositions of Wrangellia flood basalts from variousareas of Vancouver Island to assess the nature of themantle source conditions of melting and magmatic his-tory of the basalts in the context of the construction ofthis major oceanic plateau This study is part of a compre-hensive research project on the nature origin and evolu-tion of the Triassic Wrangellia flood basalt province inBritish Columbia Yukon and Alaska (Greene et al 20082009)
GEOLOGICAL SETT INGWrangellia on Vancouver IslandWrangellia flood basalts form the core of the Wrangelliaterrane orWrangellia one of the largest outboard terranesaccreted to western North America (Jones et al 1977)Middle to LateTriassic flood basalts extend in a discontin-uous belt from Vancouver and Queen Charlotte Islands(Karmutsen Formation) through SE Alaska and SWYukon and into the Wrangell Mountains Alaska Rangeand Talkeetna Mountains in east and central Alaska(Nikolai Formation Fig 1)Wrangellia covers 80 of Vancouver Island which is
460 km long by 130 km wide (Fig 1) Wrangellia is theuppermost sheet of a stack of SW-vergent thrust sheetsthat form the crust of Vancouver Island and has a cumula-tive thickness of410 km (Monger amp Journeay 1994) Pre-Karmutsen units of Wrangellia on Vancouver Islandinclude Devonian arc sequences of the Sicker Group andMississippian to Early Permian siliciclastic and carbonaterocks of the Buttle Lake Group (Muller 1980 Sutherland-Brown et al1986)These Paleozoic sequences have variableestimated thicknesses (2^5 km Muller 1980 Massey1995a) and are exposed only on central and southernVancouver Island (Fig 1) The Karmutsen Formation isoverlain by the shallow-water Quatsino limestone (30^750m) and shallow- to deeper-water Parson BayFormation (600m) which is intercalated with and over-lain by Bonanza arc volcanics (169^202 Ma Nixon et al2006b) Bonanza arc plutonic rocks (167^197 Ma) alsointrude the Karmutsen Formation and are some of theyoungest units of Wrangellia that formed prior to accretionwith North America (Nixon et al 2006b) Following accre-tionWrangellia units were intruded by the predominantlyCretaceous Coast Plutonic Complex (Wheeler amp McFeely1991)The Karmutsen Formation (19142 km2 calculation
based on digital geological compilations) is composed ofbasal sediment^sill complexes a lower unit of pillowed
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
2
and massive submarine flows a middle unit of mostlypillow breccia and hyaloclastite and an upper unit of pre-dominantly massive subaerial flows (Carlisle amp Suzuki1974) The pillow basalts directly overlie Middle Triassicand late Paleozoic marine sedimentary sequences that areintruded by Karmutsen mafic sills The boundary betweenpillow breccia^hyaloclastite and massive lava flows repre-sents the transition from a submarine to a subaerial erup-tive environment The uppermost flows of the Karmutsenare intercalated with and overlain by shallow-water lime-stone and local occurrences of pillowed flows and hyalo-clastite deposits occur within the upper subaerial memberThe Karmutsen Formation and Wrangellia on Vancouver
Island have been the focus of detailed mapping and strati-graphic studies by Carlisle (Carlisle 1963 1972 Carlisle ampSuzuki 1974) and Muller (Muller et al 1974 Muller 1977)Recent descriptions of the Karmutsen Formation on north-ern Vancouver Island have been made during regionalmapping studies (150 000 scale) by Nixon and coworkers(Nixon et al 2006b 2008 Nixon amp Orr 2007)
Age of the Karmutsen FormationThe age and duration of Karmutsen volcanism are con-strained by fossils in the underlying and overlying sedi-mentary units and by three U^Pb isotopic age
Fig 1 Simplified map of Vancouver Island showing the distribution of the Karmutsen Formation and underlying Paleozoic formations (afterMassey et al 2005a 2005b) Areas in white are mostly younger units The main areas of field study are indicated with boxes or circles withcapital letters (see legend) Ocean is light blue The inset shows the extent of theWrangellia flood basalts (green) in British Columbia Yukonand Alaska
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
3
determinations on intrusive rocks that are considered to berelated to the Karmutsen volcanics Daonella-bearing shale100^200m below the base of the pillow basalts on SchoenMountain and Halobia-rich shale interlayered with flows inthe upper part of the Karmutsen indicate eruption of mostof the flood basalts between the Middle Ladinian (c 232^235 Ma Middle Triassic) and latest Carnian or possiblyearliest Norian (c 216^227 Ma Late Triassic Carlisle ampSuzuki 1974) The only published U^Pb age is based on asingle concordant analysis of a multi-grain baddeleyitefraction from a gabbro on southernVancouver Island thatyielded a 206Pb238U age of 227326 Ma (Parrish ampMcNicoll 1992) Two unpublished 206Pb238U baddeleyiteages also from a gabbro on southern Vancouver Islandare 226805 Ma (five multi-grain fractions) and228425 Ma (two multi-grain fractions Sluggett 2003)
VOLCANIC STRAT IGRAPHY ANDPETROGRAPHYFieldwork undertaken onVancouver Island as part of thisstudy in 2004^2006 explored the volcanic stratigraphy ofthe Karmutsen flood basalts in three main areas theKarmutsen Range (between Alice and Nimpkish lakes)the area around Schoen Lake Provincial Park and atMount Arrowsmith (Greene et al 2005 2006) areasaround Holberg Inlet in northernmost Vancouver Islandand Buttle Lake in central Vancouver Island were alsoinvestigated (Fig 1) The character and thickness of theflood basalt sequences vary locally although the tripartitesuccession of the Karmutsen Formation appears to be pres-ent throughout Vancouver Island The stratigraphic thick-nesses for the pillow volcaniclastic and subaerial flowunits in the type area around Buttle Lake are estimated tobe 2600m 1100m and 2900m respectively (Surdam1967) estimated thicknesses on northernVancouver Islandare 43000m 400^1500m and 41500m respectively(Nixon et al 2008) in the vicinity of Mount Arrowsmithand nearby areas on southern Vancouver Island thick-nesses are 1100m 950m and 1200m respectively (Yorathet al 1999 Fig 1)Picritic pillow lavas west of the Karmutsen Range on
northernVancouver Island occur in a roughly triangular-shaped area (30 km across) bounded by KeoghMaynard and Sara lakes (Figs 2 and 3 Greene et al2006 Nixon et al 2008) Excellent exposures of picriticpillow lavas occur in roadcuts along the north shore ofKeogh Lake the type locality (Greene et al 2006) TheKeogh Lake picrites form pillowed and massive flow units(515m thick Fig 3) with pillows and tubes of varieddimensions (typically51m wide) Numerous thermal con-traction features in the pillows are filled with quartz^carbonate such as drain-back ledges and tortoise-shelljointing (Greene et al 2006) The picritic pillow basalts
are not readily distinguishable in the field from basaltexcept by their density non-magnetic character and com-paratively minor amounts of interpillow quartz^carbonateFieldwork and mapping indicates that the picrites occurnear the transition between the uppermost pillow lavasand the lowermost part of the volcaniclastic unit (Nixonet al 2008)In and around Schoen Lake Provincial Park where
the base of the Karmutsen Formation is exposed is asediment^sill complex consisting of Middle Triassic andPennsylvanian to Permian limestones and fine-grained sili-ciclastic sedimentary rocks intruded by mafic sills andoverlain by pillow basalt (Figs 2 and 4 Carlisle 1972) Thesediment^sill complex is 600^900m thick whichincludes 150^200m of pre-intrusive sedimentary rocks(Carlisle 1972) TheTriassic sedimentary rocks range fromthinly bedded siliceous and calcareous shales to bandedcherts and finely laminated Daonella-bearing shales(Carlisle 1972) Basal sills and pre-Karmutsen sedimentsare also exposed around Buttle Lake (Fig 1)Exposures with considerable vertical relief around
Mount Arrowsmith (see Electronic Appendix 1 availablefor downloading at httppetrologyoxfordjournalsorg)and Buttle Lake preserve thick successions of submarinepillow basalts and pillow breccias with rare intercalationsof sediment and subaerial flows Massive submarine flowsin the pillow unit are locally recognizable by their irregu-lar hackly columnar jointing The massive subaerial flowsform monotonous sequences marked mainly by amygdaloi-dal horizons brecciated flow tops are rarely observedA locality with well-preserved pahoehoe flow features iswell exposed north of Holberg Inlet (Fig 3 Nixon amp Orr2007 Nixon et al 2008) There is no evidence of significantdetrital material from a continental source in the sedi-ments associated with the flood basalts anywhere onVancouver Island The uppermost Karmutsen flows areinterbedded with thin (commonly 0^4m thick) lenses oflimestone and rarely siliciclastic sedimentary rocks(Nixon et al 2006b) Plagioclase-phyric lavas characterizethe upper part of the subaerial sequence and locallyinclude distinctive plagioclase-megacrystic (1^2 cm crys-tals) trachytoid-textured flows (Nixon et al 2006b)A total of 129 samples were collected from the
Karmutsen Formation on Vancouver Island from which63 samples were selected for geochemical analysis basedon the relative degree of alteration and geographical distri-bution of the samples Fifty-six of these samples weredivided into four main groups based on petrography(Table 1) and geochemistry tholeiitic basalt picrite high-MgO basalt and coarse-grained mafic rocks In additiontwo outlying samples and a mineralized sill are distin-guished in Table 1 The tholeiitic basalts are dominantlyglomeroporphyritic and less commonly aphanitic withan intersertal to intergranular groundmass (Table 1)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
4
Fig 2 Geological map and stratigraphy of the Schoen Lake Provincial Park and Karmutsen Range areas (locations shown in Fig 1 northernVancouver Island) (a) Stratigraphic column depicts units exposed in the Schoen Lake area derived from Carlisle (1972) and fieldwork (b)Generalized geology for the Schoen Lake area with sample locations Map derived from Massey et al (2005a) The exposures in the SchoenLake area are the lower volcanic stratigraphy and base of the Karmutsen Formation (c) Stratigraphic column for geology in the Alice^Nimpkish Lake area derived from Nixon amp Orr (2007) (d) Geological map from mapping of Nixon et al (2006a 2008) Sample sites andlithologies are denoted in the legend The Keogh Lake picrites are exposed near Keogh Sara and Maynard lakes and areas to the east ofMaynard Lake (Nixon et al 2008)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
5
Fig 3 Photographs of picritic and tholeiitic pillowed and massive basalt flows from the Karmutsen Range (Alice and Nimpkish Lake area)and Holberg Inlet (HI) areas northernVancouver Island (locations shown in Fig 1) (a) Massive submarine flow (marked with arrows) drapedover high-MgO pillow basalt of similar composition (b) Picritic pillow basalts in cross-section near Maynard Lake (Note vesicular uppermargin to pillows) (c) Close-up photograph of undulating contact of massive submarine flow draped over high-MgO pillow basalt (markedwith arrows) from photograph (a) (area of photograph indicated with a white box) (d) Photograph of margin of pahoehoe lobe (sledgehammerfor scale) (e) Cross-section of a large picritic pillow lobe with infilling of quartz^carbonate in cooling^contraction cracks (sledgehammer forscale) (f) High-MgO pillow breccia stratigraphically between submarine and subaerial flows (arrow points to lens cap 7 cm diameter forscale)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
6
Plagioclase forms most of the phenocrysts and glomero-crysts and the groundmass is fine-grained comprising pla-gioclase microlites clinopyroxene granules small grains ofFe^Ti oxide devitrified glass and secondary minerals
there is no fresh glass preserved The picritic and high-MgO basaltic lavas are characterized by spherulitic plagi-oclase and clinopyroxene with abundant euhedral olivinephenocrysts (Fig 5 Table 1) Olivine is completely replaced
Fig 4 Photographs showing field relations from the Schoen Lake Provincial Park area northernVancouver Island (a) Sediment^sill complexat the base of the Karmutsen Formation on the north side of Mt Schoen 100m below the Daonella locality (see Fig 2 for location) (b)Interbedded mafic sills and deformed finely banded chert and shale with calcareous horizons from location between Mt Adam and MtSchoen
Fig 5 Photomicrographs of picritic pillow basalts Karmutsen Formation northernVancouver Island Scale bars represent 1mm (a) Picritewith abundant euhedral olivine pseudomorphs from Keogh Lake type locality (Fig 2) in cross-polarized light (XPL) (sample 4722A4 19^198wt MgO four analyses) Samples contain dense clusters (24^42 vol ) of olivine pseudomorphs (52mm) in a groundmass of curvedand branching sheaves of acicular plagioclase and intergrown with clinopyroxene and altered glass In many cases plagioclase nucleated on theedges of the olivine phenocrysts (b) Sheaves of intergrown plagioclase and clinopyroxene in aphyric picrite pillow lava from west of MaynardLake (Fig 2) in XPL (sample 4723A2108 wt MgO)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
7
Table 1 Summary of petrographic characteristics and phenocryst proportions of Karmutsen basalts on Vancouver Island
BC
Sample Area Flow Group Texture vol Ol Plag Cpx Ox Alteration Notes
4718A1 MA PIL THOL intersertal 20 5 3 few plag glcr52mm cpx51 mm
4718A2 MA PIL THOL intersertal glomero 15 1 plag glcr52mm very fresh
4718A5 MA FLO THOL intersertal glomero 10 2 mottled few plag glcr54 mm
4723A4 KR PIL PIC intergranular intersertal 0 3 swtl plag51mm no ol phenos
4723A13 KR PIL PIC spherulitic 24 3 swtl plag51 mm
(continued)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
8
by an assemblage of talc^tremolite^serpentine^opaqueoxides there is no fresh olivine present in the Karmutsenlavas onVancouver Island Plagioclase and clinopyroxenephenocrysts are absent in the high-MgO lavas Thecoarse-grained mafic rocks are characterized by sub-ophi-tic textures with an average grain size typically 41mmand are generally not glomeroporphyritic these rocks aremainly from the interiors of massive flows although somemay represent sillsSample preparation and analytical methods for major-
and trace-element and Sr^Nd^Hf^Pb isotopic composi-tions of the Karmutsen volcanic rocks are given in theAppendix
WHOLE -ROCK CHEMISTRYMajor- and trace-element compositionsThe most abundant type of volcanic rock in theKarmutsen Formation is tholeiitic basalt with a restrictedrange of major- and trace-element compositions The tho-leiitic basalts have similar compositions to the coarse-grained mafic rocks and both groups are distinct fromthe picrites and high-MgO basalts [Fig 6 picrites412wt MgO (LeBas 2000) high-MgO basalts8^12wt MgO] The tholeiitic basalts have lower MgO(57^77wt MgO) and higher TiO2 (14^22wt
TiO2) than the picrites (130^198wt MgO05^07wt TiO2) and high-MgO basalts (91^116wt MgO 05^08wt TiO2 Fig 6 Table 2) Almost allcompositions plot within the tholeiitic field in a total alka-lis vs silica plot although there has been substantial K lossin most samples from the Karmutsen Formation (seebelow) and the tholeiitic basalts generally have higherSiO2 Na2OthornK2O CaO and FeOT (total iron expressedas FeO) than the picrites The tholeiitic basalts also havenoticeably lower loss on ignition (LOI mean1713wt ) than the picrites (mean 54 08wt )and high-MgO basalts (mean 3815wt ) (Fig 6)reflecting the presence of abundant altered olivinephenocrysts in the latter groups Ni concentrations are sig-nificantly higher for the picrites (339^755 ppm) and high-MgO basalts (122^551ppm) than the tholeiitic basalts(58^125 ppm except for one anomalous sample) andcoarse-grained rocks (70^213 ppm) (Fig 6 Table 2) Ananomalous tholeiitic pillowed flow (two samples outlierin Tables 1 and 2) from near the picrite type locality atKeogh Lake and a mineralized sill (disseminated sulfide)from the basal sediment^sill complex at Schoen Lake aredistinguished by higher TiO2 and FeOT than the maingroup of tholeiitic basalts The tholeiitic basalts are similarin major-element composition to previously publishedresults for samples from the Karmutsen Formation how-ever the Keogh Lake picrites which encompass the
Table 1 Continued
Sample Area Flow Group Texture vol Ol Plag Cpx Ox Alteration Notes
4722A5 KR FLO OUTLIER intersertal 3 mottled very fg v small pl needles
5615A11 KR PIL OUTLIER intersertal 3 mottled very fg v small pl needles
5617A4 SL SIL MIN intersertal 5 3 fg
Sample number last digit of year month day initial sample station (except 93G171) MA Mount Arrowsmith SLSchoen Lake KR Karmutsen Range PIL pillow BRE breccia FLO flow SIL sill GAB gabbro THOL tholeiiticbasalt PIC picrite HI-MG high-MgO basalt CGcoarse-grained MIN mineralized sill glomero glomeroporphyriticOlivine modes for picrites are based on area calculations from scans (see lsquoOlivine accumulation in high-MgO and picriticlavasrsquo section and Fig 11) Visual alteration index is based primarily on degree of plagioclase alteration and presence ofsecondary minerals (1 least altered 3 most altered) Plagioclase phenocrysts are commonly altered to albite pumpellyiteand chlorite olivine is altered to talc tremolite and clinochlore (determined using the Rietveld method of X-ray powderdiffraction) clinopyroxene is unaltered FendashTi oxide is commonly replaced by sphene glcr glomerocrysts fg fine-grained cg coarse-grained oik oikocryst chad chadacryst enc enclosing swtl swallow-tail ol olivinepseudomorphs plag plagioclase cpx clinopyroxene ox oxides (includes ilmenite thorn titanomagnetite)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
9
00
05
10
15
20
25
30
35
40
0 5 10 15 20 25
10
11
12
13
14
15
16
17
18
19
20
0 5 10 15 20 25
6
7
8
9
10
11
12
13
14
15
0 5 10 15 20 25
6
7
8
9
10
11
12
13
14
15
16
0 5 10 15 20 25
0
1
2
3
4
5
6
40 45 50 55 60
0
200
400
600
800
0 5 10 15 20 25
(wt)(wt)
(wt)
(wt)
+K2ONa2O TiO2
FeO (T)
Ni (ppm)
Al2O3
CaO
alkalic
tholeiitic
Karmutsen Form
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
(compiled data)
Pillowed flowMineralized sill
SiO2 (wt) MgO (wt)
MgO (wt)MgO (wt)
MgO (wt) MgO (wt)
(a) (b)
(g)
(e)
(f)
(c)
LOI (wt )
(d)
MgO (wt )0
1
2
3
4
5
6
7
0 5 10 15 20 25
(wt)
Fig 6 Whole-rock major-element Ni and LOI variation diagrams for the Karmutsen Formation New samples from this study (Table 2) areshown by symbols with black outlines (see legend) Previously published analyses are shown by small gray symbols without black outlines Theboundary of the alkaline and tholeiitic fields is that of MacDonald amp Katsura (1964) Total iron expressed as FeO LOI loss on ignition oxidesare plotted on an anhydrous normalized basis References for the 322 compiled analyses for the Karmutsen Formation are Surdam (1967)Kuniyoshi (1972) Muller et al (1974) Barker et al (1989) Lassiter et al (1995) Massey (1995a1995b)Yorath et al (1999) and G Nixon (unpublisheddata) It should be noted that the compiled dataset has not been filtered many of the samples with high SiO2 (452wt ) are probably notKarmutsen flood basalts but younger Bonanza basalts and andesites Three pillowed flow samples and a mineralized sill are distinguishedseparately because of their anomalous chemistry Coarse-grained samples are indicated separately Three tholeiitic basalt samples (4724A54718A5 4718A6) with416wt Al2O3 and4270 ppm Sr contain 10^25 modal of large plagioclase phenocrysts (7^8mm) or glomerocrysts(44mm)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
10
Table 2 Major element (wt oxide) and trace element (ppm) abundances in whole-rock samples of Karmutsen basalts
1Ni concentrations for these samples were determined by XRFTHOL tholeiitic basalt PIC picrite HI-MG high MgO basalt CG coarse-grained (sill or gabbro) MIN SIL mineralizedsill OUTLIER anomalous pillowed flow in plots MA Mount Arrowsmith SL Schoen Lake KR Karmutsen Range QIQuadra Island Sample locations are given using the Universal Transverse Mercator (UTM) coordinate system (NAD83zones 9 and 10) Analyses were performed at Activation Laboratory (ActLabs) Fe2O3
is total iron expressed as Fe2O3LOI loss on ignition All major elements Sr V and Y were measured by ICP quadrupole OES on solutions of fusedsamples Cu Ni Pb and Zn were measured by total dilution ICP Cs Ga Ge Hf Nb Rb Ta Th U Zr and REE weremeasured by magnetic-sector ICP on solutions of fused samples Co Cr and Sc were measured by INAA Blanks arebelow detection limit See Electronic Appendices 2 and 3 for complete XRF data and PCIGR trace element datarespectively Major elements for sample 93G17 are normalized anhydrous Sample 5615A7 was analyzed by XRFSamples from Quadra Island are not shown in figures
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
17
picritic and high-MgO basalt pillow lavas extend to20wt MgO (Fig 6 and references listed in thecaption)The tholeiitic basalts from the Karmutsen Range dis-
play a tight range of parallel light rare earth element(LREE)-enriched REE patterns (LaYbCNfrac14 21^24mean 2202) whereas the picrites and high-MgObasalts are characterized by sub-parallel LREE-depleted
patterns (LaYbCNfrac14 03^07 mean 06 02) with lowerREE abundances than the tholeiitic basalts (Fig 7)Tholeiitic basalts from the Schoen Lake and MountArrowsmith areas have similar REE patterns tosamples from the Karmutsen Range (Schoen Lake LaYbCNfrac14 21^26 mean 2303 Mount Arrowsmith LaYbCNfrac1419^25 mean 2204) and the coarse-grainedmafic rocks have similar REE patterns to the tholeiitic
Depleted MORB average(Salters amp Stracke 2004)
Schoen LakeSchoen Lake
Mount Arrowsmith
Karmutsen RangeS
ampl
eC
hond
rite
Sam
ple
Cho
ndrit
e
Sam
ple
Prim
itive
Man
tleS
ampl
eP
rimiti
ve M
antle
Sam
ple
Prim
itive
Man
tle
Mount Arrowsmith
Karmutsen Range
(a) (b)
(c) (d)
(e) (f )
Sam
ple
Cho
ndrit
e
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained mafic rock
Pillowed flowMineralized sill
1
10
100
1
10
100
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
01
1
10
100
1
10
100
1
10
100
Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
Fig 7 Whole-rock REE and other incompatible-element concentrations for the Karmutsen Formation (a) (c) and (e) are chondrite-normal-ized REE patterns for samples from each of the three main field areas onVancouver Island (b) (d) and (f) are primitive mantle-normalizedtrace-element patterns for samples from each of the three main field areas All normalization values are from McDonough amp Sun (1995) Theclear distinction between LREE-enriched tholeiitic basalts and LREE-depleted picrites the effects of alteration on the LILE (especially loss ofK and Rb) and the different range for the vertical scale in (b) (extends down to 01) should be noted
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
18
basalts (LaYbCNfrac14 21^29 mean 2508) Two LREE-depleted coarse-grained mafic rocks from the SchoenLake area have similar REE patterns (LaYbCNfrac14 07) tothe picrites from the Keogh Lake area indicating that thisdistinctive suite of rocks is exposed as far south asSchoen Lake that is 75 km away (Fig 7) The anomalouspillowed flow from the Karmutsen Range has lower LaYbCN (13^18) than the main group of tholeiitic basaltsfrom the Karmutsen Range (Fig 7) The mineralized sillfrom the Schoen Lake area is distinctly LREE-enriched(LaYbCNfrac14 33) with the highest REE abundances(LaCNfrac14 940) of all Karmutsen samples this may reflectcontamination by adjacent sediment also evident in thin-section where sulfide blebs are cored by quartz (Greeneet al 2006)The primitive mantle-normalized trace-element pat-
terns (Fig 7) and trace-element variations and ratios(Fig 8) highlight the differences in trace-element
concentrations between the four main groups ofKarmutsen flood basalts onVancouver Island The tholeii-tic basalts from all three areas have relatively smooth par-allel trace-element patterns some samples have smallpositive Sr anomalies relative to Nd and Sm and manysamples have prominent negative K anomalies relative toU and Nb reflecting K loss during alteration (Fig 7) Thelarge ion lithophile elements (LILE) in the tholeiiticbasalts (mainly Rb and K not Cs) are depleted relativeto the high field strength elements (HFSE) and LREELILE segments especially for the Schoen Lake area(Fig 7d) are remarkably parallel The picrites andhigh-MgO basalts form a tight band of parallel trace-element patterns and are depleted in HFSE and LREEwith positive Sr anomalies and relatively low Rb Thecoarse-grained mafic rocks from the Karmutsen Rangehave identical trace-element patterns to the tholeiiticbasalts Samples from the Schoen Lake area have similar
000
025
050
075
100
125
0 1 2 3 4 5
00
04
08
12
16
0 5 10 15 20 25
MgO (wt)
0
5
10
15
0 50 100 150
Hf (ppm)
Th (ppm)
NbLaNb (ppm)
Zr (ppm)
(a)
(c)
(b)
(d)
Th (ppm)
00
02
04
06
08
10
12
00 01 02 03 04 05
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
NbLa=1
4723A4 5615A12
U (ppm)
Fig 8 Whole-rock trace-element concentrations and ratios for the Karmutsen Formation [except (b) which is vs wt MgO] (a) Nb vs Zr(b) NbLa vs MgO (c) Th vs Hf (d) Th vs U There is a clear distinction between the tholeiitic basalts and picrites in Nb and Zr both inconcentration and the slope of each trend
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
19
trace-element patterns to the Karmutsen Range except forthe two LREE-depleted coarse-grained mafic rocks andthe mineralized sill (Fig 7d)
Sr^Nd^Hf^Pb isotopic compositionsA total of 19 samples from the four main groups ofthe Karmutsen Formation were selected for radiogenicisotopic geochemistry on the basis of major- and trace-element variation stratigraphic position and geographicallocation The samples have overlapping age-correctedHf and Nd isotope ratios and distinct ranges of Sr isotopiccompositions (Fig 9) The tholeiitic basalts picriteshigh-MgO basalt and coarse-grained mafic rocks have
initial eHffrac14thorn87 to thorn126 and eNdfrac14thorn66 to thorn88 cor-rected for in situ radioactive decay since 230 Ma (Fig 9Tables 3 and 4) All the Karmutsen samples form a well-defined linear array in a Lu^Hf isochron diagram corre-sponding to an age of 24115 Ma (Fig 9) This is withinerror of the accepted age of the Karmutsen Formationand indicates that the Hf isotope systematics behaved as aclosed system since c 230 Ma The anomalous pillowedflow has similar initial eHf (thorn98) and slightly lower initialeNd (thorn62) compared with the other Karmutsen samplesThe picrites and high-MgO basalt have higher initial87Sr86Sr (070398^070518) than the tholeiitic basalts(initial 87Sr86Srfrac14 070306^070381) and the coarse-grained mafic rocks (initial 87Sr86Srfrac14 070265^070428)
05128
05129
05130
05131
05132
015 017 019 021 023 025
02829
02830
02831
02832
02833
02834
000 002 004 006 008 010
4
6
8
10
0702 0703 0704 0705 07066
8
10
12
14
5 6 7 8 9 10
12
0
1
2
3
4
5
6
7
0702 0703 0704 0705 0706
(d)
LOI (wt )
87Sr 86Sr
4723A2
Nd
143Nd144Nd
147Sm144Nd
87Sr 86Sr (230 Ma)
(230 Ma)
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
(a)
(c)
176Hf177Hf
Age=241 plusmn 15 Ma=+990 plusmn 064
176Lu177Hf
Hf (230 Ma)
Nd (230 Ma)
(e)
Hf (i)
(b)
Fig 9 Whole-rock Sr Nd and Hf isotopic compositions for the Karmutsen Formation (a) 143Nd144Nd vs 147Sm144Nd (b) 176Hf177Hf vs176Lu177Hf The slope of the best-fit line for all samples corresponds to an age of 24115 Ma (c) Initial eNd vs 87Sr86Sr Age-correction to230 Ma (d) LOI vs 87Sr86Sr (e) Initial eHf vs eNd Average 2 error bars are shown in a corner of each panel Complete chemical duplicatesshown inTables 3 and 4 (samples 4720A7 and 4722A4) are circled in each plot
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
overlap the ranges of the picrites and tholeiitic basalts(Fig 9 Table 3)The measured Pb isotopic compositions of the tholeiitic
basalts are more radiogenic than those of the picrites andthe most magnesian Keogh Lake picrites have the leastradiogenic Pb isotopic compositions The range ofinitial Pb isotope ratios for the picrites is206Pb204Pbfrac1417868^18580 207Pb204Pbfrac1415547^15562and 208Pb204Pbfrac14 37873^38257 and the range for thetholeiitic basalts is 206Pb204Pbfrac1418782^19098207Pb204Pbfrac1415570^15584 and 208Pb204Pbfrac14 37312^38587 (Fig 10 Table 5) The coarse-grained mafic rocksoverlap the range of initial Pb isotopic compositions forthe tholeiitic basalts with 206Pb204Pbfrac1418652^19155207Pb204Pbfrac1415568^15588 and 208Pb204Pbfrac14 38153^38735 One high-MgO basalt has the highest initial206Pb204Pb (19242) and 207Pb204Pb (15590) and thelowest initial 208Pb204Pb (37676) The Pb isotopic ratiosfor Karmutsen samples define broadly linearrelationships in Pb isotope plots The age-corrected Pb
isotopic compositions of several picrites have been affectedby U and Pb (andor Th) mobility during alteration(Fig 10)
ALTERAT IONThe Karmutsen basalts have retained most of their origi-nal igneous structures and textures however secondaryalteration and low-grade metamorphism have producedzeolitic- and prehnite^pumpellyite-bearing mineral assem-blages (Table 1 Cho et al 1987) Alteration and metamor-phism have primarily affected the distribution of the LILEand the Sr isotopic systematics The Keogh Lake picriteshave been affected more by alteration than the tholeiiticbasalts and have higher LOI (up to 55wt ) variableLILE and higher measured Sr Nd and Hf and lowermeasured Pb isotope ratios than the tholeiitic basalts Srand Pb isotope compositions for picrites and tholeiiticbasalts show a relationship with LOI (Figs 9 and 10)whereas Nd and Hf isotopic compositions do not correlate
Table 3 Sr and Nd isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Rb Sr 87Sr86Sr 2m87Rb86Sr 87Sr86Srt Sm Nd 143Nd144Nd 2m eNd
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses carried out at the PCIGR thecomplete trace element analyses are given in Electronic Appendix 3
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
21
with LOI (not shown) The relatively high apparent initialSr isotopic compositions for the high-MgO lavas (up to07052) probably resulted from post-magmatic loss of Rbor an increase in 87Sr86Sr through addition of seawater Sr(eg Hauff et al 2003) High-MgO lavas from theCaribbean plateau have similarly high 87Sr86Sr comparedwith basalts (Recurren villon et al 2002)The correction for in situdecay on initial Pb isotopic ratios has been affected bymobilization of U and Th (and Pb) in the whole-rockssince their formation A thorough acid leaching duringsample preparation was used that has been shown to effec-tively remove alteration phases (Weis et al 2006 NobreSilva et al in preparation) Whole-rock SmNd ratiosand age-correction of Nd isotopes may be affected bythe leaching procedure for submarine rocks older than50 Ma (Thompson et al 2008 Fig 9) there ishowever very little variation in Nd isotopic compositionsof the picrites or tholeiitic basalts and this system does notappear to be disturbed by alteration The HFSE abun-dances for tholeiitic and picritic basalts exhibit clear
linear correlations in binary diagrams (Fig 8) whereasplots of LILE vs HFSE are highly scattered (not shown)because of the mobility of some of the alkali and alkalineearth LILE (Cs Rb Ba K) and Sr for most samplesduring alterationThe degree of alteration in the Karmutsen samples does
not appear to be related to eruption environment or depthof burial Pillow rims are compositionally different frompillow cores (Surdam1967 Kuniyoshi1972) and aquagenetuffs are chemically different from isolated pillows withinpillow breccias however submarine and subaerial basaltsexhibit a similar degree of alterationThere is no clear cor-relation between the submarine and subaerial basalts andsome commonly used chemical alteration indices (eg BaRb vs K2OP2O5 Huang amp Frey 2005) There is also nodefinitive correlation between the petrographic alterationindex (Table 1) and chemical alteration indices although10 of 13 tholeiitic basalts with the highest K and LILEabundances have the highest petrographic alterationindex of three (seeTable 1)
Table 4 Hf isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Lu Hf 176Hf177Hf 2m eHf176Lu177Hf 176Hf177Hft eHf(t)
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
OLIV INE ACCUMULAT ION INH IGH-MgO AND PICR IT IC LAVASGeochemical trends and petrographic characteristics indi-cate that accumulation of olivine played an important rolein the formation of the high-MgO basalts and picrites fromthe Keogh Lake area on northern Vancouver Island TheKeogh Lake samples show a strong linear correlation inplots of Al2O3 TiO2 ScYb and Ni vs MgO and many ofthe picrites have abundant clusters of olivine pseudomorphs(Table1 Fig11)There is a strong linear correlationbetween
modal per cent olivine and whole-rock magnesium contentmagnesium contents range between 10 and 19wt MgOand the proportion of olivine phenocrysts in these samplesvaries from 0 to 42 vol (Fig 11)With a few exceptionsboth the size range and median size of the olivine pheno-crysts in most samples are comparable The clear correla-tion between the proportion of olivine phenocrysts andwhole-rock magnesium contents indicates that accumula-tion of olivine was directly responsible for the high MgOcontents (410wt ) in most of the picritic lavas
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
207Pb204Pb
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
208Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
(a) (b)
(c) (e)
4723A24723A2
4723A2
4723A2
4722A4
4723A4
4722A4
4723A4
4723A134723A3
4723A3
4723A13
1556
1558
1560
1562
1564
1566
1568
186 190 194 198 202 206 2101550
1552
1554
1556
1558
1560
178 180 182 184 186 188 190 192 194
380
384
388
392
396
186 190 194 198 202 206 210374
378
382
386
390
178 180 182 184 186 188 190 192 194
206Pb204Pb(d)
LOI (wt )
0
1
2
3
4
5
6
7
186 190 194 198 202 206 210
4723A2
Fig 10 Pb isotopic compositions of leached whole-rock samples measured by MC-ICP-MS for the Karmutsen Formation Error bars are smal-ler than symbols (a) Measured 207Pb204Pb vs 206Pb204Pb (b) Initial 207Pb204Pb vs 206Pb204Pb Age-correction to 230 Ma (c) Measured208Pb204Pb vs 206Pb204Pb (d) LOI vs 206Pb204Pb (e) Initial 208Pb204Pb vs 206Pb204Pb Complete chemical duplicates shown in Table 5(samples 4720A7 and 4722A4) are circled in each plot The dashed lines in (b) and (e) show the differences in age-corrections for two picrites(4723A4 4722A4) using the measured UTh and Pb concentrations for each sample and the age-corrections when concentrations are used fromthe two picrites (4723A3 4723A13) that appear to be least affected by alteration
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
23
DISCUSS IONThe compositional and spatial development of the eruptedlava sequences of the Wrangellia oceanic plateau onVancouver Island provide information about the evolutionof a mantle plume upwelling beneath oceanic lithospherein an analogous way to the interpretation of continentalflood basalts The Caribbean and Ontong Java plateausare the two primary examples of accreted parts of oceanicplateaus that have been studied to date and comparisonscan be made with the Wrangellia oceanic plateauHowever the Wrangellia oceanic plateau is a distinctiveoccurrence of an oceanic plateau because it formed on thecrust of a Devonian to Mississippian intra-oceanic islandarc overlain by Mississippian to Permian limestones andpelagic sediments The nature and thickness of oceaniclithospheric mantle are considered to play a fundamentalrole in the decompression and evolution of a mantleplume and thus the compositions of the erupted lavasequences (Greene et al 2008) The geochemical and
stratigraphic relationships observed in the KarmutsenFormation onVancouver Island and the focus of the subse-quent discussion sections provide constraints on (1) thetemperature and degree of melting required to generatethe primary picritic magmas (2) the composition of themantle source (3) the depth of melting and residual min-eralogy in the source region (4) the low-pressure evolutionof the Karmutsen tholeiitic basalts (5) the growth of theWrangellia oceanic plateau
Melting conditions and major-elementcomposition of the primary magmasThe primary magmas to most flood basalt provinces arebelieved to be picritic (eg Cox 1980) and they have beenfound in many continental and oceanic flood basaltsequences worldwide (eg Siberia Karoo Paranacurren ^Etendeka Caribbean Deccan Saunders 2005) Near-pri-mary picritic lavas with low total alkali abundances
Table 5 Pb isotopic compositions of Karmutsen basaltsVancouver Island BC
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
require high-degree partial melting of the mantle source(eg Herzberg amp OrsquoHara 2002) The Keogh Lake picritesfrom northernVancouver Island are the best candidates foridentifying the least modified partial melts of the mantleplume source despite having accumulated olivine pheno-crysts and can be used to estimate conditions of meltingand the composition of the primary magmas for theKarmutsen FormationPrimary magma compositions mantle potential tem-
peratures and source melt fractions for the picritic lavas ofthe Karmutsen Formation were calculated from primitivewhole-rock compositions using PRIMELT1XLS software(Herzberg et al 2007) and are presented in Fig 12 andTable 6 A detailed discussion of the computationalmethod has been given by Herzberg amp OrsquoHara (2002)and Herzberg et al (2007) key features of the approachare outlined belowPrimary magma composition is calculated for a primi-
tive lava by incremental addition of olivine and compari-son with primary melts of mantle peridotite Lavas thathad experienced plagioclase andor clinopyroxene fractio-nation are excluded from this analysis All calculated pri-mary magma compositions are assumed to be derived byaccumulated fractional melting of fertile peridotite KR-4003 Uncertainties in fertile peridotite composition donot propagate to significant variations in melt fractionand mantle potential temperature melting of depletedperidotite propagates to calculated melt fractions that aretoo high but with a negligible error in mantle potential
temperature PRIMELT1XLS calculates the olivine liqui-dus temperature at 1atm which should be similar to theactual eruption temperature of the primary magmaand is accurate to 308C at the 2 level of confidenceMantle potential temperature is model dependent witha precision that is similar to the accuracy of the olivineliquidus temperature (C Herzberg personal communica-tion 2008)With regard to the evidence for accumulated olivine in
the Keogh Lake picrites it should not matter whether themodeled lava composition is aphyric or laden with accu-mulated olivine phenocrysts PRIMELT1 computes theprimary magma composition by adding olivine to a lavathat has experienced olivine fractionation in this way itinverts the effects of fractional crystallization The onlyrequirement is that the olivine phenocrysts are not acci-dental or exotic to the lava compositions Support for thisprocedure can be seen in the calculated olivine liquid-line-of-descent shown by the curved arrays with 5 olivineaddition increments (Fig 12c) Fractional crystallization ofolivine from model primary magmas having 85^95wt FeO and 153^175wt MgO will yield arrays of deriva-tive liquids and randomly sorted olivines that in most butnot all cases connect the picrites and high-MgO basalts Anewer version of the PRIMELT software (Herzberg ampAsimow 2008) can perform olivine addition to and sub-traction from lavas with highly variably olivine phenocrystcontents and yields results that converge with those shownin Fig 12
Fig 11 Relationship between abundance of olivine phenocrysts and whole-rock MgO contents for the Keogh Lake picrites (a) Tracings ofolivine grains (gray) in six high-MgO samples made from high resolution (2400 dpi) scans of petrographic thin-sections Scale bars represent10mm sample numbers are indicated at the upper left (b) Modal olivine vs whole-rock MgO Continuous line is the best-fit line for 10 high-MgO samplesThe areal proportion of olivine was calculated from raster images of olivine grains using ImageJ image analysis software whichprovides an acceptable estimation of the modal abundance (eg Chayes 1954)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
25
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
Gray lines = final melting pressure
L + Ol + Opx
07
08
09
Pressure (GPa)
10
3
4
5
6
1
SOLIDUS
01
04
0506
L + Ol
2
03
02
PeridotiteKR-4003
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
Thick black lines = initial melting pressure
Dashed lines = melt fraction
Melt Fraction
2
3 4 5 6
L+Ol+Opx
SolidusSpinel
SolidusGarnet Peridotite
7
L+Ol
0 10 20 30 4040
45
50
55
60
MgO (wt)
SiO2
(wt)
PeridotiteKR-4003
Melt Fraction
0 10 20 305
10
15
CaO(wt)
MgO (wt)
3
4 5 6 7
2
46
2
Peridotite Partial Melts
Thick black line = initial melting pressureGray lines = final melting pressure
Peridotite
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
02
Pressure (GPa)
0302
04
Pressure (GPa)
Picrite
High-MgO basalt
Tholeiitic basalt
(c)
(d)
Karmutsen flood basaltsOlivine addition modelOlivine addition (5 increment)Primary magma
PicriteHigh-MgO basaltTholeiitic basalt
(a)
(b)
800 850 900
910
920
930
940
950
Karmutsen flood basalts
Olivine addition modelOlivine addition (5 increment)Primary magma
Gray lines = Fo contents of olivine
Fig 12 Estimated primary magma compositions for three Keogh Lake picrites and high-MgO basalts (samples 4723A4 4723A135616A7) using the forward and inverse modeling technique of Herzberg et al (2007) Karmutsen compositions and modeling results areoverlain on diagrams provided by C Herzberg (a) Whole-rock FeO vs MgO for Karmutsen samples from this study Total iron estimatedto be FeO is 090 Gray lines show olivine compositions that would precipitate from liquid of a given MgO^FeO composition Black lineswith crosses show results of olivine addition (inverse model) using PRIMELT1 (Herzberg et al 2007) Gray field highlights olivinecompositions with Fo contents of 91^92 predicted to precipitate (b) FeO vs MgO showing Karmutsen lava compositions and results ofa forward model for accumulated fractional melting of fertile peridotite KR-4003 (Kettle River Peridotite Walter 1998) (c) SiO2 vsMgO for Karmtusen lavas and model results (d) CaO vs MgO showing Karmtusen lava compositions and model results To brieflysummarize the technique [see Herzberg et al (2007) for complete description] potential parental magma compositions for the high-MgOlava series were selected (highest MgO and appropriate CaO) and using PRIMELT1 software olivine was incrementally added tothe selected compositions to show an array of potential primary magma compositions (inverse model) Then using PRIMELT1 theresults from the inverse model were compared with a range of accumulated fractional melts for fertile peridotite derived fromparameterization of the experimental results for KR-4003 from Walter (1998) forward model Herzberg amp OrsquoHara (2002) A melt fractionwas sought that was unique to both the inverse and forward models (Herzberg et al 2007) A unique solution was found when there was a com-mon melt fraction for both models in FeO^MgO and CaO^MgO^Al2O3^SiO2 (CMAS) projection space This modeling assumes thatolivine was the only phase crystallizing and ignores chromite precipitation and possible augite fractionation in the mantle (Herzbergamp OrsquoHara 2002) Results are best for a residue of spinel lherzolite (not pyroxenite) The presence of accumulated olivine in samples ofthe Keogh Lake picrites used as starting compositions does not significantly affect the results because we are modeling addition of olivine(see text for further description) The tholeiitic basalts cannot be used for modeling because they are all saturated in plagthorn cpxthornolThick black line for initial melting pressure at 4 GPa is not shown in (c) for clarity Gray area in (a) indicates parental magmas thatwill crystallize olivine with Fo 91^92 at the surface Dashed lines in (c) indicate melt fraction Solidi for spinel and garnet peridotite arelabelled in (c)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
The PRIMELT1 results indicate that if the Karmutsenpicritic lavas were derived from accumulated fractionalmelting of fertile peridotite the primary magmas wouldhave contained 15^17wt MgO and 10wt CaO(Fig 12 Table 6) formed from 23^27 partial meltingand would have crystallized olivine with a Fo content of 91 (Fig 12 Table 6) Calculated mantle temperatures forthe Karmutsen picrites (14908C) indicate melting ofanomalously hot mantle (100^2008C above ambient) com-pared with ambient mantle that produces mid-ocean ridgebasalt (MORB 1280^14008C Herzberg et al 2007)which initiated between 3 and 4 GPa Several picritesappear to be near-primary melts and required very littleaddition of olivine (eg sample 93G171 with measured158wt MgO) These temperature and melting esti-mates of the Karmutsen picrites are consistent withdecompression of hot mantle peridotite in an actively con-vecting plume head (ie plume initiation model) Threesamples (picrite 5614A1 and high-MgO basalts 5614A3and 5614A5) are lower in iron than the other Karmutsenlavas with FeO in the 65^80wt range (Fig 12c) andhave MgO and FeO contents that are similar to primitiveMORB from the Sequeiros fracture zone of the EastPacific Rise (EPR Perfit et al 1996) These samples
however differ from EPR MORB in having higher SiO2
and lower Al2O3 and they are restricted geographicallyto one area (Sara Lake Fig 2)We therefore have evidence for primary magmas with
MgO contents that range from a possible low of 110wt to as high as 174wt with inferred mantle potential tem-peratures from 1340 to 15208C This is similar to the rangedisplayed by lavas from the Ontong Java Plateau deter-mined by Herzberg (2004) who found that primarymagma compositions assuming a peridotite sourcewould contain 17wt MgO and 10wt CaO(Table 6) and that these magmas would have formed by27 melting at a mantle potential temperature of15008C (first melting occurring at 36 GPa) Fitton ampGodard (2004) estimated 30 partial melting for theOntong Java lavas based on the Zr contents of primarymagmas of Kroenke-type basalt calculated by incrementaladdition of equilibrium olivine However if a considerableamount of eclogite was involved in the formation of theOntong Java plateau magmas the excess temperatures esti-mated by Herzberg et al (2007) may not be required(Korenaga 2005) The variability in mantle potential tem-perature for the Keogh Lake picrites is also similar to thewide range of temperatures that have been calculated for
Table 6 Estimated primary magma compositions for Karmutsen basalts and other oceanic plateaus or islands
Sample 4723A4 4723A13 5616A7 93G171 Average OJP Mauna Keay Gorgonaz
Ontong Java primary magma composition for accumulated fraction melting (AFM) from Herzberg (2004)yMauna Kea primary magma composition is average of four samples of Herzberg (2006 table 1)zGorgona primary magma composition for AFM for 1F fertile source from Herzberg amp OrsquoHara (2002 table 4)Total iron estimated to be FeO is 090
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
27
lavas from single volcanoes in the Galapagos Hawaii andelsewhere by Herzberg amp Asimow (2008) Those workersinterpreted the range to reflect the transport of magmasfrom both the cool periphery and the hotter axis of aplume A thermally heterogeneous plume will yield pri-mary magmas with different MgO contents and the rangeof primary magma compositions and their inferred mantlepotential temperatures for Karmutsen lavas may reflectmelting from different parts of the plume
Source of the Karmutsen Formation lavasThe Sr^Nd^Hf^Pb isotopic compositions of theKarmutsen volcanic rocks on Vancouver Island indicatethat they formed from plume-type mantle containing adepleted component that is compositionally similar to butdistinct from other oceanic plateaus that formed in thePacific Ocean (eg Ontong Java and Caribbean Fig 13)Trace element and isotopic constraints can be used to testwhether the depleted component of the KarmutsenFormation was intrinsic to the mantle plume and distinctfrom the source of MORB or the result of mixing or invol-vement of ambient depleted MORB mantle with plume-type mantle The Karmutsen tholeiitic basalts have initialeNd and
87Sr86Sr ratios that fall within the range of basaltsfrom the Caribbean Plateau (eg Kerr et al 1996 1997Hauff et al 2000 Kerr 2003) and a range of Hawaiian iso-tope data (eg Frey et al 2005) and lie within and abovethe range for the Ontong Java Plateau (Tejada et al 2004Fig 13a) Karmutsen tholeiitic basalts with the highest ini-tial eNd and lowest initial 87Sr86Sr lie within and justbelow the field for age-corrected northern EPR MORB at230 Ma (eg Niu et al1999 Regelous et al1999) Initial Hfand Nd isotopic compositions for the Karmutsen samplesare noticeably offset to the low side of the ocean islandbasalt (OIB) array (Vervoort et al 1999) and lie belowand slightly overlap the low eHf end of fields for Hawaiiand the Caribbean Plateau (Fig 13c) the Karmutsen sam-ples mostly fall between and slightly overlap a field forage-corrected Pacific MORB at 230 Ma and Ontong Java(Mahoney et al 1992 1994 Nowell et al 1998 Chauvel ampBlichert-Toft 2001) Karmutsen lava Pb isotope ratios over-lap the broad range for the Caribbean Plateau and thelinear trends for the Karmutsen samples intersect thefields for EPR MORB Hawaii and Ontong Java in208Pb^206Pb space but do not intersect these fields in207Pb^206Pb space (Fig 13d and e) The more radiogenic207Pb204Pb for a given 206Pb204Pb of the Karmutsenbasalts requires high UPb ratios at an ancient time when235U was abundant similar to the Caribbean Plateau andenriched in comparison with EPR MORB Hawaii andOntong JavaThe low eHf^low eNd component of the Karmutsen
basalts that places them below the OIB mantle array andpartly within the field for age-corrected Pacific MORBcan form from low ratios of LuHf and high SmNd over
time (eg Salters amp White 1998) MORB will develop aHf^Nd isotopic signature over time that falls below themantle array Low LuHf can also lead to low 176Hf177Hffrom remelting of residues produced by melting in theabsence of garnet (eg Salters amp Hart 1991 Salters ampWhite 1998) The Hf^Nd isotopic compositions of theKarmutsen Formation indicate the possible involvementof depleted MORB mantle however similar to Iceland(Fitton et al 2003) Nb^Zr^Y variations provide a way todistinguish between a depleted component within theplume and a MORB-like component The Karmutsenbasalts and picrites fall within the Iceland array and donot overlap the MORB field in a logarithmic plot of NbYvs ZrY (Fig 13b) using the fields established by Fittonet al (1997) Similar to the depleted component within theCaribbean Plateau (Thompson et al 2003) the trend ofHf^Nd isotopic compositions towards Pacific MORB-likecompositions is not followed by MORB values in Nb^Zr^Y systematics and this suggests that the depleted compo-nent in the source of the Karmutsen basalts is distinctfrom the source of MORB and probably an intrinsic partof the plume The relatively radiogenic Pb isotopic ratiosand REE patterns distinct from MORB also support thisinterpretationTrace-element characteristics suggest the possible invol-
vement of sub-arc lithospheric mantle in the generation ofthe Karmutsen picrites Lassiter et al (1995) suggested thatmixing of a plume-type source with eNdthorn 6 to thorn 7 witharc material with low NbTh could reproduce the varia-tions observed in the Karmutsen basalts however theyconcluded that the absence of low NbLa ratios in theirsample of tholeiitic basalts restricts the amount of arc litho-sphere involved The study of Lassiter et al (1995) wasbased on major and trace element and Sr Nd and Pb iso-topic data for a suite of 29 samples from Buttle Lake in theStrathcona Provincial Park on central Vancouver Island(Fig 1) Whereas there are no clear HFSE depletions inthe Karmutsen tholeiitic basalts the low NbLa (and TaLa) of the Karmutsen picritic lavas in this study indicatethe possible involvement of Paleozoic arc lithosphere onVancouver Island (Fig 8)
REE modeling dynamic melting andsource mineralogyThe combined isotopic and trace-element variations of thepicritic and tholeiitic Karmutsen lavas indicate that differ-ences in trace elements may have originated from differentmelting histories For example the LuHf ratios of theKarmutsen picritic lavas (mean 008) are considerablyhigher than those in the tholeiitic basalts (mean 002)however the small range of overlapping initial eHf valuesfor the two lava suites suggests that these differences devel-oped during melting within the plume (Fig 9b) Below wemodel the effects of source mineralogy and depth of
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
28
melting on differences in trace-element concentrations inthe tholeiitic basalts and picritesDynamic melting models simulate progressive decom-
pression melting where some of the melt fraction isretained in the residue and only when the degree of partialmelting is greater than the critical mass porosity and the
source becomes permeable is excess melt extracted fromthe residue (Zou1998) For the Karmutsen lavas evolutionof trace element concentrations was simulated using theincongruent dynamic melting model developed by Zou ampReid (2001) Melting of the mantle at pressures greater than1 GPa involves incongruent melting (eg Longhi 2002)
OIB array
PacificMORB(230 Ma)
(0 Ma)
EPR(230 Ma)
-4
-2
0
2
4
8
10
12
14
16
18
20
22
-4 -2 0 2 4 6 8 10 12 14
minus2
0
2
4
6
8
10
12
14
0702 0703 0704 0705 0706
IndianMORB
87Sr 86Sr (initial)
Hf (initial)
(a) (c)
Nd (initial)
Karmutsen tholeiitic basalt
HawaiiOntong JavaCaribbean Plateau
Karmutsen high-MgO lava
high-MgO lavasKarmutsen
1540
1545
1550
1555
1560
1565
175 180 185 190 195 200375
380
385
390
395
175 180 185 190 195 200
(d) (e)
EPR
EPR
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb
Karmutsen Formation (initial)
HawaiiOntong JavaCaribbean Plateau
East Pacific Rise
Karmutsen Formation (measured)
Juan de FucaGorda
ε Nd
(initial)
Karmutsen Formation (different age-correction for 3 picrites)
(0 Ma)
NbY
001
01
1
1 10
ZrY
OIB
NMORB
(b)
Iceland array
6
Fig 13 Comparison of age-corrected (230 Ma) Sr^Nd^Hf^Pb isotopic compositions for Karmutsen flood basalts onVancouver Island withage-corrected OIB and MORB and Nb^Zr^Y variation (a) Initial eNd vs 87Sr86Sr (b) NbYand ZrY variation (after Fitton et al 1997)(c) Initial eHf vs eNd (d) Measured and initial 207Pb204Pb vs 206Pb204Pb (e) Measured and initial 208Pb204Pb vs 206Pb204Pb References fordata sources are listed in Electronic Appendix 4 Most of the compiled data were extracted from the GEOROC database (httpgeorocmpch-mainzgwdgdegeoroc) OIB array line in (c) is fromVervoort et al (1999) EPR East Pacific Rise Several high-MgO samples affected bysecondary alteration were not plotted for clarity in Pb isotope plots Estimation of fields for age-corrected Pacific MORB and EPR at 230 Mawere made using parent^daughter concentrations (in ppm) from Salters amp Stracke (2004) of Lufrac14 0063 Hffrac14 0202 Smfrac14 027 Ndfrac14 071Rbfrac14 0086 and Srfrac14 836 In the NbYvs ZrYplot the normal (N)-MORB and OIB fields not including Icelandic OIB are from data com-pilations of Fitton et al (1997 2003) The OIB field is data from 780 basalts (MgO45wt ) from major ocean islands given by Fitton et al(2003) The N-MORB field is based on data from the Reykjanes Ridge EPR and Southwest Indian Ridge from Fitton et al (1997) The twoparallel gray lines in (b) (Iceland array) are the limits of data for Icelandic rocks
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
29
with olivine being produced during the reactioncpxthornopxthorn spfrac14meltthornol for spinel peridotites The mod-eling used coefficients for incongruent melting reactionsbased on experiments on lherzolite melting (eg Kinzleramp Grove 1992 Longhi 2002) and was separated into (1)melting of spinel lherzolite (2) two combined intervals ofmelting for a garnet lherzolite (gt lherz) and spinel lher-zolite (sp lherz) source (3) melting of garnet lherzoliteThe model solutions are non-unique and a range indegree of melting partition coefficients melt reactioncoefficients and proportion of mixed source componentsis possible to achieve acceptable solutions (see Fig 14 cap-tion for description of modeling parameters anduncertainties)The LREE-depleted high-MgO lavas require melting of
a depleted spinel lherzolite source (Fig 14a) The modelingresults indicate a high degree of melting (22^25) similarto the results from PRIMELT1 based on major-elementvariations (discussed above) of a LREE-depleted sourceThe high-MgO lavas lack a residual garnet signature Theenriched tholeiitic basalts involved melting of both garnetand spinel lherzolite and represent aggregate melts pro-duced from continuous melting throughout the depth ofthe melting column (Fig 14b) Trace-element concentra-tions in the tholeiitic and picritic lavas are not consistentwith low-degree melting of garnet lherzolite alone(Fig 14c) The modeling results indicate that the aggregatemelts that formed the tholeiitic basalts would haveinvolved lesser proportions of enriched small-degreemelts (1^6 melting) generated at depth and greateramounts of depleted high-degree melts (12^25) gener-ated at lower pressure (a ratio of garnet lherzolite tospinel lherzolite melt of 14 provides the best fits seeFig 14b and description in figure caption) The totaldegree of melting for the tholeiitic basalts was high(23^27) similar to the degree of partial melting in thespinel lherzolite facies for the primary melts that eruptedas the Keogh Lake picrites of a source that was also prob-ably depleted in LREE
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
(c)
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
Melting of garnet lherzolite
High-MgO lavas
Melting of spinel lherzolite
source (07DM+03PM)
source (07DM+03PM)
6 melting
30 melting
(b)
(a)
30 melting
Sam
ple
Cho
ndrit
eS
ampl
eC
hond
rite 20 melting
1 melting
Melting of garnet lherzoliteand spinel lherzolite
x gt lherz + 4x sp lherz
5 meltingx gt lherz + 4x sp lherz
Sam
ple
Cho
ndrit
e
Tholeiitic lavas
Tholeiitic lavas
High-MgO lavas
source (07DM+03PM)
Fig 14 Trace-element modeling results for incongruent dynamicmantle melting for picritic and tholeiitic lavas from the KarmutsenFormation Three steps of modeling are shown (a) melting of spinellherzolite compared with high-MgO lavas (b) melting of garnet lher-zolite and spinel lherzolite compared with tholeiitic lavas (c) meltingof garnet lherzolite compared with tholeiitic and picritic lavasShaded fields in each panel for tholeiitic basalts and picrites arelabeled Patterns with symbols in each panel are modeling results(patterns are 5 melting increments in (a) and (b) and 1 incre-ments in (c) PM primitive mantle DM depleted MORB mantle gtlherz garnet lherzolite sp lherz spinel lherzolite In (b) the ratios ofper cent melting for garnet and spinel lherzolite are indicated (x gtlherzthorn 4x sp lherz means that for 5 melting 1 melt in garnetfacies and 4 melt in spinel facies where x is per cent melting in theinterval of modeling) The ratio of the respective melt proportions inthe aggregate melt (x gt lherzthorn 4x sp lherz) was kept constant andthis ratio of melt contribution provided the best fit to the data Arange of proportion of source components (07DMthorn 03PM to
09DMthorn 01PM) can reproduce the variation in the high-MgOlavas Melting modeling uses the formulation of Zou amp Reid (2001)an example calculation is shown in their Appendix PM fromMcDonough amp Sun (1995) and DM from Salters amp Stracke (2004)Melt reaction coefficients of spinel lherzolite from Kinzler amp Grove(1992) and garnet lherzolite fromWalter (1998) Partition coefficientsfrom Salters amp Stracke (2004) and Shaw (2000) were kept constantSource mineralogy for spinel lherzolite (018cpx027opx052ol003sp) from Kinzler (1997) and for garnet lherzolite(034cpx008opx053ol005gt) from Salters amp Stracke (2004) Arange of source compositions for melting of spinel lherzolite(07DMthorn 03PM to 03DMthorn 07PM) reproduce REE patternssimilar to those of the high-MgO lavas with best fits for 22ndash25melting and 07DMthorn 03PM source composition Concentrationsof tholeiitic basalts cannot be reproduced from an entirelyprimitive or depleted source
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
30
The uncertainty in the composition of the source in thespinel lherzolite melting model for the high-MgO lavasprecludes determining whether the source was moredepleted in incompatible trace elements than the source ofthe tholeiitic lavas (see Fig 14 caption for description) It ispossible that early high-pressure low-degree melting left aresidue within parts of the plume (possibly the hotter inte-rior portion) that was depleted in trace elements (egElliott et al 1991) further decompression and high-degreemelting of such depleted regions could then have generatedthe depleted high-MgO Karmutsen lavas Shallow high-degree melting would preferentially sample depletedregions whereas deeper low-degree melting may notsample more refractory depleted regions (eg CaribbeanEscuder-Viruete et al 2007) The picrites lie at the highend of the range of Nd isotopic compositions and havelow NbZr similar to depleted Icelandic basalts (Fittonet al 2003) therefore the picrites may represent the partsof the mantle plume that were the most depleted prior tothe initiation of melting As Fitton et al (2003) suggestedthese depleted melts may only rarely be erupted withoutcombining with more voluminous less depleted melts andthey may be detected only as undifferentiated small-volume flows like the Keogh Lake picrites within theKarmutsen Formation on Vancouver IslandDecompression melting within the mantle plume initiatedwithin the garnet stability field (35^25 GPa) at highmantle potential temperatures (Tp414508C) and pro-ceeded beneath oceanic arc lithosphere within the spinelstability field (25^09 GPa) where more extensivedegrees of melting could occur
Magmatic evolution of Karmutsentholeiitic basaltsPrimary magmas in large igneous provinces leave exten-sive coarse-grained cumulate residues within or below thecrust these mafic and ultramafic plutonic sequences repre-sent a significant proportion of the original magma thatpartially crystallized at depth (eg Farnetani et al 1996)For example seismic and petrological studies of OntongJava (eg Farnetani et al 1996) combined with the use ofMELTS (Ghiorso amp Sack 1995) indicate that the volcanicsequence may represent only 40 of the total magmareaching the Moho Neal et al (1997) estimated that30^45 fractional crystallization took place for theOntong Java magmas and produced cumulate thicknessesof 7^14 km corresponding to 11^22 km of flood basaltsdepending on the presence of pre-existing oceanic crustand the total crustal thickness (25^35 km) Interpretationsof seismic velocity measurements suggest the presence ofpyroxene and gabbroic cumulates 9^16 km thick beneaththe Ontong Java Plateau (eg Hussong et al 1979)Wrangellia is characterized by high crustal velocities in
seismic refraction lines which is indicative of mafic
plutonic rocks extending to depth beneath VancouverIsland (Clowes et al 1995)Wrangellia crust is 25^30 kmthick beneath Vancouver Island and is underlain by astrongly reflective zone of high velocity and density thathas been interpreted as a major shear zone where lowerWrangellia lithosphere was detached (Clowes et al 1995)The vast majority of the Karmutsen flows are evolved tho-leiitic lavas (eg low MgO high FeOMgO) indicating animportant role for partial crystallization of melts at lowpressure and a significant portion of the crust beneathVancouver Island has seismic properties that are consistentwith crystalline residues from partially crystallizedKarmutsen magmas To test the proportion of the primarymagma that fractionated within the crust MELTS(Ghiorso amp Sack 1995) was used to simulate fractionalcrystallization using several estimated primary magmacompositions from the PRIMELT1 modeling results Themajor-element composition of the primary magmas couldnot be estimated for the tholeiitic basalts because of exten-sive plagthorn cpxthornol fractionation and thus the MELTSmodeling of the picrites is used as a proxy for the evolutionof major elements in the volcanic sequence A pressure of 1kbar was used with variable water contents (anhydrousand 02wt H2O) calculated at the quartz^fayalite^magnetite (QFM) oxygen buffer Experimental resultsfrom Ontong Java indicate that most crystallizationoccurs in magma chambers shallower than 6 km deep(52 kbar Sano amp Yamashita 2004) however some crys-tallization may take place at greater depths (3^5 kbarFarnetani et al 1996)A selection of the MELTS results are shown in Fig 15
Olivine and spinel crystallize at high temperature until15^20wt of the liquid mass has fractionated and theresidual magma contains 9^10wt MgO (Fig 15f) At1235^12258C plagioclase begins crystallizing and between1235 and 11908C olivine ceases crystallizing and clinopyr-oxene saturates (Fig 15f) As expected the addition of clin-opyroxene and plagioclase to the crystallization sequencecauses substantial changes in the composition of the resid-ual liquid FeO and TiO2 increase and Al2O3 and CaOdecrease while the decrease in MgO lessens considerably(Fig 15) The near-vertical trends for the Karmutsen tho-leiitic basalts especially for CaO do not match the calcu-lated liquid-line-of-descent very well which misses manyof the high-CaO high-Al2O3 low-FeO basalts A positivecorrelation between Al2O3CaO and SrNd indicates thataccumulation of plagioclase played a major role in the evo-lution of the tholeiitic basalts (Fig 15g) Increasing thecrystallization pressure slightly and the water content ofthe melts does not systematically improve the match ofthe MELTS models to the observed dataThe MELTS results indicate that a significant propor-
tion of the crystallization takes place before plagioclase(15^20 crystallization) and clinopyroxene (35^45
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
31
crystallization) join the crystallization sequence Theseresults suggest that445 crystallization of the primarymagmas occurred (estimated from the residual liquid percent in Fig 15f) and if the original flood basalt stratigra-phy on Vancouver Island was 6 km thick at least anequivalent amount of complementary plutonic rocksshould have crystallized The Karmutsen basalts and theirplutonic residues thus represent a significant addition ofcrust (perhaps 412 km thickness) previously thickenedduring the formation of the underlying Paleozoic arc
Growth of the Wrangellia oceanic plateauon Vancouver IslandThe extrusive part of the Wrangellia oceanic plateau onVancouver Island was constructed in a three-layersequence of mostly tholeiitic basalts with a restrictedrange of composition (Fig 16) The plateau overliesDevonian to Mississippian arc volcanic sequences andmarine sediments of Mississippian to Permian ages thatare intruded by gabbroic rocks related to the flood basaltsThe volcanic stratigraphy of theWrangellia plateau formed
00
05
10
15
20
25
30
35
40
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
16
0 2 4 6 8 10 12 14 16 18 20
10
11
12
13
14
15
16
17
18
19
20
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
0 2 4 6 8 10 12 14 16 18 20
MgO (wt)
0 10 20 30 40 50
percent of total mass
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
) Mineral proportions
Sp (e)
Ol
CpxPlag
PicriteHigh-MgO basaltTholeiitic basalt
TiO2 (wt)
MgO (wt)
Al2O3 (wt) CaO (wt)
FeO (wt)
MgO(a) (b)
(c) (d)
MgO (wt)MgO (wt)
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
)
Residual liquid ()
1220
1190
Ol + Sp
Plag+Ol+Sp Plag+Cpx+Sp
02 wt H2O
no H2O
Residual liquid and crystallizing phases
(f)
0102030405060708090100
plagioclaseaccumulation
plagioclaseaccumulation
plagioclase
accumulatio
n
10
12
14
16
0 10 20 30 40
Al2O3CaO
SrNd
tholeiitic basalts
(g)
Fig 15 Forward fractional crystallization modeling results for major elements from MELTS (Ghiorso amp Sack 1995) compared with picriticand tholeiitic lavas from the Karmutsen Formation (a^d) Major-element compositions of residual liquids compared with Karmutsen lavacompositions For clarity the results are shown only for one starting composition (open circle) using the composition of the estimated primarymagma for the high-MgO lava series from the modeling technique of Herzberg et al (2007) for sample 4723A13 as shown in Fig 12 The esti-mated primary magmas for samples 4723A3 and 93G171 yield similar results A pressure of 1 kbar was used with no H2O and 02wt H2Ocalculated at QFM black line is anhydrous and grey line is 02wt H2OThe end-point for the anhydrous result is at 78 fractionation and11308C and the end-point for the result with 02wt H2O is 75 fractionation and11108C (e) Mineral proportions of crystallized phases fromone MELTrun (sample 4723A13) with no H2O (f) Per cent residual liquid vs temperature labelled with intervals of crystallizing phases (g)Al2O3CaO vs SrNd for the tholeiitic basalts
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
32
in a single eruptive phase (c 231^225 Ma) as an emergentbasalt sequence during a deeper-water shallow-water andsubaerial stage and subsided below sea level relativelysoon after its formation (Fig 16) The deeper-water stagecontains pillowed and massive submarine flows (3000mthick) that built up the submarine volcanic edifice from theseafloor The lowest pillow basalts were emplaced onunconsolidated and lithified fine-grained siliciclastic andcarbonaceous sediments containing Daonella beds(c 235^232 Ma) that are intruded by abundantKarmutsen mafic sills (Fig 16a) The different pillowedand massive submarine flows resulted primarily from dif-ferent effusive rates and local topography Some of themassive submarine flows commonly recognized by radiat-ing columnar jointing represent master tubes for the deliv-ery of lava to distal parts of submarine flow fields Thepillowed flows which commonly have uneven topography
in modern sequences may have served as levees to themassive submarine flows rather than allowing unob-structed spreading and inflation typical of the massive sub-aerial flows that form on shallow slopes in continentalflood basalt provinces (eg Self et al 1997) The rapid suc-cession of flows is evident by the near absence of sedimentbetween submarine flows Minor volumes of picritic pillowlavas occur in the upper part of the pillowed flow sequencein areas of the plateau exposed on northern VancouverIsland (Fig 16 Nixon et al 2008)The growth of the plateau in shallow water led to
emplacement of an increasing proportion of volcaniclastics(pillow breccias and hyaloclastites) that accumulated up to1500m thick in some areas (Fig 16b Nixon et al 2008)Horizons of pillowed flows occur locally within the volca-niclastic stratigraphy Sedimentary structures (eg gradedbedding fluidization structures) and petrographic textures
Late Devonian-Mississippian arc
Daonella beds
Middle Triassic
deeper water
Late Devonian-Mississippian arc
pillowed flows massive submarine
flows
pillow brecciahyaloclastiteshallow water
~4 km
~6 km
Late Triassic limestoneinterflow limestonelenses
Karmutsen Formation
mafic intrusions sediment-sill complex
Karmutsen Formation
(a) (b)
(c) (d)
subaerial flows
Late Devonian-Mississippian arc Late Devonian-Mississippian arc
picritic pillow basalts
Miss to Permian marine sediments
sills
Fig 16 Growth of the Wrangellia oceanic plateau onVancouver Island (a) Deep-water eruptive phase of pillowed and massive submarineflows with almost no intervening sediments onto MiddleTriassic shale containing Daonella beds (c 235^232 Ma) The basement of the plateauis composed of Devonian to Mississippian arc sequences overlain by Mississippian to Permian carbonates and siliciclastic sediments(b) Shallow-water flow emplacement with increasing proportion of volcaniclastics resulted primarily from cooling^contraction granulationPicritic pillow basalts occur primarily near the transition between pillowed submarine flows and volcaniclastic units Pillowed and massiveflows occur locally within the volcaniclastic unit (c) Emergent stage of growth consists of overlapping subaerial flow fields similar to continen-tal flood basalts with pahoehoe flow structures Multi-tiered columnar jointing occurrs locally in response to ponding of flows (eg Bondre et al2004) Interflow carbonate deposits and plagioclase megacrystic flows developed locally during the waning stage of volcanism (d) Post-volcanicsubsidence and deposition of Late Carnian to Norian (c 212^223 Ma) platform carbonates on top of the plateau Units are labeled in the dia-grams Thicknesses of the underlying units are not to scale
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
33
within the volcaniclastic rocks indicate formation primar-ily via cooling^contraction granulation magma^water^steam interaction autobrecciation resedimentation andmass-wasting rather than pyroclastic eruption (Nixonet al 2008)A broad subaerial platform was constructed by emplace-
ment of subaerial lava flows during an emergent subaerialstage (Fig 16c) The subaerial unit predominantly com-prises well-layered sheet flows that formed in a similarway to continental flood basalts as hundred(s) of pahoehoeflow fields erupted on shallow slopes (Self et al 1997) how-ever volcaniclastic and pillowed flow units are locally pre-sent mostly in the upper parts of the stratigraphyLocalized ponding of flows led to the formation of multi-tiered columnar jointing in some subaerial flows (egBondre et al 2004) As volcanism waned local interflowcarbonate deposits developed and plagioclase megacrysticflows were erupted Subsidence of the plateau is recordedby the interflow lenses and deposits of platform carbonates(40^1500m thick) on top of the Karmutsen basalts thatcontain Late Carnian to Early Norian fossils (Fig 16d c216^227 Ma Nixon et al 2008)The evolution of the internal architecture of the
Wrangellia oceanic plateau was primarily related to theconditions of emplacement (ie water depth and eruptionenvironment) There does not appear to have been a sys-tematic coincident evolution or shift of lava compositionsonVancouver Island as has been recognized in most con-tinental flood basalts and parts of the Wrangellia oceanicplateau in Alaska and Yukon (Greene et al 2008 2009)In general with the exception of the picritic pillow lavasin the submarine stratigraphy and local variations betweenhigh-MgO and tholeiitic lavas there are no significantcompositional differences between lavas erupted duringthe different stages of growth of the plateau (ie submarineand subaerial flows) nor are substantial compositional dif-ferences generally found between massive submarine flowsin contact with pillowed flows tholeiitic basalt flows nearthe base of the Karmutsen Formation are compositionallysimilar to flows sampled near the top of the volcanicsequence Where it was possible to sample continuous sec-tions of volcanic stratigraphy greater than several hun-dreds of meters (eg Keogh^Maynard Lake area west ofthe Karmutsen Range Fig 2) there is no clear relationshipbetween stratigraphic position and compositionalvariation
CONCLUSIONSThe Karmutsen Formation covers extensive areas ofVancouver Island (20 000 km2) and forms part of a largeoceanic plateau erupted over a geologically short interval(c 225^231 Ma) The tripartite volcanic stratigraphyexposed on Vancouver Island is at least 6 km thick andcomprises a succession of submarine flows volcaniclastic
deposits and massive subaerial flowsThese volcanologicaldifferences are primarily related to the eruption environ-ment (deep-water shallow-water or subaerial) The rapidgrowth of the plateau prevented significant accumulationsof intervening sediment except in the uppermost stratigra-phy where isolated limestone lenses commonly associatedwith pillow basalt and hyaloclastite deposits preserve arecord of the final stages of subsidence of the plateauThe Wrangellia plateau on Vancouver Island was con-
structed dominantly of tholeiitic basalt with restrictedmajor- and trace-element and isotopic compositional var-iations Minor volumes of picritic pillow basalts eruptedlate in the submarine phase The Karmutsen basalts wereemplaced onto older Paleozoic arc volcanic sequences andlow NbLa for the picrites may indicate a contributionfrom the sub-arc lithospheric mantle Modeling of the pet-rogenesis of the Karmutsen picrites indicates melting ofanomalously hot mantle (15008C) and extensive degreesof partial melting (23^27) and is consistent with aplume initiation model Combined Sr^Nd^Hf^Pb isotopicsystematics indicate that the picritic and tholeiitic lavaswere derived from a depleted plume-type mantle sourcedistinct from the source of MORB that has compositionalsimilarities to the source of the lavas forming theCaribbean Plateau Picrites were generated by melting ofspinel lherzolite whereas the tholeiitic basalts formed asaggregate melts of both garnet and spinel lherzolite Thetholeiitic basalts underwent extensive low-pressure fractio-nation (52^3 kbar) and seismic studies indicate that someof theWrangellia crust beneathVancouver Island may cor-respond to the plutonic residues of the Karmutsen floodbasalts
ACKNOWLEDGEMENTSWe are grateful to Nick Arndt for helping us get this pro-ject started and Nick Massey for insights into VancouverIsland geology We greatly appreciate the assistance withPRIMELT modeling from Claude Herzberg We thankMikkel Schau for his insight and enthusiasm during field-work Jane Barling assisted with analyses by MC-ICP-MSThoughtful and thorough reviews of the manuscript byJohn Mahoney Fred Frey Wendy Bohrson DougalJerram and Teal Riley are greatly appreciated Fundingwas generously provided by the Rocks to Riches Programadministered by the BC amp Yukon Chamber of Mines in2004 by the BC Geological Survey and by NSERCDiscovery Grants to JSS and DW ARG was supportedby a University Graduate Fellowship at UBC
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
34
REFERENCESBarker F Brown A S Budahn J R amp Plafker G (1989) Back-arc
with frontal-arc component origin of Triassic Karmutsen basaltBritish Columbia Canada Chemical Geology 75 81^102
Blichert-Toft JWeis D Maerschalk C Agranier A amp Albare de F(2003) Hawaiian hot spot dynamics as inferred from the Hf and Pbisotope evolution of Mauna Kea volcano Geochemistry GeophysicsGeosystems 4(2) 1^27 doi1010292002GC000340
Bondre N R Duraiswami R A amp Dole G (2004) Morphologyand emplacement of flows from the Deccan Volcanic ProvinceIndia Bulletin of Volcanology 66 29^45
Carlisle D (1963) Pillow breccias and their aquagene tuffs QuadraIsland British Columbia Journal of Geology 71 48^71
Carlisle D (1972) Late Paleozoic to Mid-Triassic sedimentary-volcanic sequence on Northeastern Vancouver Island Geological
Society of Canada Paper Report of Activities 72-1 Part B 24^30Carlisle D amp SuzukiT (1974) Emergent basalt and submergent car-
bonate^clastic sequences including the Upper Triassic Dilleri andWelleri zones on Vancouver Island Canadian Journal of Earth
Sciences 11 254^279Chauvel C amp Blichert-Toft J (2001) A hafnium isotope and trace
element perspective on melting of the depleted mantle Earth and
Planetary Science Letters 190 137^151Chayes F (1954)The theory of thin-section analysis Journal of Geology
62 92^101Cho M Liou J G amp Maruyama S (1987) Transition from the zeo-
lite to prehnite^pumpellyite facies in the Karmutsen metabasitesVancouver Island British Columbia Journal of Petrology 27(2)467^494
Clowes R M Zelt C A Amor J R amp Ellis R M (1995)Lithospheric structure in the southern Canadian Cordillera froma network of seismic refraction lines Canadian Journal of Earth
Sciences 32(10) 1485^1513Coffin M F amp Eldholm O (1994) Large igneous provinces Crustal
structure dimensions and external consequences Reviews of
Geophysics 32(1) 1^36Cox K G (1980) A model for flood basalt volcanism Journal of
Petrology 21 629^650Eldholm O amp Coffin M F (2000) Large igneous provinces
and plate tectonics In Richards M A Gordon R G amp vander Hilst R D (eds) The History and Dynamics of Global
Plate Motions American Geophysical Union Geophysical Monograph 121309^326
Elliott T R Hawkesworth C J amp Grolaquo nvold K (1991) Dynamicmelting of the Iceland plume Nature 351 201^206
Escuder-Viruete J Pecurren rez-Estaucurren n A Contreras F Joubert MWeis D Ullrich T D amp Spadea P (2007) Plume mantle sourceheterogeneity through time Insights from the Duarte ComplexHispaniola northeastern Caribbean Journal of Geophysical Research112(B04203) doi1010292006JB004323
Farnetani C G Richards M A amp Ghiorso M S (1996)Petrological models of magma evolution and deep crustal structurebeneath hotspots and flood basalt provinces Earth and Planetary
Science Letters 143 81^94Fitton J G amp Godard M (2004) Origin and evolution of magmas
on the Ontong Java Plateau In Fitton J G Mahoney J JWallace P J amp Saunders A D (eds) Origin and Evolution of the
Ontong Java Plateau Geological Society London Special Publications 229151^178
Fitton J G Saunders A D Norry M J Hardarson B S ampTaylor R N (1997) Thermal and chemical structure of theIceland plume Earth and Planetary Science Letters 153 197^208
Fitton J G Saunders A D Kempton P D amp Hardarson B S(2003) Does depleted mantle form an intrinsic part of the Icelandplume Geochemistry Geophysics Geosystems 4(3) 1032 doi1010292002GC000424
Frey F A Huang S Blichert-Toft J Regelous M amp Boyet M(2005) Origin of depleted components in basalt related to theHawaiian hotspot Evidence from isotopic and incompatible ele-ment ratios Geochemistry Geophysics Geosystems 6(1) 23 doi1010292004GC000757
Galer S J G amp Abouchami W (1998) Practical application of leadtriple spiking for correction of instrumental mass discriminationMineralogical Magazine 62A 491^492
Ghiorso M S amp Sack R O (1995) Chemical mass transfer in mag-matic processes IV A revised and internally consistent thermody-namic model for the interpolation and extrapolation of liquid^solidequilibria in magmatic systems at elevated temperatures and pres-sures Contributions to Mineralogy and Petrology 119 197^212
Greene A R Scoates J S amp Weis D (2005)WrangelliaTerrane onVancouver Island Distribution of flood basalts with implicationsfor potential Ni^Cu^PGE mineralization In Grant B (ed)Geological Fieldwork 2004 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2005-1 209^220Greene A R Scoates J S Nixon G T amp Weis D (2006) Picritic
lavas and basal sills in the Karmutsen flood basalt provinceWrangellia northern Vancouver Island In Grant B (ed)Geological Fieldwork 2005 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2006-1 39^52Greene A R Scoates J S amp Weis D (2008) Wrangellia flood
basalts in Alaska A record of plume^lithosphere interaction in aLate Triassic accreted oceanic plateau Geochemistry Geophysics
Geosystems 9(Q12004) doi1010292008GC002092Greene A R Scoates J S Weis D amp Israel S (2009)
Geochemistry of triassic flood basalts from the Yukon (Canada)segment of the accreted Wrangellia oceanic plateau Lithosdoi101016jlithos200811010
Hauff F Hoernle K Tilton G Graham D W amp Kerr A C(2000) Large volume recycling of oceanic lithosphere over shorttime scales geochemical constraints from the Caribbean LargeIgneous Province Earth and Planetary Science Letters 174 247^263
Hauff F Hoernle K amp Schmidt A (2003) Sr^Nd^Pb compositionof Mesozoic Pacific oceanic crust (Site 1149 and 801 ODP Leg185)Implications for alteration of ocean crust and the input into theIzu^Bonin^Mariana subduction system Geochemistry Geophysics
Geosystems 4(8) doi1010292002GC000421Herzberg C (2004) Partial melting below the Ontong Java Plateau
In Fitton J G Mahoney J J Wallace P J amp Saunders A D(eds) Origin and Evolution of the Ontong Java Plateau Geological SocietyLondon Special Publications 229 179^183
Herzberg C (2006) Petrology and thermal structure of the Hawaiianplume from Mauna Kea volcano Nature 444(7119) 605
Herzberg C amp Asimow P D (2008) Petrology of some oceanicisland basalts PRIMELT2XLS software for primary magma cal-culation Geochemistry Geophysics Geosystems 9(Q09001) doi1010292008GC002057
Herzberg C amp OrsquoHara M J (2002) Plume-associated ultramaficmagmas of Phanerozoic age Journal of Petrology 43(10) 1857^1883
Herzberg C Asimow P D Arndt N Niu Y Lesher C MFitton J G Cheadle M J amp Saunders A D (2007)Temperatures in ambient mantle and plumes Constraints frombasalts picrites and komatiites Geochemistry Geophysics Geosystems8(Q02006) doi1010292006GC001390
Huang S amp Frey F A (2005) Trace element abundances of MaunaKea basalt from phase 2 of the Hawaii Scientific Drilling Project
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
35
Petrogenetic implications of correlations with major element con-tent and isotopic ratios Geochemistry Geophysics Geosystems 4(6)doi1010292002GC000322
Hussong D M Wipperman L K amp Kroenke L W (1979) Thecrustal structure of the Ontong Java and Manihiki OceanicPlateaus Journal of Geophysical Research 84(B11) 6003^6010
Jerram D A amp Widdowson M (2005) The anatomy of ContinentalFlood Basalt Provinces geological constraints on the processes andproducts of flood volcanism Lithos 79(3^4) 385^405
Jones D L Silberling N J amp Hillhouse J (1977)Wrangellia a dis-placed terrane in northwestern North America CanadianJournal ofEarth Sciences 14(11) 2565^2577
Kerr A C (2003) Oceanic Plateaus In Rudnick R (ed) The Crust
Treatise on GeochemistryVol 3 Oxford Elsevier Science pp 537^565Kerr A C amp Mahoney J J (2007) Oceanic plateaus Problematic
plumes potential paradigms Chemical Geology 241 332^353Kerr A CTarney J Marriner G F Klaver GT Saunders A D
amp Thirlwall M F (1996) The geochemistry and petrogenesis ofthe Late-Cretaceous picrites and basalts of Curacao NetherlandsAntilles a remnant of an oceanic plateau Contributions to
Mineralogy and Petrology 124(1) 29^43Kerr A C Marriner G F Tarney J Nivia A Saunders A D
Thirlwall M F amp Sinton A C (1997) Cretaceous basaltic ter-ranes in Western Colombia Elemental chronological and Sr-Ndisotopic constraints on petrogenesis Journal of Petrology 38(6)677^702
Kinzler R J (1997) Melting of mantle peridotite at pressuresapproaching the spinel to garnet transition Application to mid-ocean ridge basalt petrogenesis Journal of Geophysical Research
102(B1) 853^874Kinzler R J amp Grove T L (1992) Primary magmas of mid-ocean
ridge basalts 1 Experiments and methods Journal of Geophysical
Research 97(B5) 6885^6906Korenaga J (2005) Why did not the Ontong Java Plateau form sub-
aerially Earth and Planetary Science Letters 234 385^399Kuniyoshi S (1972) Petrology of the Karmutsen (Volcanic) Group
northeastern Vancouver Island British Columbia PhD disserta-tion University of California Los Angeles 242 pp
Lassiter J C DePaolo D J amp Mahoney J J (1995) Geochemistry ofthe Wrangellia flood basalt province Implications for the role ofcontinental and oceanic lithosphere in flood basalt genesis Journalof Petrology 36(4) 983^1009
LeBas M J (2000) IUGS reclassification of the high-Mg and picriticvolcanic rocks Journal of Petrology 41(10) 1467^1470
Longhi J (2002) Some phase equilibrium systematics of lherzolitemelting I Geochemistry Geophysics Geosystems 3(3) doi1010292001GC000204
MacDonald G A amp Katsura T (1964) Chemical composition ofHawaiian lavas Journal of Petrology 5 82^133
Mahoney J J (1987) An isotopic survey of Pacific oceanic plateausimplications for their nature and origin In Keating B HFryer P Batiza R amp Boehlert GW (eds) Seamounts Islands andAtolls American Geophysical Union Geophysical Monograph 43 207^220
Mahoney J LeRoex A P Peng Z Fisher R L amp Natland J H(1992) Southwestern limits of Indian Ocean Ridge mantle and theorigin of low 206Pb204Pb mid-ocean ridge basalt isotope systema-tics of the central Southwest Indian Ridge (178^508E) Journal ofGeophysical Research 97(B13) 19771^19790
Mahoney J J Sinton J M Kurz M D MacDougall J DSpencer K J amp Lugmair GW (1994) Isotope and trace elementcharacteristics of a super-fast spreading ridge East Pacific rise13^238S Earth and Planetary Science Letters 121 173^193
Massey N W D (1995a) Geology and mineral resources of theAlberni^Nanaimo Lakes sheet Vancouver Island 92F1W 92F2Eand part of 92F7E British Columbia Ministry of Energy Mines and
Petroleum Resources Paper 1992-2 132 ppMassey N W D (1995b) Geology and mineral resources of the
Cowichan Lake sheet Vancouver Island 92C16 British Columbia
Ministry of Energy Mines and Petroleum Resources Paper 1992-3 112 ppMassey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005a) Digital Geology Map of British Columbia Tile NM9 MidCoast BC British Columbia Ministry of Energy and Mines Geofile
2005-2Massey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005b) Digital Geology Map of British Columbia Tile NM10Southwest BC British Columbia Ministry of Energy and Mines Geofile
2005-3McDonough W F amp Sun S (1995) The composition of the Earth
Chemical Geology 120 223^253Monger JW H amp Journeay J M (1994) Basement geology and tec-
tonic evolution of theVancouver region In Monger JW H (ed)Geology and Geological Hazards of the Vancouver Region Southwestern
British Columbia Geological Survey of Canada Bulletin 481 3^25Muller J E (1977) Geology of Vancouver Island Geological Survey of
Canada Open File Map 463Muller J E (1980)The Paleozoic Sicker Group of Vancouver Island British
Columbia Geological Survey of Canada Paper 79-30 22 ppMuller J E Northcote K E amp Carlisle D (1974) Geology and
mineral deposits of Alert Bay^Cape Scott map area VancouverIsland British Columbia Geological Survey of Canada Paper 74-8 77
Neal C R Mahoney J J Kroenke L W Duncan R A ampPetterson M G (1997) The Ontong^Java Plateau InMahoney J J amp Coffin M F (eds) Large Igneous Provinces
Continental Oceanic and Planetary FloodVolcanism American Geophysical
Union Geophysical Monograph 100 183^216Niu Y Collerson K D Batiza R Wendt J I amp Regelous M
(1999) Origin of enriched-typed mid-ocean ridge basalt at ridgesfar from mantle plumes The East Pacific Rise at 11820rsquoN Journalof Geophysical Research 104 7067^7088
Nixon G T amp Orr A J (2007) Recent revisions to the EarlyMesozoic stratigraphy of Northern Vancouver Island (NTS 102I092L) and metallogenic implicatinos British Columbia InGrant B (ed) Geological Fieldwork 2006 British Columbia Ministry of
Energy Mines and Petroleum Resources Paper 2007-1 163^177Nixon GT Hammack J L KoyanagiV M Payie G J Haggart
JW Orchard M JTozerT Friedman R M Archibald D APalfy J amp Cordey F (2006a) Geology of the Quatsino^PortMcNeill area northernVancouver Island British Columbia Ministry
of Energy Mines and Petroleum Resources Geoscience Map 2006-2 scale150 000
Nixon G T Kelman M C Stevenson D Stokes L A ampJohnston K A (2006b) Preliminary geology of the Nimpkishmap area (NTS 092L07) northern Vancouver Island BritishColumbia In Grant B amp Newell J M (eds) Geological Fieldwork2005 British Columbia Ministry of Energy Mines and Petroleum Resources
Paper 2006-1 135^152Nixon G T Laroque J Pals A Styan J Greene A R amp
Scoates J S (2008) High-Mg lavas in the Karmutsen flood basaltsnorthernVancouver Island (NTS 092L) Stratigraphic setting andmetallogenic significance In Grant B (ed) Geological Fieldwork
2007 British Columbia Ministry of Energy Mines and Petroleum
Resources Paper 2008-1 175^190Nowell G M Kempton P D Noble S R Fitton J G
Saunders A Mahoney J J amp Taylor R N (1998) High precisionHf isotope measurements of MORB and OIB by thermal ionisation
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
36
mass spectrometry insights into the depleted mantle Chemical
Geology 149 211^233Parrish R R amp McNicoll V J (1992) U^Pb age determinations
from the southern Vancouver Island area British Columbia InRadiogenic Age and Isotopic Studies Report 5 Geological Survey of
Canada Paper 91-2 79^86Peate DW (1997) The Parana^Etendeka Province In Coffin M F
amp Mahoney J (eds) Large Igneous Provinces Continental Oceanic andPlanetary Flood Volcanism American Geophysical Union Geophysical
Monograph 100 217^245Perfit M R Fornari D J Ridley W I Kirk P D Casey J
Kastens K A Reynolds J R Edwards M Desonie DShuster R amp Paradis S (1996) Recent volcanism in the Siqueirostransform fault picritic basalts and implications for MORBmagma genesis Earth and Planetary Science Letters 141(1^4) 91^108
Pretorius W Weis D Williams G Hanano D Kieffer B ampScoates J S (2006) Complete trace elemental characterization ofgranitoid (USGSG-2GSP-2) reference materials by high resolutioninductively coupled plasma-mass spectrometry Geostandards and
Geoanalytical Research 30(1) 39^54Regelous M Niu Y Wendt J I Batiza R Greig A amp
Collerson K D (1999) Variations in the geochemistry of magma-tism on the East Pacific Rise at 10830rsquoN since 800 ka Earth and
Planetary Science Letters 168 45^63Recurren villon S Chauvel C Arndt N T Pik R Martineau F
Fourcade S amp Marty B (2002) Heterogeneity of the Caribbeanplateau mantle source Sr O and He isotopic compositions of oli-vine and clinopyroxene from Gorgona Island Earth and Planetary
Science Letters 205(1^2) 91Rhodes J M (1996) Geochemical stratigraphy of lava flows samples
by the Hawaii Scientific Drilling Project Journal of Geophysical
Research 101(B5) 11729^11746Rhodes J M amp Vollinger M J (2004) Composition of basaltic lavas
sampled by phase-2 of the Hawaii Scientific Drilling ProjectGeochemical stratigraphy and magma types Geochemistry
Geophysics Geosystems 5(3) doi1010292002GC000434Richards M A Jones D L Duncan R A amp DePaolo D J (1991)
A mantle plume initiation model for the Wrangellia flood basaltand other oceanic plateaus Science 254 263^267
Salters V J M amp Hart S R (1991) The mantle sources of oceanridges islands and arcs the Hf-isotope connection Earth and
Planetary Science Letters 104 364^380Salters V J M amp Stracke A (2004) Composition of the depleted
SaltersV J amp WhiteW (1998) Hf isotope constraints on mantle evo-lution Chemical Geology 145 447^460
SanoT amp Yamashita S (2004) Experimental petrology of basementlavas from Ocean Drilling Program Leg 192 implications for dif-ferentiation processes in Ontong Java Plateau magmas InFitton J G Mahoney J JWallace P J amp Saunders A D (eds)Origin and Evolution of the Ontong Java Plateau Geological Society
London Special Publications 229 185^218Saunders A D (2005) Large igneous provinces Origin and environ-
mental consequences Elements 1 259^263Self S ThordarsonT amp Keszthely L (1997) Emplacement of conti-
nental flood basalt lava flows In Mahoney J J amp Coffin M F(eds) Large Igneous Provinces Continental Oceanic and Planetary Flood
Volcanism American Geophysical Union Geophysical Monograph 100381^410
Shaw D M (2000) Continuous (dynamic) melting theory revisitedCanadian Mineralogist 38(5) 1041^1063
Sluggett C L (2003) Uranium^lead age and geochemical constraintson Paleozoic and Early Mesozoic magmatism in WrangelliaTerrane Saltspring Island British Columbia BSc thesisUniversity of British ColumbiaVancouver 84 pp
Storey M Mahoney J J amp Saunders A D (1997) Cretaceousbasalts in Madagascar and the transition between plume and con-tinental lithospheric mantle sources In Mahoney J J ampCoffin M F (eds) Large Igneous Provinces Continental Oceanic and
Planetary Flood Volcanism Aamerican Geophysical Union Geophysical
Monograph 100 95^122Surdam R C (1967) Low-grade metamorphism of the Karmutsen
Group PhD dissertation University of California Los Angeles288 pp
Sutherland-Brown A Yorath C J Anderson R G amp Dom K(1986) Geological maps of southern Vancouver IslandLITHOPROBE I 92C10 11 14 16 92F1 2 7 8 Geological Survey ofCanada Open File 1272
Tejada M L G Mahoney J J Castillo P R Ingle S P Sheth HC amp Weis D (2004) Pin-pricking the elephant evidence on theorigin of the Ontong Java Plateau from Pb^Sr^Hf^Nd isotopiccharacteristics of ODP Leg 192 basalts In Fitton J GMahoney J J Wallace P J amp Saunders A D (eds) Origin and
Evolution of the Ontong Java Plateau Geological Society London Special
Publications 229 133^150Thompson P M E Kempton P D amp Kerr A C (2008) Evaluation
of the effects of alteration and leaching on Sm^Nd and Lu^Hf sys-tematics in submarine mafic rocks Lithos 104(1^4) 164^176
Thompson P M E Kempton P D White R V Kerr A CTarney J Saunders A D Fitton J G amp McBirney A (2003)Hf-Nd isotope constraints on the origin of the CretaceousCaribbean plateau and its relationship to the Galapagos plumeEarth and Planetary Science Letters 217(1^2) 59^75
Vervoort J D Patchett P Blichert-Toft J amp Albare de F (1999)Relationships between Lu^Hf and Sm^Nd isotopic systems in theglobal sedimentary system Earth and Planetary Science Letters 16879^99
Walter M J (1998) Melting of garnet peridotite and the origin ofkomatiite and depleted lithosphere Journal of Petrology 39(1) 29^60
Weis D Kieffer B Maerschalk C Barling J de Jong JWilliams G A Hanano D Mattielli N Scoates J SGoolaerts A Friedman R A amp Mahoney J B (2006) High-pre-cision isotopic characterization of USGS reference materials byTIMS and MC-ICP-MS Geochemistry Geophysics Geosystems
7(Q08006) doi1010292006GC001283Weis D Kieffer B Hanano D Silva I N Barling J
Pretorius W Maerschalk C amp Mattielli N (2007) Hf isotopecompositions of US Geological Survey reference materialsGeochemistry Geophysics Geosystems 8(Q06006) doi1010292006GC001473
Wheeler J O amp McFeely P (1991) Tectonic assemblage map of theCanadian Cordillera and adjacent part of the United States ofAmerica Geological Survey of Canada Map 1712A
WhiteW M Albare de F amp Tecurren louk P (2000) High-precision analy-sis of Pb isotope ratios by multi-collector ICP-MS Chemical Geology167 257^270
Yorath C J Sutherland Brown A amp Massey N W D (1999)LITHOPROBE southern Vancouver Island British Columbia Geological
Survey of Canada Bulletin 498 145 ppZou H (1998) Trace element fractionation during modal and
non-model dynamic melting and open-system melting Amathematical treatment Geochimica et Cosmochimica Acta 62(11)1937^1945
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
37
Zou H amp Reid M R (2001) Quantitative modeling of trace elementfractionation during incongruent dynamic melting Geochimica et
Cosmochimica Acta 65(1) 153^162
APPENDIX SAMPLEPREPARAT ION ANDANALYT ICAL METHODSOnly the freshest rocks in the field were sampled and onlythe least altered samples were selected for chemical andisotopic analysis based on thorough petrographic inspec-tion Fifty-nine of the 129 collected samples were crushed(400 g) into small pieces 52mm in diameter in aRocklabs hydraulic piston crusher between WC plates tominimize contamination The coarse-crush was mixedand 100 g was powdered in a planetary mill using agatejars and balls cleaned with quartz sand between samples
ActLabs analytical methodsThe major- and trace-element compositions of the whole-rock powders were determined at Activation LaboratoriesLtd (Actlabs) in Ancaster Ontario Analytical techniquesand detection limits are also available from Actlabs (httpwwwactlabscommethsub_code4erehtm) The analyticalmethod for each of the elements analyzed is indicated inTable 2 For the major elements a 02 g sample was mixedwith a lithium metaborate^lithium tetraborate mixtureand fused in a graphite crucible The molten mixture waspoured into a 5 HNO3 solution and shaken until dis-solved (30min) The samples were analyzed for majoroxides and selected trace elements by inductively coupledplasma optical emission spectrometry (ICP-OES) on acombination simultaneous^sequential Thermo Jarrell-AshEnviro II system Internal calibration was achieved usinga variety of international reference materials (eg W-2BIR-1 DNC-1) and independent control samplesAdditional trace elements were analyzed by both instru-mental neutron activation analysis (INAA) and induc-tively couple plasma mass spectrometry (ICP-MS) ForINAA15^25 g of sample was weighed into small polyeth-ylene vials and irradiated with a control international ref-erence material CANMET WMS-1 and NiCr flux wires ata thermal neutron flux of 71012 neutronscm2 per s in theMcMaster Nuclear Reactor Following a 7 day waitingperiod the samples were measured on an Ortec high-purity Ge detector linked to a Canberra Series 95 multi-channel analyzer Activities for each element weredecay- and weight-corrected and compared with a detectorcalibration developed from multiple international certifiedreference materials For ICP-MS 025 g of sample wasdigested in HF followed by HNO3^HClO4 heated andtaken to dryness The samples were brought back into solu-tion with HCl Samples were analyzed using a PerkinElmer Optima 3000 ICP In-lab standards or certified ref-erence materials (eg W-2 BIR-1 DNC-1) were used for
quality control A total of 15 blind duplicates were ana-lyzed from the same sample powders to assessreproducibility
University of Massachusetts analyticalmethodsFourteen sample duplicate powders taken from the same100 g sample aliquot (high-MgO basalts and picrites)were also analyzed at the Ronald B Gilmore X-RayFluorescence (XRF) Laboratory at the University ofMassachusetts Major elements were measured on a fusedLa-bearing lithium borate glass disc using a SiemensMRS-400 spectrometer Trace element concentrations(Rb Sr Ba Ce Nb Zr Y Pb Zn Ga Ni Cr V) weremeasured on a separate powder pellet using a PhilipsPW2400 sequential spectrometer using a Rh tube LOIand ferrous iron measurements were made as describedby Rhodes amp Vollinger (2004) Precision and accuracy esti-mates for the data have been described by Rhodes (1996)and Rhodes amp Vollinger (2004) Results for each sampleare the average of two separate analyses and are shown inElectronic Appendix 2 for a comparison of analyses ofduplicate powders using XRFand ICP-MS
PCIGR analytical methodsA subset of 19 samples was selected for high-precisiontrace-element analysis and Sr Nd Pb and Hf isotopicanalysis at the Pacific Centre for Isotopic andGeochemical Research (PCIGR) at the University ofBritish Columbia (UBC) Samples were selected from the59 samples analyzed for whole-rock chemistry at ActLabsbased on major- and trace-element chemistry alteration(low LOI and petrographic alteration index) and samplelocation Samples were prepared for trace-element analysisat the PCIGR by the technique described by Pretorius et al(2006) on unleached rock powders Sample powders(100mg) were weighed in 7ml screw-top Savillex bea-kers and dissolved in 1ml 14N HNO3 and 5ml 48 HFon a hotplate for 48 h at 1308C with periodic ultrasonica-tion Samples were dried and redissolved in 6ml 6N HClon a hotplate for 24 h and then dried and redissolved in1ml concentrated HNO3 for 24 h before final dryingTrace element abundances were measured by high-resolution inductively coupled plasma-mass spectrometry(HR-ICP-MS) with aThermo Finnigan Element2 systemfollowing the procedures described by Pretorius et al(2006) within 24 h of redissolution HFSE and LILE weremeasured in medium-resolution mode at 2000 dilutionusing a PFATeflon spray chamber washed with aqua regiafor 3min between samples REE were measured in high-resolution mode and U and Pb in low-resolution modeat 2000 dilution using a glass spray chamber washedwith 2 HNO3 between samples Total procedural blanksand reference materials (BCR-2 BHVO-2) were analyzedwith the batch of samples Indium was used as an internal
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
38
standard in all samples and standard solutionsBackground and standard solutions were analyzed afterevery five samples to detect memory effects and massdrift Results for PCIGR trace-element analyses areshown in Electronic Appendix 3 for the full set of analyzedtrace elementsSample digestion for purification of Sr Nd Hf and Pb
for column chemistry began by weighing each samplepowder (400^500mg) prior to leaching All samples wereleached with 6N HCl and placed in an ultrasonic bath for15min Samples were rinsed twice with 18 M-cm H2Obetween each leaching step (15 total) until the supernatantliquid was clear [following the method of Mahoney (1987)]Samples were then dried on a hotplate for 24 h andweighed again Sample solutions were then prepared bydissolving 100^250mg of the leached powder dissolvedin 1ml 14N HNO3 and 10ml 48 HF on a hotplate for48 h at 1308C with periodic ultrasonication Samples weredried and redissolved in 6ml 6N HCl on a hotplate for24 h and then dried Pb was separated using anion exhangecolumns and the discard was used for Sr REE and Hfseparation Nd was separated from the other REE and Hfrequired two additional purification steps Detailed proce-dures for column chemistry for separating Sr Nd and Pbat the PCIGR have been described byWeis et al (2006) andHf purification has been described byWeis et al (2007) Srand Nd isotope ratios were measured by thermal ioniza-tion mass spectrometry (TIMS) on a Thermo FinniganTriton system in static mode with relay matrix rotation ona single Ta and double Re^Ta filament respectively Fourto five filaments per barrel of 21 were occupied by stan-dards (NBS 987 for Sr and LaJolla for Nd) for each barrelwhere samples were run Sample Sr and Nd isotopic com-positions were corrected for mass fractionation using86Sr88Srfrac14 01194 and 146Nd144Ndfrac14 07219 Each samplewas then normalized using the barrel average of the refer-ence material relative to the values of143Nd144Ndfrac14 0511858 and 87Sr86Srfrac14 0710248 (Weiset al 2006) During the course of the Vancouver Islandanalyses the La Jolla Nd standard gave an average valueof 05118566 (nfrac14 7) and NBS987 standard gave an aver-age of 07102408 (nfrac1411) (2 error is reported as times106) 147Sm144Nd errors are 15 or 0006 USGeological Survey (USGS) reference material BHVO-2was processed with the samples and yielded Sr and Nd iso-topic ratios of 07034607 and 05129786 respectivelyThese are in agreement with the published values of070347920 and 051298411 respectively (Weis et al2006)
Pb and Hf isotopic compositions were analyzed by mul-tiple collector-inductively coupled plasma-mass spectrome-try (MC-ICP-MS) on a static multi-collection on a NuPlasma (Nu Instruments) system The detailed analyticalprocedure for Pb isotopic analyses on the Nu Plasma atthe PCIGR has been described by Weis et al (2006) Theconfiguration for Pb analyses allows for collection of PbTl and Hg togetherTl and Hg are used to monitor instru-mental mass discrimination and isobaric overlap repec-tively All sample solutions were analyzed withapproximately the same PbTl ratio (4) as the referencematerial NIST SRM 981 To accomplish this a small ali-quot of each sample solution from the Pb columns was ana-lyzed on the Element2 to determine the precise amount ofPb available for analysis on the Nu Plasma The SRM 981standard was run after every two samples on the NuPlasma During the time samples were run analyses ofthe SRM 981 Pb reference material gave values of206Pb204Pbfrac1416940322 207Pb204Pbfrac1415495823and 208Pb204Pbfrac14 36713164 (nfrac14 61 2 error is reportedas times 104) these values are in excellent agreement withreported TIMS triple-spike values of Galer amp Abouchami(1998) Fractionation-corrected Pb isotopic ratios were fur-ther corrected by the sample-standard bracketing methodor the ln^ln correction method described by White et al(2000) and Blichert-Toft et al (2003) Leached powders ofUSGS reference material BHVO-2 yielded Pb isotopicratios of 206Pb204Pbfrac14186454 8 207Pb204Pbfrac14 1549105 and 208Pb204Pbfrac14 38222514 and BCR-2yielded 206Pb204Pbfrac141880466 207Pb204Pbfrac14 1562518 and 208Pb204Pbfrac14 3883496 (2 error isreported as times 104) These values are in agreement withthose for leached residues of BHVO-2 and BCR-2 fromWeis et al (2006)Hf isotopic compositions were analyzed following the
procedures detailed byWeis et al (2007) The configurationfor Hf analyses monitored Lu mass 175 andYb mass 172 toallow for interference correction to masses 174 and 176 Hfisotopic ratios were normalized internally for mass fractio-nation to a 179Hf177Hf ratio of 07325 using an exponentialcorrection Standards were run after every two samplesand sample results were normalized to the ratio of the in-run daily average and a 176Hf177Hf ratio for JMC-475 of0282160 During the course of analyses the Hf standardJMC-475 gave an average value 02821679 (nfrac14130)USGS reference materials BCR-2 and BHVO-2 were pro-cessed with the samples and yielded Hf isotopic ratios of02828675 and 02831005 respectively Publishedvalues for BCR-2 and BHVO-2 are 02828717 and0283104 8 respectively (Weis et al 2007)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
39
and massive submarine flows a middle unit of mostlypillow breccia and hyaloclastite and an upper unit of pre-dominantly massive subaerial flows (Carlisle amp Suzuki1974) The pillow basalts directly overlie Middle Triassicand late Paleozoic marine sedimentary sequences that areintruded by Karmutsen mafic sills The boundary betweenpillow breccia^hyaloclastite and massive lava flows repre-sents the transition from a submarine to a subaerial erup-tive environment The uppermost flows of the Karmutsenare intercalated with and overlain by shallow-water lime-stone and local occurrences of pillowed flows and hyalo-clastite deposits occur within the upper subaerial memberThe Karmutsen Formation and Wrangellia on Vancouver
Island have been the focus of detailed mapping and strati-graphic studies by Carlisle (Carlisle 1963 1972 Carlisle ampSuzuki 1974) and Muller (Muller et al 1974 Muller 1977)Recent descriptions of the Karmutsen Formation on north-ern Vancouver Island have been made during regionalmapping studies (150 000 scale) by Nixon and coworkers(Nixon et al 2006b 2008 Nixon amp Orr 2007)
Age of the Karmutsen FormationThe age and duration of Karmutsen volcanism are con-strained by fossils in the underlying and overlying sedi-mentary units and by three U^Pb isotopic age
Fig 1 Simplified map of Vancouver Island showing the distribution of the Karmutsen Formation and underlying Paleozoic formations (afterMassey et al 2005a 2005b) Areas in white are mostly younger units The main areas of field study are indicated with boxes or circles withcapital letters (see legend) Ocean is light blue The inset shows the extent of theWrangellia flood basalts (green) in British Columbia Yukonand Alaska
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
3
determinations on intrusive rocks that are considered to berelated to the Karmutsen volcanics Daonella-bearing shale100^200m below the base of the pillow basalts on SchoenMountain and Halobia-rich shale interlayered with flows inthe upper part of the Karmutsen indicate eruption of mostof the flood basalts between the Middle Ladinian (c 232^235 Ma Middle Triassic) and latest Carnian or possiblyearliest Norian (c 216^227 Ma Late Triassic Carlisle ampSuzuki 1974) The only published U^Pb age is based on asingle concordant analysis of a multi-grain baddeleyitefraction from a gabbro on southernVancouver Island thatyielded a 206Pb238U age of 227326 Ma (Parrish ampMcNicoll 1992) Two unpublished 206Pb238U baddeleyiteages also from a gabbro on southern Vancouver Islandare 226805 Ma (five multi-grain fractions) and228425 Ma (two multi-grain fractions Sluggett 2003)
VOLCANIC STRAT IGRAPHY ANDPETROGRAPHYFieldwork undertaken onVancouver Island as part of thisstudy in 2004^2006 explored the volcanic stratigraphy ofthe Karmutsen flood basalts in three main areas theKarmutsen Range (between Alice and Nimpkish lakes)the area around Schoen Lake Provincial Park and atMount Arrowsmith (Greene et al 2005 2006) areasaround Holberg Inlet in northernmost Vancouver Islandand Buttle Lake in central Vancouver Island were alsoinvestigated (Fig 1) The character and thickness of theflood basalt sequences vary locally although the tripartitesuccession of the Karmutsen Formation appears to be pres-ent throughout Vancouver Island The stratigraphic thick-nesses for the pillow volcaniclastic and subaerial flowunits in the type area around Buttle Lake are estimated tobe 2600m 1100m and 2900m respectively (Surdam1967) estimated thicknesses on northernVancouver Islandare 43000m 400^1500m and 41500m respectively(Nixon et al 2008) in the vicinity of Mount Arrowsmithand nearby areas on southern Vancouver Island thick-nesses are 1100m 950m and 1200m respectively (Yorathet al 1999 Fig 1)Picritic pillow lavas west of the Karmutsen Range on
northernVancouver Island occur in a roughly triangular-shaped area (30 km across) bounded by KeoghMaynard and Sara lakes (Figs 2 and 3 Greene et al2006 Nixon et al 2008) Excellent exposures of picriticpillow lavas occur in roadcuts along the north shore ofKeogh Lake the type locality (Greene et al 2006) TheKeogh Lake picrites form pillowed and massive flow units(515m thick Fig 3) with pillows and tubes of varieddimensions (typically51m wide) Numerous thermal con-traction features in the pillows are filled with quartz^carbonate such as drain-back ledges and tortoise-shelljointing (Greene et al 2006) The picritic pillow basalts
are not readily distinguishable in the field from basaltexcept by their density non-magnetic character and com-paratively minor amounts of interpillow quartz^carbonateFieldwork and mapping indicates that the picrites occurnear the transition between the uppermost pillow lavasand the lowermost part of the volcaniclastic unit (Nixonet al 2008)In and around Schoen Lake Provincial Park where
the base of the Karmutsen Formation is exposed is asediment^sill complex consisting of Middle Triassic andPennsylvanian to Permian limestones and fine-grained sili-ciclastic sedimentary rocks intruded by mafic sills andoverlain by pillow basalt (Figs 2 and 4 Carlisle 1972) Thesediment^sill complex is 600^900m thick whichincludes 150^200m of pre-intrusive sedimentary rocks(Carlisle 1972) TheTriassic sedimentary rocks range fromthinly bedded siliceous and calcareous shales to bandedcherts and finely laminated Daonella-bearing shales(Carlisle 1972) Basal sills and pre-Karmutsen sedimentsare also exposed around Buttle Lake (Fig 1)Exposures with considerable vertical relief around
Mount Arrowsmith (see Electronic Appendix 1 availablefor downloading at httppetrologyoxfordjournalsorg)and Buttle Lake preserve thick successions of submarinepillow basalts and pillow breccias with rare intercalationsof sediment and subaerial flows Massive submarine flowsin the pillow unit are locally recognizable by their irregu-lar hackly columnar jointing The massive subaerial flowsform monotonous sequences marked mainly by amygdaloi-dal horizons brecciated flow tops are rarely observedA locality with well-preserved pahoehoe flow features iswell exposed north of Holberg Inlet (Fig 3 Nixon amp Orr2007 Nixon et al 2008) There is no evidence of significantdetrital material from a continental source in the sedi-ments associated with the flood basalts anywhere onVancouver Island The uppermost Karmutsen flows areinterbedded with thin (commonly 0^4m thick) lenses oflimestone and rarely siliciclastic sedimentary rocks(Nixon et al 2006b) Plagioclase-phyric lavas characterizethe upper part of the subaerial sequence and locallyinclude distinctive plagioclase-megacrystic (1^2 cm crys-tals) trachytoid-textured flows (Nixon et al 2006b)A total of 129 samples were collected from the
Karmutsen Formation on Vancouver Island from which63 samples were selected for geochemical analysis basedon the relative degree of alteration and geographical distri-bution of the samples Fifty-six of these samples weredivided into four main groups based on petrography(Table 1) and geochemistry tholeiitic basalt picrite high-MgO basalt and coarse-grained mafic rocks In additiontwo outlying samples and a mineralized sill are distin-guished in Table 1 The tholeiitic basalts are dominantlyglomeroporphyritic and less commonly aphanitic withan intersertal to intergranular groundmass (Table 1)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
4
Fig 2 Geological map and stratigraphy of the Schoen Lake Provincial Park and Karmutsen Range areas (locations shown in Fig 1 northernVancouver Island) (a) Stratigraphic column depicts units exposed in the Schoen Lake area derived from Carlisle (1972) and fieldwork (b)Generalized geology for the Schoen Lake area with sample locations Map derived from Massey et al (2005a) The exposures in the SchoenLake area are the lower volcanic stratigraphy and base of the Karmutsen Formation (c) Stratigraphic column for geology in the Alice^Nimpkish Lake area derived from Nixon amp Orr (2007) (d) Geological map from mapping of Nixon et al (2006a 2008) Sample sites andlithologies are denoted in the legend The Keogh Lake picrites are exposed near Keogh Sara and Maynard lakes and areas to the east ofMaynard Lake (Nixon et al 2008)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
5
Fig 3 Photographs of picritic and tholeiitic pillowed and massive basalt flows from the Karmutsen Range (Alice and Nimpkish Lake area)and Holberg Inlet (HI) areas northernVancouver Island (locations shown in Fig 1) (a) Massive submarine flow (marked with arrows) drapedover high-MgO pillow basalt of similar composition (b) Picritic pillow basalts in cross-section near Maynard Lake (Note vesicular uppermargin to pillows) (c) Close-up photograph of undulating contact of massive submarine flow draped over high-MgO pillow basalt (markedwith arrows) from photograph (a) (area of photograph indicated with a white box) (d) Photograph of margin of pahoehoe lobe (sledgehammerfor scale) (e) Cross-section of a large picritic pillow lobe with infilling of quartz^carbonate in cooling^contraction cracks (sledgehammer forscale) (f) High-MgO pillow breccia stratigraphically between submarine and subaerial flows (arrow points to lens cap 7 cm diameter forscale)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
6
Plagioclase forms most of the phenocrysts and glomero-crysts and the groundmass is fine-grained comprising pla-gioclase microlites clinopyroxene granules small grains ofFe^Ti oxide devitrified glass and secondary minerals
there is no fresh glass preserved The picritic and high-MgO basaltic lavas are characterized by spherulitic plagi-oclase and clinopyroxene with abundant euhedral olivinephenocrysts (Fig 5 Table 1) Olivine is completely replaced
Fig 4 Photographs showing field relations from the Schoen Lake Provincial Park area northernVancouver Island (a) Sediment^sill complexat the base of the Karmutsen Formation on the north side of Mt Schoen 100m below the Daonella locality (see Fig 2 for location) (b)Interbedded mafic sills and deformed finely banded chert and shale with calcareous horizons from location between Mt Adam and MtSchoen
Fig 5 Photomicrographs of picritic pillow basalts Karmutsen Formation northernVancouver Island Scale bars represent 1mm (a) Picritewith abundant euhedral olivine pseudomorphs from Keogh Lake type locality (Fig 2) in cross-polarized light (XPL) (sample 4722A4 19^198wt MgO four analyses) Samples contain dense clusters (24^42 vol ) of olivine pseudomorphs (52mm) in a groundmass of curvedand branching sheaves of acicular plagioclase and intergrown with clinopyroxene and altered glass In many cases plagioclase nucleated on theedges of the olivine phenocrysts (b) Sheaves of intergrown plagioclase and clinopyroxene in aphyric picrite pillow lava from west of MaynardLake (Fig 2) in XPL (sample 4723A2108 wt MgO)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
7
Table 1 Summary of petrographic characteristics and phenocryst proportions of Karmutsen basalts on Vancouver Island
BC
Sample Area Flow Group Texture vol Ol Plag Cpx Ox Alteration Notes
4718A1 MA PIL THOL intersertal 20 5 3 few plag glcr52mm cpx51 mm
4718A2 MA PIL THOL intersertal glomero 15 1 plag glcr52mm very fresh
4718A5 MA FLO THOL intersertal glomero 10 2 mottled few plag glcr54 mm
4723A4 KR PIL PIC intergranular intersertal 0 3 swtl plag51mm no ol phenos
4723A13 KR PIL PIC spherulitic 24 3 swtl plag51 mm
(continued)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
8
by an assemblage of talc^tremolite^serpentine^opaqueoxides there is no fresh olivine present in the Karmutsenlavas onVancouver Island Plagioclase and clinopyroxenephenocrysts are absent in the high-MgO lavas Thecoarse-grained mafic rocks are characterized by sub-ophi-tic textures with an average grain size typically 41mmand are generally not glomeroporphyritic these rocks aremainly from the interiors of massive flows although somemay represent sillsSample preparation and analytical methods for major-
and trace-element and Sr^Nd^Hf^Pb isotopic composi-tions of the Karmutsen volcanic rocks are given in theAppendix
WHOLE -ROCK CHEMISTRYMajor- and trace-element compositionsThe most abundant type of volcanic rock in theKarmutsen Formation is tholeiitic basalt with a restrictedrange of major- and trace-element compositions The tho-leiitic basalts have similar compositions to the coarse-grained mafic rocks and both groups are distinct fromthe picrites and high-MgO basalts [Fig 6 picrites412wt MgO (LeBas 2000) high-MgO basalts8^12wt MgO] The tholeiitic basalts have lower MgO(57^77wt MgO) and higher TiO2 (14^22wt
TiO2) than the picrites (130^198wt MgO05^07wt TiO2) and high-MgO basalts (91^116wt MgO 05^08wt TiO2 Fig 6 Table 2) Almost allcompositions plot within the tholeiitic field in a total alka-lis vs silica plot although there has been substantial K lossin most samples from the Karmutsen Formation (seebelow) and the tholeiitic basalts generally have higherSiO2 Na2OthornK2O CaO and FeOT (total iron expressedas FeO) than the picrites The tholeiitic basalts also havenoticeably lower loss on ignition (LOI mean1713wt ) than the picrites (mean 54 08wt )and high-MgO basalts (mean 3815wt ) (Fig 6)reflecting the presence of abundant altered olivinephenocrysts in the latter groups Ni concentrations are sig-nificantly higher for the picrites (339^755 ppm) and high-MgO basalts (122^551ppm) than the tholeiitic basalts(58^125 ppm except for one anomalous sample) andcoarse-grained rocks (70^213 ppm) (Fig 6 Table 2) Ananomalous tholeiitic pillowed flow (two samples outlierin Tables 1 and 2) from near the picrite type locality atKeogh Lake and a mineralized sill (disseminated sulfide)from the basal sediment^sill complex at Schoen Lake aredistinguished by higher TiO2 and FeOT than the maingroup of tholeiitic basalts The tholeiitic basalts are similarin major-element composition to previously publishedresults for samples from the Karmutsen Formation how-ever the Keogh Lake picrites which encompass the
Table 1 Continued
Sample Area Flow Group Texture vol Ol Plag Cpx Ox Alteration Notes
4722A5 KR FLO OUTLIER intersertal 3 mottled very fg v small pl needles
5615A11 KR PIL OUTLIER intersertal 3 mottled very fg v small pl needles
5617A4 SL SIL MIN intersertal 5 3 fg
Sample number last digit of year month day initial sample station (except 93G171) MA Mount Arrowsmith SLSchoen Lake KR Karmutsen Range PIL pillow BRE breccia FLO flow SIL sill GAB gabbro THOL tholeiiticbasalt PIC picrite HI-MG high-MgO basalt CGcoarse-grained MIN mineralized sill glomero glomeroporphyriticOlivine modes for picrites are based on area calculations from scans (see lsquoOlivine accumulation in high-MgO and picriticlavasrsquo section and Fig 11) Visual alteration index is based primarily on degree of plagioclase alteration and presence ofsecondary minerals (1 least altered 3 most altered) Plagioclase phenocrysts are commonly altered to albite pumpellyiteand chlorite olivine is altered to talc tremolite and clinochlore (determined using the Rietveld method of X-ray powderdiffraction) clinopyroxene is unaltered FendashTi oxide is commonly replaced by sphene glcr glomerocrysts fg fine-grained cg coarse-grained oik oikocryst chad chadacryst enc enclosing swtl swallow-tail ol olivinepseudomorphs plag plagioclase cpx clinopyroxene ox oxides (includes ilmenite thorn titanomagnetite)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
9
00
05
10
15
20
25
30
35
40
0 5 10 15 20 25
10
11
12
13
14
15
16
17
18
19
20
0 5 10 15 20 25
6
7
8
9
10
11
12
13
14
15
0 5 10 15 20 25
6
7
8
9
10
11
12
13
14
15
16
0 5 10 15 20 25
0
1
2
3
4
5
6
40 45 50 55 60
0
200
400
600
800
0 5 10 15 20 25
(wt)(wt)
(wt)
(wt)
+K2ONa2O TiO2
FeO (T)
Ni (ppm)
Al2O3
CaO
alkalic
tholeiitic
Karmutsen Form
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
(compiled data)
Pillowed flowMineralized sill
SiO2 (wt) MgO (wt)
MgO (wt)MgO (wt)
MgO (wt) MgO (wt)
(a) (b)
(g)
(e)
(f)
(c)
LOI (wt )
(d)
MgO (wt )0
1
2
3
4
5
6
7
0 5 10 15 20 25
(wt)
Fig 6 Whole-rock major-element Ni and LOI variation diagrams for the Karmutsen Formation New samples from this study (Table 2) areshown by symbols with black outlines (see legend) Previously published analyses are shown by small gray symbols without black outlines Theboundary of the alkaline and tholeiitic fields is that of MacDonald amp Katsura (1964) Total iron expressed as FeO LOI loss on ignition oxidesare plotted on an anhydrous normalized basis References for the 322 compiled analyses for the Karmutsen Formation are Surdam (1967)Kuniyoshi (1972) Muller et al (1974) Barker et al (1989) Lassiter et al (1995) Massey (1995a1995b)Yorath et al (1999) and G Nixon (unpublisheddata) It should be noted that the compiled dataset has not been filtered many of the samples with high SiO2 (452wt ) are probably notKarmutsen flood basalts but younger Bonanza basalts and andesites Three pillowed flow samples and a mineralized sill are distinguishedseparately because of their anomalous chemistry Coarse-grained samples are indicated separately Three tholeiitic basalt samples (4724A54718A5 4718A6) with416wt Al2O3 and4270 ppm Sr contain 10^25 modal of large plagioclase phenocrysts (7^8mm) or glomerocrysts(44mm)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
10
Table 2 Major element (wt oxide) and trace element (ppm) abundances in whole-rock samples of Karmutsen basalts
1Ni concentrations for these samples were determined by XRFTHOL tholeiitic basalt PIC picrite HI-MG high MgO basalt CG coarse-grained (sill or gabbro) MIN SIL mineralizedsill OUTLIER anomalous pillowed flow in plots MA Mount Arrowsmith SL Schoen Lake KR Karmutsen Range QIQuadra Island Sample locations are given using the Universal Transverse Mercator (UTM) coordinate system (NAD83zones 9 and 10) Analyses were performed at Activation Laboratory (ActLabs) Fe2O3
is total iron expressed as Fe2O3LOI loss on ignition All major elements Sr V and Y were measured by ICP quadrupole OES on solutions of fusedsamples Cu Ni Pb and Zn were measured by total dilution ICP Cs Ga Ge Hf Nb Rb Ta Th U Zr and REE weremeasured by magnetic-sector ICP on solutions of fused samples Co Cr and Sc were measured by INAA Blanks arebelow detection limit See Electronic Appendices 2 and 3 for complete XRF data and PCIGR trace element datarespectively Major elements for sample 93G17 are normalized anhydrous Sample 5615A7 was analyzed by XRFSamples from Quadra Island are not shown in figures
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
17
picritic and high-MgO basalt pillow lavas extend to20wt MgO (Fig 6 and references listed in thecaption)The tholeiitic basalts from the Karmutsen Range dis-
play a tight range of parallel light rare earth element(LREE)-enriched REE patterns (LaYbCNfrac14 21^24mean 2202) whereas the picrites and high-MgObasalts are characterized by sub-parallel LREE-depleted
patterns (LaYbCNfrac14 03^07 mean 06 02) with lowerREE abundances than the tholeiitic basalts (Fig 7)Tholeiitic basalts from the Schoen Lake and MountArrowsmith areas have similar REE patterns tosamples from the Karmutsen Range (Schoen Lake LaYbCNfrac14 21^26 mean 2303 Mount Arrowsmith LaYbCNfrac1419^25 mean 2204) and the coarse-grainedmafic rocks have similar REE patterns to the tholeiitic
Depleted MORB average(Salters amp Stracke 2004)
Schoen LakeSchoen Lake
Mount Arrowsmith
Karmutsen RangeS
ampl
eC
hond
rite
Sam
ple
Cho
ndrit
e
Sam
ple
Prim
itive
Man
tleS
ampl
eP
rimiti
ve M
antle
Sam
ple
Prim
itive
Man
tle
Mount Arrowsmith
Karmutsen Range
(a) (b)
(c) (d)
(e) (f )
Sam
ple
Cho
ndrit
e
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained mafic rock
Pillowed flowMineralized sill
1
10
100
1
10
100
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
01
1
10
100
1
10
100
1
10
100
Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
Fig 7 Whole-rock REE and other incompatible-element concentrations for the Karmutsen Formation (a) (c) and (e) are chondrite-normal-ized REE patterns for samples from each of the three main field areas onVancouver Island (b) (d) and (f) are primitive mantle-normalizedtrace-element patterns for samples from each of the three main field areas All normalization values are from McDonough amp Sun (1995) Theclear distinction between LREE-enriched tholeiitic basalts and LREE-depleted picrites the effects of alteration on the LILE (especially loss ofK and Rb) and the different range for the vertical scale in (b) (extends down to 01) should be noted
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
18
basalts (LaYbCNfrac14 21^29 mean 2508) Two LREE-depleted coarse-grained mafic rocks from the SchoenLake area have similar REE patterns (LaYbCNfrac14 07) tothe picrites from the Keogh Lake area indicating that thisdistinctive suite of rocks is exposed as far south asSchoen Lake that is 75 km away (Fig 7) The anomalouspillowed flow from the Karmutsen Range has lower LaYbCN (13^18) than the main group of tholeiitic basaltsfrom the Karmutsen Range (Fig 7) The mineralized sillfrom the Schoen Lake area is distinctly LREE-enriched(LaYbCNfrac14 33) with the highest REE abundances(LaCNfrac14 940) of all Karmutsen samples this may reflectcontamination by adjacent sediment also evident in thin-section where sulfide blebs are cored by quartz (Greeneet al 2006)The primitive mantle-normalized trace-element pat-
terns (Fig 7) and trace-element variations and ratios(Fig 8) highlight the differences in trace-element
concentrations between the four main groups ofKarmutsen flood basalts onVancouver Island The tholeii-tic basalts from all three areas have relatively smooth par-allel trace-element patterns some samples have smallpositive Sr anomalies relative to Nd and Sm and manysamples have prominent negative K anomalies relative toU and Nb reflecting K loss during alteration (Fig 7) Thelarge ion lithophile elements (LILE) in the tholeiiticbasalts (mainly Rb and K not Cs) are depleted relativeto the high field strength elements (HFSE) and LREELILE segments especially for the Schoen Lake area(Fig 7d) are remarkably parallel The picrites andhigh-MgO basalts form a tight band of parallel trace-element patterns and are depleted in HFSE and LREEwith positive Sr anomalies and relatively low Rb Thecoarse-grained mafic rocks from the Karmutsen Rangehave identical trace-element patterns to the tholeiiticbasalts Samples from the Schoen Lake area have similar
000
025
050
075
100
125
0 1 2 3 4 5
00
04
08
12
16
0 5 10 15 20 25
MgO (wt)
0
5
10
15
0 50 100 150
Hf (ppm)
Th (ppm)
NbLaNb (ppm)
Zr (ppm)
(a)
(c)
(b)
(d)
Th (ppm)
00
02
04
06
08
10
12
00 01 02 03 04 05
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
NbLa=1
4723A4 5615A12
U (ppm)
Fig 8 Whole-rock trace-element concentrations and ratios for the Karmutsen Formation [except (b) which is vs wt MgO] (a) Nb vs Zr(b) NbLa vs MgO (c) Th vs Hf (d) Th vs U There is a clear distinction between the tholeiitic basalts and picrites in Nb and Zr both inconcentration and the slope of each trend
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
19
trace-element patterns to the Karmutsen Range except forthe two LREE-depleted coarse-grained mafic rocks andthe mineralized sill (Fig 7d)
Sr^Nd^Hf^Pb isotopic compositionsA total of 19 samples from the four main groups ofthe Karmutsen Formation were selected for radiogenicisotopic geochemistry on the basis of major- and trace-element variation stratigraphic position and geographicallocation The samples have overlapping age-correctedHf and Nd isotope ratios and distinct ranges of Sr isotopiccompositions (Fig 9) The tholeiitic basalts picriteshigh-MgO basalt and coarse-grained mafic rocks have
initial eHffrac14thorn87 to thorn126 and eNdfrac14thorn66 to thorn88 cor-rected for in situ radioactive decay since 230 Ma (Fig 9Tables 3 and 4) All the Karmutsen samples form a well-defined linear array in a Lu^Hf isochron diagram corre-sponding to an age of 24115 Ma (Fig 9) This is withinerror of the accepted age of the Karmutsen Formationand indicates that the Hf isotope systematics behaved as aclosed system since c 230 Ma The anomalous pillowedflow has similar initial eHf (thorn98) and slightly lower initialeNd (thorn62) compared with the other Karmutsen samplesThe picrites and high-MgO basalt have higher initial87Sr86Sr (070398^070518) than the tholeiitic basalts(initial 87Sr86Srfrac14 070306^070381) and the coarse-grained mafic rocks (initial 87Sr86Srfrac14 070265^070428)
05128
05129
05130
05131
05132
015 017 019 021 023 025
02829
02830
02831
02832
02833
02834
000 002 004 006 008 010
4
6
8
10
0702 0703 0704 0705 07066
8
10
12
14
5 6 7 8 9 10
12
0
1
2
3
4
5
6
7
0702 0703 0704 0705 0706
(d)
LOI (wt )
87Sr 86Sr
4723A2
Nd
143Nd144Nd
147Sm144Nd
87Sr 86Sr (230 Ma)
(230 Ma)
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
(a)
(c)
176Hf177Hf
Age=241 plusmn 15 Ma=+990 plusmn 064
176Lu177Hf
Hf (230 Ma)
Nd (230 Ma)
(e)
Hf (i)
(b)
Fig 9 Whole-rock Sr Nd and Hf isotopic compositions for the Karmutsen Formation (a) 143Nd144Nd vs 147Sm144Nd (b) 176Hf177Hf vs176Lu177Hf The slope of the best-fit line for all samples corresponds to an age of 24115 Ma (c) Initial eNd vs 87Sr86Sr Age-correction to230 Ma (d) LOI vs 87Sr86Sr (e) Initial eHf vs eNd Average 2 error bars are shown in a corner of each panel Complete chemical duplicatesshown inTables 3 and 4 (samples 4720A7 and 4722A4) are circled in each plot
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
overlap the ranges of the picrites and tholeiitic basalts(Fig 9 Table 3)The measured Pb isotopic compositions of the tholeiitic
basalts are more radiogenic than those of the picrites andthe most magnesian Keogh Lake picrites have the leastradiogenic Pb isotopic compositions The range ofinitial Pb isotope ratios for the picrites is206Pb204Pbfrac1417868^18580 207Pb204Pbfrac1415547^15562and 208Pb204Pbfrac14 37873^38257 and the range for thetholeiitic basalts is 206Pb204Pbfrac1418782^19098207Pb204Pbfrac1415570^15584 and 208Pb204Pbfrac14 37312^38587 (Fig 10 Table 5) The coarse-grained mafic rocksoverlap the range of initial Pb isotopic compositions forthe tholeiitic basalts with 206Pb204Pbfrac1418652^19155207Pb204Pbfrac1415568^15588 and 208Pb204Pbfrac14 38153^38735 One high-MgO basalt has the highest initial206Pb204Pb (19242) and 207Pb204Pb (15590) and thelowest initial 208Pb204Pb (37676) The Pb isotopic ratiosfor Karmutsen samples define broadly linearrelationships in Pb isotope plots The age-corrected Pb
isotopic compositions of several picrites have been affectedby U and Pb (andor Th) mobility during alteration(Fig 10)
ALTERAT IONThe Karmutsen basalts have retained most of their origi-nal igneous structures and textures however secondaryalteration and low-grade metamorphism have producedzeolitic- and prehnite^pumpellyite-bearing mineral assem-blages (Table 1 Cho et al 1987) Alteration and metamor-phism have primarily affected the distribution of the LILEand the Sr isotopic systematics The Keogh Lake picriteshave been affected more by alteration than the tholeiiticbasalts and have higher LOI (up to 55wt ) variableLILE and higher measured Sr Nd and Hf and lowermeasured Pb isotope ratios than the tholeiitic basalts Srand Pb isotope compositions for picrites and tholeiiticbasalts show a relationship with LOI (Figs 9 and 10)whereas Nd and Hf isotopic compositions do not correlate
Table 3 Sr and Nd isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Rb Sr 87Sr86Sr 2m87Rb86Sr 87Sr86Srt Sm Nd 143Nd144Nd 2m eNd
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses carried out at the PCIGR thecomplete trace element analyses are given in Electronic Appendix 3
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
21
with LOI (not shown) The relatively high apparent initialSr isotopic compositions for the high-MgO lavas (up to07052) probably resulted from post-magmatic loss of Rbor an increase in 87Sr86Sr through addition of seawater Sr(eg Hauff et al 2003) High-MgO lavas from theCaribbean plateau have similarly high 87Sr86Sr comparedwith basalts (Recurren villon et al 2002)The correction for in situdecay on initial Pb isotopic ratios has been affected bymobilization of U and Th (and Pb) in the whole-rockssince their formation A thorough acid leaching duringsample preparation was used that has been shown to effec-tively remove alteration phases (Weis et al 2006 NobreSilva et al in preparation) Whole-rock SmNd ratiosand age-correction of Nd isotopes may be affected bythe leaching procedure for submarine rocks older than50 Ma (Thompson et al 2008 Fig 9) there ishowever very little variation in Nd isotopic compositionsof the picrites or tholeiitic basalts and this system does notappear to be disturbed by alteration The HFSE abun-dances for tholeiitic and picritic basalts exhibit clear
linear correlations in binary diagrams (Fig 8) whereasplots of LILE vs HFSE are highly scattered (not shown)because of the mobility of some of the alkali and alkalineearth LILE (Cs Rb Ba K) and Sr for most samplesduring alterationThe degree of alteration in the Karmutsen samples does
not appear to be related to eruption environment or depthof burial Pillow rims are compositionally different frompillow cores (Surdam1967 Kuniyoshi1972) and aquagenetuffs are chemically different from isolated pillows withinpillow breccias however submarine and subaerial basaltsexhibit a similar degree of alterationThere is no clear cor-relation between the submarine and subaerial basalts andsome commonly used chemical alteration indices (eg BaRb vs K2OP2O5 Huang amp Frey 2005) There is also nodefinitive correlation between the petrographic alterationindex (Table 1) and chemical alteration indices although10 of 13 tholeiitic basalts with the highest K and LILEabundances have the highest petrographic alterationindex of three (seeTable 1)
Table 4 Hf isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Lu Hf 176Hf177Hf 2m eHf176Lu177Hf 176Hf177Hft eHf(t)
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
OLIV INE ACCUMULAT ION INH IGH-MgO AND PICR IT IC LAVASGeochemical trends and petrographic characteristics indi-cate that accumulation of olivine played an important rolein the formation of the high-MgO basalts and picrites fromthe Keogh Lake area on northern Vancouver Island TheKeogh Lake samples show a strong linear correlation inplots of Al2O3 TiO2 ScYb and Ni vs MgO and many ofthe picrites have abundant clusters of olivine pseudomorphs(Table1 Fig11)There is a strong linear correlationbetween
modal per cent olivine and whole-rock magnesium contentmagnesium contents range between 10 and 19wt MgOand the proportion of olivine phenocrysts in these samplesvaries from 0 to 42 vol (Fig 11)With a few exceptionsboth the size range and median size of the olivine pheno-crysts in most samples are comparable The clear correla-tion between the proportion of olivine phenocrysts andwhole-rock magnesium contents indicates that accumula-tion of olivine was directly responsible for the high MgOcontents (410wt ) in most of the picritic lavas
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
207Pb204Pb
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
208Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
(a) (b)
(c) (e)
4723A24723A2
4723A2
4723A2
4722A4
4723A4
4722A4
4723A4
4723A134723A3
4723A3
4723A13
1556
1558
1560
1562
1564
1566
1568
186 190 194 198 202 206 2101550
1552
1554
1556
1558
1560
178 180 182 184 186 188 190 192 194
380
384
388
392
396
186 190 194 198 202 206 210374
378
382
386
390
178 180 182 184 186 188 190 192 194
206Pb204Pb(d)
LOI (wt )
0
1
2
3
4
5
6
7
186 190 194 198 202 206 210
4723A2
Fig 10 Pb isotopic compositions of leached whole-rock samples measured by MC-ICP-MS for the Karmutsen Formation Error bars are smal-ler than symbols (a) Measured 207Pb204Pb vs 206Pb204Pb (b) Initial 207Pb204Pb vs 206Pb204Pb Age-correction to 230 Ma (c) Measured208Pb204Pb vs 206Pb204Pb (d) LOI vs 206Pb204Pb (e) Initial 208Pb204Pb vs 206Pb204Pb Complete chemical duplicates shown in Table 5(samples 4720A7 and 4722A4) are circled in each plot The dashed lines in (b) and (e) show the differences in age-corrections for two picrites(4723A4 4722A4) using the measured UTh and Pb concentrations for each sample and the age-corrections when concentrations are used fromthe two picrites (4723A3 4723A13) that appear to be least affected by alteration
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
23
DISCUSS IONThe compositional and spatial development of the eruptedlava sequences of the Wrangellia oceanic plateau onVancouver Island provide information about the evolutionof a mantle plume upwelling beneath oceanic lithospherein an analogous way to the interpretation of continentalflood basalts The Caribbean and Ontong Java plateausare the two primary examples of accreted parts of oceanicplateaus that have been studied to date and comparisonscan be made with the Wrangellia oceanic plateauHowever the Wrangellia oceanic plateau is a distinctiveoccurrence of an oceanic plateau because it formed on thecrust of a Devonian to Mississippian intra-oceanic islandarc overlain by Mississippian to Permian limestones andpelagic sediments The nature and thickness of oceaniclithospheric mantle are considered to play a fundamentalrole in the decompression and evolution of a mantleplume and thus the compositions of the erupted lavasequences (Greene et al 2008) The geochemical and
stratigraphic relationships observed in the KarmutsenFormation onVancouver Island and the focus of the subse-quent discussion sections provide constraints on (1) thetemperature and degree of melting required to generatethe primary picritic magmas (2) the composition of themantle source (3) the depth of melting and residual min-eralogy in the source region (4) the low-pressure evolutionof the Karmutsen tholeiitic basalts (5) the growth of theWrangellia oceanic plateau
Melting conditions and major-elementcomposition of the primary magmasThe primary magmas to most flood basalt provinces arebelieved to be picritic (eg Cox 1980) and they have beenfound in many continental and oceanic flood basaltsequences worldwide (eg Siberia Karoo Paranacurren ^Etendeka Caribbean Deccan Saunders 2005) Near-pri-mary picritic lavas with low total alkali abundances
Table 5 Pb isotopic compositions of Karmutsen basaltsVancouver Island BC
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
require high-degree partial melting of the mantle source(eg Herzberg amp OrsquoHara 2002) The Keogh Lake picritesfrom northernVancouver Island are the best candidates foridentifying the least modified partial melts of the mantleplume source despite having accumulated olivine pheno-crysts and can be used to estimate conditions of meltingand the composition of the primary magmas for theKarmutsen FormationPrimary magma compositions mantle potential tem-
peratures and source melt fractions for the picritic lavas ofthe Karmutsen Formation were calculated from primitivewhole-rock compositions using PRIMELT1XLS software(Herzberg et al 2007) and are presented in Fig 12 andTable 6 A detailed discussion of the computationalmethod has been given by Herzberg amp OrsquoHara (2002)and Herzberg et al (2007) key features of the approachare outlined belowPrimary magma composition is calculated for a primi-
tive lava by incremental addition of olivine and compari-son with primary melts of mantle peridotite Lavas thathad experienced plagioclase andor clinopyroxene fractio-nation are excluded from this analysis All calculated pri-mary magma compositions are assumed to be derived byaccumulated fractional melting of fertile peridotite KR-4003 Uncertainties in fertile peridotite composition donot propagate to significant variations in melt fractionand mantle potential temperature melting of depletedperidotite propagates to calculated melt fractions that aretoo high but with a negligible error in mantle potential
temperature PRIMELT1XLS calculates the olivine liqui-dus temperature at 1atm which should be similar to theactual eruption temperature of the primary magmaand is accurate to 308C at the 2 level of confidenceMantle potential temperature is model dependent witha precision that is similar to the accuracy of the olivineliquidus temperature (C Herzberg personal communica-tion 2008)With regard to the evidence for accumulated olivine in
the Keogh Lake picrites it should not matter whether themodeled lava composition is aphyric or laden with accu-mulated olivine phenocrysts PRIMELT1 computes theprimary magma composition by adding olivine to a lavathat has experienced olivine fractionation in this way itinverts the effects of fractional crystallization The onlyrequirement is that the olivine phenocrysts are not acci-dental or exotic to the lava compositions Support for thisprocedure can be seen in the calculated olivine liquid-line-of-descent shown by the curved arrays with 5 olivineaddition increments (Fig 12c) Fractional crystallization ofolivine from model primary magmas having 85^95wt FeO and 153^175wt MgO will yield arrays of deriva-tive liquids and randomly sorted olivines that in most butnot all cases connect the picrites and high-MgO basalts Anewer version of the PRIMELT software (Herzberg ampAsimow 2008) can perform olivine addition to and sub-traction from lavas with highly variably olivine phenocrystcontents and yields results that converge with those shownin Fig 12
Fig 11 Relationship between abundance of olivine phenocrysts and whole-rock MgO contents for the Keogh Lake picrites (a) Tracings ofolivine grains (gray) in six high-MgO samples made from high resolution (2400 dpi) scans of petrographic thin-sections Scale bars represent10mm sample numbers are indicated at the upper left (b) Modal olivine vs whole-rock MgO Continuous line is the best-fit line for 10 high-MgO samplesThe areal proportion of olivine was calculated from raster images of olivine grains using ImageJ image analysis software whichprovides an acceptable estimation of the modal abundance (eg Chayes 1954)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
25
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
Gray lines = final melting pressure
L + Ol + Opx
07
08
09
Pressure (GPa)
10
3
4
5
6
1
SOLIDUS
01
04
0506
L + Ol
2
03
02
PeridotiteKR-4003
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
Thick black lines = initial melting pressure
Dashed lines = melt fraction
Melt Fraction
2
3 4 5 6
L+Ol+Opx
SolidusSpinel
SolidusGarnet Peridotite
7
L+Ol
0 10 20 30 4040
45
50
55
60
MgO (wt)
SiO2
(wt)
PeridotiteKR-4003
Melt Fraction
0 10 20 305
10
15
CaO(wt)
MgO (wt)
3
4 5 6 7
2
46
2
Peridotite Partial Melts
Thick black line = initial melting pressureGray lines = final melting pressure
Peridotite
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
02
Pressure (GPa)
0302
04
Pressure (GPa)
Picrite
High-MgO basalt
Tholeiitic basalt
(c)
(d)
Karmutsen flood basaltsOlivine addition modelOlivine addition (5 increment)Primary magma
PicriteHigh-MgO basaltTholeiitic basalt
(a)
(b)
800 850 900
910
920
930
940
950
Karmutsen flood basalts
Olivine addition modelOlivine addition (5 increment)Primary magma
Gray lines = Fo contents of olivine
Fig 12 Estimated primary magma compositions for three Keogh Lake picrites and high-MgO basalts (samples 4723A4 4723A135616A7) using the forward and inverse modeling technique of Herzberg et al (2007) Karmutsen compositions and modeling results areoverlain on diagrams provided by C Herzberg (a) Whole-rock FeO vs MgO for Karmutsen samples from this study Total iron estimatedto be FeO is 090 Gray lines show olivine compositions that would precipitate from liquid of a given MgO^FeO composition Black lineswith crosses show results of olivine addition (inverse model) using PRIMELT1 (Herzberg et al 2007) Gray field highlights olivinecompositions with Fo contents of 91^92 predicted to precipitate (b) FeO vs MgO showing Karmutsen lava compositions and results ofa forward model for accumulated fractional melting of fertile peridotite KR-4003 (Kettle River Peridotite Walter 1998) (c) SiO2 vsMgO for Karmtusen lavas and model results (d) CaO vs MgO showing Karmtusen lava compositions and model results To brieflysummarize the technique [see Herzberg et al (2007) for complete description] potential parental magma compositions for the high-MgOlava series were selected (highest MgO and appropriate CaO) and using PRIMELT1 software olivine was incrementally added tothe selected compositions to show an array of potential primary magma compositions (inverse model) Then using PRIMELT1 theresults from the inverse model were compared with a range of accumulated fractional melts for fertile peridotite derived fromparameterization of the experimental results for KR-4003 from Walter (1998) forward model Herzberg amp OrsquoHara (2002) A melt fractionwas sought that was unique to both the inverse and forward models (Herzberg et al 2007) A unique solution was found when there was a com-mon melt fraction for both models in FeO^MgO and CaO^MgO^Al2O3^SiO2 (CMAS) projection space This modeling assumes thatolivine was the only phase crystallizing and ignores chromite precipitation and possible augite fractionation in the mantle (Herzbergamp OrsquoHara 2002) Results are best for a residue of spinel lherzolite (not pyroxenite) The presence of accumulated olivine in samples ofthe Keogh Lake picrites used as starting compositions does not significantly affect the results because we are modeling addition of olivine(see text for further description) The tholeiitic basalts cannot be used for modeling because they are all saturated in plagthorn cpxthornolThick black line for initial melting pressure at 4 GPa is not shown in (c) for clarity Gray area in (a) indicates parental magmas thatwill crystallize olivine with Fo 91^92 at the surface Dashed lines in (c) indicate melt fraction Solidi for spinel and garnet peridotite arelabelled in (c)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
The PRIMELT1 results indicate that if the Karmutsenpicritic lavas were derived from accumulated fractionalmelting of fertile peridotite the primary magmas wouldhave contained 15^17wt MgO and 10wt CaO(Fig 12 Table 6) formed from 23^27 partial meltingand would have crystallized olivine with a Fo content of 91 (Fig 12 Table 6) Calculated mantle temperatures forthe Karmutsen picrites (14908C) indicate melting ofanomalously hot mantle (100^2008C above ambient) com-pared with ambient mantle that produces mid-ocean ridgebasalt (MORB 1280^14008C Herzberg et al 2007)which initiated between 3 and 4 GPa Several picritesappear to be near-primary melts and required very littleaddition of olivine (eg sample 93G171 with measured158wt MgO) These temperature and melting esti-mates of the Karmutsen picrites are consistent withdecompression of hot mantle peridotite in an actively con-vecting plume head (ie plume initiation model) Threesamples (picrite 5614A1 and high-MgO basalts 5614A3and 5614A5) are lower in iron than the other Karmutsenlavas with FeO in the 65^80wt range (Fig 12c) andhave MgO and FeO contents that are similar to primitiveMORB from the Sequeiros fracture zone of the EastPacific Rise (EPR Perfit et al 1996) These samples
however differ from EPR MORB in having higher SiO2
and lower Al2O3 and they are restricted geographicallyto one area (Sara Lake Fig 2)We therefore have evidence for primary magmas with
MgO contents that range from a possible low of 110wt to as high as 174wt with inferred mantle potential tem-peratures from 1340 to 15208C This is similar to the rangedisplayed by lavas from the Ontong Java Plateau deter-mined by Herzberg (2004) who found that primarymagma compositions assuming a peridotite sourcewould contain 17wt MgO and 10wt CaO(Table 6) and that these magmas would have formed by27 melting at a mantle potential temperature of15008C (first melting occurring at 36 GPa) Fitton ampGodard (2004) estimated 30 partial melting for theOntong Java lavas based on the Zr contents of primarymagmas of Kroenke-type basalt calculated by incrementaladdition of equilibrium olivine However if a considerableamount of eclogite was involved in the formation of theOntong Java plateau magmas the excess temperatures esti-mated by Herzberg et al (2007) may not be required(Korenaga 2005) The variability in mantle potential tem-perature for the Keogh Lake picrites is also similar to thewide range of temperatures that have been calculated for
Table 6 Estimated primary magma compositions for Karmutsen basalts and other oceanic plateaus or islands
Sample 4723A4 4723A13 5616A7 93G171 Average OJP Mauna Keay Gorgonaz
Ontong Java primary magma composition for accumulated fraction melting (AFM) from Herzberg (2004)yMauna Kea primary magma composition is average of four samples of Herzberg (2006 table 1)zGorgona primary magma composition for AFM for 1F fertile source from Herzberg amp OrsquoHara (2002 table 4)Total iron estimated to be FeO is 090
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
27
lavas from single volcanoes in the Galapagos Hawaii andelsewhere by Herzberg amp Asimow (2008) Those workersinterpreted the range to reflect the transport of magmasfrom both the cool periphery and the hotter axis of aplume A thermally heterogeneous plume will yield pri-mary magmas with different MgO contents and the rangeof primary magma compositions and their inferred mantlepotential temperatures for Karmutsen lavas may reflectmelting from different parts of the plume
Source of the Karmutsen Formation lavasThe Sr^Nd^Hf^Pb isotopic compositions of theKarmutsen volcanic rocks on Vancouver Island indicatethat they formed from plume-type mantle containing adepleted component that is compositionally similar to butdistinct from other oceanic plateaus that formed in thePacific Ocean (eg Ontong Java and Caribbean Fig 13)Trace element and isotopic constraints can be used to testwhether the depleted component of the KarmutsenFormation was intrinsic to the mantle plume and distinctfrom the source of MORB or the result of mixing or invol-vement of ambient depleted MORB mantle with plume-type mantle The Karmutsen tholeiitic basalts have initialeNd and
87Sr86Sr ratios that fall within the range of basaltsfrom the Caribbean Plateau (eg Kerr et al 1996 1997Hauff et al 2000 Kerr 2003) and a range of Hawaiian iso-tope data (eg Frey et al 2005) and lie within and abovethe range for the Ontong Java Plateau (Tejada et al 2004Fig 13a) Karmutsen tholeiitic basalts with the highest ini-tial eNd and lowest initial 87Sr86Sr lie within and justbelow the field for age-corrected northern EPR MORB at230 Ma (eg Niu et al1999 Regelous et al1999) Initial Hfand Nd isotopic compositions for the Karmutsen samplesare noticeably offset to the low side of the ocean islandbasalt (OIB) array (Vervoort et al 1999) and lie belowand slightly overlap the low eHf end of fields for Hawaiiand the Caribbean Plateau (Fig 13c) the Karmutsen sam-ples mostly fall between and slightly overlap a field forage-corrected Pacific MORB at 230 Ma and Ontong Java(Mahoney et al 1992 1994 Nowell et al 1998 Chauvel ampBlichert-Toft 2001) Karmutsen lava Pb isotope ratios over-lap the broad range for the Caribbean Plateau and thelinear trends for the Karmutsen samples intersect thefields for EPR MORB Hawaii and Ontong Java in208Pb^206Pb space but do not intersect these fields in207Pb^206Pb space (Fig 13d and e) The more radiogenic207Pb204Pb for a given 206Pb204Pb of the Karmutsenbasalts requires high UPb ratios at an ancient time when235U was abundant similar to the Caribbean Plateau andenriched in comparison with EPR MORB Hawaii andOntong JavaThe low eHf^low eNd component of the Karmutsen
basalts that places them below the OIB mantle array andpartly within the field for age-corrected Pacific MORBcan form from low ratios of LuHf and high SmNd over
time (eg Salters amp White 1998) MORB will develop aHf^Nd isotopic signature over time that falls below themantle array Low LuHf can also lead to low 176Hf177Hffrom remelting of residues produced by melting in theabsence of garnet (eg Salters amp Hart 1991 Salters ampWhite 1998) The Hf^Nd isotopic compositions of theKarmutsen Formation indicate the possible involvementof depleted MORB mantle however similar to Iceland(Fitton et al 2003) Nb^Zr^Y variations provide a way todistinguish between a depleted component within theplume and a MORB-like component The Karmutsenbasalts and picrites fall within the Iceland array and donot overlap the MORB field in a logarithmic plot of NbYvs ZrY (Fig 13b) using the fields established by Fittonet al (1997) Similar to the depleted component within theCaribbean Plateau (Thompson et al 2003) the trend ofHf^Nd isotopic compositions towards Pacific MORB-likecompositions is not followed by MORB values in Nb^Zr^Y systematics and this suggests that the depleted compo-nent in the source of the Karmutsen basalts is distinctfrom the source of MORB and probably an intrinsic partof the plume The relatively radiogenic Pb isotopic ratiosand REE patterns distinct from MORB also support thisinterpretationTrace-element characteristics suggest the possible invol-
vement of sub-arc lithospheric mantle in the generation ofthe Karmutsen picrites Lassiter et al (1995) suggested thatmixing of a plume-type source with eNdthorn 6 to thorn 7 witharc material with low NbTh could reproduce the varia-tions observed in the Karmutsen basalts however theyconcluded that the absence of low NbLa ratios in theirsample of tholeiitic basalts restricts the amount of arc litho-sphere involved The study of Lassiter et al (1995) wasbased on major and trace element and Sr Nd and Pb iso-topic data for a suite of 29 samples from Buttle Lake in theStrathcona Provincial Park on central Vancouver Island(Fig 1) Whereas there are no clear HFSE depletions inthe Karmutsen tholeiitic basalts the low NbLa (and TaLa) of the Karmutsen picritic lavas in this study indicatethe possible involvement of Paleozoic arc lithosphere onVancouver Island (Fig 8)
REE modeling dynamic melting andsource mineralogyThe combined isotopic and trace-element variations of thepicritic and tholeiitic Karmutsen lavas indicate that differ-ences in trace elements may have originated from differentmelting histories For example the LuHf ratios of theKarmutsen picritic lavas (mean 008) are considerablyhigher than those in the tholeiitic basalts (mean 002)however the small range of overlapping initial eHf valuesfor the two lava suites suggests that these differences devel-oped during melting within the plume (Fig 9b) Below wemodel the effects of source mineralogy and depth of
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
28
melting on differences in trace-element concentrations inthe tholeiitic basalts and picritesDynamic melting models simulate progressive decom-
pression melting where some of the melt fraction isretained in the residue and only when the degree of partialmelting is greater than the critical mass porosity and the
source becomes permeable is excess melt extracted fromthe residue (Zou1998) For the Karmutsen lavas evolutionof trace element concentrations was simulated using theincongruent dynamic melting model developed by Zou ampReid (2001) Melting of the mantle at pressures greater than1 GPa involves incongruent melting (eg Longhi 2002)
OIB array
PacificMORB(230 Ma)
(0 Ma)
EPR(230 Ma)
-4
-2
0
2
4
8
10
12
14
16
18
20
22
-4 -2 0 2 4 6 8 10 12 14
minus2
0
2
4
6
8
10
12
14
0702 0703 0704 0705 0706
IndianMORB
87Sr 86Sr (initial)
Hf (initial)
(a) (c)
Nd (initial)
Karmutsen tholeiitic basalt
HawaiiOntong JavaCaribbean Plateau
Karmutsen high-MgO lava
high-MgO lavasKarmutsen
1540
1545
1550
1555
1560
1565
175 180 185 190 195 200375
380
385
390
395
175 180 185 190 195 200
(d) (e)
EPR
EPR
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb
Karmutsen Formation (initial)
HawaiiOntong JavaCaribbean Plateau
East Pacific Rise
Karmutsen Formation (measured)
Juan de FucaGorda
ε Nd
(initial)
Karmutsen Formation (different age-correction for 3 picrites)
(0 Ma)
NbY
001
01
1
1 10
ZrY
OIB
NMORB
(b)
Iceland array
6
Fig 13 Comparison of age-corrected (230 Ma) Sr^Nd^Hf^Pb isotopic compositions for Karmutsen flood basalts onVancouver Island withage-corrected OIB and MORB and Nb^Zr^Y variation (a) Initial eNd vs 87Sr86Sr (b) NbYand ZrY variation (after Fitton et al 1997)(c) Initial eHf vs eNd (d) Measured and initial 207Pb204Pb vs 206Pb204Pb (e) Measured and initial 208Pb204Pb vs 206Pb204Pb References fordata sources are listed in Electronic Appendix 4 Most of the compiled data were extracted from the GEOROC database (httpgeorocmpch-mainzgwdgdegeoroc) OIB array line in (c) is fromVervoort et al (1999) EPR East Pacific Rise Several high-MgO samples affected bysecondary alteration were not plotted for clarity in Pb isotope plots Estimation of fields for age-corrected Pacific MORB and EPR at 230 Mawere made using parent^daughter concentrations (in ppm) from Salters amp Stracke (2004) of Lufrac14 0063 Hffrac14 0202 Smfrac14 027 Ndfrac14 071Rbfrac14 0086 and Srfrac14 836 In the NbYvs ZrYplot the normal (N)-MORB and OIB fields not including Icelandic OIB are from data com-pilations of Fitton et al (1997 2003) The OIB field is data from 780 basalts (MgO45wt ) from major ocean islands given by Fitton et al(2003) The N-MORB field is based on data from the Reykjanes Ridge EPR and Southwest Indian Ridge from Fitton et al (1997) The twoparallel gray lines in (b) (Iceland array) are the limits of data for Icelandic rocks
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
29
with olivine being produced during the reactioncpxthornopxthorn spfrac14meltthornol for spinel peridotites The mod-eling used coefficients for incongruent melting reactionsbased on experiments on lherzolite melting (eg Kinzleramp Grove 1992 Longhi 2002) and was separated into (1)melting of spinel lherzolite (2) two combined intervals ofmelting for a garnet lherzolite (gt lherz) and spinel lher-zolite (sp lherz) source (3) melting of garnet lherzoliteThe model solutions are non-unique and a range indegree of melting partition coefficients melt reactioncoefficients and proportion of mixed source componentsis possible to achieve acceptable solutions (see Fig 14 cap-tion for description of modeling parameters anduncertainties)The LREE-depleted high-MgO lavas require melting of
a depleted spinel lherzolite source (Fig 14a) The modelingresults indicate a high degree of melting (22^25) similarto the results from PRIMELT1 based on major-elementvariations (discussed above) of a LREE-depleted sourceThe high-MgO lavas lack a residual garnet signature Theenriched tholeiitic basalts involved melting of both garnetand spinel lherzolite and represent aggregate melts pro-duced from continuous melting throughout the depth ofthe melting column (Fig 14b) Trace-element concentra-tions in the tholeiitic and picritic lavas are not consistentwith low-degree melting of garnet lherzolite alone(Fig 14c) The modeling results indicate that the aggregatemelts that formed the tholeiitic basalts would haveinvolved lesser proportions of enriched small-degreemelts (1^6 melting) generated at depth and greateramounts of depleted high-degree melts (12^25) gener-ated at lower pressure (a ratio of garnet lherzolite tospinel lherzolite melt of 14 provides the best fits seeFig 14b and description in figure caption) The totaldegree of melting for the tholeiitic basalts was high(23^27) similar to the degree of partial melting in thespinel lherzolite facies for the primary melts that eruptedas the Keogh Lake picrites of a source that was also prob-ably depleted in LREE
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
(c)
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
Melting of garnet lherzolite
High-MgO lavas
Melting of spinel lherzolite
source (07DM+03PM)
source (07DM+03PM)
6 melting
30 melting
(b)
(a)
30 melting
Sam
ple
Cho
ndrit
eS
ampl
eC
hond
rite 20 melting
1 melting
Melting of garnet lherzoliteand spinel lherzolite
x gt lherz + 4x sp lherz
5 meltingx gt lherz + 4x sp lherz
Sam
ple
Cho
ndrit
e
Tholeiitic lavas
Tholeiitic lavas
High-MgO lavas
source (07DM+03PM)
Fig 14 Trace-element modeling results for incongruent dynamicmantle melting for picritic and tholeiitic lavas from the KarmutsenFormation Three steps of modeling are shown (a) melting of spinellherzolite compared with high-MgO lavas (b) melting of garnet lher-zolite and spinel lherzolite compared with tholeiitic lavas (c) meltingof garnet lherzolite compared with tholeiitic and picritic lavasShaded fields in each panel for tholeiitic basalts and picrites arelabeled Patterns with symbols in each panel are modeling results(patterns are 5 melting increments in (a) and (b) and 1 incre-ments in (c) PM primitive mantle DM depleted MORB mantle gtlherz garnet lherzolite sp lherz spinel lherzolite In (b) the ratios ofper cent melting for garnet and spinel lherzolite are indicated (x gtlherzthorn 4x sp lherz means that for 5 melting 1 melt in garnetfacies and 4 melt in spinel facies where x is per cent melting in theinterval of modeling) The ratio of the respective melt proportions inthe aggregate melt (x gt lherzthorn 4x sp lherz) was kept constant andthis ratio of melt contribution provided the best fit to the data Arange of proportion of source components (07DMthorn 03PM to
09DMthorn 01PM) can reproduce the variation in the high-MgOlavas Melting modeling uses the formulation of Zou amp Reid (2001)an example calculation is shown in their Appendix PM fromMcDonough amp Sun (1995) and DM from Salters amp Stracke (2004)Melt reaction coefficients of spinel lherzolite from Kinzler amp Grove(1992) and garnet lherzolite fromWalter (1998) Partition coefficientsfrom Salters amp Stracke (2004) and Shaw (2000) were kept constantSource mineralogy for spinel lherzolite (018cpx027opx052ol003sp) from Kinzler (1997) and for garnet lherzolite(034cpx008opx053ol005gt) from Salters amp Stracke (2004) Arange of source compositions for melting of spinel lherzolite(07DMthorn 03PM to 03DMthorn 07PM) reproduce REE patternssimilar to those of the high-MgO lavas with best fits for 22ndash25melting and 07DMthorn 03PM source composition Concentrationsof tholeiitic basalts cannot be reproduced from an entirelyprimitive or depleted source
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
30
The uncertainty in the composition of the source in thespinel lherzolite melting model for the high-MgO lavasprecludes determining whether the source was moredepleted in incompatible trace elements than the source ofthe tholeiitic lavas (see Fig 14 caption for description) It ispossible that early high-pressure low-degree melting left aresidue within parts of the plume (possibly the hotter inte-rior portion) that was depleted in trace elements (egElliott et al 1991) further decompression and high-degreemelting of such depleted regions could then have generatedthe depleted high-MgO Karmutsen lavas Shallow high-degree melting would preferentially sample depletedregions whereas deeper low-degree melting may notsample more refractory depleted regions (eg CaribbeanEscuder-Viruete et al 2007) The picrites lie at the highend of the range of Nd isotopic compositions and havelow NbZr similar to depleted Icelandic basalts (Fittonet al 2003) therefore the picrites may represent the partsof the mantle plume that were the most depleted prior tothe initiation of melting As Fitton et al (2003) suggestedthese depleted melts may only rarely be erupted withoutcombining with more voluminous less depleted melts andthey may be detected only as undifferentiated small-volume flows like the Keogh Lake picrites within theKarmutsen Formation on Vancouver IslandDecompression melting within the mantle plume initiatedwithin the garnet stability field (35^25 GPa) at highmantle potential temperatures (Tp414508C) and pro-ceeded beneath oceanic arc lithosphere within the spinelstability field (25^09 GPa) where more extensivedegrees of melting could occur
Magmatic evolution of Karmutsentholeiitic basaltsPrimary magmas in large igneous provinces leave exten-sive coarse-grained cumulate residues within or below thecrust these mafic and ultramafic plutonic sequences repre-sent a significant proportion of the original magma thatpartially crystallized at depth (eg Farnetani et al 1996)For example seismic and petrological studies of OntongJava (eg Farnetani et al 1996) combined with the use ofMELTS (Ghiorso amp Sack 1995) indicate that the volcanicsequence may represent only 40 of the total magmareaching the Moho Neal et al (1997) estimated that30^45 fractional crystallization took place for theOntong Java magmas and produced cumulate thicknessesof 7^14 km corresponding to 11^22 km of flood basaltsdepending on the presence of pre-existing oceanic crustand the total crustal thickness (25^35 km) Interpretationsof seismic velocity measurements suggest the presence ofpyroxene and gabbroic cumulates 9^16 km thick beneaththe Ontong Java Plateau (eg Hussong et al 1979)Wrangellia is characterized by high crustal velocities in
seismic refraction lines which is indicative of mafic
plutonic rocks extending to depth beneath VancouverIsland (Clowes et al 1995)Wrangellia crust is 25^30 kmthick beneath Vancouver Island and is underlain by astrongly reflective zone of high velocity and density thathas been interpreted as a major shear zone where lowerWrangellia lithosphere was detached (Clowes et al 1995)The vast majority of the Karmutsen flows are evolved tho-leiitic lavas (eg low MgO high FeOMgO) indicating animportant role for partial crystallization of melts at lowpressure and a significant portion of the crust beneathVancouver Island has seismic properties that are consistentwith crystalline residues from partially crystallizedKarmutsen magmas To test the proportion of the primarymagma that fractionated within the crust MELTS(Ghiorso amp Sack 1995) was used to simulate fractionalcrystallization using several estimated primary magmacompositions from the PRIMELT1 modeling results Themajor-element composition of the primary magmas couldnot be estimated for the tholeiitic basalts because of exten-sive plagthorn cpxthornol fractionation and thus the MELTSmodeling of the picrites is used as a proxy for the evolutionof major elements in the volcanic sequence A pressure of 1kbar was used with variable water contents (anhydrousand 02wt H2O) calculated at the quartz^fayalite^magnetite (QFM) oxygen buffer Experimental resultsfrom Ontong Java indicate that most crystallizationoccurs in magma chambers shallower than 6 km deep(52 kbar Sano amp Yamashita 2004) however some crys-tallization may take place at greater depths (3^5 kbarFarnetani et al 1996)A selection of the MELTS results are shown in Fig 15
Olivine and spinel crystallize at high temperature until15^20wt of the liquid mass has fractionated and theresidual magma contains 9^10wt MgO (Fig 15f) At1235^12258C plagioclase begins crystallizing and between1235 and 11908C olivine ceases crystallizing and clinopyr-oxene saturates (Fig 15f) As expected the addition of clin-opyroxene and plagioclase to the crystallization sequencecauses substantial changes in the composition of the resid-ual liquid FeO and TiO2 increase and Al2O3 and CaOdecrease while the decrease in MgO lessens considerably(Fig 15) The near-vertical trends for the Karmutsen tho-leiitic basalts especially for CaO do not match the calcu-lated liquid-line-of-descent very well which misses manyof the high-CaO high-Al2O3 low-FeO basalts A positivecorrelation between Al2O3CaO and SrNd indicates thataccumulation of plagioclase played a major role in the evo-lution of the tholeiitic basalts (Fig 15g) Increasing thecrystallization pressure slightly and the water content ofthe melts does not systematically improve the match ofthe MELTS models to the observed dataThe MELTS results indicate that a significant propor-
tion of the crystallization takes place before plagioclase(15^20 crystallization) and clinopyroxene (35^45
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
31
crystallization) join the crystallization sequence Theseresults suggest that445 crystallization of the primarymagmas occurred (estimated from the residual liquid percent in Fig 15f) and if the original flood basalt stratigra-phy on Vancouver Island was 6 km thick at least anequivalent amount of complementary plutonic rocksshould have crystallized The Karmutsen basalts and theirplutonic residues thus represent a significant addition ofcrust (perhaps 412 km thickness) previously thickenedduring the formation of the underlying Paleozoic arc
Growth of the Wrangellia oceanic plateauon Vancouver IslandThe extrusive part of the Wrangellia oceanic plateau onVancouver Island was constructed in a three-layersequence of mostly tholeiitic basalts with a restrictedrange of composition (Fig 16) The plateau overliesDevonian to Mississippian arc volcanic sequences andmarine sediments of Mississippian to Permian ages thatare intruded by gabbroic rocks related to the flood basaltsThe volcanic stratigraphy of theWrangellia plateau formed
00
05
10
15
20
25
30
35
40
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
16
0 2 4 6 8 10 12 14 16 18 20
10
11
12
13
14
15
16
17
18
19
20
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
0 2 4 6 8 10 12 14 16 18 20
MgO (wt)
0 10 20 30 40 50
percent of total mass
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
) Mineral proportions
Sp (e)
Ol
CpxPlag
PicriteHigh-MgO basaltTholeiitic basalt
TiO2 (wt)
MgO (wt)
Al2O3 (wt) CaO (wt)
FeO (wt)
MgO(a) (b)
(c) (d)
MgO (wt)MgO (wt)
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
)
Residual liquid ()
1220
1190
Ol + Sp
Plag+Ol+Sp Plag+Cpx+Sp
02 wt H2O
no H2O
Residual liquid and crystallizing phases
(f)
0102030405060708090100
plagioclaseaccumulation
plagioclaseaccumulation
plagioclase
accumulatio
n
10
12
14
16
0 10 20 30 40
Al2O3CaO
SrNd
tholeiitic basalts
(g)
Fig 15 Forward fractional crystallization modeling results for major elements from MELTS (Ghiorso amp Sack 1995) compared with picriticand tholeiitic lavas from the Karmutsen Formation (a^d) Major-element compositions of residual liquids compared with Karmutsen lavacompositions For clarity the results are shown only for one starting composition (open circle) using the composition of the estimated primarymagma for the high-MgO lava series from the modeling technique of Herzberg et al (2007) for sample 4723A13 as shown in Fig 12 The esti-mated primary magmas for samples 4723A3 and 93G171 yield similar results A pressure of 1 kbar was used with no H2O and 02wt H2Ocalculated at QFM black line is anhydrous and grey line is 02wt H2OThe end-point for the anhydrous result is at 78 fractionation and11308C and the end-point for the result with 02wt H2O is 75 fractionation and11108C (e) Mineral proportions of crystallized phases fromone MELTrun (sample 4723A13) with no H2O (f) Per cent residual liquid vs temperature labelled with intervals of crystallizing phases (g)Al2O3CaO vs SrNd for the tholeiitic basalts
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
32
in a single eruptive phase (c 231^225 Ma) as an emergentbasalt sequence during a deeper-water shallow-water andsubaerial stage and subsided below sea level relativelysoon after its formation (Fig 16) The deeper-water stagecontains pillowed and massive submarine flows (3000mthick) that built up the submarine volcanic edifice from theseafloor The lowest pillow basalts were emplaced onunconsolidated and lithified fine-grained siliciclastic andcarbonaceous sediments containing Daonella beds(c 235^232 Ma) that are intruded by abundantKarmutsen mafic sills (Fig 16a) The different pillowedand massive submarine flows resulted primarily from dif-ferent effusive rates and local topography Some of themassive submarine flows commonly recognized by radiat-ing columnar jointing represent master tubes for the deliv-ery of lava to distal parts of submarine flow fields Thepillowed flows which commonly have uneven topography
in modern sequences may have served as levees to themassive submarine flows rather than allowing unob-structed spreading and inflation typical of the massive sub-aerial flows that form on shallow slopes in continentalflood basalt provinces (eg Self et al 1997) The rapid suc-cession of flows is evident by the near absence of sedimentbetween submarine flows Minor volumes of picritic pillowlavas occur in the upper part of the pillowed flow sequencein areas of the plateau exposed on northern VancouverIsland (Fig 16 Nixon et al 2008)The growth of the plateau in shallow water led to
emplacement of an increasing proportion of volcaniclastics(pillow breccias and hyaloclastites) that accumulated up to1500m thick in some areas (Fig 16b Nixon et al 2008)Horizons of pillowed flows occur locally within the volca-niclastic stratigraphy Sedimentary structures (eg gradedbedding fluidization structures) and petrographic textures
Late Devonian-Mississippian arc
Daonella beds
Middle Triassic
deeper water
Late Devonian-Mississippian arc
pillowed flows massive submarine
flows
pillow brecciahyaloclastiteshallow water
~4 km
~6 km
Late Triassic limestoneinterflow limestonelenses
Karmutsen Formation
mafic intrusions sediment-sill complex
Karmutsen Formation
(a) (b)
(c) (d)
subaerial flows
Late Devonian-Mississippian arc Late Devonian-Mississippian arc
picritic pillow basalts
Miss to Permian marine sediments
sills
Fig 16 Growth of the Wrangellia oceanic plateau onVancouver Island (a) Deep-water eruptive phase of pillowed and massive submarineflows with almost no intervening sediments onto MiddleTriassic shale containing Daonella beds (c 235^232 Ma) The basement of the plateauis composed of Devonian to Mississippian arc sequences overlain by Mississippian to Permian carbonates and siliciclastic sediments(b) Shallow-water flow emplacement with increasing proportion of volcaniclastics resulted primarily from cooling^contraction granulationPicritic pillow basalts occur primarily near the transition between pillowed submarine flows and volcaniclastic units Pillowed and massiveflows occur locally within the volcaniclastic unit (c) Emergent stage of growth consists of overlapping subaerial flow fields similar to continen-tal flood basalts with pahoehoe flow structures Multi-tiered columnar jointing occurrs locally in response to ponding of flows (eg Bondre et al2004) Interflow carbonate deposits and plagioclase megacrystic flows developed locally during the waning stage of volcanism (d) Post-volcanicsubsidence and deposition of Late Carnian to Norian (c 212^223 Ma) platform carbonates on top of the plateau Units are labeled in the dia-grams Thicknesses of the underlying units are not to scale
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
33
within the volcaniclastic rocks indicate formation primar-ily via cooling^contraction granulation magma^water^steam interaction autobrecciation resedimentation andmass-wasting rather than pyroclastic eruption (Nixonet al 2008)A broad subaerial platform was constructed by emplace-
ment of subaerial lava flows during an emergent subaerialstage (Fig 16c) The subaerial unit predominantly com-prises well-layered sheet flows that formed in a similarway to continental flood basalts as hundred(s) of pahoehoeflow fields erupted on shallow slopes (Self et al 1997) how-ever volcaniclastic and pillowed flow units are locally pre-sent mostly in the upper parts of the stratigraphyLocalized ponding of flows led to the formation of multi-tiered columnar jointing in some subaerial flows (egBondre et al 2004) As volcanism waned local interflowcarbonate deposits developed and plagioclase megacrysticflows were erupted Subsidence of the plateau is recordedby the interflow lenses and deposits of platform carbonates(40^1500m thick) on top of the Karmutsen basalts thatcontain Late Carnian to Early Norian fossils (Fig 16d c216^227 Ma Nixon et al 2008)The evolution of the internal architecture of the
Wrangellia oceanic plateau was primarily related to theconditions of emplacement (ie water depth and eruptionenvironment) There does not appear to have been a sys-tematic coincident evolution or shift of lava compositionsonVancouver Island as has been recognized in most con-tinental flood basalts and parts of the Wrangellia oceanicplateau in Alaska and Yukon (Greene et al 2008 2009)In general with the exception of the picritic pillow lavasin the submarine stratigraphy and local variations betweenhigh-MgO and tholeiitic lavas there are no significantcompositional differences between lavas erupted duringthe different stages of growth of the plateau (ie submarineand subaerial flows) nor are substantial compositional dif-ferences generally found between massive submarine flowsin contact with pillowed flows tholeiitic basalt flows nearthe base of the Karmutsen Formation are compositionallysimilar to flows sampled near the top of the volcanicsequence Where it was possible to sample continuous sec-tions of volcanic stratigraphy greater than several hun-dreds of meters (eg Keogh^Maynard Lake area west ofthe Karmutsen Range Fig 2) there is no clear relationshipbetween stratigraphic position and compositionalvariation
CONCLUSIONSThe Karmutsen Formation covers extensive areas ofVancouver Island (20 000 km2) and forms part of a largeoceanic plateau erupted over a geologically short interval(c 225^231 Ma) The tripartite volcanic stratigraphyexposed on Vancouver Island is at least 6 km thick andcomprises a succession of submarine flows volcaniclastic
deposits and massive subaerial flowsThese volcanologicaldifferences are primarily related to the eruption environ-ment (deep-water shallow-water or subaerial) The rapidgrowth of the plateau prevented significant accumulationsof intervening sediment except in the uppermost stratigra-phy where isolated limestone lenses commonly associatedwith pillow basalt and hyaloclastite deposits preserve arecord of the final stages of subsidence of the plateauThe Wrangellia plateau on Vancouver Island was con-
structed dominantly of tholeiitic basalt with restrictedmajor- and trace-element and isotopic compositional var-iations Minor volumes of picritic pillow basalts eruptedlate in the submarine phase The Karmutsen basalts wereemplaced onto older Paleozoic arc volcanic sequences andlow NbLa for the picrites may indicate a contributionfrom the sub-arc lithospheric mantle Modeling of the pet-rogenesis of the Karmutsen picrites indicates melting ofanomalously hot mantle (15008C) and extensive degreesof partial melting (23^27) and is consistent with aplume initiation model Combined Sr^Nd^Hf^Pb isotopicsystematics indicate that the picritic and tholeiitic lavaswere derived from a depleted plume-type mantle sourcedistinct from the source of MORB that has compositionalsimilarities to the source of the lavas forming theCaribbean Plateau Picrites were generated by melting ofspinel lherzolite whereas the tholeiitic basalts formed asaggregate melts of both garnet and spinel lherzolite Thetholeiitic basalts underwent extensive low-pressure fractio-nation (52^3 kbar) and seismic studies indicate that someof theWrangellia crust beneathVancouver Island may cor-respond to the plutonic residues of the Karmutsen floodbasalts
ACKNOWLEDGEMENTSWe are grateful to Nick Arndt for helping us get this pro-ject started and Nick Massey for insights into VancouverIsland geology We greatly appreciate the assistance withPRIMELT modeling from Claude Herzberg We thankMikkel Schau for his insight and enthusiasm during field-work Jane Barling assisted with analyses by MC-ICP-MSThoughtful and thorough reviews of the manuscript byJohn Mahoney Fred Frey Wendy Bohrson DougalJerram and Teal Riley are greatly appreciated Fundingwas generously provided by the Rocks to Riches Programadministered by the BC amp Yukon Chamber of Mines in2004 by the BC Geological Survey and by NSERCDiscovery Grants to JSS and DW ARG was supportedby a University Graduate Fellowship at UBC
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
34
REFERENCESBarker F Brown A S Budahn J R amp Plafker G (1989) Back-arc
with frontal-arc component origin of Triassic Karmutsen basaltBritish Columbia Canada Chemical Geology 75 81^102
Blichert-Toft JWeis D Maerschalk C Agranier A amp Albare de F(2003) Hawaiian hot spot dynamics as inferred from the Hf and Pbisotope evolution of Mauna Kea volcano Geochemistry GeophysicsGeosystems 4(2) 1^27 doi1010292002GC000340
Bondre N R Duraiswami R A amp Dole G (2004) Morphologyand emplacement of flows from the Deccan Volcanic ProvinceIndia Bulletin of Volcanology 66 29^45
Carlisle D (1963) Pillow breccias and their aquagene tuffs QuadraIsland British Columbia Journal of Geology 71 48^71
Carlisle D (1972) Late Paleozoic to Mid-Triassic sedimentary-volcanic sequence on Northeastern Vancouver Island Geological
Society of Canada Paper Report of Activities 72-1 Part B 24^30Carlisle D amp SuzukiT (1974) Emergent basalt and submergent car-
bonate^clastic sequences including the Upper Triassic Dilleri andWelleri zones on Vancouver Island Canadian Journal of Earth
Sciences 11 254^279Chauvel C amp Blichert-Toft J (2001) A hafnium isotope and trace
element perspective on melting of the depleted mantle Earth and
Planetary Science Letters 190 137^151Chayes F (1954)The theory of thin-section analysis Journal of Geology
62 92^101Cho M Liou J G amp Maruyama S (1987) Transition from the zeo-
lite to prehnite^pumpellyite facies in the Karmutsen metabasitesVancouver Island British Columbia Journal of Petrology 27(2)467^494
Clowes R M Zelt C A Amor J R amp Ellis R M (1995)Lithospheric structure in the southern Canadian Cordillera froma network of seismic refraction lines Canadian Journal of Earth
Sciences 32(10) 1485^1513Coffin M F amp Eldholm O (1994) Large igneous provinces Crustal
structure dimensions and external consequences Reviews of
Geophysics 32(1) 1^36Cox K G (1980) A model for flood basalt volcanism Journal of
Petrology 21 629^650Eldholm O amp Coffin M F (2000) Large igneous provinces
and plate tectonics In Richards M A Gordon R G amp vander Hilst R D (eds) The History and Dynamics of Global
Plate Motions American Geophysical Union Geophysical Monograph 121309^326
Elliott T R Hawkesworth C J amp Grolaquo nvold K (1991) Dynamicmelting of the Iceland plume Nature 351 201^206
Escuder-Viruete J Pecurren rez-Estaucurren n A Contreras F Joubert MWeis D Ullrich T D amp Spadea P (2007) Plume mantle sourceheterogeneity through time Insights from the Duarte ComplexHispaniola northeastern Caribbean Journal of Geophysical Research112(B04203) doi1010292006JB004323
Farnetani C G Richards M A amp Ghiorso M S (1996)Petrological models of magma evolution and deep crustal structurebeneath hotspots and flood basalt provinces Earth and Planetary
Science Letters 143 81^94Fitton J G amp Godard M (2004) Origin and evolution of magmas
on the Ontong Java Plateau In Fitton J G Mahoney J JWallace P J amp Saunders A D (eds) Origin and Evolution of the
Ontong Java Plateau Geological Society London Special Publications 229151^178
Fitton J G Saunders A D Norry M J Hardarson B S ampTaylor R N (1997) Thermal and chemical structure of theIceland plume Earth and Planetary Science Letters 153 197^208
Fitton J G Saunders A D Kempton P D amp Hardarson B S(2003) Does depleted mantle form an intrinsic part of the Icelandplume Geochemistry Geophysics Geosystems 4(3) 1032 doi1010292002GC000424
Frey F A Huang S Blichert-Toft J Regelous M amp Boyet M(2005) Origin of depleted components in basalt related to theHawaiian hotspot Evidence from isotopic and incompatible ele-ment ratios Geochemistry Geophysics Geosystems 6(1) 23 doi1010292004GC000757
Galer S J G amp Abouchami W (1998) Practical application of leadtriple spiking for correction of instrumental mass discriminationMineralogical Magazine 62A 491^492
Ghiorso M S amp Sack R O (1995) Chemical mass transfer in mag-matic processes IV A revised and internally consistent thermody-namic model for the interpolation and extrapolation of liquid^solidequilibria in magmatic systems at elevated temperatures and pres-sures Contributions to Mineralogy and Petrology 119 197^212
Greene A R Scoates J S amp Weis D (2005)WrangelliaTerrane onVancouver Island Distribution of flood basalts with implicationsfor potential Ni^Cu^PGE mineralization In Grant B (ed)Geological Fieldwork 2004 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2005-1 209^220Greene A R Scoates J S Nixon G T amp Weis D (2006) Picritic
lavas and basal sills in the Karmutsen flood basalt provinceWrangellia northern Vancouver Island In Grant B (ed)Geological Fieldwork 2005 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2006-1 39^52Greene A R Scoates J S amp Weis D (2008) Wrangellia flood
basalts in Alaska A record of plume^lithosphere interaction in aLate Triassic accreted oceanic plateau Geochemistry Geophysics
Geosystems 9(Q12004) doi1010292008GC002092Greene A R Scoates J S Weis D amp Israel S (2009)
Geochemistry of triassic flood basalts from the Yukon (Canada)segment of the accreted Wrangellia oceanic plateau Lithosdoi101016jlithos200811010
Hauff F Hoernle K Tilton G Graham D W amp Kerr A C(2000) Large volume recycling of oceanic lithosphere over shorttime scales geochemical constraints from the Caribbean LargeIgneous Province Earth and Planetary Science Letters 174 247^263
Hauff F Hoernle K amp Schmidt A (2003) Sr^Nd^Pb compositionof Mesozoic Pacific oceanic crust (Site 1149 and 801 ODP Leg185)Implications for alteration of ocean crust and the input into theIzu^Bonin^Mariana subduction system Geochemistry Geophysics
Geosystems 4(8) doi1010292002GC000421Herzberg C (2004) Partial melting below the Ontong Java Plateau
In Fitton J G Mahoney J J Wallace P J amp Saunders A D(eds) Origin and Evolution of the Ontong Java Plateau Geological SocietyLondon Special Publications 229 179^183
Herzberg C (2006) Petrology and thermal structure of the Hawaiianplume from Mauna Kea volcano Nature 444(7119) 605
Herzberg C amp Asimow P D (2008) Petrology of some oceanicisland basalts PRIMELT2XLS software for primary magma cal-culation Geochemistry Geophysics Geosystems 9(Q09001) doi1010292008GC002057
Herzberg C amp OrsquoHara M J (2002) Plume-associated ultramaficmagmas of Phanerozoic age Journal of Petrology 43(10) 1857^1883
Herzberg C Asimow P D Arndt N Niu Y Lesher C MFitton J G Cheadle M J amp Saunders A D (2007)Temperatures in ambient mantle and plumes Constraints frombasalts picrites and komatiites Geochemistry Geophysics Geosystems8(Q02006) doi1010292006GC001390
Huang S amp Frey F A (2005) Trace element abundances of MaunaKea basalt from phase 2 of the Hawaii Scientific Drilling Project
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
35
Petrogenetic implications of correlations with major element con-tent and isotopic ratios Geochemistry Geophysics Geosystems 4(6)doi1010292002GC000322
Hussong D M Wipperman L K amp Kroenke L W (1979) Thecrustal structure of the Ontong Java and Manihiki OceanicPlateaus Journal of Geophysical Research 84(B11) 6003^6010
Jerram D A amp Widdowson M (2005) The anatomy of ContinentalFlood Basalt Provinces geological constraints on the processes andproducts of flood volcanism Lithos 79(3^4) 385^405
Jones D L Silberling N J amp Hillhouse J (1977)Wrangellia a dis-placed terrane in northwestern North America CanadianJournal ofEarth Sciences 14(11) 2565^2577
Kerr A C (2003) Oceanic Plateaus In Rudnick R (ed) The Crust
Treatise on GeochemistryVol 3 Oxford Elsevier Science pp 537^565Kerr A C amp Mahoney J J (2007) Oceanic plateaus Problematic
plumes potential paradigms Chemical Geology 241 332^353Kerr A CTarney J Marriner G F Klaver GT Saunders A D
amp Thirlwall M F (1996) The geochemistry and petrogenesis ofthe Late-Cretaceous picrites and basalts of Curacao NetherlandsAntilles a remnant of an oceanic plateau Contributions to
Mineralogy and Petrology 124(1) 29^43Kerr A C Marriner G F Tarney J Nivia A Saunders A D
Thirlwall M F amp Sinton A C (1997) Cretaceous basaltic ter-ranes in Western Colombia Elemental chronological and Sr-Ndisotopic constraints on petrogenesis Journal of Petrology 38(6)677^702
Kinzler R J (1997) Melting of mantle peridotite at pressuresapproaching the spinel to garnet transition Application to mid-ocean ridge basalt petrogenesis Journal of Geophysical Research
102(B1) 853^874Kinzler R J amp Grove T L (1992) Primary magmas of mid-ocean
ridge basalts 1 Experiments and methods Journal of Geophysical
Research 97(B5) 6885^6906Korenaga J (2005) Why did not the Ontong Java Plateau form sub-
aerially Earth and Planetary Science Letters 234 385^399Kuniyoshi S (1972) Petrology of the Karmutsen (Volcanic) Group
northeastern Vancouver Island British Columbia PhD disserta-tion University of California Los Angeles 242 pp
Lassiter J C DePaolo D J amp Mahoney J J (1995) Geochemistry ofthe Wrangellia flood basalt province Implications for the role ofcontinental and oceanic lithosphere in flood basalt genesis Journalof Petrology 36(4) 983^1009
LeBas M J (2000) IUGS reclassification of the high-Mg and picriticvolcanic rocks Journal of Petrology 41(10) 1467^1470
Longhi J (2002) Some phase equilibrium systematics of lherzolitemelting I Geochemistry Geophysics Geosystems 3(3) doi1010292001GC000204
MacDonald G A amp Katsura T (1964) Chemical composition ofHawaiian lavas Journal of Petrology 5 82^133
Mahoney J J (1987) An isotopic survey of Pacific oceanic plateausimplications for their nature and origin In Keating B HFryer P Batiza R amp Boehlert GW (eds) Seamounts Islands andAtolls American Geophysical Union Geophysical Monograph 43 207^220
Mahoney J LeRoex A P Peng Z Fisher R L amp Natland J H(1992) Southwestern limits of Indian Ocean Ridge mantle and theorigin of low 206Pb204Pb mid-ocean ridge basalt isotope systema-tics of the central Southwest Indian Ridge (178^508E) Journal ofGeophysical Research 97(B13) 19771^19790
Mahoney J J Sinton J M Kurz M D MacDougall J DSpencer K J amp Lugmair GW (1994) Isotope and trace elementcharacteristics of a super-fast spreading ridge East Pacific rise13^238S Earth and Planetary Science Letters 121 173^193
Massey N W D (1995a) Geology and mineral resources of theAlberni^Nanaimo Lakes sheet Vancouver Island 92F1W 92F2Eand part of 92F7E British Columbia Ministry of Energy Mines and
Petroleum Resources Paper 1992-2 132 ppMassey N W D (1995b) Geology and mineral resources of the
Cowichan Lake sheet Vancouver Island 92C16 British Columbia
Ministry of Energy Mines and Petroleum Resources Paper 1992-3 112 ppMassey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005a) Digital Geology Map of British Columbia Tile NM9 MidCoast BC British Columbia Ministry of Energy and Mines Geofile
2005-2Massey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005b) Digital Geology Map of British Columbia Tile NM10Southwest BC British Columbia Ministry of Energy and Mines Geofile
2005-3McDonough W F amp Sun S (1995) The composition of the Earth
Chemical Geology 120 223^253Monger JW H amp Journeay J M (1994) Basement geology and tec-
tonic evolution of theVancouver region In Monger JW H (ed)Geology and Geological Hazards of the Vancouver Region Southwestern
British Columbia Geological Survey of Canada Bulletin 481 3^25Muller J E (1977) Geology of Vancouver Island Geological Survey of
Canada Open File Map 463Muller J E (1980)The Paleozoic Sicker Group of Vancouver Island British
Columbia Geological Survey of Canada Paper 79-30 22 ppMuller J E Northcote K E amp Carlisle D (1974) Geology and
mineral deposits of Alert Bay^Cape Scott map area VancouverIsland British Columbia Geological Survey of Canada Paper 74-8 77
Neal C R Mahoney J J Kroenke L W Duncan R A ampPetterson M G (1997) The Ontong^Java Plateau InMahoney J J amp Coffin M F (eds) Large Igneous Provinces
Continental Oceanic and Planetary FloodVolcanism American Geophysical
Union Geophysical Monograph 100 183^216Niu Y Collerson K D Batiza R Wendt J I amp Regelous M
(1999) Origin of enriched-typed mid-ocean ridge basalt at ridgesfar from mantle plumes The East Pacific Rise at 11820rsquoN Journalof Geophysical Research 104 7067^7088
Nixon G T amp Orr A J (2007) Recent revisions to the EarlyMesozoic stratigraphy of Northern Vancouver Island (NTS 102I092L) and metallogenic implicatinos British Columbia InGrant B (ed) Geological Fieldwork 2006 British Columbia Ministry of
Energy Mines and Petroleum Resources Paper 2007-1 163^177Nixon GT Hammack J L KoyanagiV M Payie G J Haggart
JW Orchard M JTozerT Friedman R M Archibald D APalfy J amp Cordey F (2006a) Geology of the Quatsino^PortMcNeill area northernVancouver Island British Columbia Ministry
of Energy Mines and Petroleum Resources Geoscience Map 2006-2 scale150 000
Nixon G T Kelman M C Stevenson D Stokes L A ampJohnston K A (2006b) Preliminary geology of the Nimpkishmap area (NTS 092L07) northern Vancouver Island BritishColumbia In Grant B amp Newell J M (eds) Geological Fieldwork2005 British Columbia Ministry of Energy Mines and Petroleum Resources
Paper 2006-1 135^152Nixon G T Laroque J Pals A Styan J Greene A R amp
Scoates J S (2008) High-Mg lavas in the Karmutsen flood basaltsnorthernVancouver Island (NTS 092L) Stratigraphic setting andmetallogenic significance In Grant B (ed) Geological Fieldwork
2007 British Columbia Ministry of Energy Mines and Petroleum
Resources Paper 2008-1 175^190Nowell G M Kempton P D Noble S R Fitton J G
Saunders A Mahoney J J amp Taylor R N (1998) High precisionHf isotope measurements of MORB and OIB by thermal ionisation
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
36
mass spectrometry insights into the depleted mantle Chemical
Geology 149 211^233Parrish R R amp McNicoll V J (1992) U^Pb age determinations
from the southern Vancouver Island area British Columbia InRadiogenic Age and Isotopic Studies Report 5 Geological Survey of
Canada Paper 91-2 79^86Peate DW (1997) The Parana^Etendeka Province In Coffin M F
amp Mahoney J (eds) Large Igneous Provinces Continental Oceanic andPlanetary Flood Volcanism American Geophysical Union Geophysical
Monograph 100 217^245Perfit M R Fornari D J Ridley W I Kirk P D Casey J
Kastens K A Reynolds J R Edwards M Desonie DShuster R amp Paradis S (1996) Recent volcanism in the Siqueirostransform fault picritic basalts and implications for MORBmagma genesis Earth and Planetary Science Letters 141(1^4) 91^108
Pretorius W Weis D Williams G Hanano D Kieffer B ampScoates J S (2006) Complete trace elemental characterization ofgranitoid (USGSG-2GSP-2) reference materials by high resolutioninductively coupled plasma-mass spectrometry Geostandards and
Geoanalytical Research 30(1) 39^54Regelous M Niu Y Wendt J I Batiza R Greig A amp
Collerson K D (1999) Variations in the geochemistry of magma-tism on the East Pacific Rise at 10830rsquoN since 800 ka Earth and
Planetary Science Letters 168 45^63Recurren villon S Chauvel C Arndt N T Pik R Martineau F
Fourcade S amp Marty B (2002) Heterogeneity of the Caribbeanplateau mantle source Sr O and He isotopic compositions of oli-vine and clinopyroxene from Gorgona Island Earth and Planetary
Science Letters 205(1^2) 91Rhodes J M (1996) Geochemical stratigraphy of lava flows samples
by the Hawaii Scientific Drilling Project Journal of Geophysical
Research 101(B5) 11729^11746Rhodes J M amp Vollinger M J (2004) Composition of basaltic lavas
sampled by phase-2 of the Hawaii Scientific Drilling ProjectGeochemical stratigraphy and magma types Geochemistry
Geophysics Geosystems 5(3) doi1010292002GC000434Richards M A Jones D L Duncan R A amp DePaolo D J (1991)
A mantle plume initiation model for the Wrangellia flood basaltand other oceanic plateaus Science 254 263^267
Salters V J M amp Hart S R (1991) The mantle sources of oceanridges islands and arcs the Hf-isotope connection Earth and
Planetary Science Letters 104 364^380Salters V J M amp Stracke A (2004) Composition of the depleted
SaltersV J amp WhiteW (1998) Hf isotope constraints on mantle evo-lution Chemical Geology 145 447^460
SanoT amp Yamashita S (2004) Experimental petrology of basementlavas from Ocean Drilling Program Leg 192 implications for dif-ferentiation processes in Ontong Java Plateau magmas InFitton J G Mahoney J JWallace P J amp Saunders A D (eds)Origin and Evolution of the Ontong Java Plateau Geological Society
London Special Publications 229 185^218Saunders A D (2005) Large igneous provinces Origin and environ-
mental consequences Elements 1 259^263Self S ThordarsonT amp Keszthely L (1997) Emplacement of conti-
nental flood basalt lava flows In Mahoney J J amp Coffin M F(eds) Large Igneous Provinces Continental Oceanic and Planetary Flood
Volcanism American Geophysical Union Geophysical Monograph 100381^410
Shaw D M (2000) Continuous (dynamic) melting theory revisitedCanadian Mineralogist 38(5) 1041^1063
Sluggett C L (2003) Uranium^lead age and geochemical constraintson Paleozoic and Early Mesozoic magmatism in WrangelliaTerrane Saltspring Island British Columbia BSc thesisUniversity of British ColumbiaVancouver 84 pp
Storey M Mahoney J J amp Saunders A D (1997) Cretaceousbasalts in Madagascar and the transition between plume and con-tinental lithospheric mantle sources In Mahoney J J ampCoffin M F (eds) Large Igneous Provinces Continental Oceanic and
Planetary Flood Volcanism Aamerican Geophysical Union Geophysical
Monograph 100 95^122Surdam R C (1967) Low-grade metamorphism of the Karmutsen
Group PhD dissertation University of California Los Angeles288 pp
Sutherland-Brown A Yorath C J Anderson R G amp Dom K(1986) Geological maps of southern Vancouver IslandLITHOPROBE I 92C10 11 14 16 92F1 2 7 8 Geological Survey ofCanada Open File 1272
Tejada M L G Mahoney J J Castillo P R Ingle S P Sheth HC amp Weis D (2004) Pin-pricking the elephant evidence on theorigin of the Ontong Java Plateau from Pb^Sr^Hf^Nd isotopiccharacteristics of ODP Leg 192 basalts In Fitton J GMahoney J J Wallace P J amp Saunders A D (eds) Origin and
Evolution of the Ontong Java Plateau Geological Society London Special
Publications 229 133^150Thompson P M E Kempton P D amp Kerr A C (2008) Evaluation
of the effects of alteration and leaching on Sm^Nd and Lu^Hf sys-tematics in submarine mafic rocks Lithos 104(1^4) 164^176
Thompson P M E Kempton P D White R V Kerr A CTarney J Saunders A D Fitton J G amp McBirney A (2003)Hf-Nd isotope constraints on the origin of the CretaceousCaribbean plateau and its relationship to the Galapagos plumeEarth and Planetary Science Letters 217(1^2) 59^75
Vervoort J D Patchett P Blichert-Toft J amp Albare de F (1999)Relationships between Lu^Hf and Sm^Nd isotopic systems in theglobal sedimentary system Earth and Planetary Science Letters 16879^99
Walter M J (1998) Melting of garnet peridotite and the origin ofkomatiite and depleted lithosphere Journal of Petrology 39(1) 29^60
Weis D Kieffer B Maerschalk C Barling J de Jong JWilliams G A Hanano D Mattielli N Scoates J SGoolaerts A Friedman R A amp Mahoney J B (2006) High-pre-cision isotopic characterization of USGS reference materials byTIMS and MC-ICP-MS Geochemistry Geophysics Geosystems
7(Q08006) doi1010292006GC001283Weis D Kieffer B Hanano D Silva I N Barling J
Pretorius W Maerschalk C amp Mattielli N (2007) Hf isotopecompositions of US Geological Survey reference materialsGeochemistry Geophysics Geosystems 8(Q06006) doi1010292006GC001473
Wheeler J O amp McFeely P (1991) Tectonic assemblage map of theCanadian Cordillera and adjacent part of the United States ofAmerica Geological Survey of Canada Map 1712A
WhiteW M Albare de F amp Tecurren louk P (2000) High-precision analy-sis of Pb isotope ratios by multi-collector ICP-MS Chemical Geology167 257^270
Yorath C J Sutherland Brown A amp Massey N W D (1999)LITHOPROBE southern Vancouver Island British Columbia Geological
Survey of Canada Bulletin 498 145 ppZou H (1998) Trace element fractionation during modal and
non-model dynamic melting and open-system melting Amathematical treatment Geochimica et Cosmochimica Acta 62(11)1937^1945
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
37
Zou H amp Reid M R (2001) Quantitative modeling of trace elementfractionation during incongruent dynamic melting Geochimica et
Cosmochimica Acta 65(1) 153^162
APPENDIX SAMPLEPREPARAT ION ANDANALYT ICAL METHODSOnly the freshest rocks in the field were sampled and onlythe least altered samples were selected for chemical andisotopic analysis based on thorough petrographic inspec-tion Fifty-nine of the 129 collected samples were crushed(400 g) into small pieces 52mm in diameter in aRocklabs hydraulic piston crusher between WC plates tominimize contamination The coarse-crush was mixedand 100 g was powdered in a planetary mill using agatejars and balls cleaned with quartz sand between samples
ActLabs analytical methodsThe major- and trace-element compositions of the whole-rock powders were determined at Activation LaboratoriesLtd (Actlabs) in Ancaster Ontario Analytical techniquesand detection limits are also available from Actlabs (httpwwwactlabscommethsub_code4erehtm) The analyticalmethod for each of the elements analyzed is indicated inTable 2 For the major elements a 02 g sample was mixedwith a lithium metaborate^lithium tetraborate mixtureand fused in a graphite crucible The molten mixture waspoured into a 5 HNO3 solution and shaken until dis-solved (30min) The samples were analyzed for majoroxides and selected trace elements by inductively coupledplasma optical emission spectrometry (ICP-OES) on acombination simultaneous^sequential Thermo Jarrell-AshEnviro II system Internal calibration was achieved usinga variety of international reference materials (eg W-2BIR-1 DNC-1) and independent control samplesAdditional trace elements were analyzed by both instru-mental neutron activation analysis (INAA) and induc-tively couple plasma mass spectrometry (ICP-MS) ForINAA15^25 g of sample was weighed into small polyeth-ylene vials and irradiated with a control international ref-erence material CANMET WMS-1 and NiCr flux wires ata thermal neutron flux of 71012 neutronscm2 per s in theMcMaster Nuclear Reactor Following a 7 day waitingperiod the samples were measured on an Ortec high-purity Ge detector linked to a Canberra Series 95 multi-channel analyzer Activities for each element weredecay- and weight-corrected and compared with a detectorcalibration developed from multiple international certifiedreference materials For ICP-MS 025 g of sample wasdigested in HF followed by HNO3^HClO4 heated andtaken to dryness The samples were brought back into solu-tion with HCl Samples were analyzed using a PerkinElmer Optima 3000 ICP In-lab standards or certified ref-erence materials (eg W-2 BIR-1 DNC-1) were used for
quality control A total of 15 blind duplicates were ana-lyzed from the same sample powders to assessreproducibility
University of Massachusetts analyticalmethodsFourteen sample duplicate powders taken from the same100 g sample aliquot (high-MgO basalts and picrites)were also analyzed at the Ronald B Gilmore X-RayFluorescence (XRF) Laboratory at the University ofMassachusetts Major elements were measured on a fusedLa-bearing lithium borate glass disc using a SiemensMRS-400 spectrometer Trace element concentrations(Rb Sr Ba Ce Nb Zr Y Pb Zn Ga Ni Cr V) weremeasured on a separate powder pellet using a PhilipsPW2400 sequential spectrometer using a Rh tube LOIand ferrous iron measurements were made as describedby Rhodes amp Vollinger (2004) Precision and accuracy esti-mates for the data have been described by Rhodes (1996)and Rhodes amp Vollinger (2004) Results for each sampleare the average of two separate analyses and are shown inElectronic Appendix 2 for a comparison of analyses ofduplicate powders using XRFand ICP-MS
PCIGR analytical methodsA subset of 19 samples was selected for high-precisiontrace-element analysis and Sr Nd Pb and Hf isotopicanalysis at the Pacific Centre for Isotopic andGeochemical Research (PCIGR) at the University ofBritish Columbia (UBC) Samples were selected from the59 samples analyzed for whole-rock chemistry at ActLabsbased on major- and trace-element chemistry alteration(low LOI and petrographic alteration index) and samplelocation Samples were prepared for trace-element analysisat the PCIGR by the technique described by Pretorius et al(2006) on unleached rock powders Sample powders(100mg) were weighed in 7ml screw-top Savillex bea-kers and dissolved in 1ml 14N HNO3 and 5ml 48 HFon a hotplate for 48 h at 1308C with periodic ultrasonica-tion Samples were dried and redissolved in 6ml 6N HClon a hotplate for 24 h and then dried and redissolved in1ml concentrated HNO3 for 24 h before final dryingTrace element abundances were measured by high-resolution inductively coupled plasma-mass spectrometry(HR-ICP-MS) with aThermo Finnigan Element2 systemfollowing the procedures described by Pretorius et al(2006) within 24 h of redissolution HFSE and LILE weremeasured in medium-resolution mode at 2000 dilutionusing a PFATeflon spray chamber washed with aqua regiafor 3min between samples REE were measured in high-resolution mode and U and Pb in low-resolution modeat 2000 dilution using a glass spray chamber washedwith 2 HNO3 between samples Total procedural blanksand reference materials (BCR-2 BHVO-2) were analyzedwith the batch of samples Indium was used as an internal
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
38
standard in all samples and standard solutionsBackground and standard solutions were analyzed afterevery five samples to detect memory effects and massdrift Results for PCIGR trace-element analyses areshown in Electronic Appendix 3 for the full set of analyzedtrace elementsSample digestion for purification of Sr Nd Hf and Pb
for column chemistry began by weighing each samplepowder (400^500mg) prior to leaching All samples wereleached with 6N HCl and placed in an ultrasonic bath for15min Samples were rinsed twice with 18 M-cm H2Obetween each leaching step (15 total) until the supernatantliquid was clear [following the method of Mahoney (1987)]Samples were then dried on a hotplate for 24 h andweighed again Sample solutions were then prepared bydissolving 100^250mg of the leached powder dissolvedin 1ml 14N HNO3 and 10ml 48 HF on a hotplate for48 h at 1308C with periodic ultrasonication Samples weredried and redissolved in 6ml 6N HCl on a hotplate for24 h and then dried Pb was separated using anion exhangecolumns and the discard was used for Sr REE and Hfseparation Nd was separated from the other REE and Hfrequired two additional purification steps Detailed proce-dures for column chemistry for separating Sr Nd and Pbat the PCIGR have been described byWeis et al (2006) andHf purification has been described byWeis et al (2007) Srand Nd isotope ratios were measured by thermal ioniza-tion mass spectrometry (TIMS) on a Thermo FinniganTriton system in static mode with relay matrix rotation ona single Ta and double Re^Ta filament respectively Fourto five filaments per barrel of 21 were occupied by stan-dards (NBS 987 for Sr and LaJolla for Nd) for each barrelwhere samples were run Sample Sr and Nd isotopic com-positions were corrected for mass fractionation using86Sr88Srfrac14 01194 and 146Nd144Ndfrac14 07219 Each samplewas then normalized using the barrel average of the refer-ence material relative to the values of143Nd144Ndfrac14 0511858 and 87Sr86Srfrac14 0710248 (Weiset al 2006) During the course of the Vancouver Islandanalyses the La Jolla Nd standard gave an average valueof 05118566 (nfrac14 7) and NBS987 standard gave an aver-age of 07102408 (nfrac1411) (2 error is reported as times106) 147Sm144Nd errors are 15 or 0006 USGeological Survey (USGS) reference material BHVO-2was processed with the samples and yielded Sr and Nd iso-topic ratios of 07034607 and 05129786 respectivelyThese are in agreement with the published values of070347920 and 051298411 respectively (Weis et al2006)
Pb and Hf isotopic compositions were analyzed by mul-tiple collector-inductively coupled plasma-mass spectrome-try (MC-ICP-MS) on a static multi-collection on a NuPlasma (Nu Instruments) system The detailed analyticalprocedure for Pb isotopic analyses on the Nu Plasma atthe PCIGR has been described by Weis et al (2006) Theconfiguration for Pb analyses allows for collection of PbTl and Hg togetherTl and Hg are used to monitor instru-mental mass discrimination and isobaric overlap repec-tively All sample solutions were analyzed withapproximately the same PbTl ratio (4) as the referencematerial NIST SRM 981 To accomplish this a small ali-quot of each sample solution from the Pb columns was ana-lyzed on the Element2 to determine the precise amount ofPb available for analysis on the Nu Plasma The SRM 981standard was run after every two samples on the NuPlasma During the time samples were run analyses ofthe SRM 981 Pb reference material gave values of206Pb204Pbfrac1416940322 207Pb204Pbfrac1415495823and 208Pb204Pbfrac14 36713164 (nfrac14 61 2 error is reportedas times 104) these values are in excellent agreement withreported TIMS triple-spike values of Galer amp Abouchami(1998) Fractionation-corrected Pb isotopic ratios were fur-ther corrected by the sample-standard bracketing methodor the ln^ln correction method described by White et al(2000) and Blichert-Toft et al (2003) Leached powders ofUSGS reference material BHVO-2 yielded Pb isotopicratios of 206Pb204Pbfrac14186454 8 207Pb204Pbfrac14 1549105 and 208Pb204Pbfrac14 38222514 and BCR-2yielded 206Pb204Pbfrac141880466 207Pb204Pbfrac14 1562518 and 208Pb204Pbfrac14 3883496 (2 error isreported as times 104) These values are in agreement withthose for leached residues of BHVO-2 and BCR-2 fromWeis et al (2006)Hf isotopic compositions were analyzed following the
procedures detailed byWeis et al (2007) The configurationfor Hf analyses monitored Lu mass 175 andYb mass 172 toallow for interference correction to masses 174 and 176 Hfisotopic ratios were normalized internally for mass fractio-nation to a 179Hf177Hf ratio of 07325 using an exponentialcorrection Standards were run after every two samplesand sample results were normalized to the ratio of the in-run daily average and a 176Hf177Hf ratio for JMC-475 of0282160 During the course of analyses the Hf standardJMC-475 gave an average value 02821679 (nfrac14130)USGS reference materials BCR-2 and BHVO-2 were pro-cessed with the samples and yielded Hf isotopic ratios of02828675 and 02831005 respectively Publishedvalues for BCR-2 and BHVO-2 are 02828717 and0283104 8 respectively (Weis et al 2007)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
39
determinations on intrusive rocks that are considered to berelated to the Karmutsen volcanics Daonella-bearing shale100^200m below the base of the pillow basalts on SchoenMountain and Halobia-rich shale interlayered with flows inthe upper part of the Karmutsen indicate eruption of mostof the flood basalts between the Middle Ladinian (c 232^235 Ma Middle Triassic) and latest Carnian or possiblyearliest Norian (c 216^227 Ma Late Triassic Carlisle ampSuzuki 1974) The only published U^Pb age is based on asingle concordant analysis of a multi-grain baddeleyitefraction from a gabbro on southernVancouver Island thatyielded a 206Pb238U age of 227326 Ma (Parrish ampMcNicoll 1992) Two unpublished 206Pb238U baddeleyiteages also from a gabbro on southern Vancouver Islandare 226805 Ma (five multi-grain fractions) and228425 Ma (two multi-grain fractions Sluggett 2003)
VOLCANIC STRAT IGRAPHY ANDPETROGRAPHYFieldwork undertaken onVancouver Island as part of thisstudy in 2004^2006 explored the volcanic stratigraphy ofthe Karmutsen flood basalts in three main areas theKarmutsen Range (between Alice and Nimpkish lakes)the area around Schoen Lake Provincial Park and atMount Arrowsmith (Greene et al 2005 2006) areasaround Holberg Inlet in northernmost Vancouver Islandand Buttle Lake in central Vancouver Island were alsoinvestigated (Fig 1) The character and thickness of theflood basalt sequences vary locally although the tripartitesuccession of the Karmutsen Formation appears to be pres-ent throughout Vancouver Island The stratigraphic thick-nesses for the pillow volcaniclastic and subaerial flowunits in the type area around Buttle Lake are estimated tobe 2600m 1100m and 2900m respectively (Surdam1967) estimated thicknesses on northernVancouver Islandare 43000m 400^1500m and 41500m respectively(Nixon et al 2008) in the vicinity of Mount Arrowsmithand nearby areas on southern Vancouver Island thick-nesses are 1100m 950m and 1200m respectively (Yorathet al 1999 Fig 1)Picritic pillow lavas west of the Karmutsen Range on
northernVancouver Island occur in a roughly triangular-shaped area (30 km across) bounded by KeoghMaynard and Sara lakes (Figs 2 and 3 Greene et al2006 Nixon et al 2008) Excellent exposures of picriticpillow lavas occur in roadcuts along the north shore ofKeogh Lake the type locality (Greene et al 2006) TheKeogh Lake picrites form pillowed and massive flow units(515m thick Fig 3) with pillows and tubes of varieddimensions (typically51m wide) Numerous thermal con-traction features in the pillows are filled with quartz^carbonate such as drain-back ledges and tortoise-shelljointing (Greene et al 2006) The picritic pillow basalts
are not readily distinguishable in the field from basaltexcept by their density non-magnetic character and com-paratively minor amounts of interpillow quartz^carbonateFieldwork and mapping indicates that the picrites occurnear the transition between the uppermost pillow lavasand the lowermost part of the volcaniclastic unit (Nixonet al 2008)In and around Schoen Lake Provincial Park where
the base of the Karmutsen Formation is exposed is asediment^sill complex consisting of Middle Triassic andPennsylvanian to Permian limestones and fine-grained sili-ciclastic sedimentary rocks intruded by mafic sills andoverlain by pillow basalt (Figs 2 and 4 Carlisle 1972) Thesediment^sill complex is 600^900m thick whichincludes 150^200m of pre-intrusive sedimentary rocks(Carlisle 1972) TheTriassic sedimentary rocks range fromthinly bedded siliceous and calcareous shales to bandedcherts and finely laminated Daonella-bearing shales(Carlisle 1972) Basal sills and pre-Karmutsen sedimentsare also exposed around Buttle Lake (Fig 1)Exposures with considerable vertical relief around
Mount Arrowsmith (see Electronic Appendix 1 availablefor downloading at httppetrologyoxfordjournalsorg)and Buttle Lake preserve thick successions of submarinepillow basalts and pillow breccias with rare intercalationsof sediment and subaerial flows Massive submarine flowsin the pillow unit are locally recognizable by their irregu-lar hackly columnar jointing The massive subaerial flowsform monotonous sequences marked mainly by amygdaloi-dal horizons brecciated flow tops are rarely observedA locality with well-preserved pahoehoe flow features iswell exposed north of Holberg Inlet (Fig 3 Nixon amp Orr2007 Nixon et al 2008) There is no evidence of significantdetrital material from a continental source in the sedi-ments associated with the flood basalts anywhere onVancouver Island The uppermost Karmutsen flows areinterbedded with thin (commonly 0^4m thick) lenses oflimestone and rarely siliciclastic sedimentary rocks(Nixon et al 2006b) Plagioclase-phyric lavas characterizethe upper part of the subaerial sequence and locallyinclude distinctive plagioclase-megacrystic (1^2 cm crys-tals) trachytoid-textured flows (Nixon et al 2006b)A total of 129 samples were collected from the
Karmutsen Formation on Vancouver Island from which63 samples were selected for geochemical analysis basedon the relative degree of alteration and geographical distri-bution of the samples Fifty-six of these samples weredivided into four main groups based on petrography(Table 1) and geochemistry tholeiitic basalt picrite high-MgO basalt and coarse-grained mafic rocks In additiontwo outlying samples and a mineralized sill are distin-guished in Table 1 The tholeiitic basalts are dominantlyglomeroporphyritic and less commonly aphanitic withan intersertal to intergranular groundmass (Table 1)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
4
Fig 2 Geological map and stratigraphy of the Schoen Lake Provincial Park and Karmutsen Range areas (locations shown in Fig 1 northernVancouver Island) (a) Stratigraphic column depicts units exposed in the Schoen Lake area derived from Carlisle (1972) and fieldwork (b)Generalized geology for the Schoen Lake area with sample locations Map derived from Massey et al (2005a) The exposures in the SchoenLake area are the lower volcanic stratigraphy and base of the Karmutsen Formation (c) Stratigraphic column for geology in the Alice^Nimpkish Lake area derived from Nixon amp Orr (2007) (d) Geological map from mapping of Nixon et al (2006a 2008) Sample sites andlithologies are denoted in the legend The Keogh Lake picrites are exposed near Keogh Sara and Maynard lakes and areas to the east ofMaynard Lake (Nixon et al 2008)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
5
Fig 3 Photographs of picritic and tholeiitic pillowed and massive basalt flows from the Karmutsen Range (Alice and Nimpkish Lake area)and Holberg Inlet (HI) areas northernVancouver Island (locations shown in Fig 1) (a) Massive submarine flow (marked with arrows) drapedover high-MgO pillow basalt of similar composition (b) Picritic pillow basalts in cross-section near Maynard Lake (Note vesicular uppermargin to pillows) (c) Close-up photograph of undulating contact of massive submarine flow draped over high-MgO pillow basalt (markedwith arrows) from photograph (a) (area of photograph indicated with a white box) (d) Photograph of margin of pahoehoe lobe (sledgehammerfor scale) (e) Cross-section of a large picritic pillow lobe with infilling of quartz^carbonate in cooling^contraction cracks (sledgehammer forscale) (f) High-MgO pillow breccia stratigraphically between submarine and subaerial flows (arrow points to lens cap 7 cm diameter forscale)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
6
Plagioclase forms most of the phenocrysts and glomero-crysts and the groundmass is fine-grained comprising pla-gioclase microlites clinopyroxene granules small grains ofFe^Ti oxide devitrified glass and secondary minerals
there is no fresh glass preserved The picritic and high-MgO basaltic lavas are characterized by spherulitic plagi-oclase and clinopyroxene with abundant euhedral olivinephenocrysts (Fig 5 Table 1) Olivine is completely replaced
Fig 4 Photographs showing field relations from the Schoen Lake Provincial Park area northernVancouver Island (a) Sediment^sill complexat the base of the Karmutsen Formation on the north side of Mt Schoen 100m below the Daonella locality (see Fig 2 for location) (b)Interbedded mafic sills and deformed finely banded chert and shale with calcareous horizons from location between Mt Adam and MtSchoen
Fig 5 Photomicrographs of picritic pillow basalts Karmutsen Formation northernVancouver Island Scale bars represent 1mm (a) Picritewith abundant euhedral olivine pseudomorphs from Keogh Lake type locality (Fig 2) in cross-polarized light (XPL) (sample 4722A4 19^198wt MgO four analyses) Samples contain dense clusters (24^42 vol ) of olivine pseudomorphs (52mm) in a groundmass of curvedand branching sheaves of acicular plagioclase and intergrown with clinopyroxene and altered glass In many cases plagioclase nucleated on theedges of the olivine phenocrysts (b) Sheaves of intergrown plagioclase and clinopyroxene in aphyric picrite pillow lava from west of MaynardLake (Fig 2) in XPL (sample 4723A2108 wt MgO)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
7
Table 1 Summary of petrographic characteristics and phenocryst proportions of Karmutsen basalts on Vancouver Island
BC
Sample Area Flow Group Texture vol Ol Plag Cpx Ox Alteration Notes
4718A1 MA PIL THOL intersertal 20 5 3 few plag glcr52mm cpx51 mm
4718A2 MA PIL THOL intersertal glomero 15 1 plag glcr52mm very fresh
4718A5 MA FLO THOL intersertal glomero 10 2 mottled few plag glcr54 mm
4723A4 KR PIL PIC intergranular intersertal 0 3 swtl plag51mm no ol phenos
4723A13 KR PIL PIC spherulitic 24 3 swtl plag51 mm
(continued)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
8
by an assemblage of talc^tremolite^serpentine^opaqueoxides there is no fresh olivine present in the Karmutsenlavas onVancouver Island Plagioclase and clinopyroxenephenocrysts are absent in the high-MgO lavas Thecoarse-grained mafic rocks are characterized by sub-ophi-tic textures with an average grain size typically 41mmand are generally not glomeroporphyritic these rocks aremainly from the interiors of massive flows although somemay represent sillsSample preparation and analytical methods for major-
and trace-element and Sr^Nd^Hf^Pb isotopic composi-tions of the Karmutsen volcanic rocks are given in theAppendix
WHOLE -ROCK CHEMISTRYMajor- and trace-element compositionsThe most abundant type of volcanic rock in theKarmutsen Formation is tholeiitic basalt with a restrictedrange of major- and trace-element compositions The tho-leiitic basalts have similar compositions to the coarse-grained mafic rocks and both groups are distinct fromthe picrites and high-MgO basalts [Fig 6 picrites412wt MgO (LeBas 2000) high-MgO basalts8^12wt MgO] The tholeiitic basalts have lower MgO(57^77wt MgO) and higher TiO2 (14^22wt
TiO2) than the picrites (130^198wt MgO05^07wt TiO2) and high-MgO basalts (91^116wt MgO 05^08wt TiO2 Fig 6 Table 2) Almost allcompositions plot within the tholeiitic field in a total alka-lis vs silica plot although there has been substantial K lossin most samples from the Karmutsen Formation (seebelow) and the tholeiitic basalts generally have higherSiO2 Na2OthornK2O CaO and FeOT (total iron expressedas FeO) than the picrites The tholeiitic basalts also havenoticeably lower loss on ignition (LOI mean1713wt ) than the picrites (mean 54 08wt )and high-MgO basalts (mean 3815wt ) (Fig 6)reflecting the presence of abundant altered olivinephenocrysts in the latter groups Ni concentrations are sig-nificantly higher for the picrites (339^755 ppm) and high-MgO basalts (122^551ppm) than the tholeiitic basalts(58^125 ppm except for one anomalous sample) andcoarse-grained rocks (70^213 ppm) (Fig 6 Table 2) Ananomalous tholeiitic pillowed flow (two samples outlierin Tables 1 and 2) from near the picrite type locality atKeogh Lake and a mineralized sill (disseminated sulfide)from the basal sediment^sill complex at Schoen Lake aredistinguished by higher TiO2 and FeOT than the maingroup of tholeiitic basalts The tholeiitic basalts are similarin major-element composition to previously publishedresults for samples from the Karmutsen Formation how-ever the Keogh Lake picrites which encompass the
Table 1 Continued
Sample Area Flow Group Texture vol Ol Plag Cpx Ox Alteration Notes
4722A5 KR FLO OUTLIER intersertal 3 mottled very fg v small pl needles
5615A11 KR PIL OUTLIER intersertal 3 mottled very fg v small pl needles
5617A4 SL SIL MIN intersertal 5 3 fg
Sample number last digit of year month day initial sample station (except 93G171) MA Mount Arrowsmith SLSchoen Lake KR Karmutsen Range PIL pillow BRE breccia FLO flow SIL sill GAB gabbro THOL tholeiiticbasalt PIC picrite HI-MG high-MgO basalt CGcoarse-grained MIN mineralized sill glomero glomeroporphyriticOlivine modes for picrites are based on area calculations from scans (see lsquoOlivine accumulation in high-MgO and picriticlavasrsquo section and Fig 11) Visual alteration index is based primarily on degree of plagioclase alteration and presence ofsecondary minerals (1 least altered 3 most altered) Plagioclase phenocrysts are commonly altered to albite pumpellyiteand chlorite olivine is altered to talc tremolite and clinochlore (determined using the Rietveld method of X-ray powderdiffraction) clinopyroxene is unaltered FendashTi oxide is commonly replaced by sphene glcr glomerocrysts fg fine-grained cg coarse-grained oik oikocryst chad chadacryst enc enclosing swtl swallow-tail ol olivinepseudomorphs plag plagioclase cpx clinopyroxene ox oxides (includes ilmenite thorn titanomagnetite)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
9
00
05
10
15
20
25
30
35
40
0 5 10 15 20 25
10
11
12
13
14
15
16
17
18
19
20
0 5 10 15 20 25
6
7
8
9
10
11
12
13
14
15
0 5 10 15 20 25
6
7
8
9
10
11
12
13
14
15
16
0 5 10 15 20 25
0
1
2
3
4
5
6
40 45 50 55 60
0
200
400
600
800
0 5 10 15 20 25
(wt)(wt)
(wt)
(wt)
+K2ONa2O TiO2
FeO (T)
Ni (ppm)
Al2O3
CaO
alkalic
tholeiitic
Karmutsen Form
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
(compiled data)
Pillowed flowMineralized sill
SiO2 (wt) MgO (wt)
MgO (wt)MgO (wt)
MgO (wt) MgO (wt)
(a) (b)
(g)
(e)
(f)
(c)
LOI (wt )
(d)
MgO (wt )0
1
2
3
4
5
6
7
0 5 10 15 20 25
(wt)
Fig 6 Whole-rock major-element Ni and LOI variation diagrams for the Karmutsen Formation New samples from this study (Table 2) areshown by symbols with black outlines (see legend) Previously published analyses are shown by small gray symbols without black outlines Theboundary of the alkaline and tholeiitic fields is that of MacDonald amp Katsura (1964) Total iron expressed as FeO LOI loss on ignition oxidesare plotted on an anhydrous normalized basis References for the 322 compiled analyses for the Karmutsen Formation are Surdam (1967)Kuniyoshi (1972) Muller et al (1974) Barker et al (1989) Lassiter et al (1995) Massey (1995a1995b)Yorath et al (1999) and G Nixon (unpublisheddata) It should be noted that the compiled dataset has not been filtered many of the samples with high SiO2 (452wt ) are probably notKarmutsen flood basalts but younger Bonanza basalts and andesites Three pillowed flow samples and a mineralized sill are distinguishedseparately because of their anomalous chemistry Coarse-grained samples are indicated separately Three tholeiitic basalt samples (4724A54718A5 4718A6) with416wt Al2O3 and4270 ppm Sr contain 10^25 modal of large plagioclase phenocrysts (7^8mm) or glomerocrysts(44mm)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
10
Table 2 Major element (wt oxide) and trace element (ppm) abundances in whole-rock samples of Karmutsen basalts
1Ni concentrations for these samples were determined by XRFTHOL tholeiitic basalt PIC picrite HI-MG high MgO basalt CG coarse-grained (sill or gabbro) MIN SIL mineralizedsill OUTLIER anomalous pillowed flow in plots MA Mount Arrowsmith SL Schoen Lake KR Karmutsen Range QIQuadra Island Sample locations are given using the Universal Transverse Mercator (UTM) coordinate system (NAD83zones 9 and 10) Analyses were performed at Activation Laboratory (ActLabs) Fe2O3
is total iron expressed as Fe2O3LOI loss on ignition All major elements Sr V and Y were measured by ICP quadrupole OES on solutions of fusedsamples Cu Ni Pb and Zn were measured by total dilution ICP Cs Ga Ge Hf Nb Rb Ta Th U Zr and REE weremeasured by magnetic-sector ICP on solutions of fused samples Co Cr and Sc were measured by INAA Blanks arebelow detection limit See Electronic Appendices 2 and 3 for complete XRF data and PCIGR trace element datarespectively Major elements for sample 93G17 are normalized anhydrous Sample 5615A7 was analyzed by XRFSamples from Quadra Island are not shown in figures
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
17
picritic and high-MgO basalt pillow lavas extend to20wt MgO (Fig 6 and references listed in thecaption)The tholeiitic basalts from the Karmutsen Range dis-
play a tight range of parallel light rare earth element(LREE)-enriched REE patterns (LaYbCNfrac14 21^24mean 2202) whereas the picrites and high-MgObasalts are characterized by sub-parallel LREE-depleted
patterns (LaYbCNfrac14 03^07 mean 06 02) with lowerREE abundances than the tholeiitic basalts (Fig 7)Tholeiitic basalts from the Schoen Lake and MountArrowsmith areas have similar REE patterns tosamples from the Karmutsen Range (Schoen Lake LaYbCNfrac14 21^26 mean 2303 Mount Arrowsmith LaYbCNfrac1419^25 mean 2204) and the coarse-grainedmafic rocks have similar REE patterns to the tholeiitic
Depleted MORB average(Salters amp Stracke 2004)
Schoen LakeSchoen Lake
Mount Arrowsmith
Karmutsen RangeS
ampl
eC
hond
rite
Sam
ple
Cho
ndrit
e
Sam
ple
Prim
itive
Man
tleS
ampl
eP
rimiti
ve M
antle
Sam
ple
Prim
itive
Man
tle
Mount Arrowsmith
Karmutsen Range
(a) (b)
(c) (d)
(e) (f )
Sam
ple
Cho
ndrit
e
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained mafic rock
Pillowed flowMineralized sill
1
10
100
1
10
100
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
01
1
10
100
1
10
100
1
10
100
Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
Fig 7 Whole-rock REE and other incompatible-element concentrations for the Karmutsen Formation (a) (c) and (e) are chondrite-normal-ized REE patterns for samples from each of the three main field areas onVancouver Island (b) (d) and (f) are primitive mantle-normalizedtrace-element patterns for samples from each of the three main field areas All normalization values are from McDonough amp Sun (1995) Theclear distinction between LREE-enriched tholeiitic basalts and LREE-depleted picrites the effects of alteration on the LILE (especially loss ofK and Rb) and the different range for the vertical scale in (b) (extends down to 01) should be noted
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
18
basalts (LaYbCNfrac14 21^29 mean 2508) Two LREE-depleted coarse-grained mafic rocks from the SchoenLake area have similar REE patterns (LaYbCNfrac14 07) tothe picrites from the Keogh Lake area indicating that thisdistinctive suite of rocks is exposed as far south asSchoen Lake that is 75 km away (Fig 7) The anomalouspillowed flow from the Karmutsen Range has lower LaYbCN (13^18) than the main group of tholeiitic basaltsfrom the Karmutsen Range (Fig 7) The mineralized sillfrom the Schoen Lake area is distinctly LREE-enriched(LaYbCNfrac14 33) with the highest REE abundances(LaCNfrac14 940) of all Karmutsen samples this may reflectcontamination by adjacent sediment also evident in thin-section where sulfide blebs are cored by quartz (Greeneet al 2006)The primitive mantle-normalized trace-element pat-
terns (Fig 7) and trace-element variations and ratios(Fig 8) highlight the differences in trace-element
concentrations between the four main groups ofKarmutsen flood basalts onVancouver Island The tholeii-tic basalts from all three areas have relatively smooth par-allel trace-element patterns some samples have smallpositive Sr anomalies relative to Nd and Sm and manysamples have prominent negative K anomalies relative toU and Nb reflecting K loss during alteration (Fig 7) Thelarge ion lithophile elements (LILE) in the tholeiiticbasalts (mainly Rb and K not Cs) are depleted relativeto the high field strength elements (HFSE) and LREELILE segments especially for the Schoen Lake area(Fig 7d) are remarkably parallel The picrites andhigh-MgO basalts form a tight band of parallel trace-element patterns and are depleted in HFSE and LREEwith positive Sr anomalies and relatively low Rb Thecoarse-grained mafic rocks from the Karmutsen Rangehave identical trace-element patterns to the tholeiiticbasalts Samples from the Schoen Lake area have similar
000
025
050
075
100
125
0 1 2 3 4 5
00
04
08
12
16
0 5 10 15 20 25
MgO (wt)
0
5
10
15
0 50 100 150
Hf (ppm)
Th (ppm)
NbLaNb (ppm)
Zr (ppm)
(a)
(c)
(b)
(d)
Th (ppm)
00
02
04
06
08
10
12
00 01 02 03 04 05
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
NbLa=1
4723A4 5615A12
U (ppm)
Fig 8 Whole-rock trace-element concentrations and ratios for the Karmutsen Formation [except (b) which is vs wt MgO] (a) Nb vs Zr(b) NbLa vs MgO (c) Th vs Hf (d) Th vs U There is a clear distinction between the tholeiitic basalts and picrites in Nb and Zr both inconcentration and the slope of each trend
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
19
trace-element patterns to the Karmutsen Range except forthe two LREE-depleted coarse-grained mafic rocks andthe mineralized sill (Fig 7d)
Sr^Nd^Hf^Pb isotopic compositionsA total of 19 samples from the four main groups ofthe Karmutsen Formation were selected for radiogenicisotopic geochemistry on the basis of major- and trace-element variation stratigraphic position and geographicallocation The samples have overlapping age-correctedHf and Nd isotope ratios and distinct ranges of Sr isotopiccompositions (Fig 9) The tholeiitic basalts picriteshigh-MgO basalt and coarse-grained mafic rocks have
initial eHffrac14thorn87 to thorn126 and eNdfrac14thorn66 to thorn88 cor-rected for in situ radioactive decay since 230 Ma (Fig 9Tables 3 and 4) All the Karmutsen samples form a well-defined linear array in a Lu^Hf isochron diagram corre-sponding to an age of 24115 Ma (Fig 9) This is withinerror of the accepted age of the Karmutsen Formationand indicates that the Hf isotope systematics behaved as aclosed system since c 230 Ma The anomalous pillowedflow has similar initial eHf (thorn98) and slightly lower initialeNd (thorn62) compared with the other Karmutsen samplesThe picrites and high-MgO basalt have higher initial87Sr86Sr (070398^070518) than the tholeiitic basalts(initial 87Sr86Srfrac14 070306^070381) and the coarse-grained mafic rocks (initial 87Sr86Srfrac14 070265^070428)
05128
05129
05130
05131
05132
015 017 019 021 023 025
02829
02830
02831
02832
02833
02834
000 002 004 006 008 010
4
6
8
10
0702 0703 0704 0705 07066
8
10
12
14
5 6 7 8 9 10
12
0
1
2
3
4
5
6
7
0702 0703 0704 0705 0706
(d)
LOI (wt )
87Sr 86Sr
4723A2
Nd
143Nd144Nd
147Sm144Nd
87Sr 86Sr (230 Ma)
(230 Ma)
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
(a)
(c)
176Hf177Hf
Age=241 plusmn 15 Ma=+990 plusmn 064
176Lu177Hf
Hf (230 Ma)
Nd (230 Ma)
(e)
Hf (i)
(b)
Fig 9 Whole-rock Sr Nd and Hf isotopic compositions for the Karmutsen Formation (a) 143Nd144Nd vs 147Sm144Nd (b) 176Hf177Hf vs176Lu177Hf The slope of the best-fit line for all samples corresponds to an age of 24115 Ma (c) Initial eNd vs 87Sr86Sr Age-correction to230 Ma (d) LOI vs 87Sr86Sr (e) Initial eHf vs eNd Average 2 error bars are shown in a corner of each panel Complete chemical duplicatesshown inTables 3 and 4 (samples 4720A7 and 4722A4) are circled in each plot
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
overlap the ranges of the picrites and tholeiitic basalts(Fig 9 Table 3)The measured Pb isotopic compositions of the tholeiitic
basalts are more radiogenic than those of the picrites andthe most magnesian Keogh Lake picrites have the leastradiogenic Pb isotopic compositions The range ofinitial Pb isotope ratios for the picrites is206Pb204Pbfrac1417868^18580 207Pb204Pbfrac1415547^15562and 208Pb204Pbfrac14 37873^38257 and the range for thetholeiitic basalts is 206Pb204Pbfrac1418782^19098207Pb204Pbfrac1415570^15584 and 208Pb204Pbfrac14 37312^38587 (Fig 10 Table 5) The coarse-grained mafic rocksoverlap the range of initial Pb isotopic compositions forthe tholeiitic basalts with 206Pb204Pbfrac1418652^19155207Pb204Pbfrac1415568^15588 and 208Pb204Pbfrac14 38153^38735 One high-MgO basalt has the highest initial206Pb204Pb (19242) and 207Pb204Pb (15590) and thelowest initial 208Pb204Pb (37676) The Pb isotopic ratiosfor Karmutsen samples define broadly linearrelationships in Pb isotope plots The age-corrected Pb
isotopic compositions of several picrites have been affectedby U and Pb (andor Th) mobility during alteration(Fig 10)
ALTERAT IONThe Karmutsen basalts have retained most of their origi-nal igneous structures and textures however secondaryalteration and low-grade metamorphism have producedzeolitic- and prehnite^pumpellyite-bearing mineral assem-blages (Table 1 Cho et al 1987) Alteration and metamor-phism have primarily affected the distribution of the LILEand the Sr isotopic systematics The Keogh Lake picriteshave been affected more by alteration than the tholeiiticbasalts and have higher LOI (up to 55wt ) variableLILE and higher measured Sr Nd and Hf and lowermeasured Pb isotope ratios than the tholeiitic basalts Srand Pb isotope compositions for picrites and tholeiiticbasalts show a relationship with LOI (Figs 9 and 10)whereas Nd and Hf isotopic compositions do not correlate
Table 3 Sr and Nd isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Rb Sr 87Sr86Sr 2m87Rb86Sr 87Sr86Srt Sm Nd 143Nd144Nd 2m eNd
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses carried out at the PCIGR thecomplete trace element analyses are given in Electronic Appendix 3
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
21
with LOI (not shown) The relatively high apparent initialSr isotopic compositions for the high-MgO lavas (up to07052) probably resulted from post-magmatic loss of Rbor an increase in 87Sr86Sr through addition of seawater Sr(eg Hauff et al 2003) High-MgO lavas from theCaribbean plateau have similarly high 87Sr86Sr comparedwith basalts (Recurren villon et al 2002)The correction for in situdecay on initial Pb isotopic ratios has been affected bymobilization of U and Th (and Pb) in the whole-rockssince their formation A thorough acid leaching duringsample preparation was used that has been shown to effec-tively remove alteration phases (Weis et al 2006 NobreSilva et al in preparation) Whole-rock SmNd ratiosand age-correction of Nd isotopes may be affected bythe leaching procedure for submarine rocks older than50 Ma (Thompson et al 2008 Fig 9) there ishowever very little variation in Nd isotopic compositionsof the picrites or tholeiitic basalts and this system does notappear to be disturbed by alteration The HFSE abun-dances for tholeiitic and picritic basalts exhibit clear
linear correlations in binary diagrams (Fig 8) whereasplots of LILE vs HFSE are highly scattered (not shown)because of the mobility of some of the alkali and alkalineearth LILE (Cs Rb Ba K) and Sr for most samplesduring alterationThe degree of alteration in the Karmutsen samples does
not appear to be related to eruption environment or depthof burial Pillow rims are compositionally different frompillow cores (Surdam1967 Kuniyoshi1972) and aquagenetuffs are chemically different from isolated pillows withinpillow breccias however submarine and subaerial basaltsexhibit a similar degree of alterationThere is no clear cor-relation between the submarine and subaerial basalts andsome commonly used chemical alteration indices (eg BaRb vs K2OP2O5 Huang amp Frey 2005) There is also nodefinitive correlation between the petrographic alterationindex (Table 1) and chemical alteration indices although10 of 13 tholeiitic basalts with the highest K and LILEabundances have the highest petrographic alterationindex of three (seeTable 1)
Table 4 Hf isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Lu Hf 176Hf177Hf 2m eHf176Lu177Hf 176Hf177Hft eHf(t)
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
OLIV INE ACCUMULAT ION INH IGH-MgO AND PICR IT IC LAVASGeochemical trends and petrographic characteristics indi-cate that accumulation of olivine played an important rolein the formation of the high-MgO basalts and picrites fromthe Keogh Lake area on northern Vancouver Island TheKeogh Lake samples show a strong linear correlation inplots of Al2O3 TiO2 ScYb and Ni vs MgO and many ofthe picrites have abundant clusters of olivine pseudomorphs(Table1 Fig11)There is a strong linear correlationbetween
modal per cent olivine and whole-rock magnesium contentmagnesium contents range between 10 and 19wt MgOand the proportion of olivine phenocrysts in these samplesvaries from 0 to 42 vol (Fig 11)With a few exceptionsboth the size range and median size of the olivine pheno-crysts in most samples are comparable The clear correla-tion between the proportion of olivine phenocrysts andwhole-rock magnesium contents indicates that accumula-tion of olivine was directly responsible for the high MgOcontents (410wt ) in most of the picritic lavas
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
207Pb204Pb
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
208Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
(a) (b)
(c) (e)
4723A24723A2
4723A2
4723A2
4722A4
4723A4
4722A4
4723A4
4723A134723A3
4723A3
4723A13
1556
1558
1560
1562
1564
1566
1568
186 190 194 198 202 206 2101550
1552
1554
1556
1558
1560
178 180 182 184 186 188 190 192 194
380
384
388
392
396
186 190 194 198 202 206 210374
378
382
386
390
178 180 182 184 186 188 190 192 194
206Pb204Pb(d)
LOI (wt )
0
1
2
3
4
5
6
7
186 190 194 198 202 206 210
4723A2
Fig 10 Pb isotopic compositions of leached whole-rock samples measured by MC-ICP-MS for the Karmutsen Formation Error bars are smal-ler than symbols (a) Measured 207Pb204Pb vs 206Pb204Pb (b) Initial 207Pb204Pb vs 206Pb204Pb Age-correction to 230 Ma (c) Measured208Pb204Pb vs 206Pb204Pb (d) LOI vs 206Pb204Pb (e) Initial 208Pb204Pb vs 206Pb204Pb Complete chemical duplicates shown in Table 5(samples 4720A7 and 4722A4) are circled in each plot The dashed lines in (b) and (e) show the differences in age-corrections for two picrites(4723A4 4722A4) using the measured UTh and Pb concentrations for each sample and the age-corrections when concentrations are used fromthe two picrites (4723A3 4723A13) that appear to be least affected by alteration
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
23
DISCUSS IONThe compositional and spatial development of the eruptedlava sequences of the Wrangellia oceanic plateau onVancouver Island provide information about the evolutionof a mantle plume upwelling beneath oceanic lithospherein an analogous way to the interpretation of continentalflood basalts The Caribbean and Ontong Java plateausare the two primary examples of accreted parts of oceanicplateaus that have been studied to date and comparisonscan be made with the Wrangellia oceanic plateauHowever the Wrangellia oceanic plateau is a distinctiveoccurrence of an oceanic plateau because it formed on thecrust of a Devonian to Mississippian intra-oceanic islandarc overlain by Mississippian to Permian limestones andpelagic sediments The nature and thickness of oceaniclithospheric mantle are considered to play a fundamentalrole in the decompression and evolution of a mantleplume and thus the compositions of the erupted lavasequences (Greene et al 2008) The geochemical and
stratigraphic relationships observed in the KarmutsenFormation onVancouver Island and the focus of the subse-quent discussion sections provide constraints on (1) thetemperature and degree of melting required to generatethe primary picritic magmas (2) the composition of themantle source (3) the depth of melting and residual min-eralogy in the source region (4) the low-pressure evolutionof the Karmutsen tholeiitic basalts (5) the growth of theWrangellia oceanic plateau
Melting conditions and major-elementcomposition of the primary magmasThe primary magmas to most flood basalt provinces arebelieved to be picritic (eg Cox 1980) and they have beenfound in many continental and oceanic flood basaltsequences worldwide (eg Siberia Karoo Paranacurren ^Etendeka Caribbean Deccan Saunders 2005) Near-pri-mary picritic lavas with low total alkali abundances
Table 5 Pb isotopic compositions of Karmutsen basaltsVancouver Island BC
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
require high-degree partial melting of the mantle source(eg Herzberg amp OrsquoHara 2002) The Keogh Lake picritesfrom northernVancouver Island are the best candidates foridentifying the least modified partial melts of the mantleplume source despite having accumulated olivine pheno-crysts and can be used to estimate conditions of meltingand the composition of the primary magmas for theKarmutsen FormationPrimary magma compositions mantle potential tem-
peratures and source melt fractions for the picritic lavas ofthe Karmutsen Formation were calculated from primitivewhole-rock compositions using PRIMELT1XLS software(Herzberg et al 2007) and are presented in Fig 12 andTable 6 A detailed discussion of the computationalmethod has been given by Herzberg amp OrsquoHara (2002)and Herzberg et al (2007) key features of the approachare outlined belowPrimary magma composition is calculated for a primi-
tive lava by incremental addition of olivine and compari-son with primary melts of mantle peridotite Lavas thathad experienced plagioclase andor clinopyroxene fractio-nation are excluded from this analysis All calculated pri-mary magma compositions are assumed to be derived byaccumulated fractional melting of fertile peridotite KR-4003 Uncertainties in fertile peridotite composition donot propagate to significant variations in melt fractionand mantle potential temperature melting of depletedperidotite propagates to calculated melt fractions that aretoo high but with a negligible error in mantle potential
temperature PRIMELT1XLS calculates the olivine liqui-dus temperature at 1atm which should be similar to theactual eruption temperature of the primary magmaand is accurate to 308C at the 2 level of confidenceMantle potential temperature is model dependent witha precision that is similar to the accuracy of the olivineliquidus temperature (C Herzberg personal communica-tion 2008)With regard to the evidence for accumulated olivine in
the Keogh Lake picrites it should not matter whether themodeled lava composition is aphyric or laden with accu-mulated olivine phenocrysts PRIMELT1 computes theprimary magma composition by adding olivine to a lavathat has experienced olivine fractionation in this way itinverts the effects of fractional crystallization The onlyrequirement is that the olivine phenocrysts are not acci-dental or exotic to the lava compositions Support for thisprocedure can be seen in the calculated olivine liquid-line-of-descent shown by the curved arrays with 5 olivineaddition increments (Fig 12c) Fractional crystallization ofolivine from model primary magmas having 85^95wt FeO and 153^175wt MgO will yield arrays of deriva-tive liquids and randomly sorted olivines that in most butnot all cases connect the picrites and high-MgO basalts Anewer version of the PRIMELT software (Herzberg ampAsimow 2008) can perform olivine addition to and sub-traction from lavas with highly variably olivine phenocrystcontents and yields results that converge with those shownin Fig 12
Fig 11 Relationship between abundance of olivine phenocrysts and whole-rock MgO contents for the Keogh Lake picrites (a) Tracings ofolivine grains (gray) in six high-MgO samples made from high resolution (2400 dpi) scans of petrographic thin-sections Scale bars represent10mm sample numbers are indicated at the upper left (b) Modal olivine vs whole-rock MgO Continuous line is the best-fit line for 10 high-MgO samplesThe areal proportion of olivine was calculated from raster images of olivine grains using ImageJ image analysis software whichprovides an acceptable estimation of the modal abundance (eg Chayes 1954)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
25
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
Gray lines = final melting pressure
L + Ol + Opx
07
08
09
Pressure (GPa)
10
3
4
5
6
1
SOLIDUS
01
04
0506
L + Ol
2
03
02
PeridotiteKR-4003
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
Thick black lines = initial melting pressure
Dashed lines = melt fraction
Melt Fraction
2
3 4 5 6
L+Ol+Opx
SolidusSpinel
SolidusGarnet Peridotite
7
L+Ol
0 10 20 30 4040
45
50
55
60
MgO (wt)
SiO2
(wt)
PeridotiteKR-4003
Melt Fraction
0 10 20 305
10
15
CaO(wt)
MgO (wt)
3
4 5 6 7
2
46
2
Peridotite Partial Melts
Thick black line = initial melting pressureGray lines = final melting pressure
Peridotite
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
02
Pressure (GPa)
0302
04
Pressure (GPa)
Picrite
High-MgO basalt
Tholeiitic basalt
(c)
(d)
Karmutsen flood basaltsOlivine addition modelOlivine addition (5 increment)Primary magma
PicriteHigh-MgO basaltTholeiitic basalt
(a)
(b)
800 850 900
910
920
930
940
950
Karmutsen flood basalts
Olivine addition modelOlivine addition (5 increment)Primary magma
Gray lines = Fo contents of olivine
Fig 12 Estimated primary magma compositions for three Keogh Lake picrites and high-MgO basalts (samples 4723A4 4723A135616A7) using the forward and inverse modeling technique of Herzberg et al (2007) Karmutsen compositions and modeling results areoverlain on diagrams provided by C Herzberg (a) Whole-rock FeO vs MgO for Karmutsen samples from this study Total iron estimatedto be FeO is 090 Gray lines show olivine compositions that would precipitate from liquid of a given MgO^FeO composition Black lineswith crosses show results of olivine addition (inverse model) using PRIMELT1 (Herzberg et al 2007) Gray field highlights olivinecompositions with Fo contents of 91^92 predicted to precipitate (b) FeO vs MgO showing Karmutsen lava compositions and results ofa forward model for accumulated fractional melting of fertile peridotite KR-4003 (Kettle River Peridotite Walter 1998) (c) SiO2 vsMgO for Karmtusen lavas and model results (d) CaO vs MgO showing Karmtusen lava compositions and model results To brieflysummarize the technique [see Herzberg et al (2007) for complete description] potential parental magma compositions for the high-MgOlava series were selected (highest MgO and appropriate CaO) and using PRIMELT1 software olivine was incrementally added tothe selected compositions to show an array of potential primary magma compositions (inverse model) Then using PRIMELT1 theresults from the inverse model were compared with a range of accumulated fractional melts for fertile peridotite derived fromparameterization of the experimental results for KR-4003 from Walter (1998) forward model Herzberg amp OrsquoHara (2002) A melt fractionwas sought that was unique to both the inverse and forward models (Herzberg et al 2007) A unique solution was found when there was a com-mon melt fraction for both models in FeO^MgO and CaO^MgO^Al2O3^SiO2 (CMAS) projection space This modeling assumes thatolivine was the only phase crystallizing and ignores chromite precipitation and possible augite fractionation in the mantle (Herzbergamp OrsquoHara 2002) Results are best for a residue of spinel lherzolite (not pyroxenite) The presence of accumulated olivine in samples ofthe Keogh Lake picrites used as starting compositions does not significantly affect the results because we are modeling addition of olivine(see text for further description) The tholeiitic basalts cannot be used for modeling because they are all saturated in plagthorn cpxthornolThick black line for initial melting pressure at 4 GPa is not shown in (c) for clarity Gray area in (a) indicates parental magmas thatwill crystallize olivine with Fo 91^92 at the surface Dashed lines in (c) indicate melt fraction Solidi for spinel and garnet peridotite arelabelled in (c)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
The PRIMELT1 results indicate that if the Karmutsenpicritic lavas were derived from accumulated fractionalmelting of fertile peridotite the primary magmas wouldhave contained 15^17wt MgO and 10wt CaO(Fig 12 Table 6) formed from 23^27 partial meltingand would have crystallized olivine with a Fo content of 91 (Fig 12 Table 6) Calculated mantle temperatures forthe Karmutsen picrites (14908C) indicate melting ofanomalously hot mantle (100^2008C above ambient) com-pared with ambient mantle that produces mid-ocean ridgebasalt (MORB 1280^14008C Herzberg et al 2007)which initiated between 3 and 4 GPa Several picritesappear to be near-primary melts and required very littleaddition of olivine (eg sample 93G171 with measured158wt MgO) These temperature and melting esti-mates of the Karmutsen picrites are consistent withdecompression of hot mantle peridotite in an actively con-vecting plume head (ie plume initiation model) Threesamples (picrite 5614A1 and high-MgO basalts 5614A3and 5614A5) are lower in iron than the other Karmutsenlavas with FeO in the 65^80wt range (Fig 12c) andhave MgO and FeO contents that are similar to primitiveMORB from the Sequeiros fracture zone of the EastPacific Rise (EPR Perfit et al 1996) These samples
however differ from EPR MORB in having higher SiO2
and lower Al2O3 and they are restricted geographicallyto one area (Sara Lake Fig 2)We therefore have evidence for primary magmas with
MgO contents that range from a possible low of 110wt to as high as 174wt with inferred mantle potential tem-peratures from 1340 to 15208C This is similar to the rangedisplayed by lavas from the Ontong Java Plateau deter-mined by Herzberg (2004) who found that primarymagma compositions assuming a peridotite sourcewould contain 17wt MgO and 10wt CaO(Table 6) and that these magmas would have formed by27 melting at a mantle potential temperature of15008C (first melting occurring at 36 GPa) Fitton ampGodard (2004) estimated 30 partial melting for theOntong Java lavas based on the Zr contents of primarymagmas of Kroenke-type basalt calculated by incrementaladdition of equilibrium olivine However if a considerableamount of eclogite was involved in the formation of theOntong Java plateau magmas the excess temperatures esti-mated by Herzberg et al (2007) may not be required(Korenaga 2005) The variability in mantle potential tem-perature for the Keogh Lake picrites is also similar to thewide range of temperatures that have been calculated for
Table 6 Estimated primary magma compositions for Karmutsen basalts and other oceanic plateaus or islands
Sample 4723A4 4723A13 5616A7 93G171 Average OJP Mauna Keay Gorgonaz
Ontong Java primary magma composition for accumulated fraction melting (AFM) from Herzberg (2004)yMauna Kea primary magma composition is average of four samples of Herzberg (2006 table 1)zGorgona primary magma composition for AFM for 1F fertile source from Herzberg amp OrsquoHara (2002 table 4)Total iron estimated to be FeO is 090
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
27
lavas from single volcanoes in the Galapagos Hawaii andelsewhere by Herzberg amp Asimow (2008) Those workersinterpreted the range to reflect the transport of magmasfrom both the cool periphery and the hotter axis of aplume A thermally heterogeneous plume will yield pri-mary magmas with different MgO contents and the rangeof primary magma compositions and their inferred mantlepotential temperatures for Karmutsen lavas may reflectmelting from different parts of the plume
Source of the Karmutsen Formation lavasThe Sr^Nd^Hf^Pb isotopic compositions of theKarmutsen volcanic rocks on Vancouver Island indicatethat they formed from plume-type mantle containing adepleted component that is compositionally similar to butdistinct from other oceanic plateaus that formed in thePacific Ocean (eg Ontong Java and Caribbean Fig 13)Trace element and isotopic constraints can be used to testwhether the depleted component of the KarmutsenFormation was intrinsic to the mantle plume and distinctfrom the source of MORB or the result of mixing or invol-vement of ambient depleted MORB mantle with plume-type mantle The Karmutsen tholeiitic basalts have initialeNd and
87Sr86Sr ratios that fall within the range of basaltsfrom the Caribbean Plateau (eg Kerr et al 1996 1997Hauff et al 2000 Kerr 2003) and a range of Hawaiian iso-tope data (eg Frey et al 2005) and lie within and abovethe range for the Ontong Java Plateau (Tejada et al 2004Fig 13a) Karmutsen tholeiitic basalts with the highest ini-tial eNd and lowest initial 87Sr86Sr lie within and justbelow the field for age-corrected northern EPR MORB at230 Ma (eg Niu et al1999 Regelous et al1999) Initial Hfand Nd isotopic compositions for the Karmutsen samplesare noticeably offset to the low side of the ocean islandbasalt (OIB) array (Vervoort et al 1999) and lie belowand slightly overlap the low eHf end of fields for Hawaiiand the Caribbean Plateau (Fig 13c) the Karmutsen sam-ples mostly fall between and slightly overlap a field forage-corrected Pacific MORB at 230 Ma and Ontong Java(Mahoney et al 1992 1994 Nowell et al 1998 Chauvel ampBlichert-Toft 2001) Karmutsen lava Pb isotope ratios over-lap the broad range for the Caribbean Plateau and thelinear trends for the Karmutsen samples intersect thefields for EPR MORB Hawaii and Ontong Java in208Pb^206Pb space but do not intersect these fields in207Pb^206Pb space (Fig 13d and e) The more radiogenic207Pb204Pb for a given 206Pb204Pb of the Karmutsenbasalts requires high UPb ratios at an ancient time when235U was abundant similar to the Caribbean Plateau andenriched in comparison with EPR MORB Hawaii andOntong JavaThe low eHf^low eNd component of the Karmutsen
basalts that places them below the OIB mantle array andpartly within the field for age-corrected Pacific MORBcan form from low ratios of LuHf and high SmNd over
time (eg Salters amp White 1998) MORB will develop aHf^Nd isotopic signature over time that falls below themantle array Low LuHf can also lead to low 176Hf177Hffrom remelting of residues produced by melting in theabsence of garnet (eg Salters amp Hart 1991 Salters ampWhite 1998) The Hf^Nd isotopic compositions of theKarmutsen Formation indicate the possible involvementof depleted MORB mantle however similar to Iceland(Fitton et al 2003) Nb^Zr^Y variations provide a way todistinguish between a depleted component within theplume and a MORB-like component The Karmutsenbasalts and picrites fall within the Iceland array and donot overlap the MORB field in a logarithmic plot of NbYvs ZrY (Fig 13b) using the fields established by Fittonet al (1997) Similar to the depleted component within theCaribbean Plateau (Thompson et al 2003) the trend ofHf^Nd isotopic compositions towards Pacific MORB-likecompositions is not followed by MORB values in Nb^Zr^Y systematics and this suggests that the depleted compo-nent in the source of the Karmutsen basalts is distinctfrom the source of MORB and probably an intrinsic partof the plume The relatively radiogenic Pb isotopic ratiosand REE patterns distinct from MORB also support thisinterpretationTrace-element characteristics suggest the possible invol-
vement of sub-arc lithospheric mantle in the generation ofthe Karmutsen picrites Lassiter et al (1995) suggested thatmixing of a plume-type source with eNdthorn 6 to thorn 7 witharc material with low NbTh could reproduce the varia-tions observed in the Karmutsen basalts however theyconcluded that the absence of low NbLa ratios in theirsample of tholeiitic basalts restricts the amount of arc litho-sphere involved The study of Lassiter et al (1995) wasbased on major and trace element and Sr Nd and Pb iso-topic data for a suite of 29 samples from Buttle Lake in theStrathcona Provincial Park on central Vancouver Island(Fig 1) Whereas there are no clear HFSE depletions inthe Karmutsen tholeiitic basalts the low NbLa (and TaLa) of the Karmutsen picritic lavas in this study indicatethe possible involvement of Paleozoic arc lithosphere onVancouver Island (Fig 8)
REE modeling dynamic melting andsource mineralogyThe combined isotopic and trace-element variations of thepicritic and tholeiitic Karmutsen lavas indicate that differ-ences in trace elements may have originated from differentmelting histories For example the LuHf ratios of theKarmutsen picritic lavas (mean 008) are considerablyhigher than those in the tholeiitic basalts (mean 002)however the small range of overlapping initial eHf valuesfor the two lava suites suggests that these differences devel-oped during melting within the plume (Fig 9b) Below wemodel the effects of source mineralogy and depth of
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
28
melting on differences in trace-element concentrations inthe tholeiitic basalts and picritesDynamic melting models simulate progressive decom-
pression melting where some of the melt fraction isretained in the residue and only when the degree of partialmelting is greater than the critical mass porosity and the
source becomes permeable is excess melt extracted fromthe residue (Zou1998) For the Karmutsen lavas evolutionof trace element concentrations was simulated using theincongruent dynamic melting model developed by Zou ampReid (2001) Melting of the mantle at pressures greater than1 GPa involves incongruent melting (eg Longhi 2002)
OIB array
PacificMORB(230 Ma)
(0 Ma)
EPR(230 Ma)
-4
-2
0
2
4
8
10
12
14
16
18
20
22
-4 -2 0 2 4 6 8 10 12 14
minus2
0
2
4
6
8
10
12
14
0702 0703 0704 0705 0706
IndianMORB
87Sr 86Sr (initial)
Hf (initial)
(a) (c)
Nd (initial)
Karmutsen tholeiitic basalt
HawaiiOntong JavaCaribbean Plateau
Karmutsen high-MgO lava
high-MgO lavasKarmutsen
1540
1545
1550
1555
1560
1565
175 180 185 190 195 200375
380
385
390
395
175 180 185 190 195 200
(d) (e)
EPR
EPR
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb
Karmutsen Formation (initial)
HawaiiOntong JavaCaribbean Plateau
East Pacific Rise
Karmutsen Formation (measured)
Juan de FucaGorda
ε Nd
(initial)
Karmutsen Formation (different age-correction for 3 picrites)
(0 Ma)
NbY
001
01
1
1 10
ZrY
OIB
NMORB
(b)
Iceland array
6
Fig 13 Comparison of age-corrected (230 Ma) Sr^Nd^Hf^Pb isotopic compositions for Karmutsen flood basalts onVancouver Island withage-corrected OIB and MORB and Nb^Zr^Y variation (a) Initial eNd vs 87Sr86Sr (b) NbYand ZrY variation (after Fitton et al 1997)(c) Initial eHf vs eNd (d) Measured and initial 207Pb204Pb vs 206Pb204Pb (e) Measured and initial 208Pb204Pb vs 206Pb204Pb References fordata sources are listed in Electronic Appendix 4 Most of the compiled data were extracted from the GEOROC database (httpgeorocmpch-mainzgwdgdegeoroc) OIB array line in (c) is fromVervoort et al (1999) EPR East Pacific Rise Several high-MgO samples affected bysecondary alteration were not plotted for clarity in Pb isotope plots Estimation of fields for age-corrected Pacific MORB and EPR at 230 Mawere made using parent^daughter concentrations (in ppm) from Salters amp Stracke (2004) of Lufrac14 0063 Hffrac14 0202 Smfrac14 027 Ndfrac14 071Rbfrac14 0086 and Srfrac14 836 In the NbYvs ZrYplot the normal (N)-MORB and OIB fields not including Icelandic OIB are from data com-pilations of Fitton et al (1997 2003) The OIB field is data from 780 basalts (MgO45wt ) from major ocean islands given by Fitton et al(2003) The N-MORB field is based on data from the Reykjanes Ridge EPR and Southwest Indian Ridge from Fitton et al (1997) The twoparallel gray lines in (b) (Iceland array) are the limits of data for Icelandic rocks
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
29
with olivine being produced during the reactioncpxthornopxthorn spfrac14meltthornol for spinel peridotites The mod-eling used coefficients for incongruent melting reactionsbased on experiments on lherzolite melting (eg Kinzleramp Grove 1992 Longhi 2002) and was separated into (1)melting of spinel lherzolite (2) two combined intervals ofmelting for a garnet lherzolite (gt lherz) and spinel lher-zolite (sp lherz) source (3) melting of garnet lherzoliteThe model solutions are non-unique and a range indegree of melting partition coefficients melt reactioncoefficients and proportion of mixed source componentsis possible to achieve acceptable solutions (see Fig 14 cap-tion for description of modeling parameters anduncertainties)The LREE-depleted high-MgO lavas require melting of
a depleted spinel lherzolite source (Fig 14a) The modelingresults indicate a high degree of melting (22^25) similarto the results from PRIMELT1 based on major-elementvariations (discussed above) of a LREE-depleted sourceThe high-MgO lavas lack a residual garnet signature Theenriched tholeiitic basalts involved melting of both garnetand spinel lherzolite and represent aggregate melts pro-duced from continuous melting throughout the depth ofthe melting column (Fig 14b) Trace-element concentra-tions in the tholeiitic and picritic lavas are not consistentwith low-degree melting of garnet lherzolite alone(Fig 14c) The modeling results indicate that the aggregatemelts that formed the tholeiitic basalts would haveinvolved lesser proportions of enriched small-degreemelts (1^6 melting) generated at depth and greateramounts of depleted high-degree melts (12^25) gener-ated at lower pressure (a ratio of garnet lherzolite tospinel lherzolite melt of 14 provides the best fits seeFig 14b and description in figure caption) The totaldegree of melting for the tholeiitic basalts was high(23^27) similar to the degree of partial melting in thespinel lherzolite facies for the primary melts that eruptedas the Keogh Lake picrites of a source that was also prob-ably depleted in LREE
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
(c)
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
Melting of garnet lherzolite
High-MgO lavas
Melting of spinel lherzolite
source (07DM+03PM)
source (07DM+03PM)
6 melting
30 melting
(b)
(a)
30 melting
Sam
ple
Cho
ndrit
eS
ampl
eC
hond
rite 20 melting
1 melting
Melting of garnet lherzoliteand spinel lherzolite
x gt lherz + 4x sp lherz
5 meltingx gt lherz + 4x sp lherz
Sam
ple
Cho
ndrit
e
Tholeiitic lavas
Tholeiitic lavas
High-MgO lavas
source (07DM+03PM)
Fig 14 Trace-element modeling results for incongruent dynamicmantle melting for picritic and tholeiitic lavas from the KarmutsenFormation Three steps of modeling are shown (a) melting of spinellherzolite compared with high-MgO lavas (b) melting of garnet lher-zolite and spinel lherzolite compared with tholeiitic lavas (c) meltingof garnet lherzolite compared with tholeiitic and picritic lavasShaded fields in each panel for tholeiitic basalts and picrites arelabeled Patterns with symbols in each panel are modeling results(patterns are 5 melting increments in (a) and (b) and 1 incre-ments in (c) PM primitive mantle DM depleted MORB mantle gtlherz garnet lherzolite sp lherz spinel lherzolite In (b) the ratios ofper cent melting for garnet and spinel lherzolite are indicated (x gtlherzthorn 4x sp lherz means that for 5 melting 1 melt in garnetfacies and 4 melt in spinel facies where x is per cent melting in theinterval of modeling) The ratio of the respective melt proportions inthe aggregate melt (x gt lherzthorn 4x sp lherz) was kept constant andthis ratio of melt contribution provided the best fit to the data Arange of proportion of source components (07DMthorn 03PM to
09DMthorn 01PM) can reproduce the variation in the high-MgOlavas Melting modeling uses the formulation of Zou amp Reid (2001)an example calculation is shown in their Appendix PM fromMcDonough amp Sun (1995) and DM from Salters amp Stracke (2004)Melt reaction coefficients of spinel lherzolite from Kinzler amp Grove(1992) and garnet lherzolite fromWalter (1998) Partition coefficientsfrom Salters amp Stracke (2004) and Shaw (2000) were kept constantSource mineralogy for spinel lherzolite (018cpx027opx052ol003sp) from Kinzler (1997) and for garnet lherzolite(034cpx008opx053ol005gt) from Salters amp Stracke (2004) Arange of source compositions for melting of spinel lherzolite(07DMthorn 03PM to 03DMthorn 07PM) reproduce REE patternssimilar to those of the high-MgO lavas with best fits for 22ndash25melting and 07DMthorn 03PM source composition Concentrationsof tholeiitic basalts cannot be reproduced from an entirelyprimitive or depleted source
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
30
The uncertainty in the composition of the source in thespinel lherzolite melting model for the high-MgO lavasprecludes determining whether the source was moredepleted in incompatible trace elements than the source ofthe tholeiitic lavas (see Fig 14 caption for description) It ispossible that early high-pressure low-degree melting left aresidue within parts of the plume (possibly the hotter inte-rior portion) that was depleted in trace elements (egElliott et al 1991) further decompression and high-degreemelting of such depleted regions could then have generatedthe depleted high-MgO Karmutsen lavas Shallow high-degree melting would preferentially sample depletedregions whereas deeper low-degree melting may notsample more refractory depleted regions (eg CaribbeanEscuder-Viruete et al 2007) The picrites lie at the highend of the range of Nd isotopic compositions and havelow NbZr similar to depleted Icelandic basalts (Fittonet al 2003) therefore the picrites may represent the partsof the mantle plume that were the most depleted prior tothe initiation of melting As Fitton et al (2003) suggestedthese depleted melts may only rarely be erupted withoutcombining with more voluminous less depleted melts andthey may be detected only as undifferentiated small-volume flows like the Keogh Lake picrites within theKarmutsen Formation on Vancouver IslandDecompression melting within the mantle plume initiatedwithin the garnet stability field (35^25 GPa) at highmantle potential temperatures (Tp414508C) and pro-ceeded beneath oceanic arc lithosphere within the spinelstability field (25^09 GPa) where more extensivedegrees of melting could occur
Magmatic evolution of Karmutsentholeiitic basaltsPrimary magmas in large igneous provinces leave exten-sive coarse-grained cumulate residues within or below thecrust these mafic and ultramafic plutonic sequences repre-sent a significant proportion of the original magma thatpartially crystallized at depth (eg Farnetani et al 1996)For example seismic and petrological studies of OntongJava (eg Farnetani et al 1996) combined with the use ofMELTS (Ghiorso amp Sack 1995) indicate that the volcanicsequence may represent only 40 of the total magmareaching the Moho Neal et al (1997) estimated that30^45 fractional crystallization took place for theOntong Java magmas and produced cumulate thicknessesof 7^14 km corresponding to 11^22 km of flood basaltsdepending on the presence of pre-existing oceanic crustand the total crustal thickness (25^35 km) Interpretationsof seismic velocity measurements suggest the presence ofpyroxene and gabbroic cumulates 9^16 km thick beneaththe Ontong Java Plateau (eg Hussong et al 1979)Wrangellia is characterized by high crustal velocities in
seismic refraction lines which is indicative of mafic
plutonic rocks extending to depth beneath VancouverIsland (Clowes et al 1995)Wrangellia crust is 25^30 kmthick beneath Vancouver Island and is underlain by astrongly reflective zone of high velocity and density thathas been interpreted as a major shear zone where lowerWrangellia lithosphere was detached (Clowes et al 1995)The vast majority of the Karmutsen flows are evolved tho-leiitic lavas (eg low MgO high FeOMgO) indicating animportant role for partial crystallization of melts at lowpressure and a significant portion of the crust beneathVancouver Island has seismic properties that are consistentwith crystalline residues from partially crystallizedKarmutsen magmas To test the proportion of the primarymagma that fractionated within the crust MELTS(Ghiorso amp Sack 1995) was used to simulate fractionalcrystallization using several estimated primary magmacompositions from the PRIMELT1 modeling results Themajor-element composition of the primary magmas couldnot be estimated for the tholeiitic basalts because of exten-sive plagthorn cpxthornol fractionation and thus the MELTSmodeling of the picrites is used as a proxy for the evolutionof major elements in the volcanic sequence A pressure of 1kbar was used with variable water contents (anhydrousand 02wt H2O) calculated at the quartz^fayalite^magnetite (QFM) oxygen buffer Experimental resultsfrom Ontong Java indicate that most crystallizationoccurs in magma chambers shallower than 6 km deep(52 kbar Sano amp Yamashita 2004) however some crys-tallization may take place at greater depths (3^5 kbarFarnetani et al 1996)A selection of the MELTS results are shown in Fig 15
Olivine and spinel crystallize at high temperature until15^20wt of the liquid mass has fractionated and theresidual magma contains 9^10wt MgO (Fig 15f) At1235^12258C plagioclase begins crystallizing and between1235 and 11908C olivine ceases crystallizing and clinopyr-oxene saturates (Fig 15f) As expected the addition of clin-opyroxene and plagioclase to the crystallization sequencecauses substantial changes in the composition of the resid-ual liquid FeO and TiO2 increase and Al2O3 and CaOdecrease while the decrease in MgO lessens considerably(Fig 15) The near-vertical trends for the Karmutsen tho-leiitic basalts especially for CaO do not match the calcu-lated liquid-line-of-descent very well which misses manyof the high-CaO high-Al2O3 low-FeO basalts A positivecorrelation between Al2O3CaO and SrNd indicates thataccumulation of plagioclase played a major role in the evo-lution of the tholeiitic basalts (Fig 15g) Increasing thecrystallization pressure slightly and the water content ofthe melts does not systematically improve the match ofthe MELTS models to the observed dataThe MELTS results indicate that a significant propor-
tion of the crystallization takes place before plagioclase(15^20 crystallization) and clinopyroxene (35^45
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
31
crystallization) join the crystallization sequence Theseresults suggest that445 crystallization of the primarymagmas occurred (estimated from the residual liquid percent in Fig 15f) and if the original flood basalt stratigra-phy on Vancouver Island was 6 km thick at least anequivalent amount of complementary plutonic rocksshould have crystallized The Karmutsen basalts and theirplutonic residues thus represent a significant addition ofcrust (perhaps 412 km thickness) previously thickenedduring the formation of the underlying Paleozoic arc
Growth of the Wrangellia oceanic plateauon Vancouver IslandThe extrusive part of the Wrangellia oceanic plateau onVancouver Island was constructed in a three-layersequence of mostly tholeiitic basalts with a restrictedrange of composition (Fig 16) The plateau overliesDevonian to Mississippian arc volcanic sequences andmarine sediments of Mississippian to Permian ages thatare intruded by gabbroic rocks related to the flood basaltsThe volcanic stratigraphy of theWrangellia plateau formed
00
05
10
15
20
25
30
35
40
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
16
0 2 4 6 8 10 12 14 16 18 20
10
11
12
13
14
15
16
17
18
19
20
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
0 2 4 6 8 10 12 14 16 18 20
MgO (wt)
0 10 20 30 40 50
percent of total mass
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
) Mineral proportions
Sp (e)
Ol
CpxPlag
PicriteHigh-MgO basaltTholeiitic basalt
TiO2 (wt)
MgO (wt)
Al2O3 (wt) CaO (wt)
FeO (wt)
MgO(a) (b)
(c) (d)
MgO (wt)MgO (wt)
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
)
Residual liquid ()
1220
1190
Ol + Sp
Plag+Ol+Sp Plag+Cpx+Sp
02 wt H2O
no H2O
Residual liquid and crystallizing phases
(f)
0102030405060708090100
plagioclaseaccumulation
plagioclaseaccumulation
plagioclase
accumulatio
n
10
12
14
16
0 10 20 30 40
Al2O3CaO
SrNd
tholeiitic basalts
(g)
Fig 15 Forward fractional crystallization modeling results for major elements from MELTS (Ghiorso amp Sack 1995) compared with picriticand tholeiitic lavas from the Karmutsen Formation (a^d) Major-element compositions of residual liquids compared with Karmutsen lavacompositions For clarity the results are shown only for one starting composition (open circle) using the composition of the estimated primarymagma for the high-MgO lava series from the modeling technique of Herzberg et al (2007) for sample 4723A13 as shown in Fig 12 The esti-mated primary magmas for samples 4723A3 and 93G171 yield similar results A pressure of 1 kbar was used with no H2O and 02wt H2Ocalculated at QFM black line is anhydrous and grey line is 02wt H2OThe end-point for the anhydrous result is at 78 fractionation and11308C and the end-point for the result with 02wt H2O is 75 fractionation and11108C (e) Mineral proportions of crystallized phases fromone MELTrun (sample 4723A13) with no H2O (f) Per cent residual liquid vs temperature labelled with intervals of crystallizing phases (g)Al2O3CaO vs SrNd for the tholeiitic basalts
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
32
in a single eruptive phase (c 231^225 Ma) as an emergentbasalt sequence during a deeper-water shallow-water andsubaerial stage and subsided below sea level relativelysoon after its formation (Fig 16) The deeper-water stagecontains pillowed and massive submarine flows (3000mthick) that built up the submarine volcanic edifice from theseafloor The lowest pillow basalts were emplaced onunconsolidated and lithified fine-grained siliciclastic andcarbonaceous sediments containing Daonella beds(c 235^232 Ma) that are intruded by abundantKarmutsen mafic sills (Fig 16a) The different pillowedand massive submarine flows resulted primarily from dif-ferent effusive rates and local topography Some of themassive submarine flows commonly recognized by radiat-ing columnar jointing represent master tubes for the deliv-ery of lava to distal parts of submarine flow fields Thepillowed flows which commonly have uneven topography
in modern sequences may have served as levees to themassive submarine flows rather than allowing unob-structed spreading and inflation typical of the massive sub-aerial flows that form on shallow slopes in continentalflood basalt provinces (eg Self et al 1997) The rapid suc-cession of flows is evident by the near absence of sedimentbetween submarine flows Minor volumes of picritic pillowlavas occur in the upper part of the pillowed flow sequencein areas of the plateau exposed on northern VancouverIsland (Fig 16 Nixon et al 2008)The growth of the plateau in shallow water led to
emplacement of an increasing proportion of volcaniclastics(pillow breccias and hyaloclastites) that accumulated up to1500m thick in some areas (Fig 16b Nixon et al 2008)Horizons of pillowed flows occur locally within the volca-niclastic stratigraphy Sedimentary structures (eg gradedbedding fluidization structures) and petrographic textures
Late Devonian-Mississippian arc
Daonella beds
Middle Triassic
deeper water
Late Devonian-Mississippian arc
pillowed flows massive submarine
flows
pillow brecciahyaloclastiteshallow water
~4 km
~6 km
Late Triassic limestoneinterflow limestonelenses
Karmutsen Formation
mafic intrusions sediment-sill complex
Karmutsen Formation
(a) (b)
(c) (d)
subaerial flows
Late Devonian-Mississippian arc Late Devonian-Mississippian arc
picritic pillow basalts
Miss to Permian marine sediments
sills
Fig 16 Growth of the Wrangellia oceanic plateau onVancouver Island (a) Deep-water eruptive phase of pillowed and massive submarineflows with almost no intervening sediments onto MiddleTriassic shale containing Daonella beds (c 235^232 Ma) The basement of the plateauis composed of Devonian to Mississippian arc sequences overlain by Mississippian to Permian carbonates and siliciclastic sediments(b) Shallow-water flow emplacement with increasing proportion of volcaniclastics resulted primarily from cooling^contraction granulationPicritic pillow basalts occur primarily near the transition between pillowed submarine flows and volcaniclastic units Pillowed and massiveflows occur locally within the volcaniclastic unit (c) Emergent stage of growth consists of overlapping subaerial flow fields similar to continen-tal flood basalts with pahoehoe flow structures Multi-tiered columnar jointing occurrs locally in response to ponding of flows (eg Bondre et al2004) Interflow carbonate deposits and plagioclase megacrystic flows developed locally during the waning stage of volcanism (d) Post-volcanicsubsidence and deposition of Late Carnian to Norian (c 212^223 Ma) platform carbonates on top of the plateau Units are labeled in the dia-grams Thicknesses of the underlying units are not to scale
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
33
within the volcaniclastic rocks indicate formation primar-ily via cooling^contraction granulation magma^water^steam interaction autobrecciation resedimentation andmass-wasting rather than pyroclastic eruption (Nixonet al 2008)A broad subaerial platform was constructed by emplace-
ment of subaerial lava flows during an emergent subaerialstage (Fig 16c) The subaerial unit predominantly com-prises well-layered sheet flows that formed in a similarway to continental flood basalts as hundred(s) of pahoehoeflow fields erupted on shallow slopes (Self et al 1997) how-ever volcaniclastic and pillowed flow units are locally pre-sent mostly in the upper parts of the stratigraphyLocalized ponding of flows led to the formation of multi-tiered columnar jointing in some subaerial flows (egBondre et al 2004) As volcanism waned local interflowcarbonate deposits developed and plagioclase megacrysticflows were erupted Subsidence of the plateau is recordedby the interflow lenses and deposits of platform carbonates(40^1500m thick) on top of the Karmutsen basalts thatcontain Late Carnian to Early Norian fossils (Fig 16d c216^227 Ma Nixon et al 2008)The evolution of the internal architecture of the
Wrangellia oceanic plateau was primarily related to theconditions of emplacement (ie water depth and eruptionenvironment) There does not appear to have been a sys-tematic coincident evolution or shift of lava compositionsonVancouver Island as has been recognized in most con-tinental flood basalts and parts of the Wrangellia oceanicplateau in Alaska and Yukon (Greene et al 2008 2009)In general with the exception of the picritic pillow lavasin the submarine stratigraphy and local variations betweenhigh-MgO and tholeiitic lavas there are no significantcompositional differences between lavas erupted duringthe different stages of growth of the plateau (ie submarineand subaerial flows) nor are substantial compositional dif-ferences generally found between massive submarine flowsin contact with pillowed flows tholeiitic basalt flows nearthe base of the Karmutsen Formation are compositionallysimilar to flows sampled near the top of the volcanicsequence Where it was possible to sample continuous sec-tions of volcanic stratigraphy greater than several hun-dreds of meters (eg Keogh^Maynard Lake area west ofthe Karmutsen Range Fig 2) there is no clear relationshipbetween stratigraphic position and compositionalvariation
CONCLUSIONSThe Karmutsen Formation covers extensive areas ofVancouver Island (20 000 km2) and forms part of a largeoceanic plateau erupted over a geologically short interval(c 225^231 Ma) The tripartite volcanic stratigraphyexposed on Vancouver Island is at least 6 km thick andcomprises a succession of submarine flows volcaniclastic
deposits and massive subaerial flowsThese volcanologicaldifferences are primarily related to the eruption environ-ment (deep-water shallow-water or subaerial) The rapidgrowth of the plateau prevented significant accumulationsof intervening sediment except in the uppermost stratigra-phy where isolated limestone lenses commonly associatedwith pillow basalt and hyaloclastite deposits preserve arecord of the final stages of subsidence of the plateauThe Wrangellia plateau on Vancouver Island was con-
structed dominantly of tholeiitic basalt with restrictedmajor- and trace-element and isotopic compositional var-iations Minor volumes of picritic pillow basalts eruptedlate in the submarine phase The Karmutsen basalts wereemplaced onto older Paleozoic arc volcanic sequences andlow NbLa for the picrites may indicate a contributionfrom the sub-arc lithospheric mantle Modeling of the pet-rogenesis of the Karmutsen picrites indicates melting ofanomalously hot mantle (15008C) and extensive degreesof partial melting (23^27) and is consistent with aplume initiation model Combined Sr^Nd^Hf^Pb isotopicsystematics indicate that the picritic and tholeiitic lavaswere derived from a depleted plume-type mantle sourcedistinct from the source of MORB that has compositionalsimilarities to the source of the lavas forming theCaribbean Plateau Picrites were generated by melting ofspinel lherzolite whereas the tholeiitic basalts formed asaggregate melts of both garnet and spinel lherzolite Thetholeiitic basalts underwent extensive low-pressure fractio-nation (52^3 kbar) and seismic studies indicate that someof theWrangellia crust beneathVancouver Island may cor-respond to the plutonic residues of the Karmutsen floodbasalts
ACKNOWLEDGEMENTSWe are grateful to Nick Arndt for helping us get this pro-ject started and Nick Massey for insights into VancouverIsland geology We greatly appreciate the assistance withPRIMELT modeling from Claude Herzberg We thankMikkel Schau for his insight and enthusiasm during field-work Jane Barling assisted with analyses by MC-ICP-MSThoughtful and thorough reviews of the manuscript byJohn Mahoney Fred Frey Wendy Bohrson DougalJerram and Teal Riley are greatly appreciated Fundingwas generously provided by the Rocks to Riches Programadministered by the BC amp Yukon Chamber of Mines in2004 by the BC Geological Survey and by NSERCDiscovery Grants to JSS and DW ARG was supportedby a University Graduate Fellowship at UBC
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
34
REFERENCESBarker F Brown A S Budahn J R amp Plafker G (1989) Back-arc
with frontal-arc component origin of Triassic Karmutsen basaltBritish Columbia Canada Chemical Geology 75 81^102
Blichert-Toft JWeis D Maerschalk C Agranier A amp Albare de F(2003) Hawaiian hot spot dynamics as inferred from the Hf and Pbisotope evolution of Mauna Kea volcano Geochemistry GeophysicsGeosystems 4(2) 1^27 doi1010292002GC000340
Bondre N R Duraiswami R A amp Dole G (2004) Morphologyand emplacement of flows from the Deccan Volcanic ProvinceIndia Bulletin of Volcanology 66 29^45
Carlisle D (1963) Pillow breccias and their aquagene tuffs QuadraIsland British Columbia Journal of Geology 71 48^71
Carlisle D (1972) Late Paleozoic to Mid-Triassic sedimentary-volcanic sequence on Northeastern Vancouver Island Geological
Society of Canada Paper Report of Activities 72-1 Part B 24^30Carlisle D amp SuzukiT (1974) Emergent basalt and submergent car-
bonate^clastic sequences including the Upper Triassic Dilleri andWelleri zones on Vancouver Island Canadian Journal of Earth
Sciences 11 254^279Chauvel C amp Blichert-Toft J (2001) A hafnium isotope and trace
element perspective on melting of the depleted mantle Earth and
Planetary Science Letters 190 137^151Chayes F (1954)The theory of thin-section analysis Journal of Geology
62 92^101Cho M Liou J G amp Maruyama S (1987) Transition from the zeo-
lite to prehnite^pumpellyite facies in the Karmutsen metabasitesVancouver Island British Columbia Journal of Petrology 27(2)467^494
Clowes R M Zelt C A Amor J R amp Ellis R M (1995)Lithospheric structure in the southern Canadian Cordillera froma network of seismic refraction lines Canadian Journal of Earth
Sciences 32(10) 1485^1513Coffin M F amp Eldholm O (1994) Large igneous provinces Crustal
structure dimensions and external consequences Reviews of
Geophysics 32(1) 1^36Cox K G (1980) A model for flood basalt volcanism Journal of
Petrology 21 629^650Eldholm O amp Coffin M F (2000) Large igneous provinces
and plate tectonics In Richards M A Gordon R G amp vander Hilst R D (eds) The History and Dynamics of Global
Plate Motions American Geophysical Union Geophysical Monograph 121309^326
Elliott T R Hawkesworth C J amp Grolaquo nvold K (1991) Dynamicmelting of the Iceland plume Nature 351 201^206
Escuder-Viruete J Pecurren rez-Estaucurren n A Contreras F Joubert MWeis D Ullrich T D amp Spadea P (2007) Plume mantle sourceheterogeneity through time Insights from the Duarte ComplexHispaniola northeastern Caribbean Journal of Geophysical Research112(B04203) doi1010292006JB004323
Farnetani C G Richards M A amp Ghiorso M S (1996)Petrological models of magma evolution and deep crustal structurebeneath hotspots and flood basalt provinces Earth and Planetary
Science Letters 143 81^94Fitton J G amp Godard M (2004) Origin and evolution of magmas
on the Ontong Java Plateau In Fitton J G Mahoney J JWallace P J amp Saunders A D (eds) Origin and Evolution of the
Ontong Java Plateau Geological Society London Special Publications 229151^178
Fitton J G Saunders A D Norry M J Hardarson B S ampTaylor R N (1997) Thermal and chemical structure of theIceland plume Earth and Planetary Science Letters 153 197^208
Fitton J G Saunders A D Kempton P D amp Hardarson B S(2003) Does depleted mantle form an intrinsic part of the Icelandplume Geochemistry Geophysics Geosystems 4(3) 1032 doi1010292002GC000424
Frey F A Huang S Blichert-Toft J Regelous M amp Boyet M(2005) Origin of depleted components in basalt related to theHawaiian hotspot Evidence from isotopic and incompatible ele-ment ratios Geochemistry Geophysics Geosystems 6(1) 23 doi1010292004GC000757
Galer S J G amp Abouchami W (1998) Practical application of leadtriple spiking for correction of instrumental mass discriminationMineralogical Magazine 62A 491^492
Ghiorso M S amp Sack R O (1995) Chemical mass transfer in mag-matic processes IV A revised and internally consistent thermody-namic model for the interpolation and extrapolation of liquid^solidequilibria in magmatic systems at elevated temperatures and pres-sures Contributions to Mineralogy and Petrology 119 197^212
Greene A R Scoates J S amp Weis D (2005)WrangelliaTerrane onVancouver Island Distribution of flood basalts with implicationsfor potential Ni^Cu^PGE mineralization In Grant B (ed)Geological Fieldwork 2004 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2005-1 209^220Greene A R Scoates J S Nixon G T amp Weis D (2006) Picritic
lavas and basal sills in the Karmutsen flood basalt provinceWrangellia northern Vancouver Island In Grant B (ed)Geological Fieldwork 2005 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2006-1 39^52Greene A R Scoates J S amp Weis D (2008) Wrangellia flood
basalts in Alaska A record of plume^lithosphere interaction in aLate Triassic accreted oceanic plateau Geochemistry Geophysics
Geosystems 9(Q12004) doi1010292008GC002092Greene A R Scoates J S Weis D amp Israel S (2009)
Geochemistry of triassic flood basalts from the Yukon (Canada)segment of the accreted Wrangellia oceanic plateau Lithosdoi101016jlithos200811010
Hauff F Hoernle K Tilton G Graham D W amp Kerr A C(2000) Large volume recycling of oceanic lithosphere over shorttime scales geochemical constraints from the Caribbean LargeIgneous Province Earth and Planetary Science Letters 174 247^263
Hauff F Hoernle K amp Schmidt A (2003) Sr^Nd^Pb compositionof Mesozoic Pacific oceanic crust (Site 1149 and 801 ODP Leg185)Implications for alteration of ocean crust and the input into theIzu^Bonin^Mariana subduction system Geochemistry Geophysics
Geosystems 4(8) doi1010292002GC000421Herzberg C (2004) Partial melting below the Ontong Java Plateau
In Fitton J G Mahoney J J Wallace P J amp Saunders A D(eds) Origin and Evolution of the Ontong Java Plateau Geological SocietyLondon Special Publications 229 179^183
Herzberg C (2006) Petrology and thermal structure of the Hawaiianplume from Mauna Kea volcano Nature 444(7119) 605
Herzberg C amp Asimow P D (2008) Petrology of some oceanicisland basalts PRIMELT2XLS software for primary magma cal-culation Geochemistry Geophysics Geosystems 9(Q09001) doi1010292008GC002057
Herzberg C amp OrsquoHara M J (2002) Plume-associated ultramaficmagmas of Phanerozoic age Journal of Petrology 43(10) 1857^1883
Herzberg C Asimow P D Arndt N Niu Y Lesher C MFitton J G Cheadle M J amp Saunders A D (2007)Temperatures in ambient mantle and plumes Constraints frombasalts picrites and komatiites Geochemistry Geophysics Geosystems8(Q02006) doi1010292006GC001390
Huang S amp Frey F A (2005) Trace element abundances of MaunaKea basalt from phase 2 of the Hawaii Scientific Drilling Project
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
35
Petrogenetic implications of correlations with major element con-tent and isotopic ratios Geochemistry Geophysics Geosystems 4(6)doi1010292002GC000322
Hussong D M Wipperman L K amp Kroenke L W (1979) Thecrustal structure of the Ontong Java and Manihiki OceanicPlateaus Journal of Geophysical Research 84(B11) 6003^6010
Jerram D A amp Widdowson M (2005) The anatomy of ContinentalFlood Basalt Provinces geological constraints on the processes andproducts of flood volcanism Lithos 79(3^4) 385^405
Jones D L Silberling N J amp Hillhouse J (1977)Wrangellia a dis-placed terrane in northwestern North America CanadianJournal ofEarth Sciences 14(11) 2565^2577
Kerr A C (2003) Oceanic Plateaus In Rudnick R (ed) The Crust
Treatise on GeochemistryVol 3 Oxford Elsevier Science pp 537^565Kerr A C amp Mahoney J J (2007) Oceanic plateaus Problematic
plumes potential paradigms Chemical Geology 241 332^353Kerr A CTarney J Marriner G F Klaver GT Saunders A D
amp Thirlwall M F (1996) The geochemistry and petrogenesis ofthe Late-Cretaceous picrites and basalts of Curacao NetherlandsAntilles a remnant of an oceanic plateau Contributions to
Mineralogy and Petrology 124(1) 29^43Kerr A C Marriner G F Tarney J Nivia A Saunders A D
Thirlwall M F amp Sinton A C (1997) Cretaceous basaltic ter-ranes in Western Colombia Elemental chronological and Sr-Ndisotopic constraints on petrogenesis Journal of Petrology 38(6)677^702
Kinzler R J (1997) Melting of mantle peridotite at pressuresapproaching the spinel to garnet transition Application to mid-ocean ridge basalt petrogenesis Journal of Geophysical Research
102(B1) 853^874Kinzler R J amp Grove T L (1992) Primary magmas of mid-ocean
ridge basalts 1 Experiments and methods Journal of Geophysical
Research 97(B5) 6885^6906Korenaga J (2005) Why did not the Ontong Java Plateau form sub-
aerially Earth and Planetary Science Letters 234 385^399Kuniyoshi S (1972) Petrology of the Karmutsen (Volcanic) Group
northeastern Vancouver Island British Columbia PhD disserta-tion University of California Los Angeles 242 pp
Lassiter J C DePaolo D J amp Mahoney J J (1995) Geochemistry ofthe Wrangellia flood basalt province Implications for the role ofcontinental and oceanic lithosphere in flood basalt genesis Journalof Petrology 36(4) 983^1009
LeBas M J (2000) IUGS reclassification of the high-Mg and picriticvolcanic rocks Journal of Petrology 41(10) 1467^1470
Longhi J (2002) Some phase equilibrium systematics of lherzolitemelting I Geochemistry Geophysics Geosystems 3(3) doi1010292001GC000204
MacDonald G A amp Katsura T (1964) Chemical composition ofHawaiian lavas Journal of Petrology 5 82^133
Mahoney J J (1987) An isotopic survey of Pacific oceanic plateausimplications for their nature and origin In Keating B HFryer P Batiza R amp Boehlert GW (eds) Seamounts Islands andAtolls American Geophysical Union Geophysical Monograph 43 207^220
Mahoney J LeRoex A P Peng Z Fisher R L amp Natland J H(1992) Southwestern limits of Indian Ocean Ridge mantle and theorigin of low 206Pb204Pb mid-ocean ridge basalt isotope systema-tics of the central Southwest Indian Ridge (178^508E) Journal ofGeophysical Research 97(B13) 19771^19790
Mahoney J J Sinton J M Kurz M D MacDougall J DSpencer K J amp Lugmair GW (1994) Isotope and trace elementcharacteristics of a super-fast spreading ridge East Pacific rise13^238S Earth and Planetary Science Letters 121 173^193
Massey N W D (1995a) Geology and mineral resources of theAlberni^Nanaimo Lakes sheet Vancouver Island 92F1W 92F2Eand part of 92F7E British Columbia Ministry of Energy Mines and
Petroleum Resources Paper 1992-2 132 ppMassey N W D (1995b) Geology and mineral resources of the
Cowichan Lake sheet Vancouver Island 92C16 British Columbia
Ministry of Energy Mines and Petroleum Resources Paper 1992-3 112 ppMassey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005a) Digital Geology Map of British Columbia Tile NM9 MidCoast BC British Columbia Ministry of Energy and Mines Geofile
2005-2Massey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005b) Digital Geology Map of British Columbia Tile NM10Southwest BC British Columbia Ministry of Energy and Mines Geofile
2005-3McDonough W F amp Sun S (1995) The composition of the Earth
Chemical Geology 120 223^253Monger JW H amp Journeay J M (1994) Basement geology and tec-
tonic evolution of theVancouver region In Monger JW H (ed)Geology and Geological Hazards of the Vancouver Region Southwestern
British Columbia Geological Survey of Canada Bulletin 481 3^25Muller J E (1977) Geology of Vancouver Island Geological Survey of
Canada Open File Map 463Muller J E (1980)The Paleozoic Sicker Group of Vancouver Island British
Columbia Geological Survey of Canada Paper 79-30 22 ppMuller J E Northcote K E amp Carlisle D (1974) Geology and
mineral deposits of Alert Bay^Cape Scott map area VancouverIsland British Columbia Geological Survey of Canada Paper 74-8 77
Neal C R Mahoney J J Kroenke L W Duncan R A ampPetterson M G (1997) The Ontong^Java Plateau InMahoney J J amp Coffin M F (eds) Large Igneous Provinces
Continental Oceanic and Planetary FloodVolcanism American Geophysical
Union Geophysical Monograph 100 183^216Niu Y Collerson K D Batiza R Wendt J I amp Regelous M
(1999) Origin of enriched-typed mid-ocean ridge basalt at ridgesfar from mantle plumes The East Pacific Rise at 11820rsquoN Journalof Geophysical Research 104 7067^7088
Nixon G T amp Orr A J (2007) Recent revisions to the EarlyMesozoic stratigraphy of Northern Vancouver Island (NTS 102I092L) and metallogenic implicatinos British Columbia InGrant B (ed) Geological Fieldwork 2006 British Columbia Ministry of
Energy Mines and Petroleum Resources Paper 2007-1 163^177Nixon GT Hammack J L KoyanagiV M Payie G J Haggart
JW Orchard M JTozerT Friedman R M Archibald D APalfy J amp Cordey F (2006a) Geology of the Quatsino^PortMcNeill area northernVancouver Island British Columbia Ministry
of Energy Mines and Petroleum Resources Geoscience Map 2006-2 scale150 000
Nixon G T Kelman M C Stevenson D Stokes L A ampJohnston K A (2006b) Preliminary geology of the Nimpkishmap area (NTS 092L07) northern Vancouver Island BritishColumbia In Grant B amp Newell J M (eds) Geological Fieldwork2005 British Columbia Ministry of Energy Mines and Petroleum Resources
Paper 2006-1 135^152Nixon G T Laroque J Pals A Styan J Greene A R amp
Scoates J S (2008) High-Mg lavas in the Karmutsen flood basaltsnorthernVancouver Island (NTS 092L) Stratigraphic setting andmetallogenic significance In Grant B (ed) Geological Fieldwork
2007 British Columbia Ministry of Energy Mines and Petroleum
Resources Paper 2008-1 175^190Nowell G M Kempton P D Noble S R Fitton J G
Saunders A Mahoney J J amp Taylor R N (1998) High precisionHf isotope measurements of MORB and OIB by thermal ionisation
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
36
mass spectrometry insights into the depleted mantle Chemical
Geology 149 211^233Parrish R R amp McNicoll V J (1992) U^Pb age determinations
from the southern Vancouver Island area British Columbia InRadiogenic Age and Isotopic Studies Report 5 Geological Survey of
Canada Paper 91-2 79^86Peate DW (1997) The Parana^Etendeka Province In Coffin M F
amp Mahoney J (eds) Large Igneous Provinces Continental Oceanic andPlanetary Flood Volcanism American Geophysical Union Geophysical
Monograph 100 217^245Perfit M R Fornari D J Ridley W I Kirk P D Casey J
Kastens K A Reynolds J R Edwards M Desonie DShuster R amp Paradis S (1996) Recent volcanism in the Siqueirostransform fault picritic basalts and implications for MORBmagma genesis Earth and Planetary Science Letters 141(1^4) 91^108
Pretorius W Weis D Williams G Hanano D Kieffer B ampScoates J S (2006) Complete trace elemental characterization ofgranitoid (USGSG-2GSP-2) reference materials by high resolutioninductively coupled plasma-mass spectrometry Geostandards and
Geoanalytical Research 30(1) 39^54Regelous M Niu Y Wendt J I Batiza R Greig A amp
Collerson K D (1999) Variations in the geochemistry of magma-tism on the East Pacific Rise at 10830rsquoN since 800 ka Earth and
Planetary Science Letters 168 45^63Recurren villon S Chauvel C Arndt N T Pik R Martineau F
Fourcade S amp Marty B (2002) Heterogeneity of the Caribbeanplateau mantle source Sr O and He isotopic compositions of oli-vine and clinopyroxene from Gorgona Island Earth and Planetary
Science Letters 205(1^2) 91Rhodes J M (1996) Geochemical stratigraphy of lava flows samples
by the Hawaii Scientific Drilling Project Journal of Geophysical
Research 101(B5) 11729^11746Rhodes J M amp Vollinger M J (2004) Composition of basaltic lavas
sampled by phase-2 of the Hawaii Scientific Drilling ProjectGeochemical stratigraphy and magma types Geochemistry
Geophysics Geosystems 5(3) doi1010292002GC000434Richards M A Jones D L Duncan R A amp DePaolo D J (1991)
A mantle plume initiation model for the Wrangellia flood basaltand other oceanic plateaus Science 254 263^267
Salters V J M amp Hart S R (1991) The mantle sources of oceanridges islands and arcs the Hf-isotope connection Earth and
Planetary Science Letters 104 364^380Salters V J M amp Stracke A (2004) Composition of the depleted
SaltersV J amp WhiteW (1998) Hf isotope constraints on mantle evo-lution Chemical Geology 145 447^460
SanoT amp Yamashita S (2004) Experimental petrology of basementlavas from Ocean Drilling Program Leg 192 implications for dif-ferentiation processes in Ontong Java Plateau magmas InFitton J G Mahoney J JWallace P J amp Saunders A D (eds)Origin and Evolution of the Ontong Java Plateau Geological Society
London Special Publications 229 185^218Saunders A D (2005) Large igneous provinces Origin and environ-
mental consequences Elements 1 259^263Self S ThordarsonT amp Keszthely L (1997) Emplacement of conti-
nental flood basalt lava flows In Mahoney J J amp Coffin M F(eds) Large Igneous Provinces Continental Oceanic and Planetary Flood
Volcanism American Geophysical Union Geophysical Monograph 100381^410
Shaw D M (2000) Continuous (dynamic) melting theory revisitedCanadian Mineralogist 38(5) 1041^1063
Sluggett C L (2003) Uranium^lead age and geochemical constraintson Paleozoic and Early Mesozoic magmatism in WrangelliaTerrane Saltspring Island British Columbia BSc thesisUniversity of British ColumbiaVancouver 84 pp
Storey M Mahoney J J amp Saunders A D (1997) Cretaceousbasalts in Madagascar and the transition between plume and con-tinental lithospheric mantle sources In Mahoney J J ampCoffin M F (eds) Large Igneous Provinces Continental Oceanic and
Planetary Flood Volcanism Aamerican Geophysical Union Geophysical
Monograph 100 95^122Surdam R C (1967) Low-grade metamorphism of the Karmutsen
Group PhD dissertation University of California Los Angeles288 pp
Sutherland-Brown A Yorath C J Anderson R G amp Dom K(1986) Geological maps of southern Vancouver IslandLITHOPROBE I 92C10 11 14 16 92F1 2 7 8 Geological Survey ofCanada Open File 1272
Tejada M L G Mahoney J J Castillo P R Ingle S P Sheth HC amp Weis D (2004) Pin-pricking the elephant evidence on theorigin of the Ontong Java Plateau from Pb^Sr^Hf^Nd isotopiccharacteristics of ODP Leg 192 basalts In Fitton J GMahoney J J Wallace P J amp Saunders A D (eds) Origin and
Evolution of the Ontong Java Plateau Geological Society London Special
Publications 229 133^150Thompson P M E Kempton P D amp Kerr A C (2008) Evaluation
of the effects of alteration and leaching on Sm^Nd and Lu^Hf sys-tematics in submarine mafic rocks Lithos 104(1^4) 164^176
Thompson P M E Kempton P D White R V Kerr A CTarney J Saunders A D Fitton J G amp McBirney A (2003)Hf-Nd isotope constraints on the origin of the CretaceousCaribbean plateau and its relationship to the Galapagos plumeEarth and Planetary Science Letters 217(1^2) 59^75
Vervoort J D Patchett P Blichert-Toft J amp Albare de F (1999)Relationships between Lu^Hf and Sm^Nd isotopic systems in theglobal sedimentary system Earth and Planetary Science Letters 16879^99
Walter M J (1998) Melting of garnet peridotite and the origin ofkomatiite and depleted lithosphere Journal of Petrology 39(1) 29^60
Weis D Kieffer B Maerschalk C Barling J de Jong JWilliams G A Hanano D Mattielli N Scoates J SGoolaerts A Friedman R A amp Mahoney J B (2006) High-pre-cision isotopic characterization of USGS reference materials byTIMS and MC-ICP-MS Geochemistry Geophysics Geosystems
7(Q08006) doi1010292006GC001283Weis D Kieffer B Hanano D Silva I N Barling J
Pretorius W Maerschalk C amp Mattielli N (2007) Hf isotopecompositions of US Geological Survey reference materialsGeochemistry Geophysics Geosystems 8(Q06006) doi1010292006GC001473
Wheeler J O amp McFeely P (1991) Tectonic assemblage map of theCanadian Cordillera and adjacent part of the United States ofAmerica Geological Survey of Canada Map 1712A
WhiteW M Albare de F amp Tecurren louk P (2000) High-precision analy-sis of Pb isotope ratios by multi-collector ICP-MS Chemical Geology167 257^270
Yorath C J Sutherland Brown A amp Massey N W D (1999)LITHOPROBE southern Vancouver Island British Columbia Geological
Survey of Canada Bulletin 498 145 ppZou H (1998) Trace element fractionation during modal and
non-model dynamic melting and open-system melting Amathematical treatment Geochimica et Cosmochimica Acta 62(11)1937^1945
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
37
Zou H amp Reid M R (2001) Quantitative modeling of trace elementfractionation during incongruent dynamic melting Geochimica et
Cosmochimica Acta 65(1) 153^162
APPENDIX SAMPLEPREPARAT ION ANDANALYT ICAL METHODSOnly the freshest rocks in the field were sampled and onlythe least altered samples were selected for chemical andisotopic analysis based on thorough petrographic inspec-tion Fifty-nine of the 129 collected samples were crushed(400 g) into small pieces 52mm in diameter in aRocklabs hydraulic piston crusher between WC plates tominimize contamination The coarse-crush was mixedand 100 g was powdered in a planetary mill using agatejars and balls cleaned with quartz sand between samples
ActLabs analytical methodsThe major- and trace-element compositions of the whole-rock powders were determined at Activation LaboratoriesLtd (Actlabs) in Ancaster Ontario Analytical techniquesand detection limits are also available from Actlabs (httpwwwactlabscommethsub_code4erehtm) The analyticalmethod for each of the elements analyzed is indicated inTable 2 For the major elements a 02 g sample was mixedwith a lithium metaborate^lithium tetraborate mixtureand fused in a graphite crucible The molten mixture waspoured into a 5 HNO3 solution and shaken until dis-solved (30min) The samples were analyzed for majoroxides and selected trace elements by inductively coupledplasma optical emission spectrometry (ICP-OES) on acombination simultaneous^sequential Thermo Jarrell-AshEnviro II system Internal calibration was achieved usinga variety of international reference materials (eg W-2BIR-1 DNC-1) and independent control samplesAdditional trace elements were analyzed by both instru-mental neutron activation analysis (INAA) and induc-tively couple plasma mass spectrometry (ICP-MS) ForINAA15^25 g of sample was weighed into small polyeth-ylene vials and irradiated with a control international ref-erence material CANMET WMS-1 and NiCr flux wires ata thermal neutron flux of 71012 neutronscm2 per s in theMcMaster Nuclear Reactor Following a 7 day waitingperiod the samples were measured on an Ortec high-purity Ge detector linked to a Canberra Series 95 multi-channel analyzer Activities for each element weredecay- and weight-corrected and compared with a detectorcalibration developed from multiple international certifiedreference materials For ICP-MS 025 g of sample wasdigested in HF followed by HNO3^HClO4 heated andtaken to dryness The samples were brought back into solu-tion with HCl Samples were analyzed using a PerkinElmer Optima 3000 ICP In-lab standards or certified ref-erence materials (eg W-2 BIR-1 DNC-1) were used for
quality control A total of 15 blind duplicates were ana-lyzed from the same sample powders to assessreproducibility
University of Massachusetts analyticalmethodsFourteen sample duplicate powders taken from the same100 g sample aliquot (high-MgO basalts and picrites)were also analyzed at the Ronald B Gilmore X-RayFluorescence (XRF) Laboratory at the University ofMassachusetts Major elements were measured on a fusedLa-bearing lithium borate glass disc using a SiemensMRS-400 spectrometer Trace element concentrations(Rb Sr Ba Ce Nb Zr Y Pb Zn Ga Ni Cr V) weremeasured on a separate powder pellet using a PhilipsPW2400 sequential spectrometer using a Rh tube LOIand ferrous iron measurements were made as describedby Rhodes amp Vollinger (2004) Precision and accuracy esti-mates for the data have been described by Rhodes (1996)and Rhodes amp Vollinger (2004) Results for each sampleare the average of two separate analyses and are shown inElectronic Appendix 2 for a comparison of analyses ofduplicate powders using XRFand ICP-MS
PCIGR analytical methodsA subset of 19 samples was selected for high-precisiontrace-element analysis and Sr Nd Pb and Hf isotopicanalysis at the Pacific Centre for Isotopic andGeochemical Research (PCIGR) at the University ofBritish Columbia (UBC) Samples were selected from the59 samples analyzed for whole-rock chemistry at ActLabsbased on major- and trace-element chemistry alteration(low LOI and petrographic alteration index) and samplelocation Samples were prepared for trace-element analysisat the PCIGR by the technique described by Pretorius et al(2006) on unleached rock powders Sample powders(100mg) were weighed in 7ml screw-top Savillex bea-kers and dissolved in 1ml 14N HNO3 and 5ml 48 HFon a hotplate for 48 h at 1308C with periodic ultrasonica-tion Samples were dried and redissolved in 6ml 6N HClon a hotplate for 24 h and then dried and redissolved in1ml concentrated HNO3 for 24 h before final dryingTrace element abundances were measured by high-resolution inductively coupled plasma-mass spectrometry(HR-ICP-MS) with aThermo Finnigan Element2 systemfollowing the procedures described by Pretorius et al(2006) within 24 h of redissolution HFSE and LILE weremeasured in medium-resolution mode at 2000 dilutionusing a PFATeflon spray chamber washed with aqua regiafor 3min between samples REE were measured in high-resolution mode and U and Pb in low-resolution modeat 2000 dilution using a glass spray chamber washedwith 2 HNO3 between samples Total procedural blanksand reference materials (BCR-2 BHVO-2) were analyzedwith the batch of samples Indium was used as an internal
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
38
standard in all samples and standard solutionsBackground and standard solutions were analyzed afterevery five samples to detect memory effects and massdrift Results for PCIGR trace-element analyses areshown in Electronic Appendix 3 for the full set of analyzedtrace elementsSample digestion for purification of Sr Nd Hf and Pb
for column chemistry began by weighing each samplepowder (400^500mg) prior to leaching All samples wereleached with 6N HCl and placed in an ultrasonic bath for15min Samples were rinsed twice with 18 M-cm H2Obetween each leaching step (15 total) until the supernatantliquid was clear [following the method of Mahoney (1987)]Samples were then dried on a hotplate for 24 h andweighed again Sample solutions were then prepared bydissolving 100^250mg of the leached powder dissolvedin 1ml 14N HNO3 and 10ml 48 HF on a hotplate for48 h at 1308C with periodic ultrasonication Samples weredried and redissolved in 6ml 6N HCl on a hotplate for24 h and then dried Pb was separated using anion exhangecolumns and the discard was used for Sr REE and Hfseparation Nd was separated from the other REE and Hfrequired two additional purification steps Detailed proce-dures for column chemistry for separating Sr Nd and Pbat the PCIGR have been described byWeis et al (2006) andHf purification has been described byWeis et al (2007) Srand Nd isotope ratios were measured by thermal ioniza-tion mass spectrometry (TIMS) on a Thermo FinniganTriton system in static mode with relay matrix rotation ona single Ta and double Re^Ta filament respectively Fourto five filaments per barrel of 21 were occupied by stan-dards (NBS 987 for Sr and LaJolla for Nd) for each barrelwhere samples were run Sample Sr and Nd isotopic com-positions were corrected for mass fractionation using86Sr88Srfrac14 01194 and 146Nd144Ndfrac14 07219 Each samplewas then normalized using the barrel average of the refer-ence material relative to the values of143Nd144Ndfrac14 0511858 and 87Sr86Srfrac14 0710248 (Weiset al 2006) During the course of the Vancouver Islandanalyses the La Jolla Nd standard gave an average valueof 05118566 (nfrac14 7) and NBS987 standard gave an aver-age of 07102408 (nfrac1411) (2 error is reported as times106) 147Sm144Nd errors are 15 or 0006 USGeological Survey (USGS) reference material BHVO-2was processed with the samples and yielded Sr and Nd iso-topic ratios of 07034607 and 05129786 respectivelyThese are in agreement with the published values of070347920 and 051298411 respectively (Weis et al2006)
Pb and Hf isotopic compositions were analyzed by mul-tiple collector-inductively coupled plasma-mass spectrome-try (MC-ICP-MS) on a static multi-collection on a NuPlasma (Nu Instruments) system The detailed analyticalprocedure for Pb isotopic analyses on the Nu Plasma atthe PCIGR has been described by Weis et al (2006) Theconfiguration for Pb analyses allows for collection of PbTl and Hg togetherTl and Hg are used to monitor instru-mental mass discrimination and isobaric overlap repec-tively All sample solutions were analyzed withapproximately the same PbTl ratio (4) as the referencematerial NIST SRM 981 To accomplish this a small ali-quot of each sample solution from the Pb columns was ana-lyzed on the Element2 to determine the precise amount ofPb available for analysis on the Nu Plasma The SRM 981standard was run after every two samples on the NuPlasma During the time samples were run analyses ofthe SRM 981 Pb reference material gave values of206Pb204Pbfrac1416940322 207Pb204Pbfrac1415495823and 208Pb204Pbfrac14 36713164 (nfrac14 61 2 error is reportedas times 104) these values are in excellent agreement withreported TIMS triple-spike values of Galer amp Abouchami(1998) Fractionation-corrected Pb isotopic ratios were fur-ther corrected by the sample-standard bracketing methodor the ln^ln correction method described by White et al(2000) and Blichert-Toft et al (2003) Leached powders ofUSGS reference material BHVO-2 yielded Pb isotopicratios of 206Pb204Pbfrac14186454 8 207Pb204Pbfrac14 1549105 and 208Pb204Pbfrac14 38222514 and BCR-2yielded 206Pb204Pbfrac141880466 207Pb204Pbfrac14 1562518 and 208Pb204Pbfrac14 3883496 (2 error isreported as times 104) These values are in agreement withthose for leached residues of BHVO-2 and BCR-2 fromWeis et al (2006)Hf isotopic compositions were analyzed following the
procedures detailed byWeis et al (2007) The configurationfor Hf analyses monitored Lu mass 175 andYb mass 172 toallow for interference correction to masses 174 and 176 Hfisotopic ratios were normalized internally for mass fractio-nation to a 179Hf177Hf ratio of 07325 using an exponentialcorrection Standards were run after every two samplesand sample results were normalized to the ratio of the in-run daily average and a 176Hf177Hf ratio for JMC-475 of0282160 During the course of analyses the Hf standardJMC-475 gave an average value 02821679 (nfrac14130)USGS reference materials BCR-2 and BHVO-2 were pro-cessed with the samples and yielded Hf isotopic ratios of02828675 and 02831005 respectively Publishedvalues for BCR-2 and BHVO-2 are 02828717 and0283104 8 respectively (Weis et al 2007)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
39
Fig 2 Geological map and stratigraphy of the Schoen Lake Provincial Park and Karmutsen Range areas (locations shown in Fig 1 northernVancouver Island) (a) Stratigraphic column depicts units exposed in the Schoen Lake area derived from Carlisle (1972) and fieldwork (b)Generalized geology for the Schoen Lake area with sample locations Map derived from Massey et al (2005a) The exposures in the SchoenLake area are the lower volcanic stratigraphy and base of the Karmutsen Formation (c) Stratigraphic column for geology in the Alice^Nimpkish Lake area derived from Nixon amp Orr (2007) (d) Geological map from mapping of Nixon et al (2006a 2008) Sample sites andlithologies are denoted in the legend The Keogh Lake picrites are exposed near Keogh Sara and Maynard lakes and areas to the east ofMaynard Lake (Nixon et al 2008)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
5
Fig 3 Photographs of picritic and tholeiitic pillowed and massive basalt flows from the Karmutsen Range (Alice and Nimpkish Lake area)and Holberg Inlet (HI) areas northernVancouver Island (locations shown in Fig 1) (a) Massive submarine flow (marked with arrows) drapedover high-MgO pillow basalt of similar composition (b) Picritic pillow basalts in cross-section near Maynard Lake (Note vesicular uppermargin to pillows) (c) Close-up photograph of undulating contact of massive submarine flow draped over high-MgO pillow basalt (markedwith arrows) from photograph (a) (area of photograph indicated with a white box) (d) Photograph of margin of pahoehoe lobe (sledgehammerfor scale) (e) Cross-section of a large picritic pillow lobe with infilling of quartz^carbonate in cooling^contraction cracks (sledgehammer forscale) (f) High-MgO pillow breccia stratigraphically between submarine and subaerial flows (arrow points to lens cap 7 cm diameter forscale)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
6
Plagioclase forms most of the phenocrysts and glomero-crysts and the groundmass is fine-grained comprising pla-gioclase microlites clinopyroxene granules small grains ofFe^Ti oxide devitrified glass and secondary minerals
there is no fresh glass preserved The picritic and high-MgO basaltic lavas are characterized by spherulitic plagi-oclase and clinopyroxene with abundant euhedral olivinephenocrysts (Fig 5 Table 1) Olivine is completely replaced
Fig 4 Photographs showing field relations from the Schoen Lake Provincial Park area northernVancouver Island (a) Sediment^sill complexat the base of the Karmutsen Formation on the north side of Mt Schoen 100m below the Daonella locality (see Fig 2 for location) (b)Interbedded mafic sills and deformed finely banded chert and shale with calcareous horizons from location between Mt Adam and MtSchoen
Fig 5 Photomicrographs of picritic pillow basalts Karmutsen Formation northernVancouver Island Scale bars represent 1mm (a) Picritewith abundant euhedral olivine pseudomorphs from Keogh Lake type locality (Fig 2) in cross-polarized light (XPL) (sample 4722A4 19^198wt MgO four analyses) Samples contain dense clusters (24^42 vol ) of olivine pseudomorphs (52mm) in a groundmass of curvedand branching sheaves of acicular plagioclase and intergrown with clinopyroxene and altered glass In many cases plagioclase nucleated on theedges of the olivine phenocrysts (b) Sheaves of intergrown plagioclase and clinopyroxene in aphyric picrite pillow lava from west of MaynardLake (Fig 2) in XPL (sample 4723A2108 wt MgO)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
7
Table 1 Summary of petrographic characteristics and phenocryst proportions of Karmutsen basalts on Vancouver Island
BC
Sample Area Flow Group Texture vol Ol Plag Cpx Ox Alteration Notes
4718A1 MA PIL THOL intersertal 20 5 3 few plag glcr52mm cpx51 mm
4718A2 MA PIL THOL intersertal glomero 15 1 plag glcr52mm very fresh
4718A5 MA FLO THOL intersertal glomero 10 2 mottled few plag glcr54 mm
4723A4 KR PIL PIC intergranular intersertal 0 3 swtl plag51mm no ol phenos
4723A13 KR PIL PIC spherulitic 24 3 swtl plag51 mm
(continued)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
8
by an assemblage of talc^tremolite^serpentine^opaqueoxides there is no fresh olivine present in the Karmutsenlavas onVancouver Island Plagioclase and clinopyroxenephenocrysts are absent in the high-MgO lavas Thecoarse-grained mafic rocks are characterized by sub-ophi-tic textures with an average grain size typically 41mmand are generally not glomeroporphyritic these rocks aremainly from the interiors of massive flows although somemay represent sillsSample preparation and analytical methods for major-
and trace-element and Sr^Nd^Hf^Pb isotopic composi-tions of the Karmutsen volcanic rocks are given in theAppendix
WHOLE -ROCK CHEMISTRYMajor- and trace-element compositionsThe most abundant type of volcanic rock in theKarmutsen Formation is tholeiitic basalt with a restrictedrange of major- and trace-element compositions The tho-leiitic basalts have similar compositions to the coarse-grained mafic rocks and both groups are distinct fromthe picrites and high-MgO basalts [Fig 6 picrites412wt MgO (LeBas 2000) high-MgO basalts8^12wt MgO] The tholeiitic basalts have lower MgO(57^77wt MgO) and higher TiO2 (14^22wt
TiO2) than the picrites (130^198wt MgO05^07wt TiO2) and high-MgO basalts (91^116wt MgO 05^08wt TiO2 Fig 6 Table 2) Almost allcompositions plot within the tholeiitic field in a total alka-lis vs silica plot although there has been substantial K lossin most samples from the Karmutsen Formation (seebelow) and the tholeiitic basalts generally have higherSiO2 Na2OthornK2O CaO and FeOT (total iron expressedas FeO) than the picrites The tholeiitic basalts also havenoticeably lower loss on ignition (LOI mean1713wt ) than the picrites (mean 54 08wt )and high-MgO basalts (mean 3815wt ) (Fig 6)reflecting the presence of abundant altered olivinephenocrysts in the latter groups Ni concentrations are sig-nificantly higher for the picrites (339^755 ppm) and high-MgO basalts (122^551ppm) than the tholeiitic basalts(58^125 ppm except for one anomalous sample) andcoarse-grained rocks (70^213 ppm) (Fig 6 Table 2) Ananomalous tholeiitic pillowed flow (two samples outlierin Tables 1 and 2) from near the picrite type locality atKeogh Lake and a mineralized sill (disseminated sulfide)from the basal sediment^sill complex at Schoen Lake aredistinguished by higher TiO2 and FeOT than the maingroup of tholeiitic basalts The tholeiitic basalts are similarin major-element composition to previously publishedresults for samples from the Karmutsen Formation how-ever the Keogh Lake picrites which encompass the
Table 1 Continued
Sample Area Flow Group Texture vol Ol Plag Cpx Ox Alteration Notes
4722A5 KR FLO OUTLIER intersertal 3 mottled very fg v small pl needles
5615A11 KR PIL OUTLIER intersertal 3 mottled very fg v small pl needles
5617A4 SL SIL MIN intersertal 5 3 fg
Sample number last digit of year month day initial sample station (except 93G171) MA Mount Arrowsmith SLSchoen Lake KR Karmutsen Range PIL pillow BRE breccia FLO flow SIL sill GAB gabbro THOL tholeiiticbasalt PIC picrite HI-MG high-MgO basalt CGcoarse-grained MIN mineralized sill glomero glomeroporphyriticOlivine modes for picrites are based on area calculations from scans (see lsquoOlivine accumulation in high-MgO and picriticlavasrsquo section and Fig 11) Visual alteration index is based primarily on degree of plagioclase alteration and presence ofsecondary minerals (1 least altered 3 most altered) Plagioclase phenocrysts are commonly altered to albite pumpellyiteand chlorite olivine is altered to talc tremolite and clinochlore (determined using the Rietveld method of X-ray powderdiffraction) clinopyroxene is unaltered FendashTi oxide is commonly replaced by sphene glcr glomerocrysts fg fine-grained cg coarse-grained oik oikocryst chad chadacryst enc enclosing swtl swallow-tail ol olivinepseudomorphs plag plagioclase cpx clinopyroxene ox oxides (includes ilmenite thorn titanomagnetite)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
9
00
05
10
15
20
25
30
35
40
0 5 10 15 20 25
10
11
12
13
14
15
16
17
18
19
20
0 5 10 15 20 25
6
7
8
9
10
11
12
13
14
15
0 5 10 15 20 25
6
7
8
9
10
11
12
13
14
15
16
0 5 10 15 20 25
0
1
2
3
4
5
6
40 45 50 55 60
0
200
400
600
800
0 5 10 15 20 25
(wt)(wt)
(wt)
(wt)
+K2ONa2O TiO2
FeO (T)
Ni (ppm)
Al2O3
CaO
alkalic
tholeiitic
Karmutsen Form
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
(compiled data)
Pillowed flowMineralized sill
SiO2 (wt) MgO (wt)
MgO (wt)MgO (wt)
MgO (wt) MgO (wt)
(a) (b)
(g)
(e)
(f)
(c)
LOI (wt )
(d)
MgO (wt )0
1
2
3
4
5
6
7
0 5 10 15 20 25
(wt)
Fig 6 Whole-rock major-element Ni and LOI variation diagrams for the Karmutsen Formation New samples from this study (Table 2) areshown by symbols with black outlines (see legend) Previously published analyses are shown by small gray symbols without black outlines Theboundary of the alkaline and tholeiitic fields is that of MacDonald amp Katsura (1964) Total iron expressed as FeO LOI loss on ignition oxidesare plotted on an anhydrous normalized basis References for the 322 compiled analyses for the Karmutsen Formation are Surdam (1967)Kuniyoshi (1972) Muller et al (1974) Barker et al (1989) Lassiter et al (1995) Massey (1995a1995b)Yorath et al (1999) and G Nixon (unpublisheddata) It should be noted that the compiled dataset has not been filtered many of the samples with high SiO2 (452wt ) are probably notKarmutsen flood basalts but younger Bonanza basalts and andesites Three pillowed flow samples and a mineralized sill are distinguishedseparately because of their anomalous chemistry Coarse-grained samples are indicated separately Three tholeiitic basalt samples (4724A54718A5 4718A6) with416wt Al2O3 and4270 ppm Sr contain 10^25 modal of large plagioclase phenocrysts (7^8mm) or glomerocrysts(44mm)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
10
Table 2 Major element (wt oxide) and trace element (ppm) abundances in whole-rock samples of Karmutsen basalts
1Ni concentrations for these samples were determined by XRFTHOL tholeiitic basalt PIC picrite HI-MG high MgO basalt CG coarse-grained (sill or gabbro) MIN SIL mineralizedsill OUTLIER anomalous pillowed flow in plots MA Mount Arrowsmith SL Schoen Lake KR Karmutsen Range QIQuadra Island Sample locations are given using the Universal Transverse Mercator (UTM) coordinate system (NAD83zones 9 and 10) Analyses were performed at Activation Laboratory (ActLabs) Fe2O3
is total iron expressed as Fe2O3LOI loss on ignition All major elements Sr V and Y were measured by ICP quadrupole OES on solutions of fusedsamples Cu Ni Pb and Zn were measured by total dilution ICP Cs Ga Ge Hf Nb Rb Ta Th U Zr and REE weremeasured by magnetic-sector ICP on solutions of fused samples Co Cr and Sc were measured by INAA Blanks arebelow detection limit See Electronic Appendices 2 and 3 for complete XRF data and PCIGR trace element datarespectively Major elements for sample 93G17 are normalized anhydrous Sample 5615A7 was analyzed by XRFSamples from Quadra Island are not shown in figures
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
17
picritic and high-MgO basalt pillow lavas extend to20wt MgO (Fig 6 and references listed in thecaption)The tholeiitic basalts from the Karmutsen Range dis-
play a tight range of parallel light rare earth element(LREE)-enriched REE patterns (LaYbCNfrac14 21^24mean 2202) whereas the picrites and high-MgObasalts are characterized by sub-parallel LREE-depleted
patterns (LaYbCNfrac14 03^07 mean 06 02) with lowerREE abundances than the tholeiitic basalts (Fig 7)Tholeiitic basalts from the Schoen Lake and MountArrowsmith areas have similar REE patterns tosamples from the Karmutsen Range (Schoen Lake LaYbCNfrac14 21^26 mean 2303 Mount Arrowsmith LaYbCNfrac1419^25 mean 2204) and the coarse-grainedmafic rocks have similar REE patterns to the tholeiitic
Depleted MORB average(Salters amp Stracke 2004)
Schoen LakeSchoen Lake
Mount Arrowsmith
Karmutsen RangeS
ampl
eC
hond
rite
Sam
ple
Cho
ndrit
e
Sam
ple
Prim
itive
Man
tleS
ampl
eP
rimiti
ve M
antle
Sam
ple
Prim
itive
Man
tle
Mount Arrowsmith
Karmutsen Range
(a) (b)
(c) (d)
(e) (f )
Sam
ple
Cho
ndrit
e
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained mafic rock
Pillowed flowMineralized sill
1
10
100
1
10
100
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
01
1
10
100
1
10
100
1
10
100
Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
Fig 7 Whole-rock REE and other incompatible-element concentrations for the Karmutsen Formation (a) (c) and (e) are chondrite-normal-ized REE patterns for samples from each of the three main field areas onVancouver Island (b) (d) and (f) are primitive mantle-normalizedtrace-element patterns for samples from each of the three main field areas All normalization values are from McDonough amp Sun (1995) Theclear distinction between LREE-enriched tholeiitic basalts and LREE-depleted picrites the effects of alteration on the LILE (especially loss ofK and Rb) and the different range for the vertical scale in (b) (extends down to 01) should be noted
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
18
basalts (LaYbCNfrac14 21^29 mean 2508) Two LREE-depleted coarse-grained mafic rocks from the SchoenLake area have similar REE patterns (LaYbCNfrac14 07) tothe picrites from the Keogh Lake area indicating that thisdistinctive suite of rocks is exposed as far south asSchoen Lake that is 75 km away (Fig 7) The anomalouspillowed flow from the Karmutsen Range has lower LaYbCN (13^18) than the main group of tholeiitic basaltsfrom the Karmutsen Range (Fig 7) The mineralized sillfrom the Schoen Lake area is distinctly LREE-enriched(LaYbCNfrac14 33) with the highest REE abundances(LaCNfrac14 940) of all Karmutsen samples this may reflectcontamination by adjacent sediment also evident in thin-section where sulfide blebs are cored by quartz (Greeneet al 2006)The primitive mantle-normalized trace-element pat-
terns (Fig 7) and trace-element variations and ratios(Fig 8) highlight the differences in trace-element
concentrations between the four main groups ofKarmutsen flood basalts onVancouver Island The tholeii-tic basalts from all three areas have relatively smooth par-allel trace-element patterns some samples have smallpositive Sr anomalies relative to Nd and Sm and manysamples have prominent negative K anomalies relative toU and Nb reflecting K loss during alteration (Fig 7) Thelarge ion lithophile elements (LILE) in the tholeiiticbasalts (mainly Rb and K not Cs) are depleted relativeto the high field strength elements (HFSE) and LREELILE segments especially for the Schoen Lake area(Fig 7d) are remarkably parallel The picrites andhigh-MgO basalts form a tight band of parallel trace-element patterns and are depleted in HFSE and LREEwith positive Sr anomalies and relatively low Rb Thecoarse-grained mafic rocks from the Karmutsen Rangehave identical trace-element patterns to the tholeiiticbasalts Samples from the Schoen Lake area have similar
000
025
050
075
100
125
0 1 2 3 4 5
00
04
08
12
16
0 5 10 15 20 25
MgO (wt)
0
5
10
15
0 50 100 150
Hf (ppm)
Th (ppm)
NbLaNb (ppm)
Zr (ppm)
(a)
(c)
(b)
(d)
Th (ppm)
00
02
04
06
08
10
12
00 01 02 03 04 05
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
NbLa=1
4723A4 5615A12
U (ppm)
Fig 8 Whole-rock trace-element concentrations and ratios for the Karmutsen Formation [except (b) which is vs wt MgO] (a) Nb vs Zr(b) NbLa vs MgO (c) Th vs Hf (d) Th vs U There is a clear distinction between the tholeiitic basalts and picrites in Nb and Zr both inconcentration and the slope of each trend
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
19
trace-element patterns to the Karmutsen Range except forthe two LREE-depleted coarse-grained mafic rocks andthe mineralized sill (Fig 7d)
Sr^Nd^Hf^Pb isotopic compositionsA total of 19 samples from the four main groups ofthe Karmutsen Formation were selected for radiogenicisotopic geochemistry on the basis of major- and trace-element variation stratigraphic position and geographicallocation The samples have overlapping age-correctedHf and Nd isotope ratios and distinct ranges of Sr isotopiccompositions (Fig 9) The tholeiitic basalts picriteshigh-MgO basalt and coarse-grained mafic rocks have
initial eHffrac14thorn87 to thorn126 and eNdfrac14thorn66 to thorn88 cor-rected for in situ radioactive decay since 230 Ma (Fig 9Tables 3 and 4) All the Karmutsen samples form a well-defined linear array in a Lu^Hf isochron diagram corre-sponding to an age of 24115 Ma (Fig 9) This is withinerror of the accepted age of the Karmutsen Formationand indicates that the Hf isotope systematics behaved as aclosed system since c 230 Ma The anomalous pillowedflow has similar initial eHf (thorn98) and slightly lower initialeNd (thorn62) compared with the other Karmutsen samplesThe picrites and high-MgO basalt have higher initial87Sr86Sr (070398^070518) than the tholeiitic basalts(initial 87Sr86Srfrac14 070306^070381) and the coarse-grained mafic rocks (initial 87Sr86Srfrac14 070265^070428)
05128
05129
05130
05131
05132
015 017 019 021 023 025
02829
02830
02831
02832
02833
02834
000 002 004 006 008 010
4
6
8
10
0702 0703 0704 0705 07066
8
10
12
14
5 6 7 8 9 10
12
0
1
2
3
4
5
6
7
0702 0703 0704 0705 0706
(d)
LOI (wt )
87Sr 86Sr
4723A2
Nd
143Nd144Nd
147Sm144Nd
87Sr 86Sr (230 Ma)
(230 Ma)
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
(a)
(c)
176Hf177Hf
Age=241 plusmn 15 Ma=+990 plusmn 064
176Lu177Hf
Hf (230 Ma)
Nd (230 Ma)
(e)
Hf (i)
(b)
Fig 9 Whole-rock Sr Nd and Hf isotopic compositions for the Karmutsen Formation (a) 143Nd144Nd vs 147Sm144Nd (b) 176Hf177Hf vs176Lu177Hf The slope of the best-fit line for all samples corresponds to an age of 24115 Ma (c) Initial eNd vs 87Sr86Sr Age-correction to230 Ma (d) LOI vs 87Sr86Sr (e) Initial eHf vs eNd Average 2 error bars are shown in a corner of each panel Complete chemical duplicatesshown inTables 3 and 4 (samples 4720A7 and 4722A4) are circled in each plot
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
overlap the ranges of the picrites and tholeiitic basalts(Fig 9 Table 3)The measured Pb isotopic compositions of the tholeiitic
basalts are more radiogenic than those of the picrites andthe most magnesian Keogh Lake picrites have the leastradiogenic Pb isotopic compositions The range ofinitial Pb isotope ratios for the picrites is206Pb204Pbfrac1417868^18580 207Pb204Pbfrac1415547^15562and 208Pb204Pbfrac14 37873^38257 and the range for thetholeiitic basalts is 206Pb204Pbfrac1418782^19098207Pb204Pbfrac1415570^15584 and 208Pb204Pbfrac14 37312^38587 (Fig 10 Table 5) The coarse-grained mafic rocksoverlap the range of initial Pb isotopic compositions forthe tholeiitic basalts with 206Pb204Pbfrac1418652^19155207Pb204Pbfrac1415568^15588 and 208Pb204Pbfrac14 38153^38735 One high-MgO basalt has the highest initial206Pb204Pb (19242) and 207Pb204Pb (15590) and thelowest initial 208Pb204Pb (37676) The Pb isotopic ratiosfor Karmutsen samples define broadly linearrelationships in Pb isotope plots The age-corrected Pb
isotopic compositions of several picrites have been affectedby U and Pb (andor Th) mobility during alteration(Fig 10)
ALTERAT IONThe Karmutsen basalts have retained most of their origi-nal igneous structures and textures however secondaryalteration and low-grade metamorphism have producedzeolitic- and prehnite^pumpellyite-bearing mineral assem-blages (Table 1 Cho et al 1987) Alteration and metamor-phism have primarily affected the distribution of the LILEand the Sr isotopic systematics The Keogh Lake picriteshave been affected more by alteration than the tholeiiticbasalts and have higher LOI (up to 55wt ) variableLILE and higher measured Sr Nd and Hf and lowermeasured Pb isotope ratios than the tholeiitic basalts Srand Pb isotope compositions for picrites and tholeiiticbasalts show a relationship with LOI (Figs 9 and 10)whereas Nd and Hf isotopic compositions do not correlate
Table 3 Sr and Nd isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Rb Sr 87Sr86Sr 2m87Rb86Sr 87Sr86Srt Sm Nd 143Nd144Nd 2m eNd
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses carried out at the PCIGR thecomplete trace element analyses are given in Electronic Appendix 3
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
21
with LOI (not shown) The relatively high apparent initialSr isotopic compositions for the high-MgO lavas (up to07052) probably resulted from post-magmatic loss of Rbor an increase in 87Sr86Sr through addition of seawater Sr(eg Hauff et al 2003) High-MgO lavas from theCaribbean plateau have similarly high 87Sr86Sr comparedwith basalts (Recurren villon et al 2002)The correction for in situdecay on initial Pb isotopic ratios has been affected bymobilization of U and Th (and Pb) in the whole-rockssince their formation A thorough acid leaching duringsample preparation was used that has been shown to effec-tively remove alteration phases (Weis et al 2006 NobreSilva et al in preparation) Whole-rock SmNd ratiosand age-correction of Nd isotopes may be affected bythe leaching procedure for submarine rocks older than50 Ma (Thompson et al 2008 Fig 9) there ishowever very little variation in Nd isotopic compositionsof the picrites or tholeiitic basalts and this system does notappear to be disturbed by alteration The HFSE abun-dances for tholeiitic and picritic basalts exhibit clear
linear correlations in binary diagrams (Fig 8) whereasplots of LILE vs HFSE are highly scattered (not shown)because of the mobility of some of the alkali and alkalineearth LILE (Cs Rb Ba K) and Sr for most samplesduring alterationThe degree of alteration in the Karmutsen samples does
not appear to be related to eruption environment or depthof burial Pillow rims are compositionally different frompillow cores (Surdam1967 Kuniyoshi1972) and aquagenetuffs are chemically different from isolated pillows withinpillow breccias however submarine and subaerial basaltsexhibit a similar degree of alterationThere is no clear cor-relation between the submarine and subaerial basalts andsome commonly used chemical alteration indices (eg BaRb vs K2OP2O5 Huang amp Frey 2005) There is also nodefinitive correlation between the petrographic alterationindex (Table 1) and chemical alteration indices although10 of 13 tholeiitic basalts with the highest K and LILEabundances have the highest petrographic alterationindex of three (seeTable 1)
Table 4 Hf isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Lu Hf 176Hf177Hf 2m eHf176Lu177Hf 176Hf177Hft eHf(t)
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
OLIV INE ACCUMULAT ION INH IGH-MgO AND PICR IT IC LAVASGeochemical trends and petrographic characteristics indi-cate that accumulation of olivine played an important rolein the formation of the high-MgO basalts and picrites fromthe Keogh Lake area on northern Vancouver Island TheKeogh Lake samples show a strong linear correlation inplots of Al2O3 TiO2 ScYb and Ni vs MgO and many ofthe picrites have abundant clusters of olivine pseudomorphs(Table1 Fig11)There is a strong linear correlationbetween
modal per cent olivine and whole-rock magnesium contentmagnesium contents range between 10 and 19wt MgOand the proportion of olivine phenocrysts in these samplesvaries from 0 to 42 vol (Fig 11)With a few exceptionsboth the size range and median size of the olivine pheno-crysts in most samples are comparable The clear correla-tion between the proportion of olivine phenocrysts andwhole-rock magnesium contents indicates that accumula-tion of olivine was directly responsible for the high MgOcontents (410wt ) in most of the picritic lavas
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
207Pb204Pb
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
208Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
(a) (b)
(c) (e)
4723A24723A2
4723A2
4723A2
4722A4
4723A4
4722A4
4723A4
4723A134723A3
4723A3
4723A13
1556
1558
1560
1562
1564
1566
1568
186 190 194 198 202 206 2101550
1552
1554
1556
1558
1560
178 180 182 184 186 188 190 192 194
380
384
388
392
396
186 190 194 198 202 206 210374
378
382
386
390
178 180 182 184 186 188 190 192 194
206Pb204Pb(d)
LOI (wt )
0
1
2
3
4
5
6
7
186 190 194 198 202 206 210
4723A2
Fig 10 Pb isotopic compositions of leached whole-rock samples measured by MC-ICP-MS for the Karmutsen Formation Error bars are smal-ler than symbols (a) Measured 207Pb204Pb vs 206Pb204Pb (b) Initial 207Pb204Pb vs 206Pb204Pb Age-correction to 230 Ma (c) Measured208Pb204Pb vs 206Pb204Pb (d) LOI vs 206Pb204Pb (e) Initial 208Pb204Pb vs 206Pb204Pb Complete chemical duplicates shown in Table 5(samples 4720A7 and 4722A4) are circled in each plot The dashed lines in (b) and (e) show the differences in age-corrections for two picrites(4723A4 4722A4) using the measured UTh and Pb concentrations for each sample and the age-corrections when concentrations are used fromthe two picrites (4723A3 4723A13) that appear to be least affected by alteration
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
23
DISCUSS IONThe compositional and spatial development of the eruptedlava sequences of the Wrangellia oceanic plateau onVancouver Island provide information about the evolutionof a mantle plume upwelling beneath oceanic lithospherein an analogous way to the interpretation of continentalflood basalts The Caribbean and Ontong Java plateausare the two primary examples of accreted parts of oceanicplateaus that have been studied to date and comparisonscan be made with the Wrangellia oceanic plateauHowever the Wrangellia oceanic plateau is a distinctiveoccurrence of an oceanic plateau because it formed on thecrust of a Devonian to Mississippian intra-oceanic islandarc overlain by Mississippian to Permian limestones andpelagic sediments The nature and thickness of oceaniclithospheric mantle are considered to play a fundamentalrole in the decompression and evolution of a mantleplume and thus the compositions of the erupted lavasequences (Greene et al 2008) The geochemical and
stratigraphic relationships observed in the KarmutsenFormation onVancouver Island and the focus of the subse-quent discussion sections provide constraints on (1) thetemperature and degree of melting required to generatethe primary picritic magmas (2) the composition of themantle source (3) the depth of melting and residual min-eralogy in the source region (4) the low-pressure evolutionof the Karmutsen tholeiitic basalts (5) the growth of theWrangellia oceanic plateau
Melting conditions and major-elementcomposition of the primary magmasThe primary magmas to most flood basalt provinces arebelieved to be picritic (eg Cox 1980) and they have beenfound in many continental and oceanic flood basaltsequences worldwide (eg Siberia Karoo Paranacurren ^Etendeka Caribbean Deccan Saunders 2005) Near-pri-mary picritic lavas with low total alkali abundances
Table 5 Pb isotopic compositions of Karmutsen basaltsVancouver Island BC
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
require high-degree partial melting of the mantle source(eg Herzberg amp OrsquoHara 2002) The Keogh Lake picritesfrom northernVancouver Island are the best candidates foridentifying the least modified partial melts of the mantleplume source despite having accumulated olivine pheno-crysts and can be used to estimate conditions of meltingand the composition of the primary magmas for theKarmutsen FormationPrimary magma compositions mantle potential tem-
peratures and source melt fractions for the picritic lavas ofthe Karmutsen Formation were calculated from primitivewhole-rock compositions using PRIMELT1XLS software(Herzberg et al 2007) and are presented in Fig 12 andTable 6 A detailed discussion of the computationalmethod has been given by Herzberg amp OrsquoHara (2002)and Herzberg et al (2007) key features of the approachare outlined belowPrimary magma composition is calculated for a primi-
tive lava by incremental addition of olivine and compari-son with primary melts of mantle peridotite Lavas thathad experienced plagioclase andor clinopyroxene fractio-nation are excluded from this analysis All calculated pri-mary magma compositions are assumed to be derived byaccumulated fractional melting of fertile peridotite KR-4003 Uncertainties in fertile peridotite composition donot propagate to significant variations in melt fractionand mantle potential temperature melting of depletedperidotite propagates to calculated melt fractions that aretoo high but with a negligible error in mantle potential
temperature PRIMELT1XLS calculates the olivine liqui-dus temperature at 1atm which should be similar to theactual eruption temperature of the primary magmaand is accurate to 308C at the 2 level of confidenceMantle potential temperature is model dependent witha precision that is similar to the accuracy of the olivineliquidus temperature (C Herzberg personal communica-tion 2008)With regard to the evidence for accumulated olivine in
the Keogh Lake picrites it should not matter whether themodeled lava composition is aphyric or laden with accu-mulated olivine phenocrysts PRIMELT1 computes theprimary magma composition by adding olivine to a lavathat has experienced olivine fractionation in this way itinverts the effects of fractional crystallization The onlyrequirement is that the olivine phenocrysts are not acci-dental or exotic to the lava compositions Support for thisprocedure can be seen in the calculated olivine liquid-line-of-descent shown by the curved arrays with 5 olivineaddition increments (Fig 12c) Fractional crystallization ofolivine from model primary magmas having 85^95wt FeO and 153^175wt MgO will yield arrays of deriva-tive liquids and randomly sorted olivines that in most butnot all cases connect the picrites and high-MgO basalts Anewer version of the PRIMELT software (Herzberg ampAsimow 2008) can perform olivine addition to and sub-traction from lavas with highly variably olivine phenocrystcontents and yields results that converge with those shownin Fig 12
Fig 11 Relationship between abundance of olivine phenocrysts and whole-rock MgO contents for the Keogh Lake picrites (a) Tracings ofolivine grains (gray) in six high-MgO samples made from high resolution (2400 dpi) scans of petrographic thin-sections Scale bars represent10mm sample numbers are indicated at the upper left (b) Modal olivine vs whole-rock MgO Continuous line is the best-fit line for 10 high-MgO samplesThe areal proportion of olivine was calculated from raster images of olivine grains using ImageJ image analysis software whichprovides an acceptable estimation of the modal abundance (eg Chayes 1954)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
25
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
Gray lines = final melting pressure
L + Ol + Opx
07
08
09
Pressure (GPa)
10
3
4
5
6
1
SOLIDUS
01
04
0506
L + Ol
2
03
02
PeridotiteKR-4003
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
Thick black lines = initial melting pressure
Dashed lines = melt fraction
Melt Fraction
2
3 4 5 6
L+Ol+Opx
SolidusSpinel
SolidusGarnet Peridotite
7
L+Ol
0 10 20 30 4040
45
50
55
60
MgO (wt)
SiO2
(wt)
PeridotiteKR-4003
Melt Fraction
0 10 20 305
10
15
CaO(wt)
MgO (wt)
3
4 5 6 7
2
46
2
Peridotite Partial Melts
Thick black line = initial melting pressureGray lines = final melting pressure
Peridotite
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
02
Pressure (GPa)
0302
04
Pressure (GPa)
Picrite
High-MgO basalt
Tholeiitic basalt
(c)
(d)
Karmutsen flood basaltsOlivine addition modelOlivine addition (5 increment)Primary magma
PicriteHigh-MgO basaltTholeiitic basalt
(a)
(b)
800 850 900
910
920
930
940
950
Karmutsen flood basalts
Olivine addition modelOlivine addition (5 increment)Primary magma
Gray lines = Fo contents of olivine
Fig 12 Estimated primary magma compositions for three Keogh Lake picrites and high-MgO basalts (samples 4723A4 4723A135616A7) using the forward and inverse modeling technique of Herzberg et al (2007) Karmutsen compositions and modeling results areoverlain on diagrams provided by C Herzberg (a) Whole-rock FeO vs MgO for Karmutsen samples from this study Total iron estimatedto be FeO is 090 Gray lines show olivine compositions that would precipitate from liquid of a given MgO^FeO composition Black lineswith crosses show results of olivine addition (inverse model) using PRIMELT1 (Herzberg et al 2007) Gray field highlights olivinecompositions with Fo contents of 91^92 predicted to precipitate (b) FeO vs MgO showing Karmutsen lava compositions and results ofa forward model for accumulated fractional melting of fertile peridotite KR-4003 (Kettle River Peridotite Walter 1998) (c) SiO2 vsMgO for Karmtusen lavas and model results (d) CaO vs MgO showing Karmtusen lava compositions and model results To brieflysummarize the technique [see Herzberg et al (2007) for complete description] potential parental magma compositions for the high-MgOlava series were selected (highest MgO and appropriate CaO) and using PRIMELT1 software olivine was incrementally added tothe selected compositions to show an array of potential primary magma compositions (inverse model) Then using PRIMELT1 theresults from the inverse model were compared with a range of accumulated fractional melts for fertile peridotite derived fromparameterization of the experimental results for KR-4003 from Walter (1998) forward model Herzberg amp OrsquoHara (2002) A melt fractionwas sought that was unique to both the inverse and forward models (Herzberg et al 2007) A unique solution was found when there was a com-mon melt fraction for both models in FeO^MgO and CaO^MgO^Al2O3^SiO2 (CMAS) projection space This modeling assumes thatolivine was the only phase crystallizing and ignores chromite precipitation and possible augite fractionation in the mantle (Herzbergamp OrsquoHara 2002) Results are best for a residue of spinel lherzolite (not pyroxenite) The presence of accumulated olivine in samples ofthe Keogh Lake picrites used as starting compositions does not significantly affect the results because we are modeling addition of olivine(see text for further description) The tholeiitic basalts cannot be used for modeling because they are all saturated in plagthorn cpxthornolThick black line for initial melting pressure at 4 GPa is not shown in (c) for clarity Gray area in (a) indicates parental magmas thatwill crystallize olivine with Fo 91^92 at the surface Dashed lines in (c) indicate melt fraction Solidi for spinel and garnet peridotite arelabelled in (c)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
The PRIMELT1 results indicate that if the Karmutsenpicritic lavas were derived from accumulated fractionalmelting of fertile peridotite the primary magmas wouldhave contained 15^17wt MgO and 10wt CaO(Fig 12 Table 6) formed from 23^27 partial meltingand would have crystallized olivine with a Fo content of 91 (Fig 12 Table 6) Calculated mantle temperatures forthe Karmutsen picrites (14908C) indicate melting ofanomalously hot mantle (100^2008C above ambient) com-pared with ambient mantle that produces mid-ocean ridgebasalt (MORB 1280^14008C Herzberg et al 2007)which initiated between 3 and 4 GPa Several picritesappear to be near-primary melts and required very littleaddition of olivine (eg sample 93G171 with measured158wt MgO) These temperature and melting esti-mates of the Karmutsen picrites are consistent withdecompression of hot mantle peridotite in an actively con-vecting plume head (ie plume initiation model) Threesamples (picrite 5614A1 and high-MgO basalts 5614A3and 5614A5) are lower in iron than the other Karmutsenlavas with FeO in the 65^80wt range (Fig 12c) andhave MgO and FeO contents that are similar to primitiveMORB from the Sequeiros fracture zone of the EastPacific Rise (EPR Perfit et al 1996) These samples
however differ from EPR MORB in having higher SiO2
and lower Al2O3 and they are restricted geographicallyto one area (Sara Lake Fig 2)We therefore have evidence for primary magmas with
MgO contents that range from a possible low of 110wt to as high as 174wt with inferred mantle potential tem-peratures from 1340 to 15208C This is similar to the rangedisplayed by lavas from the Ontong Java Plateau deter-mined by Herzberg (2004) who found that primarymagma compositions assuming a peridotite sourcewould contain 17wt MgO and 10wt CaO(Table 6) and that these magmas would have formed by27 melting at a mantle potential temperature of15008C (first melting occurring at 36 GPa) Fitton ampGodard (2004) estimated 30 partial melting for theOntong Java lavas based on the Zr contents of primarymagmas of Kroenke-type basalt calculated by incrementaladdition of equilibrium olivine However if a considerableamount of eclogite was involved in the formation of theOntong Java plateau magmas the excess temperatures esti-mated by Herzberg et al (2007) may not be required(Korenaga 2005) The variability in mantle potential tem-perature for the Keogh Lake picrites is also similar to thewide range of temperatures that have been calculated for
Table 6 Estimated primary magma compositions for Karmutsen basalts and other oceanic plateaus or islands
Sample 4723A4 4723A13 5616A7 93G171 Average OJP Mauna Keay Gorgonaz
Ontong Java primary magma composition for accumulated fraction melting (AFM) from Herzberg (2004)yMauna Kea primary magma composition is average of four samples of Herzberg (2006 table 1)zGorgona primary magma composition for AFM for 1F fertile source from Herzberg amp OrsquoHara (2002 table 4)Total iron estimated to be FeO is 090
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
27
lavas from single volcanoes in the Galapagos Hawaii andelsewhere by Herzberg amp Asimow (2008) Those workersinterpreted the range to reflect the transport of magmasfrom both the cool periphery and the hotter axis of aplume A thermally heterogeneous plume will yield pri-mary magmas with different MgO contents and the rangeof primary magma compositions and their inferred mantlepotential temperatures for Karmutsen lavas may reflectmelting from different parts of the plume
Source of the Karmutsen Formation lavasThe Sr^Nd^Hf^Pb isotopic compositions of theKarmutsen volcanic rocks on Vancouver Island indicatethat they formed from plume-type mantle containing adepleted component that is compositionally similar to butdistinct from other oceanic plateaus that formed in thePacific Ocean (eg Ontong Java and Caribbean Fig 13)Trace element and isotopic constraints can be used to testwhether the depleted component of the KarmutsenFormation was intrinsic to the mantle plume and distinctfrom the source of MORB or the result of mixing or invol-vement of ambient depleted MORB mantle with plume-type mantle The Karmutsen tholeiitic basalts have initialeNd and
87Sr86Sr ratios that fall within the range of basaltsfrom the Caribbean Plateau (eg Kerr et al 1996 1997Hauff et al 2000 Kerr 2003) and a range of Hawaiian iso-tope data (eg Frey et al 2005) and lie within and abovethe range for the Ontong Java Plateau (Tejada et al 2004Fig 13a) Karmutsen tholeiitic basalts with the highest ini-tial eNd and lowest initial 87Sr86Sr lie within and justbelow the field for age-corrected northern EPR MORB at230 Ma (eg Niu et al1999 Regelous et al1999) Initial Hfand Nd isotopic compositions for the Karmutsen samplesare noticeably offset to the low side of the ocean islandbasalt (OIB) array (Vervoort et al 1999) and lie belowand slightly overlap the low eHf end of fields for Hawaiiand the Caribbean Plateau (Fig 13c) the Karmutsen sam-ples mostly fall between and slightly overlap a field forage-corrected Pacific MORB at 230 Ma and Ontong Java(Mahoney et al 1992 1994 Nowell et al 1998 Chauvel ampBlichert-Toft 2001) Karmutsen lava Pb isotope ratios over-lap the broad range for the Caribbean Plateau and thelinear trends for the Karmutsen samples intersect thefields for EPR MORB Hawaii and Ontong Java in208Pb^206Pb space but do not intersect these fields in207Pb^206Pb space (Fig 13d and e) The more radiogenic207Pb204Pb for a given 206Pb204Pb of the Karmutsenbasalts requires high UPb ratios at an ancient time when235U was abundant similar to the Caribbean Plateau andenriched in comparison with EPR MORB Hawaii andOntong JavaThe low eHf^low eNd component of the Karmutsen
basalts that places them below the OIB mantle array andpartly within the field for age-corrected Pacific MORBcan form from low ratios of LuHf and high SmNd over
time (eg Salters amp White 1998) MORB will develop aHf^Nd isotopic signature over time that falls below themantle array Low LuHf can also lead to low 176Hf177Hffrom remelting of residues produced by melting in theabsence of garnet (eg Salters amp Hart 1991 Salters ampWhite 1998) The Hf^Nd isotopic compositions of theKarmutsen Formation indicate the possible involvementof depleted MORB mantle however similar to Iceland(Fitton et al 2003) Nb^Zr^Y variations provide a way todistinguish between a depleted component within theplume and a MORB-like component The Karmutsenbasalts and picrites fall within the Iceland array and donot overlap the MORB field in a logarithmic plot of NbYvs ZrY (Fig 13b) using the fields established by Fittonet al (1997) Similar to the depleted component within theCaribbean Plateau (Thompson et al 2003) the trend ofHf^Nd isotopic compositions towards Pacific MORB-likecompositions is not followed by MORB values in Nb^Zr^Y systematics and this suggests that the depleted compo-nent in the source of the Karmutsen basalts is distinctfrom the source of MORB and probably an intrinsic partof the plume The relatively radiogenic Pb isotopic ratiosand REE patterns distinct from MORB also support thisinterpretationTrace-element characteristics suggest the possible invol-
vement of sub-arc lithospheric mantle in the generation ofthe Karmutsen picrites Lassiter et al (1995) suggested thatmixing of a plume-type source with eNdthorn 6 to thorn 7 witharc material with low NbTh could reproduce the varia-tions observed in the Karmutsen basalts however theyconcluded that the absence of low NbLa ratios in theirsample of tholeiitic basalts restricts the amount of arc litho-sphere involved The study of Lassiter et al (1995) wasbased on major and trace element and Sr Nd and Pb iso-topic data for a suite of 29 samples from Buttle Lake in theStrathcona Provincial Park on central Vancouver Island(Fig 1) Whereas there are no clear HFSE depletions inthe Karmutsen tholeiitic basalts the low NbLa (and TaLa) of the Karmutsen picritic lavas in this study indicatethe possible involvement of Paleozoic arc lithosphere onVancouver Island (Fig 8)
REE modeling dynamic melting andsource mineralogyThe combined isotopic and trace-element variations of thepicritic and tholeiitic Karmutsen lavas indicate that differ-ences in trace elements may have originated from differentmelting histories For example the LuHf ratios of theKarmutsen picritic lavas (mean 008) are considerablyhigher than those in the tholeiitic basalts (mean 002)however the small range of overlapping initial eHf valuesfor the two lava suites suggests that these differences devel-oped during melting within the plume (Fig 9b) Below wemodel the effects of source mineralogy and depth of
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
28
melting on differences in trace-element concentrations inthe tholeiitic basalts and picritesDynamic melting models simulate progressive decom-
pression melting where some of the melt fraction isretained in the residue and only when the degree of partialmelting is greater than the critical mass porosity and the
source becomes permeable is excess melt extracted fromthe residue (Zou1998) For the Karmutsen lavas evolutionof trace element concentrations was simulated using theincongruent dynamic melting model developed by Zou ampReid (2001) Melting of the mantle at pressures greater than1 GPa involves incongruent melting (eg Longhi 2002)
OIB array
PacificMORB(230 Ma)
(0 Ma)
EPR(230 Ma)
-4
-2
0
2
4
8
10
12
14
16
18
20
22
-4 -2 0 2 4 6 8 10 12 14
minus2
0
2
4
6
8
10
12
14
0702 0703 0704 0705 0706
IndianMORB
87Sr 86Sr (initial)
Hf (initial)
(a) (c)
Nd (initial)
Karmutsen tholeiitic basalt
HawaiiOntong JavaCaribbean Plateau
Karmutsen high-MgO lava
high-MgO lavasKarmutsen
1540
1545
1550
1555
1560
1565
175 180 185 190 195 200375
380
385
390
395
175 180 185 190 195 200
(d) (e)
EPR
EPR
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb
Karmutsen Formation (initial)
HawaiiOntong JavaCaribbean Plateau
East Pacific Rise
Karmutsen Formation (measured)
Juan de FucaGorda
ε Nd
(initial)
Karmutsen Formation (different age-correction for 3 picrites)
(0 Ma)
NbY
001
01
1
1 10
ZrY
OIB
NMORB
(b)
Iceland array
6
Fig 13 Comparison of age-corrected (230 Ma) Sr^Nd^Hf^Pb isotopic compositions for Karmutsen flood basalts onVancouver Island withage-corrected OIB and MORB and Nb^Zr^Y variation (a) Initial eNd vs 87Sr86Sr (b) NbYand ZrY variation (after Fitton et al 1997)(c) Initial eHf vs eNd (d) Measured and initial 207Pb204Pb vs 206Pb204Pb (e) Measured and initial 208Pb204Pb vs 206Pb204Pb References fordata sources are listed in Electronic Appendix 4 Most of the compiled data were extracted from the GEOROC database (httpgeorocmpch-mainzgwdgdegeoroc) OIB array line in (c) is fromVervoort et al (1999) EPR East Pacific Rise Several high-MgO samples affected bysecondary alteration were not plotted for clarity in Pb isotope plots Estimation of fields for age-corrected Pacific MORB and EPR at 230 Mawere made using parent^daughter concentrations (in ppm) from Salters amp Stracke (2004) of Lufrac14 0063 Hffrac14 0202 Smfrac14 027 Ndfrac14 071Rbfrac14 0086 and Srfrac14 836 In the NbYvs ZrYplot the normal (N)-MORB and OIB fields not including Icelandic OIB are from data com-pilations of Fitton et al (1997 2003) The OIB field is data from 780 basalts (MgO45wt ) from major ocean islands given by Fitton et al(2003) The N-MORB field is based on data from the Reykjanes Ridge EPR and Southwest Indian Ridge from Fitton et al (1997) The twoparallel gray lines in (b) (Iceland array) are the limits of data for Icelandic rocks
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
29
with olivine being produced during the reactioncpxthornopxthorn spfrac14meltthornol for spinel peridotites The mod-eling used coefficients for incongruent melting reactionsbased on experiments on lherzolite melting (eg Kinzleramp Grove 1992 Longhi 2002) and was separated into (1)melting of spinel lherzolite (2) two combined intervals ofmelting for a garnet lherzolite (gt lherz) and spinel lher-zolite (sp lherz) source (3) melting of garnet lherzoliteThe model solutions are non-unique and a range indegree of melting partition coefficients melt reactioncoefficients and proportion of mixed source componentsis possible to achieve acceptable solutions (see Fig 14 cap-tion for description of modeling parameters anduncertainties)The LREE-depleted high-MgO lavas require melting of
a depleted spinel lherzolite source (Fig 14a) The modelingresults indicate a high degree of melting (22^25) similarto the results from PRIMELT1 based on major-elementvariations (discussed above) of a LREE-depleted sourceThe high-MgO lavas lack a residual garnet signature Theenriched tholeiitic basalts involved melting of both garnetand spinel lherzolite and represent aggregate melts pro-duced from continuous melting throughout the depth ofthe melting column (Fig 14b) Trace-element concentra-tions in the tholeiitic and picritic lavas are not consistentwith low-degree melting of garnet lherzolite alone(Fig 14c) The modeling results indicate that the aggregatemelts that formed the tholeiitic basalts would haveinvolved lesser proportions of enriched small-degreemelts (1^6 melting) generated at depth and greateramounts of depleted high-degree melts (12^25) gener-ated at lower pressure (a ratio of garnet lherzolite tospinel lherzolite melt of 14 provides the best fits seeFig 14b and description in figure caption) The totaldegree of melting for the tholeiitic basalts was high(23^27) similar to the degree of partial melting in thespinel lherzolite facies for the primary melts that eruptedas the Keogh Lake picrites of a source that was also prob-ably depleted in LREE
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
(c)
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
Melting of garnet lherzolite
High-MgO lavas
Melting of spinel lherzolite
source (07DM+03PM)
source (07DM+03PM)
6 melting
30 melting
(b)
(a)
30 melting
Sam
ple
Cho
ndrit
eS
ampl
eC
hond
rite 20 melting
1 melting
Melting of garnet lherzoliteand spinel lherzolite
x gt lherz + 4x sp lherz
5 meltingx gt lherz + 4x sp lherz
Sam
ple
Cho
ndrit
e
Tholeiitic lavas
Tholeiitic lavas
High-MgO lavas
source (07DM+03PM)
Fig 14 Trace-element modeling results for incongruent dynamicmantle melting for picritic and tholeiitic lavas from the KarmutsenFormation Three steps of modeling are shown (a) melting of spinellherzolite compared with high-MgO lavas (b) melting of garnet lher-zolite and spinel lherzolite compared with tholeiitic lavas (c) meltingof garnet lherzolite compared with tholeiitic and picritic lavasShaded fields in each panel for tholeiitic basalts and picrites arelabeled Patterns with symbols in each panel are modeling results(patterns are 5 melting increments in (a) and (b) and 1 incre-ments in (c) PM primitive mantle DM depleted MORB mantle gtlherz garnet lherzolite sp lherz spinel lherzolite In (b) the ratios ofper cent melting for garnet and spinel lherzolite are indicated (x gtlherzthorn 4x sp lherz means that for 5 melting 1 melt in garnetfacies and 4 melt in spinel facies where x is per cent melting in theinterval of modeling) The ratio of the respective melt proportions inthe aggregate melt (x gt lherzthorn 4x sp lherz) was kept constant andthis ratio of melt contribution provided the best fit to the data Arange of proportion of source components (07DMthorn 03PM to
09DMthorn 01PM) can reproduce the variation in the high-MgOlavas Melting modeling uses the formulation of Zou amp Reid (2001)an example calculation is shown in their Appendix PM fromMcDonough amp Sun (1995) and DM from Salters amp Stracke (2004)Melt reaction coefficients of spinel lherzolite from Kinzler amp Grove(1992) and garnet lherzolite fromWalter (1998) Partition coefficientsfrom Salters amp Stracke (2004) and Shaw (2000) were kept constantSource mineralogy for spinel lherzolite (018cpx027opx052ol003sp) from Kinzler (1997) and for garnet lherzolite(034cpx008opx053ol005gt) from Salters amp Stracke (2004) Arange of source compositions for melting of spinel lherzolite(07DMthorn 03PM to 03DMthorn 07PM) reproduce REE patternssimilar to those of the high-MgO lavas with best fits for 22ndash25melting and 07DMthorn 03PM source composition Concentrationsof tholeiitic basalts cannot be reproduced from an entirelyprimitive or depleted source
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
30
The uncertainty in the composition of the source in thespinel lherzolite melting model for the high-MgO lavasprecludes determining whether the source was moredepleted in incompatible trace elements than the source ofthe tholeiitic lavas (see Fig 14 caption for description) It ispossible that early high-pressure low-degree melting left aresidue within parts of the plume (possibly the hotter inte-rior portion) that was depleted in trace elements (egElliott et al 1991) further decompression and high-degreemelting of such depleted regions could then have generatedthe depleted high-MgO Karmutsen lavas Shallow high-degree melting would preferentially sample depletedregions whereas deeper low-degree melting may notsample more refractory depleted regions (eg CaribbeanEscuder-Viruete et al 2007) The picrites lie at the highend of the range of Nd isotopic compositions and havelow NbZr similar to depleted Icelandic basalts (Fittonet al 2003) therefore the picrites may represent the partsof the mantle plume that were the most depleted prior tothe initiation of melting As Fitton et al (2003) suggestedthese depleted melts may only rarely be erupted withoutcombining with more voluminous less depleted melts andthey may be detected only as undifferentiated small-volume flows like the Keogh Lake picrites within theKarmutsen Formation on Vancouver IslandDecompression melting within the mantle plume initiatedwithin the garnet stability field (35^25 GPa) at highmantle potential temperatures (Tp414508C) and pro-ceeded beneath oceanic arc lithosphere within the spinelstability field (25^09 GPa) where more extensivedegrees of melting could occur
Magmatic evolution of Karmutsentholeiitic basaltsPrimary magmas in large igneous provinces leave exten-sive coarse-grained cumulate residues within or below thecrust these mafic and ultramafic plutonic sequences repre-sent a significant proportion of the original magma thatpartially crystallized at depth (eg Farnetani et al 1996)For example seismic and petrological studies of OntongJava (eg Farnetani et al 1996) combined with the use ofMELTS (Ghiorso amp Sack 1995) indicate that the volcanicsequence may represent only 40 of the total magmareaching the Moho Neal et al (1997) estimated that30^45 fractional crystallization took place for theOntong Java magmas and produced cumulate thicknessesof 7^14 km corresponding to 11^22 km of flood basaltsdepending on the presence of pre-existing oceanic crustand the total crustal thickness (25^35 km) Interpretationsof seismic velocity measurements suggest the presence ofpyroxene and gabbroic cumulates 9^16 km thick beneaththe Ontong Java Plateau (eg Hussong et al 1979)Wrangellia is characterized by high crustal velocities in
seismic refraction lines which is indicative of mafic
plutonic rocks extending to depth beneath VancouverIsland (Clowes et al 1995)Wrangellia crust is 25^30 kmthick beneath Vancouver Island and is underlain by astrongly reflective zone of high velocity and density thathas been interpreted as a major shear zone where lowerWrangellia lithosphere was detached (Clowes et al 1995)The vast majority of the Karmutsen flows are evolved tho-leiitic lavas (eg low MgO high FeOMgO) indicating animportant role for partial crystallization of melts at lowpressure and a significant portion of the crust beneathVancouver Island has seismic properties that are consistentwith crystalline residues from partially crystallizedKarmutsen magmas To test the proportion of the primarymagma that fractionated within the crust MELTS(Ghiorso amp Sack 1995) was used to simulate fractionalcrystallization using several estimated primary magmacompositions from the PRIMELT1 modeling results Themajor-element composition of the primary magmas couldnot be estimated for the tholeiitic basalts because of exten-sive plagthorn cpxthornol fractionation and thus the MELTSmodeling of the picrites is used as a proxy for the evolutionof major elements in the volcanic sequence A pressure of 1kbar was used with variable water contents (anhydrousand 02wt H2O) calculated at the quartz^fayalite^magnetite (QFM) oxygen buffer Experimental resultsfrom Ontong Java indicate that most crystallizationoccurs in magma chambers shallower than 6 km deep(52 kbar Sano amp Yamashita 2004) however some crys-tallization may take place at greater depths (3^5 kbarFarnetani et al 1996)A selection of the MELTS results are shown in Fig 15
Olivine and spinel crystallize at high temperature until15^20wt of the liquid mass has fractionated and theresidual magma contains 9^10wt MgO (Fig 15f) At1235^12258C plagioclase begins crystallizing and between1235 and 11908C olivine ceases crystallizing and clinopyr-oxene saturates (Fig 15f) As expected the addition of clin-opyroxene and plagioclase to the crystallization sequencecauses substantial changes in the composition of the resid-ual liquid FeO and TiO2 increase and Al2O3 and CaOdecrease while the decrease in MgO lessens considerably(Fig 15) The near-vertical trends for the Karmutsen tho-leiitic basalts especially for CaO do not match the calcu-lated liquid-line-of-descent very well which misses manyof the high-CaO high-Al2O3 low-FeO basalts A positivecorrelation between Al2O3CaO and SrNd indicates thataccumulation of plagioclase played a major role in the evo-lution of the tholeiitic basalts (Fig 15g) Increasing thecrystallization pressure slightly and the water content ofthe melts does not systematically improve the match ofthe MELTS models to the observed dataThe MELTS results indicate that a significant propor-
tion of the crystallization takes place before plagioclase(15^20 crystallization) and clinopyroxene (35^45
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
31
crystallization) join the crystallization sequence Theseresults suggest that445 crystallization of the primarymagmas occurred (estimated from the residual liquid percent in Fig 15f) and if the original flood basalt stratigra-phy on Vancouver Island was 6 km thick at least anequivalent amount of complementary plutonic rocksshould have crystallized The Karmutsen basalts and theirplutonic residues thus represent a significant addition ofcrust (perhaps 412 km thickness) previously thickenedduring the formation of the underlying Paleozoic arc
Growth of the Wrangellia oceanic plateauon Vancouver IslandThe extrusive part of the Wrangellia oceanic plateau onVancouver Island was constructed in a three-layersequence of mostly tholeiitic basalts with a restrictedrange of composition (Fig 16) The plateau overliesDevonian to Mississippian arc volcanic sequences andmarine sediments of Mississippian to Permian ages thatare intruded by gabbroic rocks related to the flood basaltsThe volcanic stratigraphy of theWrangellia plateau formed
00
05
10
15
20
25
30
35
40
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
16
0 2 4 6 8 10 12 14 16 18 20
10
11
12
13
14
15
16
17
18
19
20
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
0 2 4 6 8 10 12 14 16 18 20
MgO (wt)
0 10 20 30 40 50
percent of total mass
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
) Mineral proportions
Sp (e)
Ol
CpxPlag
PicriteHigh-MgO basaltTholeiitic basalt
TiO2 (wt)
MgO (wt)
Al2O3 (wt) CaO (wt)
FeO (wt)
MgO(a) (b)
(c) (d)
MgO (wt)MgO (wt)
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
)
Residual liquid ()
1220
1190
Ol + Sp
Plag+Ol+Sp Plag+Cpx+Sp
02 wt H2O
no H2O
Residual liquid and crystallizing phases
(f)
0102030405060708090100
plagioclaseaccumulation
plagioclaseaccumulation
plagioclase
accumulatio
n
10
12
14
16
0 10 20 30 40
Al2O3CaO
SrNd
tholeiitic basalts
(g)
Fig 15 Forward fractional crystallization modeling results for major elements from MELTS (Ghiorso amp Sack 1995) compared with picriticand tholeiitic lavas from the Karmutsen Formation (a^d) Major-element compositions of residual liquids compared with Karmutsen lavacompositions For clarity the results are shown only for one starting composition (open circle) using the composition of the estimated primarymagma for the high-MgO lava series from the modeling technique of Herzberg et al (2007) for sample 4723A13 as shown in Fig 12 The esti-mated primary magmas for samples 4723A3 and 93G171 yield similar results A pressure of 1 kbar was used with no H2O and 02wt H2Ocalculated at QFM black line is anhydrous and grey line is 02wt H2OThe end-point for the anhydrous result is at 78 fractionation and11308C and the end-point for the result with 02wt H2O is 75 fractionation and11108C (e) Mineral proportions of crystallized phases fromone MELTrun (sample 4723A13) with no H2O (f) Per cent residual liquid vs temperature labelled with intervals of crystallizing phases (g)Al2O3CaO vs SrNd for the tholeiitic basalts
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
32
in a single eruptive phase (c 231^225 Ma) as an emergentbasalt sequence during a deeper-water shallow-water andsubaerial stage and subsided below sea level relativelysoon after its formation (Fig 16) The deeper-water stagecontains pillowed and massive submarine flows (3000mthick) that built up the submarine volcanic edifice from theseafloor The lowest pillow basalts were emplaced onunconsolidated and lithified fine-grained siliciclastic andcarbonaceous sediments containing Daonella beds(c 235^232 Ma) that are intruded by abundantKarmutsen mafic sills (Fig 16a) The different pillowedand massive submarine flows resulted primarily from dif-ferent effusive rates and local topography Some of themassive submarine flows commonly recognized by radiat-ing columnar jointing represent master tubes for the deliv-ery of lava to distal parts of submarine flow fields Thepillowed flows which commonly have uneven topography
in modern sequences may have served as levees to themassive submarine flows rather than allowing unob-structed spreading and inflation typical of the massive sub-aerial flows that form on shallow slopes in continentalflood basalt provinces (eg Self et al 1997) The rapid suc-cession of flows is evident by the near absence of sedimentbetween submarine flows Minor volumes of picritic pillowlavas occur in the upper part of the pillowed flow sequencein areas of the plateau exposed on northern VancouverIsland (Fig 16 Nixon et al 2008)The growth of the plateau in shallow water led to
emplacement of an increasing proportion of volcaniclastics(pillow breccias and hyaloclastites) that accumulated up to1500m thick in some areas (Fig 16b Nixon et al 2008)Horizons of pillowed flows occur locally within the volca-niclastic stratigraphy Sedimentary structures (eg gradedbedding fluidization structures) and petrographic textures
Late Devonian-Mississippian arc
Daonella beds
Middle Triassic
deeper water
Late Devonian-Mississippian arc
pillowed flows massive submarine
flows
pillow brecciahyaloclastiteshallow water
~4 km
~6 km
Late Triassic limestoneinterflow limestonelenses
Karmutsen Formation
mafic intrusions sediment-sill complex
Karmutsen Formation
(a) (b)
(c) (d)
subaerial flows
Late Devonian-Mississippian arc Late Devonian-Mississippian arc
picritic pillow basalts
Miss to Permian marine sediments
sills
Fig 16 Growth of the Wrangellia oceanic plateau onVancouver Island (a) Deep-water eruptive phase of pillowed and massive submarineflows with almost no intervening sediments onto MiddleTriassic shale containing Daonella beds (c 235^232 Ma) The basement of the plateauis composed of Devonian to Mississippian arc sequences overlain by Mississippian to Permian carbonates and siliciclastic sediments(b) Shallow-water flow emplacement with increasing proportion of volcaniclastics resulted primarily from cooling^contraction granulationPicritic pillow basalts occur primarily near the transition between pillowed submarine flows and volcaniclastic units Pillowed and massiveflows occur locally within the volcaniclastic unit (c) Emergent stage of growth consists of overlapping subaerial flow fields similar to continen-tal flood basalts with pahoehoe flow structures Multi-tiered columnar jointing occurrs locally in response to ponding of flows (eg Bondre et al2004) Interflow carbonate deposits and plagioclase megacrystic flows developed locally during the waning stage of volcanism (d) Post-volcanicsubsidence and deposition of Late Carnian to Norian (c 212^223 Ma) platform carbonates on top of the plateau Units are labeled in the dia-grams Thicknesses of the underlying units are not to scale
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
33
within the volcaniclastic rocks indicate formation primar-ily via cooling^contraction granulation magma^water^steam interaction autobrecciation resedimentation andmass-wasting rather than pyroclastic eruption (Nixonet al 2008)A broad subaerial platform was constructed by emplace-
ment of subaerial lava flows during an emergent subaerialstage (Fig 16c) The subaerial unit predominantly com-prises well-layered sheet flows that formed in a similarway to continental flood basalts as hundred(s) of pahoehoeflow fields erupted on shallow slopes (Self et al 1997) how-ever volcaniclastic and pillowed flow units are locally pre-sent mostly in the upper parts of the stratigraphyLocalized ponding of flows led to the formation of multi-tiered columnar jointing in some subaerial flows (egBondre et al 2004) As volcanism waned local interflowcarbonate deposits developed and plagioclase megacrysticflows were erupted Subsidence of the plateau is recordedby the interflow lenses and deposits of platform carbonates(40^1500m thick) on top of the Karmutsen basalts thatcontain Late Carnian to Early Norian fossils (Fig 16d c216^227 Ma Nixon et al 2008)The evolution of the internal architecture of the
Wrangellia oceanic plateau was primarily related to theconditions of emplacement (ie water depth and eruptionenvironment) There does not appear to have been a sys-tematic coincident evolution or shift of lava compositionsonVancouver Island as has been recognized in most con-tinental flood basalts and parts of the Wrangellia oceanicplateau in Alaska and Yukon (Greene et al 2008 2009)In general with the exception of the picritic pillow lavasin the submarine stratigraphy and local variations betweenhigh-MgO and tholeiitic lavas there are no significantcompositional differences between lavas erupted duringthe different stages of growth of the plateau (ie submarineand subaerial flows) nor are substantial compositional dif-ferences generally found between massive submarine flowsin contact with pillowed flows tholeiitic basalt flows nearthe base of the Karmutsen Formation are compositionallysimilar to flows sampled near the top of the volcanicsequence Where it was possible to sample continuous sec-tions of volcanic stratigraphy greater than several hun-dreds of meters (eg Keogh^Maynard Lake area west ofthe Karmutsen Range Fig 2) there is no clear relationshipbetween stratigraphic position and compositionalvariation
CONCLUSIONSThe Karmutsen Formation covers extensive areas ofVancouver Island (20 000 km2) and forms part of a largeoceanic plateau erupted over a geologically short interval(c 225^231 Ma) The tripartite volcanic stratigraphyexposed on Vancouver Island is at least 6 km thick andcomprises a succession of submarine flows volcaniclastic
deposits and massive subaerial flowsThese volcanologicaldifferences are primarily related to the eruption environ-ment (deep-water shallow-water or subaerial) The rapidgrowth of the plateau prevented significant accumulationsof intervening sediment except in the uppermost stratigra-phy where isolated limestone lenses commonly associatedwith pillow basalt and hyaloclastite deposits preserve arecord of the final stages of subsidence of the plateauThe Wrangellia plateau on Vancouver Island was con-
structed dominantly of tholeiitic basalt with restrictedmajor- and trace-element and isotopic compositional var-iations Minor volumes of picritic pillow basalts eruptedlate in the submarine phase The Karmutsen basalts wereemplaced onto older Paleozoic arc volcanic sequences andlow NbLa for the picrites may indicate a contributionfrom the sub-arc lithospheric mantle Modeling of the pet-rogenesis of the Karmutsen picrites indicates melting ofanomalously hot mantle (15008C) and extensive degreesof partial melting (23^27) and is consistent with aplume initiation model Combined Sr^Nd^Hf^Pb isotopicsystematics indicate that the picritic and tholeiitic lavaswere derived from a depleted plume-type mantle sourcedistinct from the source of MORB that has compositionalsimilarities to the source of the lavas forming theCaribbean Plateau Picrites were generated by melting ofspinel lherzolite whereas the tholeiitic basalts formed asaggregate melts of both garnet and spinel lherzolite Thetholeiitic basalts underwent extensive low-pressure fractio-nation (52^3 kbar) and seismic studies indicate that someof theWrangellia crust beneathVancouver Island may cor-respond to the plutonic residues of the Karmutsen floodbasalts
ACKNOWLEDGEMENTSWe are grateful to Nick Arndt for helping us get this pro-ject started and Nick Massey for insights into VancouverIsland geology We greatly appreciate the assistance withPRIMELT modeling from Claude Herzberg We thankMikkel Schau for his insight and enthusiasm during field-work Jane Barling assisted with analyses by MC-ICP-MSThoughtful and thorough reviews of the manuscript byJohn Mahoney Fred Frey Wendy Bohrson DougalJerram and Teal Riley are greatly appreciated Fundingwas generously provided by the Rocks to Riches Programadministered by the BC amp Yukon Chamber of Mines in2004 by the BC Geological Survey and by NSERCDiscovery Grants to JSS and DW ARG was supportedby a University Graduate Fellowship at UBC
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
34
REFERENCESBarker F Brown A S Budahn J R amp Plafker G (1989) Back-arc
with frontal-arc component origin of Triassic Karmutsen basaltBritish Columbia Canada Chemical Geology 75 81^102
Blichert-Toft JWeis D Maerschalk C Agranier A amp Albare de F(2003) Hawaiian hot spot dynamics as inferred from the Hf and Pbisotope evolution of Mauna Kea volcano Geochemistry GeophysicsGeosystems 4(2) 1^27 doi1010292002GC000340
Bondre N R Duraiswami R A amp Dole G (2004) Morphologyand emplacement of flows from the Deccan Volcanic ProvinceIndia Bulletin of Volcanology 66 29^45
Carlisle D (1963) Pillow breccias and their aquagene tuffs QuadraIsland British Columbia Journal of Geology 71 48^71
Carlisle D (1972) Late Paleozoic to Mid-Triassic sedimentary-volcanic sequence on Northeastern Vancouver Island Geological
Society of Canada Paper Report of Activities 72-1 Part B 24^30Carlisle D amp SuzukiT (1974) Emergent basalt and submergent car-
bonate^clastic sequences including the Upper Triassic Dilleri andWelleri zones on Vancouver Island Canadian Journal of Earth
Sciences 11 254^279Chauvel C amp Blichert-Toft J (2001) A hafnium isotope and trace
element perspective on melting of the depleted mantle Earth and
Planetary Science Letters 190 137^151Chayes F (1954)The theory of thin-section analysis Journal of Geology
62 92^101Cho M Liou J G amp Maruyama S (1987) Transition from the zeo-
lite to prehnite^pumpellyite facies in the Karmutsen metabasitesVancouver Island British Columbia Journal of Petrology 27(2)467^494
Clowes R M Zelt C A Amor J R amp Ellis R M (1995)Lithospheric structure in the southern Canadian Cordillera froma network of seismic refraction lines Canadian Journal of Earth
Sciences 32(10) 1485^1513Coffin M F amp Eldholm O (1994) Large igneous provinces Crustal
structure dimensions and external consequences Reviews of
Geophysics 32(1) 1^36Cox K G (1980) A model for flood basalt volcanism Journal of
Petrology 21 629^650Eldholm O amp Coffin M F (2000) Large igneous provinces
and plate tectonics In Richards M A Gordon R G amp vander Hilst R D (eds) The History and Dynamics of Global
Plate Motions American Geophysical Union Geophysical Monograph 121309^326
Elliott T R Hawkesworth C J amp Grolaquo nvold K (1991) Dynamicmelting of the Iceland plume Nature 351 201^206
Escuder-Viruete J Pecurren rez-Estaucurren n A Contreras F Joubert MWeis D Ullrich T D amp Spadea P (2007) Plume mantle sourceheterogeneity through time Insights from the Duarte ComplexHispaniola northeastern Caribbean Journal of Geophysical Research112(B04203) doi1010292006JB004323
Farnetani C G Richards M A amp Ghiorso M S (1996)Petrological models of magma evolution and deep crustal structurebeneath hotspots and flood basalt provinces Earth and Planetary
Science Letters 143 81^94Fitton J G amp Godard M (2004) Origin and evolution of magmas
on the Ontong Java Plateau In Fitton J G Mahoney J JWallace P J amp Saunders A D (eds) Origin and Evolution of the
Ontong Java Plateau Geological Society London Special Publications 229151^178
Fitton J G Saunders A D Norry M J Hardarson B S ampTaylor R N (1997) Thermal and chemical structure of theIceland plume Earth and Planetary Science Letters 153 197^208
Fitton J G Saunders A D Kempton P D amp Hardarson B S(2003) Does depleted mantle form an intrinsic part of the Icelandplume Geochemistry Geophysics Geosystems 4(3) 1032 doi1010292002GC000424
Frey F A Huang S Blichert-Toft J Regelous M amp Boyet M(2005) Origin of depleted components in basalt related to theHawaiian hotspot Evidence from isotopic and incompatible ele-ment ratios Geochemistry Geophysics Geosystems 6(1) 23 doi1010292004GC000757
Galer S J G amp Abouchami W (1998) Practical application of leadtriple spiking for correction of instrumental mass discriminationMineralogical Magazine 62A 491^492
Ghiorso M S amp Sack R O (1995) Chemical mass transfer in mag-matic processes IV A revised and internally consistent thermody-namic model for the interpolation and extrapolation of liquid^solidequilibria in magmatic systems at elevated temperatures and pres-sures Contributions to Mineralogy and Petrology 119 197^212
Greene A R Scoates J S amp Weis D (2005)WrangelliaTerrane onVancouver Island Distribution of flood basalts with implicationsfor potential Ni^Cu^PGE mineralization In Grant B (ed)Geological Fieldwork 2004 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2005-1 209^220Greene A R Scoates J S Nixon G T amp Weis D (2006) Picritic
lavas and basal sills in the Karmutsen flood basalt provinceWrangellia northern Vancouver Island In Grant B (ed)Geological Fieldwork 2005 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2006-1 39^52Greene A R Scoates J S amp Weis D (2008) Wrangellia flood
basalts in Alaska A record of plume^lithosphere interaction in aLate Triassic accreted oceanic plateau Geochemistry Geophysics
Geosystems 9(Q12004) doi1010292008GC002092Greene A R Scoates J S Weis D amp Israel S (2009)
Geochemistry of triassic flood basalts from the Yukon (Canada)segment of the accreted Wrangellia oceanic plateau Lithosdoi101016jlithos200811010
Hauff F Hoernle K Tilton G Graham D W amp Kerr A C(2000) Large volume recycling of oceanic lithosphere over shorttime scales geochemical constraints from the Caribbean LargeIgneous Province Earth and Planetary Science Letters 174 247^263
Hauff F Hoernle K amp Schmidt A (2003) Sr^Nd^Pb compositionof Mesozoic Pacific oceanic crust (Site 1149 and 801 ODP Leg185)Implications for alteration of ocean crust and the input into theIzu^Bonin^Mariana subduction system Geochemistry Geophysics
Geosystems 4(8) doi1010292002GC000421Herzberg C (2004) Partial melting below the Ontong Java Plateau
In Fitton J G Mahoney J J Wallace P J amp Saunders A D(eds) Origin and Evolution of the Ontong Java Plateau Geological SocietyLondon Special Publications 229 179^183
Herzberg C (2006) Petrology and thermal structure of the Hawaiianplume from Mauna Kea volcano Nature 444(7119) 605
Herzberg C amp Asimow P D (2008) Petrology of some oceanicisland basalts PRIMELT2XLS software for primary magma cal-culation Geochemistry Geophysics Geosystems 9(Q09001) doi1010292008GC002057
Herzberg C amp OrsquoHara M J (2002) Plume-associated ultramaficmagmas of Phanerozoic age Journal of Petrology 43(10) 1857^1883
Herzberg C Asimow P D Arndt N Niu Y Lesher C MFitton J G Cheadle M J amp Saunders A D (2007)Temperatures in ambient mantle and plumes Constraints frombasalts picrites and komatiites Geochemistry Geophysics Geosystems8(Q02006) doi1010292006GC001390
Huang S amp Frey F A (2005) Trace element abundances of MaunaKea basalt from phase 2 of the Hawaii Scientific Drilling Project
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
35
Petrogenetic implications of correlations with major element con-tent and isotopic ratios Geochemistry Geophysics Geosystems 4(6)doi1010292002GC000322
Hussong D M Wipperman L K amp Kroenke L W (1979) Thecrustal structure of the Ontong Java and Manihiki OceanicPlateaus Journal of Geophysical Research 84(B11) 6003^6010
Jerram D A amp Widdowson M (2005) The anatomy of ContinentalFlood Basalt Provinces geological constraints on the processes andproducts of flood volcanism Lithos 79(3^4) 385^405
Jones D L Silberling N J amp Hillhouse J (1977)Wrangellia a dis-placed terrane in northwestern North America CanadianJournal ofEarth Sciences 14(11) 2565^2577
Kerr A C (2003) Oceanic Plateaus In Rudnick R (ed) The Crust
Treatise on GeochemistryVol 3 Oxford Elsevier Science pp 537^565Kerr A C amp Mahoney J J (2007) Oceanic plateaus Problematic
plumes potential paradigms Chemical Geology 241 332^353Kerr A CTarney J Marriner G F Klaver GT Saunders A D
amp Thirlwall M F (1996) The geochemistry and petrogenesis ofthe Late-Cretaceous picrites and basalts of Curacao NetherlandsAntilles a remnant of an oceanic plateau Contributions to
Mineralogy and Petrology 124(1) 29^43Kerr A C Marriner G F Tarney J Nivia A Saunders A D
Thirlwall M F amp Sinton A C (1997) Cretaceous basaltic ter-ranes in Western Colombia Elemental chronological and Sr-Ndisotopic constraints on petrogenesis Journal of Petrology 38(6)677^702
Kinzler R J (1997) Melting of mantle peridotite at pressuresapproaching the spinel to garnet transition Application to mid-ocean ridge basalt petrogenesis Journal of Geophysical Research
102(B1) 853^874Kinzler R J amp Grove T L (1992) Primary magmas of mid-ocean
ridge basalts 1 Experiments and methods Journal of Geophysical
Research 97(B5) 6885^6906Korenaga J (2005) Why did not the Ontong Java Plateau form sub-
aerially Earth and Planetary Science Letters 234 385^399Kuniyoshi S (1972) Petrology of the Karmutsen (Volcanic) Group
northeastern Vancouver Island British Columbia PhD disserta-tion University of California Los Angeles 242 pp
Lassiter J C DePaolo D J amp Mahoney J J (1995) Geochemistry ofthe Wrangellia flood basalt province Implications for the role ofcontinental and oceanic lithosphere in flood basalt genesis Journalof Petrology 36(4) 983^1009
LeBas M J (2000) IUGS reclassification of the high-Mg and picriticvolcanic rocks Journal of Petrology 41(10) 1467^1470
Longhi J (2002) Some phase equilibrium systematics of lherzolitemelting I Geochemistry Geophysics Geosystems 3(3) doi1010292001GC000204
MacDonald G A amp Katsura T (1964) Chemical composition ofHawaiian lavas Journal of Petrology 5 82^133
Mahoney J J (1987) An isotopic survey of Pacific oceanic plateausimplications for their nature and origin In Keating B HFryer P Batiza R amp Boehlert GW (eds) Seamounts Islands andAtolls American Geophysical Union Geophysical Monograph 43 207^220
Mahoney J LeRoex A P Peng Z Fisher R L amp Natland J H(1992) Southwestern limits of Indian Ocean Ridge mantle and theorigin of low 206Pb204Pb mid-ocean ridge basalt isotope systema-tics of the central Southwest Indian Ridge (178^508E) Journal ofGeophysical Research 97(B13) 19771^19790
Mahoney J J Sinton J M Kurz M D MacDougall J DSpencer K J amp Lugmair GW (1994) Isotope and trace elementcharacteristics of a super-fast spreading ridge East Pacific rise13^238S Earth and Planetary Science Letters 121 173^193
Massey N W D (1995a) Geology and mineral resources of theAlberni^Nanaimo Lakes sheet Vancouver Island 92F1W 92F2Eand part of 92F7E British Columbia Ministry of Energy Mines and
Petroleum Resources Paper 1992-2 132 ppMassey N W D (1995b) Geology and mineral resources of the
Cowichan Lake sheet Vancouver Island 92C16 British Columbia
Ministry of Energy Mines and Petroleum Resources Paper 1992-3 112 ppMassey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005a) Digital Geology Map of British Columbia Tile NM9 MidCoast BC British Columbia Ministry of Energy and Mines Geofile
2005-2Massey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005b) Digital Geology Map of British Columbia Tile NM10Southwest BC British Columbia Ministry of Energy and Mines Geofile
2005-3McDonough W F amp Sun S (1995) The composition of the Earth
Chemical Geology 120 223^253Monger JW H amp Journeay J M (1994) Basement geology and tec-
tonic evolution of theVancouver region In Monger JW H (ed)Geology and Geological Hazards of the Vancouver Region Southwestern
British Columbia Geological Survey of Canada Bulletin 481 3^25Muller J E (1977) Geology of Vancouver Island Geological Survey of
Canada Open File Map 463Muller J E (1980)The Paleozoic Sicker Group of Vancouver Island British
Columbia Geological Survey of Canada Paper 79-30 22 ppMuller J E Northcote K E amp Carlisle D (1974) Geology and
mineral deposits of Alert Bay^Cape Scott map area VancouverIsland British Columbia Geological Survey of Canada Paper 74-8 77
Neal C R Mahoney J J Kroenke L W Duncan R A ampPetterson M G (1997) The Ontong^Java Plateau InMahoney J J amp Coffin M F (eds) Large Igneous Provinces
Continental Oceanic and Planetary FloodVolcanism American Geophysical
Union Geophysical Monograph 100 183^216Niu Y Collerson K D Batiza R Wendt J I amp Regelous M
(1999) Origin of enriched-typed mid-ocean ridge basalt at ridgesfar from mantle plumes The East Pacific Rise at 11820rsquoN Journalof Geophysical Research 104 7067^7088
Nixon G T amp Orr A J (2007) Recent revisions to the EarlyMesozoic stratigraphy of Northern Vancouver Island (NTS 102I092L) and metallogenic implicatinos British Columbia InGrant B (ed) Geological Fieldwork 2006 British Columbia Ministry of
Energy Mines and Petroleum Resources Paper 2007-1 163^177Nixon GT Hammack J L KoyanagiV M Payie G J Haggart
JW Orchard M JTozerT Friedman R M Archibald D APalfy J amp Cordey F (2006a) Geology of the Quatsino^PortMcNeill area northernVancouver Island British Columbia Ministry
of Energy Mines and Petroleum Resources Geoscience Map 2006-2 scale150 000
Nixon G T Kelman M C Stevenson D Stokes L A ampJohnston K A (2006b) Preliminary geology of the Nimpkishmap area (NTS 092L07) northern Vancouver Island BritishColumbia In Grant B amp Newell J M (eds) Geological Fieldwork2005 British Columbia Ministry of Energy Mines and Petroleum Resources
Paper 2006-1 135^152Nixon G T Laroque J Pals A Styan J Greene A R amp
Scoates J S (2008) High-Mg lavas in the Karmutsen flood basaltsnorthernVancouver Island (NTS 092L) Stratigraphic setting andmetallogenic significance In Grant B (ed) Geological Fieldwork
2007 British Columbia Ministry of Energy Mines and Petroleum
Resources Paper 2008-1 175^190Nowell G M Kempton P D Noble S R Fitton J G
Saunders A Mahoney J J amp Taylor R N (1998) High precisionHf isotope measurements of MORB and OIB by thermal ionisation
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
36
mass spectrometry insights into the depleted mantle Chemical
Geology 149 211^233Parrish R R amp McNicoll V J (1992) U^Pb age determinations
from the southern Vancouver Island area British Columbia InRadiogenic Age and Isotopic Studies Report 5 Geological Survey of
Canada Paper 91-2 79^86Peate DW (1997) The Parana^Etendeka Province In Coffin M F
amp Mahoney J (eds) Large Igneous Provinces Continental Oceanic andPlanetary Flood Volcanism American Geophysical Union Geophysical
Monograph 100 217^245Perfit M R Fornari D J Ridley W I Kirk P D Casey J
Kastens K A Reynolds J R Edwards M Desonie DShuster R amp Paradis S (1996) Recent volcanism in the Siqueirostransform fault picritic basalts and implications for MORBmagma genesis Earth and Planetary Science Letters 141(1^4) 91^108
Pretorius W Weis D Williams G Hanano D Kieffer B ampScoates J S (2006) Complete trace elemental characterization ofgranitoid (USGSG-2GSP-2) reference materials by high resolutioninductively coupled plasma-mass spectrometry Geostandards and
Geoanalytical Research 30(1) 39^54Regelous M Niu Y Wendt J I Batiza R Greig A amp
Collerson K D (1999) Variations in the geochemistry of magma-tism on the East Pacific Rise at 10830rsquoN since 800 ka Earth and
Planetary Science Letters 168 45^63Recurren villon S Chauvel C Arndt N T Pik R Martineau F
Fourcade S amp Marty B (2002) Heterogeneity of the Caribbeanplateau mantle source Sr O and He isotopic compositions of oli-vine and clinopyroxene from Gorgona Island Earth and Planetary
Science Letters 205(1^2) 91Rhodes J M (1996) Geochemical stratigraphy of lava flows samples
by the Hawaii Scientific Drilling Project Journal of Geophysical
Research 101(B5) 11729^11746Rhodes J M amp Vollinger M J (2004) Composition of basaltic lavas
sampled by phase-2 of the Hawaii Scientific Drilling ProjectGeochemical stratigraphy and magma types Geochemistry
Geophysics Geosystems 5(3) doi1010292002GC000434Richards M A Jones D L Duncan R A amp DePaolo D J (1991)
A mantle plume initiation model for the Wrangellia flood basaltand other oceanic plateaus Science 254 263^267
Salters V J M amp Hart S R (1991) The mantle sources of oceanridges islands and arcs the Hf-isotope connection Earth and
Planetary Science Letters 104 364^380Salters V J M amp Stracke A (2004) Composition of the depleted
SaltersV J amp WhiteW (1998) Hf isotope constraints on mantle evo-lution Chemical Geology 145 447^460
SanoT amp Yamashita S (2004) Experimental petrology of basementlavas from Ocean Drilling Program Leg 192 implications for dif-ferentiation processes in Ontong Java Plateau magmas InFitton J G Mahoney J JWallace P J amp Saunders A D (eds)Origin and Evolution of the Ontong Java Plateau Geological Society
London Special Publications 229 185^218Saunders A D (2005) Large igneous provinces Origin and environ-
mental consequences Elements 1 259^263Self S ThordarsonT amp Keszthely L (1997) Emplacement of conti-
nental flood basalt lava flows In Mahoney J J amp Coffin M F(eds) Large Igneous Provinces Continental Oceanic and Planetary Flood
Volcanism American Geophysical Union Geophysical Monograph 100381^410
Shaw D M (2000) Continuous (dynamic) melting theory revisitedCanadian Mineralogist 38(5) 1041^1063
Sluggett C L (2003) Uranium^lead age and geochemical constraintson Paleozoic and Early Mesozoic magmatism in WrangelliaTerrane Saltspring Island British Columbia BSc thesisUniversity of British ColumbiaVancouver 84 pp
Storey M Mahoney J J amp Saunders A D (1997) Cretaceousbasalts in Madagascar and the transition between plume and con-tinental lithospheric mantle sources In Mahoney J J ampCoffin M F (eds) Large Igneous Provinces Continental Oceanic and
Planetary Flood Volcanism Aamerican Geophysical Union Geophysical
Monograph 100 95^122Surdam R C (1967) Low-grade metamorphism of the Karmutsen
Group PhD dissertation University of California Los Angeles288 pp
Sutherland-Brown A Yorath C J Anderson R G amp Dom K(1986) Geological maps of southern Vancouver IslandLITHOPROBE I 92C10 11 14 16 92F1 2 7 8 Geological Survey ofCanada Open File 1272
Tejada M L G Mahoney J J Castillo P R Ingle S P Sheth HC amp Weis D (2004) Pin-pricking the elephant evidence on theorigin of the Ontong Java Plateau from Pb^Sr^Hf^Nd isotopiccharacteristics of ODP Leg 192 basalts In Fitton J GMahoney J J Wallace P J amp Saunders A D (eds) Origin and
Evolution of the Ontong Java Plateau Geological Society London Special
Publications 229 133^150Thompson P M E Kempton P D amp Kerr A C (2008) Evaluation
of the effects of alteration and leaching on Sm^Nd and Lu^Hf sys-tematics in submarine mafic rocks Lithos 104(1^4) 164^176
Thompson P M E Kempton P D White R V Kerr A CTarney J Saunders A D Fitton J G amp McBirney A (2003)Hf-Nd isotope constraints on the origin of the CretaceousCaribbean plateau and its relationship to the Galapagos plumeEarth and Planetary Science Letters 217(1^2) 59^75
Vervoort J D Patchett P Blichert-Toft J amp Albare de F (1999)Relationships between Lu^Hf and Sm^Nd isotopic systems in theglobal sedimentary system Earth and Planetary Science Letters 16879^99
Walter M J (1998) Melting of garnet peridotite and the origin ofkomatiite and depleted lithosphere Journal of Petrology 39(1) 29^60
Weis D Kieffer B Maerschalk C Barling J de Jong JWilliams G A Hanano D Mattielli N Scoates J SGoolaerts A Friedman R A amp Mahoney J B (2006) High-pre-cision isotopic characterization of USGS reference materials byTIMS and MC-ICP-MS Geochemistry Geophysics Geosystems
7(Q08006) doi1010292006GC001283Weis D Kieffer B Hanano D Silva I N Barling J
Pretorius W Maerschalk C amp Mattielli N (2007) Hf isotopecompositions of US Geological Survey reference materialsGeochemistry Geophysics Geosystems 8(Q06006) doi1010292006GC001473
Wheeler J O amp McFeely P (1991) Tectonic assemblage map of theCanadian Cordillera and adjacent part of the United States ofAmerica Geological Survey of Canada Map 1712A
WhiteW M Albare de F amp Tecurren louk P (2000) High-precision analy-sis of Pb isotope ratios by multi-collector ICP-MS Chemical Geology167 257^270
Yorath C J Sutherland Brown A amp Massey N W D (1999)LITHOPROBE southern Vancouver Island British Columbia Geological
Survey of Canada Bulletin 498 145 ppZou H (1998) Trace element fractionation during modal and
non-model dynamic melting and open-system melting Amathematical treatment Geochimica et Cosmochimica Acta 62(11)1937^1945
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
37
Zou H amp Reid M R (2001) Quantitative modeling of trace elementfractionation during incongruent dynamic melting Geochimica et
Cosmochimica Acta 65(1) 153^162
APPENDIX SAMPLEPREPARAT ION ANDANALYT ICAL METHODSOnly the freshest rocks in the field were sampled and onlythe least altered samples were selected for chemical andisotopic analysis based on thorough petrographic inspec-tion Fifty-nine of the 129 collected samples were crushed(400 g) into small pieces 52mm in diameter in aRocklabs hydraulic piston crusher between WC plates tominimize contamination The coarse-crush was mixedand 100 g was powdered in a planetary mill using agatejars and balls cleaned with quartz sand between samples
ActLabs analytical methodsThe major- and trace-element compositions of the whole-rock powders were determined at Activation LaboratoriesLtd (Actlabs) in Ancaster Ontario Analytical techniquesand detection limits are also available from Actlabs (httpwwwactlabscommethsub_code4erehtm) The analyticalmethod for each of the elements analyzed is indicated inTable 2 For the major elements a 02 g sample was mixedwith a lithium metaborate^lithium tetraborate mixtureand fused in a graphite crucible The molten mixture waspoured into a 5 HNO3 solution and shaken until dis-solved (30min) The samples were analyzed for majoroxides and selected trace elements by inductively coupledplasma optical emission spectrometry (ICP-OES) on acombination simultaneous^sequential Thermo Jarrell-AshEnviro II system Internal calibration was achieved usinga variety of international reference materials (eg W-2BIR-1 DNC-1) and independent control samplesAdditional trace elements were analyzed by both instru-mental neutron activation analysis (INAA) and induc-tively couple plasma mass spectrometry (ICP-MS) ForINAA15^25 g of sample was weighed into small polyeth-ylene vials and irradiated with a control international ref-erence material CANMET WMS-1 and NiCr flux wires ata thermal neutron flux of 71012 neutronscm2 per s in theMcMaster Nuclear Reactor Following a 7 day waitingperiod the samples were measured on an Ortec high-purity Ge detector linked to a Canberra Series 95 multi-channel analyzer Activities for each element weredecay- and weight-corrected and compared with a detectorcalibration developed from multiple international certifiedreference materials For ICP-MS 025 g of sample wasdigested in HF followed by HNO3^HClO4 heated andtaken to dryness The samples were brought back into solu-tion with HCl Samples were analyzed using a PerkinElmer Optima 3000 ICP In-lab standards or certified ref-erence materials (eg W-2 BIR-1 DNC-1) were used for
quality control A total of 15 blind duplicates were ana-lyzed from the same sample powders to assessreproducibility
University of Massachusetts analyticalmethodsFourteen sample duplicate powders taken from the same100 g sample aliquot (high-MgO basalts and picrites)were also analyzed at the Ronald B Gilmore X-RayFluorescence (XRF) Laboratory at the University ofMassachusetts Major elements were measured on a fusedLa-bearing lithium borate glass disc using a SiemensMRS-400 spectrometer Trace element concentrations(Rb Sr Ba Ce Nb Zr Y Pb Zn Ga Ni Cr V) weremeasured on a separate powder pellet using a PhilipsPW2400 sequential spectrometer using a Rh tube LOIand ferrous iron measurements were made as describedby Rhodes amp Vollinger (2004) Precision and accuracy esti-mates for the data have been described by Rhodes (1996)and Rhodes amp Vollinger (2004) Results for each sampleare the average of two separate analyses and are shown inElectronic Appendix 2 for a comparison of analyses ofduplicate powders using XRFand ICP-MS
PCIGR analytical methodsA subset of 19 samples was selected for high-precisiontrace-element analysis and Sr Nd Pb and Hf isotopicanalysis at the Pacific Centre for Isotopic andGeochemical Research (PCIGR) at the University ofBritish Columbia (UBC) Samples were selected from the59 samples analyzed for whole-rock chemistry at ActLabsbased on major- and trace-element chemistry alteration(low LOI and petrographic alteration index) and samplelocation Samples were prepared for trace-element analysisat the PCIGR by the technique described by Pretorius et al(2006) on unleached rock powders Sample powders(100mg) were weighed in 7ml screw-top Savillex bea-kers and dissolved in 1ml 14N HNO3 and 5ml 48 HFon a hotplate for 48 h at 1308C with periodic ultrasonica-tion Samples were dried and redissolved in 6ml 6N HClon a hotplate for 24 h and then dried and redissolved in1ml concentrated HNO3 for 24 h before final dryingTrace element abundances were measured by high-resolution inductively coupled plasma-mass spectrometry(HR-ICP-MS) with aThermo Finnigan Element2 systemfollowing the procedures described by Pretorius et al(2006) within 24 h of redissolution HFSE and LILE weremeasured in medium-resolution mode at 2000 dilutionusing a PFATeflon spray chamber washed with aqua regiafor 3min between samples REE were measured in high-resolution mode and U and Pb in low-resolution modeat 2000 dilution using a glass spray chamber washedwith 2 HNO3 between samples Total procedural blanksand reference materials (BCR-2 BHVO-2) were analyzedwith the batch of samples Indium was used as an internal
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
38
standard in all samples and standard solutionsBackground and standard solutions were analyzed afterevery five samples to detect memory effects and massdrift Results for PCIGR trace-element analyses areshown in Electronic Appendix 3 for the full set of analyzedtrace elementsSample digestion for purification of Sr Nd Hf and Pb
for column chemistry began by weighing each samplepowder (400^500mg) prior to leaching All samples wereleached with 6N HCl and placed in an ultrasonic bath for15min Samples were rinsed twice with 18 M-cm H2Obetween each leaching step (15 total) until the supernatantliquid was clear [following the method of Mahoney (1987)]Samples were then dried on a hotplate for 24 h andweighed again Sample solutions were then prepared bydissolving 100^250mg of the leached powder dissolvedin 1ml 14N HNO3 and 10ml 48 HF on a hotplate for48 h at 1308C with periodic ultrasonication Samples weredried and redissolved in 6ml 6N HCl on a hotplate for24 h and then dried Pb was separated using anion exhangecolumns and the discard was used for Sr REE and Hfseparation Nd was separated from the other REE and Hfrequired two additional purification steps Detailed proce-dures for column chemistry for separating Sr Nd and Pbat the PCIGR have been described byWeis et al (2006) andHf purification has been described byWeis et al (2007) Srand Nd isotope ratios were measured by thermal ioniza-tion mass spectrometry (TIMS) on a Thermo FinniganTriton system in static mode with relay matrix rotation ona single Ta and double Re^Ta filament respectively Fourto five filaments per barrel of 21 were occupied by stan-dards (NBS 987 for Sr and LaJolla for Nd) for each barrelwhere samples were run Sample Sr and Nd isotopic com-positions were corrected for mass fractionation using86Sr88Srfrac14 01194 and 146Nd144Ndfrac14 07219 Each samplewas then normalized using the barrel average of the refer-ence material relative to the values of143Nd144Ndfrac14 0511858 and 87Sr86Srfrac14 0710248 (Weiset al 2006) During the course of the Vancouver Islandanalyses the La Jolla Nd standard gave an average valueof 05118566 (nfrac14 7) and NBS987 standard gave an aver-age of 07102408 (nfrac1411) (2 error is reported as times106) 147Sm144Nd errors are 15 or 0006 USGeological Survey (USGS) reference material BHVO-2was processed with the samples and yielded Sr and Nd iso-topic ratios of 07034607 and 05129786 respectivelyThese are in agreement with the published values of070347920 and 051298411 respectively (Weis et al2006)
Pb and Hf isotopic compositions were analyzed by mul-tiple collector-inductively coupled plasma-mass spectrome-try (MC-ICP-MS) on a static multi-collection on a NuPlasma (Nu Instruments) system The detailed analyticalprocedure for Pb isotopic analyses on the Nu Plasma atthe PCIGR has been described by Weis et al (2006) Theconfiguration for Pb analyses allows for collection of PbTl and Hg togetherTl and Hg are used to monitor instru-mental mass discrimination and isobaric overlap repec-tively All sample solutions were analyzed withapproximately the same PbTl ratio (4) as the referencematerial NIST SRM 981 To accomplish this a small ali-quot of each sample solution from the Pb columns was ana-lyzed on the Element2 to determine the precise amount ofPb available for analysis on the Nu Plasma The SRM 981standard was run after every two samples on the NuPlasma During the time samples were run analyses ofthe SRM 981 Pb reference material gave values of206Pb204Pbfrac1416940322 207Pb204Pbfrac1415495823and 208Pb204Pbfrac14 36713164 (nfrac14 61 2 error is reportedas times 104) these values are in excellent agreement withreported TIMS triple-spike values of Galer amp Abouchami(1998) Fractionation-corrected Pb isotopic ratios were fur-ther corrected by the sample-standard bracketing methodor the ln^ln correction method described by White et al(2000) and Blichert-Toft et al (2003) Leached powders ofUSGS reference material BHVO-2 yielded Pb isotopicratios of 206Pb204Pbfrac14186454 8 207Pb204Pbfrac14 1549105 and 208Pb204Pbfrac14 38222514 and BCR-2yielded 206Pb204Pbfrac141880466 207Pb204Pbfrac14 1562518 and 208Pb204Pbfrac14 3883496 (2 error isreported as times 104) These values are in agreement withthose for leached residues of BHVO-2 and BCR-2 fromWeis et al (2006)Hf isotopic compositions were analyzed following the
procedures detailed byWeis et al (2007) The configurationfor Hf analyses monitored Lu mass 175 andYb mass 172 toallow for interference correction to masses 174 and 176 Hfisotopic ratios were normalized internally for mass fractio-nation to a 179Hf177Hf ratio of 07325 using an exponentialcorrection Standards were run after every two samplesand sample results were normalized to the ratio of the in-run daily average and a 176Hf177Hf ratio for JMC-475 of0282160 During the course of analyses the Hf standardJMC-475 gave an average value 02821679 (nfrac14130)USGS reference materials BCR-2 and BHVO-2 were pro-cessed with the samples and yielded Hf isotopic ratios of02828675 and 02831005 respectively Publishedvalues for BCR-2 and BHVO-2 are 02828717 and0283104 8 respectively (Weis et al 2007)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
39
Fig 3 Photographs of picritic and tholeiitic pillowed and massive basalt flows from the Karmutsen Range (Alice and Nimpkish Lake area)and Holberg Inlet (HI) areas northernVancouver Island (locations shown in Fig 1) (a) Massive submarine flow (marked with arrows) drapedover high-MgO pillow basalt of similar composition (b) Picritic pillow basalts in cross-section near Maynard Lake (Note vesicular uppermargin to pillows) (c) Close-up photograph of undulating contact of massive submarine flow draped over high-MgO pillow basalt (markedwith arrows) from photograph (a) (area of photograph indicated with a white box) (d) Photograph of margin of pahoehoe lobe (sledgehammerfor scale) (e) Cross-section of a large picritic pillow lobe with infilling of quartz^carbonate in cooling^contraction cracks (sledgehammer forscale) (f) High-MgO pillow breccia stratigraphically between submarine and subaerial flows (arrow points to lens cap 7 cm diameter forscale)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
6
Plagioclase forms most of the phenocrysts and glomero-crysts and the groundmass is fine-grained comprising pla-gioclase microlites clinopyroxene granules small grains ofFe^Ti oxide devitrified glass and secondary minerals
there is no fresh glass preserved The picritic and high-MgO basaltic lavas are characterized by spherulitic plagi-oclase and clinopyroxene with abundant euhedral olivinephenocrysts (Fig 5 Table 1) Olivine is completely replaced
Fig 4 Photographs showing field relations from the Schoen Lake Provincial Park area northernVancouver Island (a) Sediment^sill complexat the base of the Karmutsen Formation on the north side of Mt Schoen 100m below the Daonella locality (see Fig 2 for location) (b)Interbedded mafic sills and deformed finely banded chert and shale with calcareous horizons from location between Mt Adam and MtSchoen
Fig 5 Photomicrographs of picritic pillow basalts Karmutsen Formation northernVancouver Island Scale bars represent 1mm (a) Picritewith abundant euhedral olivine pseudomorphs from Keogh Lake type locality (Fig 2) in cross-polarized light (XPL) (sample 4722A4 19^198wt MgO four analyses) Samples contain dense clusters (24^42 vol ) of olivine pseudomorphs (52mm) in a groundmass of curvedand branching sheaves of acicular plagioclase and intergrown with clinopyroxene and altered glass In many cases plagioclase nucleated on theedges of the olivine phenocrysts (b) Sheaves of intergrown plagioclase and clinopyroxene in aphyric picrite pillow lava from west of MaynardLake (Fig 2) in XPL (sample 4723A2108 wt MgO)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
7
Table 1 Summary of petrographic characteristics and phenocryst proportions of Karmutsen basalts on Vancouver Island
BC
Sample Area Flow Group Texture vol Ol Plag Cpx Ox Alteration Notes
4718A1 MA PIL THOL intersertal 20 5 3 few plag glcr52mm cpx51 mm
4718A2 MA PIL THOL intersertal glomero 15 1 plag glcr52mm very fresh
4718A5 MA FLO THOL intersertal glomero 10 2 mottled few plag glcr54 mm
4723A4 KR PIL PIC intergranular intersertal 0 3 swtl plag51mm no ol phenos
4723A13 KR PIL PIC spherulitic 24 3 swtl plag51 mm
(continued)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
8
by an assemblage of talc^tremolite^serpentine^opaqueoxides there is no fresh olivine present in the Karmutsenlavas onVancouver Island Plagioclase and clinopyroxenephenocrysts are absent in the high-MgO lavas Thecoarse-grained mafic rocks are characterized by sub-ophi-tic textures with an average grain size typically 41mmand are generally not glomeroporphyritic these rocks aremainly from the interiors of massive flows although somemay represent sillsSample preparation and analytical methods for major-
and trace-element and Sr^Nd^Hf^Pb isotopic composi-tions of the Karmutsen volcanic rocks are given in theAppendix
WHOLE -ROCK CHEMISTRYMajor- and trace-element compositionsThe most abundant type of volcanic rock in theKarmutsen Formation is tholeiitic basalt with a restrictedrange of major- and trace-element compositions The tho-leiitic basalts have similar compositions to the coarse-grained mafic rocks and both groups are distinct fromthe picrites and high-MgO basalts [Fig 6 picrites412wt MgO (LeBas 2000) high-MgO basalts8^12wt MgO] The tholeiitic basalts have lower MgO(57^77wt MgO) and higher TiO2 (14^22wt
TiO2) than the picrites (130^198wt MgO05^07wt TiO2) and high-MgO basalts (91^116wt MgO 05^08wt TiO2 Fig 6 Table 2) Almost allcompositions plot within the tholeiitic field in a total alka-lis vs silica plot although there has been substantial K lossin most samples from the Karmutsen Formation (seebelow) and the tholeiitic basalts generally have higherSiO2 Na2OthornK2O CaO and FeOT (total iron expressedas FeO) than the picrites The tholeiitic basalts also havenoticeably lower loss on ignition (LOI mean1713wt ) than the picrites (mean 54 08wt )and high-MgO basalts (mean 3815wt ) (Fig 6)reflecting the presence of abundant altered olivinephenocrysts in the latter groups Ni concentrations are sig-nificantly higher for the picrites (339^755 ppm) and high-MgO basalts (122^551ppm) than the tholeiitic basalts(58^125 ppm except for one anomalous sample) andcoarse-grained rocks (70^213 ppm) (Fig 6 Table 2) Ananomalous tholeiitic pillowed flow (two samples outlierin Tables 1 and 2) from near the picrite type locality atKeogh Lake and a mineralized sill (disseminated sulfide)from the basal sediment^sill complex at Schoen Lake aredistinguished by higher TiO2 and FeOT than the maingroup of tholeiitic basalts The tholeiitic basalts are similarin major-element composition to previously publishedresults for samples from the Karmutsen Formation how-ever the Keogh Lake picrites which encompass the
Table 1 Continued
Sample Area Flow Group Texture vol Ol Plag Cpx Ox Alteration Notes
4722A5 KR FLO OUTLIER intersertal 3 mottled very fg v small pl needles
5615A11 KR PIL OUTLIER intersertal 3 mottled very fg v small pl needles
5617A4 SL SIL MIN intersertal 5 3 fg
Sample number last digit of year month day initial sample station (except 93G171) MA Mount Arrowsmith SLSchoen Lake KR Karmutsen Range PIL pillow BRE breccia FLO flow SIL sill GAB gabbro THOL tholeiiticbasalt PIC picrite HI-MG high-MgO basalt CGcoarse-grained MIN mineralized sill glomero glomeroporphyriticOlivine modes for picrites are based on area calculations from scans (see lsquoOlivine accumulation in high-MgO and picriticlavasrsquo section and Fig 11) Visual alteration index is based primarily on degree of plagioclase alteration and presence ofsecondary minerals (1 least altered 3 most altered) Plagioclase phenocrysts are commonly altered to albite pumpellyiteand chlorite olivine is altered to talc tremolite and clinochlore (determined using the Rietveld method of X-ray powderdiffraction) clinopyroxene is unaltered FendashTi oxide is commonly replaced by sphene glcr glomerocrysts fg fine-grained cg coarse-grained oik oikocryst chad chadacryst enc enclosing swtl swallow-tail ol olivinepseudomorphs plag plagioclase cpx clinopyroxene ox oxides (includes ilmenite thorn titanomagnetite)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
9
00
05
10
15
20
25
30
35
40
0 5 10 15 20 25
10
11
12
13
14
15
16
17
18
19
20
0 5 10 15 20 25
6
7
8
9
10
11
12
13
14
15
0 5 10 15 20 25
6
7
8
9
10
11
12
13
14
15
16
0 5 10 15 20 25
0
1
2
3
4
5
6
40 45 50 55 60
0
200
400
600
800
0 5 10 15 20 25
(wt)(wt)
(wt)
(wt)
+K2ONa2O TiO2
FeO (T)
Ni (ppm)
Al2O3
CaO
alkalic
tholeiitic
Karmutsen Form
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
(compiled data)
Pillowed flowMineralized sill
SiO2 (wt) MgO (wt)
MgO (wt)MgO (wt)
MgO (wt) MgO (wt)
(a) (b)
(g)
(e)
(f)
(c)
LOI (wt )
(d)
MgO (wt )0
1
2
3
4
5
6
7
0 5 10 15 20 25
(wt)
Fig 6 Whole-rock major-element Ni and LOI variation diagrams for the Karmutsen Formation New samples from this study (Table 2) areshown by symbols with black outlines (see legend) Previously published analyses are shown by small gray symbols without black outlines Theboundary of the alkaline and tholeiitic fields is that of MacDonald amp Katsura (1964) Total iron expressed as FeO LOI loss on ignition oxidesare plotted on an anhydrous normalized basis References for the 322 compiled analyses for the Karmutsen Formation are Surdam (1967)Kuniyoshi (1972) Muller et al (1974) Barker et al (1989) Lassiter et al (1995) Massey (1995a1995b)Yorath et al (1999) and G Nixon (unpublisheddata) It should be noted that the compiled dataset has not been filtered many of the samples with high SiO2 (452wt ) are probably notKarmutsen flood basalts but younger Bonanza basalts and andesites Three pillowed flow samples and a mineralized sill are distinguishedseparately because of their anomalous chemistry Coarse-grained samples are indicated separately Three tholeiitic basalt samples (4724A54718A5 4718A6) with416wt Al2O3 and4270 ppm Sr contain 10^25 modal of large plagioclase phenocrysts (7^8mm) or glomerocrysts(44mm)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
10
Table 2 Major element (wt oxide) and trace element (ppm) abundances in whole-rock samples of Karmutsen basalts
1Ni concentrations for these samples were determined by XRFTHOL tholeiitic basalt PIC picrite HI-MG high MgO basalt CG coarse-grained (sill or gabbro) MIN SIL mineralizedsill OUTLIER anomalous pillowed flow in plots MA Mount Arrowsmith SL Schoen Lake KR Karmutsen Range QIQuadra Island Sample locations are given using the Universal Transverse Mercator (UTM) coordinate system (NAD83zones 9 and 10) Analyses were performed at Activation Laboratory (ActLabs) Fe2O3
is total iron expressed as Fe2O3LOI loss on ignition All major elements Sr V and Y were measured by ICP quadrupole OES on solutions of fusedsamples Cu Ni Pb and Zn were measured by total dilution ICP Cs Ga Ge Hf Nb Rb Ta Th U Zr and REE weremeasured by magnetic-sector ICP on solutions of fused samples Co Cr and Sc were measured by INAA Blanks arebelow detection limit See Electronic Appendices 2 and 3 for complete XRF data and PCIGR trace element datarespectively Major elements for sample 93G17 are normalized anhydrous Sample 5615A7 was analyzed by XRFSamples from Quadra Island are not shown in figures
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
17
picritic and high-MgO basalt pillow lavas extend to20wt MgO (Fig 6 and references listed in thecaption)The tholeiitic basalts from the Karmutsen Range dis-
play a tight range of parallel light rare earth element(LREE)-enriched REE patterns (LaYbCNfrac14 21^24mean 2202) whereas the picrites and high-MgObasalts are characterized by sub-parallel LREE-depleted
patterns (LaYbCNfrac14 03^07 mean 06 02) with lowerREE abundances than the tholeiitic basalts (Fig 7)Tholeiitic basalts from the Schoen Lake and MountArrowsmith areas have similar REE patterns tosamples from the Karmutsen Range (Schoen Lake LaYbCNfrac14 21^26 mean 2303 Mount Arrowsmith LaYbCNfrac1419^25 mean 2204) and the coarse-grainedmafic rocks have similar REE patterns to the tholeiitic
Depleted MORB average(Salters amp Stracke 2004)
Schoen LakeSchoen Lake
Mount Arrowsmith
Karmutsen RangeS
ampl
eC
hond
rite
Sam
ple
Cho
ndrit
e
Sam
ple
Prim
itive
Man
tleS
ampl
eP
rimiti
ve M
antle
Sam
ple
Prim
itive
Man
tle
Mount Arrowsmith
Karmutsen Range
(a) (b)
(c) (d)
(e) (f )
Sam
ple
Cho
ndrit
e
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained mafic rock
Pillowed flowMineralized sill
1
10
100
1
10
100
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
01
1
10
100
1
10
100
1
10
100
Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
Fig 7 Whole-rock REE and other incompatible-element concentrations for the Karmutsen Formation (a) (c) and (e) are chondrite-normal-ized REE patterns for samples from each of the three main field areas onVancouver Island (b) (d) and (f) are primitive mantle-normalizedtrace-element patterns for samples from each of the three main field areas All normalization values are from McDonough amp Sun (1995) Theclear distinction between LREE-enriched tholeiitic basalts and LREE-depleted picrites the effects of alteration on the LILE (especially loss ofK and Rb) and the different range for the vertical scale in (b) (extends down to 01) should be noted
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
18
basalts (LaYbCNfrac14 21^29 mean 2508) Two LREE-depleted coarse-grained mafic rocks from the SchoenLake area have similar REE patterns (LaYbCNfrac14 07) tothe picrites from the Keogh Lake area indicating that thisdistinctive suite of rocks is exposed as far south asSchoen Lake that is 75 km away (Fig 7) The anomalouspillowed flow from the Karmutsen Range has lower LaYbCN (13^18) than the main group of tholeiitic basaltsfrom the Karmutsen Range (Fig 7) The mineralized sillfrom the Schoen Lake area is distinctly LREE-enriched(LaYbCNfrac14 33) with the highest REE abundances(LaCNfrac14 940) of all Karmutsen samples this may reflectcontamination by adjacent sediment also evident in thin-section where sulfide blebs are cored by quartz (Greeneet al 2006)The primitive mantle-normalized trace-element pat-
terns (Fig 7) and trace-element variations and ratios(Fig 8) highlight the differences in trace-element
concentrations between the four main groups ofKarmutsen flood basalts onVancouver Island The tholeii-tic basalts from all three areas have relatively smooth par-allel trace-element patterns some samples have smallpositive Sr anomalies relative to Nd and Sm and manysamples have prominent negative K anomalies relative toU and Nb reflecting K loss during alteration (Fig 7) Thelarge ion lithophile elements (LILE) in the tholeiiticbasalts (mainly Rb and K not Cs) are depleted relativeto the high field strength elements (HFSE) and LREELILE segments especially for the Schoen Lake area(Fig 7d) are remarkably parallel The picrites andhigh-MgO basalts form a tight band of parallel trace-element patterns and are depleted in HFSE and LREEwith positive Sr anomalies and relatively low Rb Thecoarse-grained mafic rocks from the Karmutsen Rangehave identical trace-element patterns to the tholeiiticbasalts Samples from the Schoen Lake area have similar
000
025
050
075
100
125
0 1 2 3 4 5
00
04
08
12
16
0 5 10 15 20 25
MgO (wt)
0
5
10
15
0 50 100 150
Hf (ppm)
Th (ppm)
NbLaNb (ppm)
Zr (ppm)
(a)
(c)
(b)
(d)
Th (ppm)
00
02
04
06
08
10
12
00 01 02 03 04 05
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
NbLa=1
4723A4 5615A12
U (ppm)
Fig 8 Whole-rock trace-element concentrations and ratios for the Karmutsen Formation [except (b) which is vs wt MgO] (a) Nb vs Zr(b) NbLa vs MgO (c) Th vs Hf (d) Th vs U There is a clear distinction between the tholeiitic basalts and picrites in Nb and Zr both inconcentration and the slope of each trend
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
19
trace-element patterns to the Karmutsen Range except forthe two LREE-depleted coarse-grained mafic rocks andthe mineralized sill (Fig 7d)
Sr^Nd^Hf^Pb isotopic compositionsA total of 19 samples from the four main groups ofthe Karmutsen Formation were selected for radiogenicisotopic geochemistry on the basis of major- and trace-element variation stratigraphic position and geographicallocation The samples have overlapping age-correctedHf and Nd isotope ratios and distinct ranges of Sr isotopiccompositions (Fig 9) The tholeiitic basalts picriteshigh-MgO basalt and coarse-grained mafic rocks have
initial eHffrac14thorn87 to thorn126 and eNdfrac14thorn66 to thorn88 cor-rected for in situ radioactive decay since 230 Ma (Fig 9Tables 3 and 4) All the Karmutsen samples form a well-defined linear array in a Lu^Hf isochron diagram corre-sponding to an age of 24115 Ma (Fig 9) This is withinerror of the accepted age of the Karmutsen Formationand indicates that the Hf isotope systematics behaved as aclosed system since c 230 Ma The anomalous pillowedflow has similar initial eHf (thorn98) and slightly lower initialeNd (thorn62) compared with the other Karmutsen samplesThe picrites and high-MgO basalt have higher initial87Sr86Sr (070398^070518) than the tholeiitic basalts(initial 87Sr86Srfrac14 070306^070381) and the coarse-grained mafic rocks (initial 87Sr86Srfrac14 070265^070428)
05128
05129
05130
05131
05132
015 017 019 021 023 025
02829
02830
02831
02832
02833
02834
000 002 004 006 008 010
4
6
8
10
0702 0703 0704 0705 07066
8
10
12
14
5 6 7 8 9 10
12
0
1
2
3
4
5
6
7
0702 0703 0704 0705 0706
(d)
LOI (wt )
87Sr 86Sr
4723A2
Nd
143Nd144Nd
147Sm144Nd
87Sr 86Sr (230 Ma)
(230 Ma)
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
(a)
(c)
176Hf177Hf
Age=241 plusmn 15 Ma=+990 plusmn 064
176Lu177Hf
Hf (230 Ma)
Nd (230 Ma)
(e)
Hf (i)
(b)
Fig 9 Whole-rock Sr Nd and Hf isotopic compositions for the Karmutsen Formation (a) 143Nd144Nd vs 147Sm144Nd (b) 176Hf177Hf vs176Lu177Hf The slope of the best-fit line for all samples corresponds to an age of 24115 Ma (c) Initial eNd vs 87Sr86Sr Age-correction to230 Ma (d) LOI vs 87Sr86Sr (e) Initial eHf vs eNd Average 2 error bars are shown in a corner of each panel Complete chemical duplicatesshown inTables 3 and 4 (samples 4720A7 and 4722A4) are circled in each plot
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
overlap the ranges of the picrites and tholeiitic basalts(Fig 9 Table 3)The measured Pb isotopic compositions of the tholeiitic
basalts are more radiogenic than those of the picrites andthe most magnesian Keogh Lake picrites have the leastradiogenic Pb isotopic compositions The range ofinitial Pb isotope ratios for the picrites is206Pb204Pbfrac1417868^18580 207Pb204Pbfrac1415547^15562and 208Pb204Pbfrac14 37873^38257 and the range for thetholeiitic basalts is 206Pb204Pbfrac1418782^19098207Pb204Pbfrac1415570^15584 and 208Pb204Pbfrac14 37312^38587 (Fig 10 Table 5) The coarse-grained mafic rocksoverlap the range of initial Pb isotopic compositions forthe tholeiitic basalts with 206Pb204Pbfrac1418652^19155207Pb204Pbfrac1415568^15588 and 208Pb204Pbfrac14 38153^38735 One high-MgO basalt has the highest initial206Pb204Pb (19242) and 207Pb204Pb (15590) and thelowest initial 208Pb204Pb (37676) The Pb isotopic ratiosfor Karmutsen samples define broadly linearrelationships in Pb isotope plots The age-corrected Pb
isotopic compositions of several picrites have been affectedby U and Pb (andor Th) mobility during alteration(Fig 10)
ALTERAT IONThe Karmutsen basalts have retained most of their origi-nal igneous structures and textures however secondaryalteration and low-grade metamorphism have producedzeolitic- and prehnite^pumpellyite-bearing mineral assem-blages (Table 1 Cho et al 1987) Alteration and metamor-phism have primarily affected the distribution of the LILEand the Sr isotopic systematics The Keogh Lake picriteshave been affected more by alteration than the tholeiiticbasalts and have higher LOI (up to 55wt ) variableLILE and higher measured Sr Nd and Hf and lowermeasured Pb isotope ratios than the tholeiitic basalts Srand Pb isotope compositions for picrites and tholeiiticbasalts show a relationship with LOI (Figs 9 and 10)whereas Nd and Hf isotopic compositions do not correlate
Table 3 Sr and Nd isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Rb Sr 87Sr86Sr 2m87Rb86Sr 87Sr86Srt Sm Nd 143Nd144Nd 2m eNd
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses carried out at the PCIGR thecomplete trace element analyses are given in Electronic Appendix 3
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
21
with LOI (not shown) The relatively high apparent initialSr isotopic compositions for the high-MgO lavas (up to07052) probably resulted from post-magmatic loss of Rbor an increase in 87Sr86Sr through addition of seawater Sr(eg Hauff et al 2003) High-MgO lavas from theCaribbean plateau have similarly high 87Sr86Sr comparedwith basalts (Recurren villon et al 2002)The correction for in situdecay on initial Pb isotopic ratios has been affected bymobilization of U and Th (and Pb) in the whole-rockssince their formation A thorough acid leaching duringsample preparation was used that has been shown to effec-tively remove alteration phases (Weis et al 2006 NobreSilva et al in preparation) Whole-rock SmNd ratiosand age-correction of Nd isotopes may be affected bythe leaching procedure for submarine rocks older than50 Ma (Thompson et al 2008 Fig 9) there ishowever very little variation in Nd isotopic compositionsof the picrites or tholeiitic basalts and this system does notappear to be disturbed by alteration The HFSE abun-dances for tholeiitic and picritic basalts exhibit clear
linear correlations in binary diagrams (Fig 8) whereasplots of LILE vs HFSE are highly scattered (not shown)because of the mobility of some of the alkali and alkalineearth LILE (Cs Rb Ba K) and Sr for most samplesduring alterationThe degree of alteration in the Karmutsen samples does
not appear to be related to eruption environment or depthof burial Pillow rims are compositionally different frompillow cores (Surdam1967 Kuniyoshi1972) and aquagenetuffs are chemically different from isolated pillows withinpillow breccias however submarine and subaerial basaltsexhibit a similar degree of alterationThere is no clear cor-relation between the submarine and subaerial basalts andsome commonly used chemical alteration indices (eg BaRb vs K2OP2O5 Huang amp Frey 2005) There is also nodefinitive correlation between the petrographic alterationindex (Table 1) and chemical alteration indices although10 of 13 tholeiitic basalts with the highest K and LILEabundances have the highest petrographic alterationindex of three (seeTable 1)
Table 4 Hf isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Lu Hf 176Hf177Hf 2m eHf176Lu177Hf 176Hf177Hft eHf(t)
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
OLIV INE ACCUMULAT ION INH IGH-MgO AND PICR IT IC LAVASGeochemical trends and petrographic characteristics indi-cate that accumulation of olivine played an important rolein the formation of the high-MgO basalts and picrites fromthe Keogh Lake area on northern Vancouver Island TheKeogh Lake samples show a strong linear correlation inplots of Al2O3 TiO2 ScYb and Ni vs MgO and many ofthe picrites have abundant clusters of olivine pseudomorphs(Table1 Fig11)There is a strong linear correlationbetween
modal per cent olivine and whole-rock magnesium contentmagnesium contents range between 10 and 19wt MgOand the proportion of olivine phenocrysts in these samplesvaries from 0 to 42 vol (Fig 11)With a few exceptionsboth the size range and median size of the olivine pheno-crysts in most samples are comparable The clear correla-tion between the proportion of olivine phenocrysts andwhole-rock magnesium contents indicates that accumula-tion of olivine was directly responsible for the high MgOcontents (410wt ) in most of the picritic lavas
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
207Pb204Pb
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
208Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
(a) (b)
(c) (e)
4723A24723A2
4723A2
4723A2
4722A4
4723A4
4722A4
4723A4
4723A134723A3
4723A3
4723A13
1556
1558
1560
1562
1564
1566
1568
186 190 194 198 202 206 2101550
1552
1554
1556
1558
1560
178 180 182 184 186 188 190 192 194
380
384
388
392
396
186 190 194 198 202 206 210374
378
382
386
390
178 180 182 184 186 188 190 192 194
206Pb204Pb(d)
LOI (wt )
0
1
2
3
4
5
6
7
186 190 194 198 202 206 210
4723A2
Fig 10 Pb isotopic compositions of leached whole-rock samples measured by MC-ICP-MS for the Karmutsen Formation Error bars are smal-ler than symbols (a) Measured 207Pb204Pb vs 206Pb204Pb (b) Initial 207Pb204Pb vs 206Pb204Pb Age-correction to 230 Ma (c) Measured208Pb204Pb vs 206Pb204Pb (d) LOI vs 206Pb204Pb (e) Initial 208Pb204Pb vs 206Pb204Pb Complete chemical duplicates shown in Table 5(samples 4720A7 and 4722A4) are circled in each plot The dashed lines in (b) and (e) show the differences in age-corrections for two picrites(4723A4 4722A4) using the measured UTh and Pb concentrations for each sample and the age-corrections when concentrations are used fromthe two picrites (4723A3 4723A13) that appear to be least affected by alteration
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
23
DISCUSS IONThe compositional and spatial development of the eruptedlava sequences of the Wrangellia oceanic plateau onVancouver Island provide information about the evolutionof a mantle plume upwelling beneath oceanic lithospherein an analogous way to the interpretation of continentalflood basalts The Caribbean and Ontong Java plateausare the two primary examples of accreted parts of oceanicplateaus that have been studied to date and comparisonscan be made with the Wrangellia oceanic plateauHowever the Wrangellia oceanic plateau is a distinctiveoccurrence of an oceanic plateau because it formed on thecrust of a Devonian to Mississippian intra-oceanic islandarc overlain by Mississippian to Permian limestones andpelagic sediments The nature and thickness of oceaniclithospheric mantle are considered to play a fundamentalrole in the decompression and evolution of a mantleplume and thus the compositions of the erupted lavasequences (Greene et al 2008) The geochemical and
stratigraphic relationships observed in the KarmutsenFormation onVancouver Island and the focus of the subse-quent discussion sections provide constraints on (1) thetemperature and degree of melting required to generatethe primary picritic magmas (2) the composition of themantle source (3) the depth of melting and residual min-eralogy in the source region (4) the low-pressure evolutionof the Karmutsen tholeiitic basalts (5) the growth of theWrangellia oceanic plateau
Melting conditions and major-elementcomposition of the primary magmasThe primary magmas to most flood basalt provinces arebelieved to be picritic (eg Cox 1980) and they have beenfound in many continental and oceanic flood basaltsequences worldwide (eg Siberia Karoo Paranacurren ^Etendeka Caribbean Deccan Saunders 2005) Near-pri-mary picritic lavas with low total alkali abundances
Table 5 Pb isotopic compositions of Karmutsen basaltsVancouver Island BC
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
require high-degree partial melting of the mantle source(eg Herzberg amp OrsquoHara 2002) The Keogh Lake picritesfrom northernVancouver Island are the best candidates foridentifying the least modified partial melts of the mantleplume source despite having accumulated olivine pheno-crysts and can be used to estimate conditions of meltingand the composition of the primary magmas for theKarmutsen FormationPrimary magma compositions mantle potential tem-
peratures and source melt fractions for the picritic lavas ofthe Karmutsen Formation were calculated from primitivewhole-rock compositions using PRIMELT1XLS software(Herzberg et al 2007) and are presented in Fig 12 andTable 6 A detailed discussion of the computationalmethod has been given by Herzberg amp OrsquoHara (2002)and Herzberg et al (2007) key features of the approachare outlined belowPrimary magma composition is calculated for a primi-
tive lava by incremental addition of olivine and compari-son with primary melts of mantle peridotite Lavas thathad experienced plagioclase andor clinopyroxene fractio-nation are excluded from this analysis All calculated pri-mary magma compositions are assumed to be derived byaccumulated fractional melting of fertile peridotite KR-4003 Uncertainties in fertile peridotite composition donot propagate to significant variations in melt fractionand mantle potential temperature melting of depletedperidotite propagates to calculated melt fractions that aretoo high but with a negligible error in mantle potential
temperature PRIMELT1XLS calculates the olivine liqui-dus temperature at 1atm which should be similar to theactual eruption temperature of the primary magmaand is accurate to 308C at the 2 level of confidenceMantle potential temperature is model dependent witha precision that is similar to the accuracy of the olivineliquidus temperature (C Herzberg personal communica-tion 2008)With regard to the evidence for accumulated olivine in
the Keogh Lake picrites it should not matter whether themodeled lava composition is aphyric or laden with accu-mulated olivine phenocrysts PRIMELT1 computes theprimary magma composition by adding olivine to a lavathat has experienced olivine fractionation in this way itinverts the effects of fractional crystallization The onlyrequirement is that the olivine phenocrysts are not acci-dental or exotic to the lava compositions Support for thisprocedure can be seen in the calculated olivine liquid-line-of-descent shown by the curved arrays with 5 olivineaddition increments (Fig 12c) Fractional crystallization ofolivine from model primary magmas having 85^95wt FeO and 153^175wt MgO will yield arrays of deriva-tive liquids and randomly sorted olivines that in most butnot all cases connect the picrites and high-MgO basalts Anewer version of the PRIMELT software (Herzberg ampAsimow 2008) can perform olivine addition to and sub-traction from lavas with highly variably olivine phenocrystcontents and yields results that converge with those shownin Fig 12
Fig 11 Relationship between abundance of olivine phenocrysts and whole-rock MgO contents for the Keogh Lake picrites (a) Tracings ofolivine grains (gray) in six high-MgO samples made from high resolution (2400 dpi) scans of petrographic thin-sections Scale bars represent10mm sample numbers are indicated at the upper left (b) Modal olivine vs whole-rock MgO Continuous line is the best-fit line for 10 high-MgO samplesThe areal proportion of olivine was calculated from raster images of olivine grains using ImageJ image analysis software whichprovides an acceptable estimation of the modal abundance (eg Chayes 1954)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
25
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
Gray lines = final melting pressure
L + Ol + Opx
07
08
09
Pressure (GPa)
10
3
4
5
6
1
SOLIDUS
01
04
0506
L + Ol
2
03
02
PeridotiteKR-4003
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
Thick black lines = initial melting pressure
Dashed lines = melt fraction
Melt Fraction
2
3 4 5 6
L+Ol+Opx
SolidusSpinel
SolidusGarnet Peridotite
7
L+Ol
0 10 20 30 4040
45
50
55
60
MgO (wt)
SiO2
(wt)
PeridotiteKR-4003
Melt Fraction
0 10 20 305
10
15
CaO(wt)
MgO (wt)
3
4 5 6 7
2
46
2
Peridotite Partial Melts
Thick black line = initial melting pressureGray lines = final melting pressure
Peridotite
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
02
Pressure (GPa)
0302
04
Pressure (GPa)
Picrite
High-MgO basalt
Tholeiitic basalt
(c)
(d)
Karmutsen flood basaltsOlivine addition modelOlivine addition (5 increment)Primary magma
PicriteHigh-MgO basaltTholeiitic basalt
(a)
(b)
800 850 900
910
920
930
940
950
Karmutsen flood basalts
Olivine addition modelOlivine addition (5 increment)Primary magma
Gray lines = Fo contents of olivine
Fig 12 Estimated primary magma compositions for three Keogh Lake picrites and high-MgO basalts (samples 4723A4 4723A135616A7) using the forward and inverse modeling technique of Herzberg et al (2007) Karmutsen compositions and modeling results areoverlain on diagrams provided by C Herzberg (a) Whole-rock FeO vs MgO for Karmutsen samples from this study Total iron estimatedto be FeO is 090 Gray lines show olivine compositions that would precipitate from liquid of a given MgO^FeO composition Black lineswith crosses show results of olivine addition (inverse model) using PRIMELT1 (Herzberg et al 2007) Gray field highlights olivinecompositions with Fo contents of 91^92 predicted to precipitate (b) FeO vs MgO showing Karmutsen lava compositions and results ofa forward model for accumulated fractional melting of fertile peridotite KR-4003 (Kettle River Peridotite Walter 1998) (c) SiO2 vsMgO for Karmtusen lavas and model results (d) CaO vs MgO showing Karmtusen lava compositions and model results To brieflysummarize the technique [see Herzberg et al (2007) for complete description] potential parental magma compositions for the high-MgOlava series were selected (highest MgO and appropriate CaO) and using PRIMELT1 software olivine was incrementally added tothe selected compositions to show an array of potential primary magma compositions (inverse model) Then using PRIMELT1 theresults from the inverse model were compared with a range of accumulated fractional melts for fertile peridotite derived fromparameterization of the experimental results for KR-4003 from Walter (1998) forward model Herzberg amp OrsquoHara (2002) A melt fractionwas sought that was unique to both the inverse and forward models (Herzberg et al 2007) A unique solution was found when there was a com-mon melt fraction for both models in FeO^MgO and CaO^MgO^Al2O3^SiO2 (CMAS) projection space This modeling assumes thatolivine was the only phase crystallizing and ignores chromite precipitation and possible augite fractionation in the mantle (Herzbergamp OrsquoHara 2002) Results are best for a residue of spinel lherzolite (not pyroxenite) The presence of accumulated olivine in samples ofthe Keogh Lake picrites used as starting compositions does not significantly affect the results because we are modeling addition of olivine(see text for further description) The tholeiitic basalts cannot be used for modeling because they are all saturated in plagthorn cpxthornolThick black line for initial melting pressure at 4 GPa is not shown in (c) for clarity Gray area in (a) indicates parental magmas thatwill crystallize olivine with Fo 91^92 at the surface Dashed lines in (c) indicate melt fraction Solidi for spinel and garnet peridotite arelabelled in (c)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
The PRIMELT1 results indicate that if the Karmutsenpicritic lavas were derived from accumulated fractionalmelting of fertile peridotite the primary magmas wouldhave contained 15^17wt MgO and 10wt CaO(Fig 12 Table 6) formed from 23^27 partial meltingand would have crystallized olivine with a Fo content of 91 (Fig 12 Table 6) Calculated mantle temperatures forthe Karmutsen picrites (14908C) indicate melting ofanomalously hot mantle (100^2008C above ambient) com-pared with ambient mantle that produces mid-ocean ridgebasalt (MORB 1280^14008C Herzberg et al 2007)which initiated between 3 and 4 GPa Several picritesappear to be near-primary melts and required very littleaddition of olivine (eg sample 93G171 with measured158wt MgO) These temperature and melting esti-mates of the Karmutsen picrites are consistent withdecompression of hot mantle peridotite in an actively con-vecting plume head (ie plume initiation model) Threesamples (picrite 5614A1 and high-MgO basalts 5614A3and 5614A5) are lower in iron than the other Karmutsenlavas with FeO in the 65^80wt range (Fig 12c) andhave MgO and FeO contents that are similar to primitiveMORB from the Sequeiros fracture zone of the EastPacific Rise (EPR Perfit et al 1996) These samples
however differ from EPR MORB in having higher SiO2
and lower Al2O3 and they are restricted geographicallyto one area (Sara Lake Fig 2)We therefore have evidence for primary magmas with
MgO contents that range from a possible low of 110wt to as high as 174wt with inferred mantle potential tem-peratures from 1340 to 15208C This is similar to the rangedisplayed by lavas from the Ontong Java Plateau deter-mined by Herzberg (2004) who found that primarymagma compositions assuming a peridotite sourcewould contain 17wt MgO and 10wt CaO(Table 6) and that these magmas would have formed by27 melting at a mantle potential temperature of15008C (first melting occurring at 36 GPa) Fitton ampGodard (2004) estimated 30 partial melting for theOntong Java lavas based on the Zr contents of primarymagmas of Kroenke-type basalt calculated by incrementaladdition of equilibrium olivine However if a considerableamount of eclogite was involved in the formation of theOntong Java plateau magmas the excess temperatures esti-mated by Herzberg et al (2007) may not be required(Korenaga 2005) The variability in mantle potential tem-perature for the Keogh Lake picrites is also similar to thewide range of temperatures that have been calculated for
Table 6 Estimated primary magma compositions for Karmutsen basalts and other oceanic plateaus or islands
Sample 4723A4 4723A13 5616A7 93G171 Average OJP Mauna Keay Gorgonaz
Ontong Java primary magma composition for accumulated fraction melting (AFM) from Herzberg (2004)yMauna Kea primary magma composition is average of four samples of Herzberg (2006 table 1)zGorgona primary magma composition for AFM for 1F fertile source from Herzberg amp OrsquoHara (2002 table 4)Total iron estimated to be FeO is 090
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
27
lavas from single volcanoes in the Galapagos Hawaii andelsewhere by Herzberg amp Asimow (2008) Those workersinterpreted the range to reflect the transport of magmasfrom both the cool periphery and the hotter axis of aplume A thermally heterogeneous plume will yield pri-mary magmas with different MgO contents and the rangeof primary magma compositions and their inferred mantlepotential temperatures for Karmutsen lavas may reflectmelting from different parts of the plume
Source of the Karmutsen Formation lavasThe Sr^Nd^Hf^Pb isotopic compositions of theKarmutsen volcanic rocks on Vancouver Island indicatethat they formed from plume-type mantle containing adepleted component that is compositionally similar to butdistinct from other oceanic plateaus that formed in thePacific Ocean (eg Ontong Java and Caribbean Fig 13)Trace element and isotopic constraints can be used to testwhether the depleted component of the KarmutsenFormation was intrinsic to the mantle plume and distinctfrom the source of MORB or the result of mixing or invol-vement of ambient depleted MORB mantle with plume-type mantle The Karmutsen tholeiitic basalts have initialeNd and
87Sr86Sr ratios that fall within the range of basaltsfrom the Caribbean Plateau (eg Kerr et al 1996 1997Hauff et al 2000 Kerr 2003) and a range of Hawaiian iso-tope data (eg Frey et al 2005) and lie within and abovethe range for the Ontong Java Plateau (Tejada et al 2004Fig 13a) Karmutsen tholeiitic basalts with the highest ini-tial eNd and lowest initial 87Sr86Sr lie within and justbelow the field for age-corrected northern EPR MORB at230 Ma (eg Niu et al1999 Regelous et al1999) Initial Hfand Nd isotopic compositions for the Karmutsen samplesare noticeably offset to the low side of the ocean islandbasalt (OIB) array (Vervoort et al 1999) and lie belowand slightly overlap the low eHf end of fields for Hawaiiand the Caribbean Plateau (Fig 13c) the Karmutsen sam-ples mostly fall between and slightly overlap a field forage-corrected Pacific MORB at 230 Ma and Ontong Java(Mahoney et al 1992 1994 Nowell et al 1998 Chauvel ampBlichert-Toft 2001) Karmutsen lava Pb isotope ratios over-lap the broad range for the Caribbean Plateau and thelinear trends for the Karmutsen samples intersect thefields for EPR MORB Hawaii and Ontong Java in208Pb^206Pb space but do not intersect these fields in207Pb^206Pb space (Fig 13d and e) The more radiogenic207Pb204Pb for a given 206Pb204Pb of the Karmutsenbasalts requires high UPb ratios at an ancient time when235U was abundant similar to the Caribbean Plateau andenriched in comparison with EPR MORB Hawaii andOntong JavaThe low eHf^low eNd component of the Karmutsen
basalts that places them below the OIB mantle array andpartly within the field for age-corrected Pacific MORBcan form from low ratios of LuHf and high SmNd over
time (eg Salters amp White 1998) MORB will develop aHf^Nd isotopic signature over time that falls below themantle array Low LuHf can also lead to low 176Hf177Hffrom remelting of residues produced by melting in theabsence of garnet (eg Salters amp Hart 1991 Salters ampWhite 1998) The Hf^Nd isotopic compositions of theKarmutsen Formation indicate the possible involvementof depleted MORB mantle however similar to Iceland(Fitton et al 2003) Nb^Zr^Y variations provide a way todistinguish between a depleted component within theplume and a MORB-like component The Karmutsenbasalts and picrites fall within the Iceland array and donot overlap the MORB field in a logarithmic plot of NbYvs ZrY (Fig 13b) using the fields established by Fittonet al (1997) Similar to the depleted component within theCaribbean Plateau (Thompson et al 2003) the trend ofHf^Nd isotopic compositions towards Pacific MORB-likecompositions is not followed by MORB values in Nb^Zr^Y systematics and this suggests that the depleted compo-nent in the source of the Karmutsen basalts is distinctfrom the source of MORB and probably an intrinsic partof the plume The relatively radiogenic Pb isotopic ratiosand REE patterns distinct from MORB also support thisinterpretationTrace-element characteristics suggest the possible invol-
vement of sub-arc lithospheric mantle in the generation ofthe Karmutsen picrites Lassiter et al (1995) suggested thatmixing of a plume-type source with eNdthorn 6 to thorn 7 witharc material with low NbTh could reproduce the varia-tions observed in the Karmutsen basalts however theyconcluded that the absence of low NbLa ratios in theirsample of tholeiitic basalts restricts the amount of arc litho-sphere involved The study of Lassiter et al (1995) wasbased on major and trace element and Sr Nd and Pb iso-topic data for a suite of 29 samples from Buttle Lake in theStrathcona Provincial Park on central Vancouver Island(Fig 1) Whereas there are no clear HFSE depletions inthe Karmutsen tholeiitic basalts the low NbLa (and TaLa) of the Karmutsen picritic lavas in this study indicatethe possible involvement of Paleozoic arc lithosphere onVancouver Island (Fig 8)
REE modeling dynamic melting andsource mineralogyThe combined isotopic and trace-element variations of thepicritic and tholeiitic Karmutsen lavas indicate that differ-ences in trace elements may have originated from differentmelting histories For example the LuHf ratios of theKarmutsen picritic lavas (mean 008) are considerablyhigher than those in the tholeiitic basalts (mean 002)however the small range of overlapping initial eHf valuesfor the two lava suites suggests that these differences devel-oped during melting within the plume (Fig 9b) Below wemodel the effects of source mineralogy and depth of
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
28
melting on differences in trace-element concentrations inthe tholeiitic basalts and picritesDynamic melting models simulate progressive decom-
pression melting where some of the melt fraction isretained in the residue and only when the degree of partialmelting is greater than the critical mass porosity and the
source becomes permeable is excess melt extracted fromthe residue (Zou1998) For the Karmutsen lavas evolutionof trace element concentrations was simulated using theincongruent dynamic melting model developed by Zou ampReid (2001) Melting of the mantle at pressures greater than1 GPa involves incongruent melting (eg Longhi 2002)
OIB array
PacificMORB(230 Ma)
(0 Ma)
EPR(230 Ma)
-4
-2
0
2
4
8
10
12
14
16
18
20
22
-4 -2 0 2 4 6 8 10 12 14
minus2
0
2
4
6
8
10
12
14
0702 0703 0704 0705 0706
IndianMORB
87Sr 86Sr (initial)
Hf (initial)
(a) (c)
Nd (initial)
Karmutsen tholeiitic basalt
HawaiiOntong JavaCaribbean Plateau
Karmutsen high-MgO lava
high-MgO lavasKarmutsen
1540
1545
1550
1555
1560
1565
175 180 185 190 195 200375
380
385
390
395
175 180 185 190 195 200
(d) (e)
EPR
EPR
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb
Karmutsen Formation (initial)
HawaiiOntong JavaCaribbean Plateau
East Pacific Rise
Karmutsen Formation (measured)
Juan de FucaGorda
ε Nd
(initial)
Karmutsen Formation (different age-correction for 3 picrites)
(0 Ma)
NbY
001
01
1
1 10
ZrY
OIB
NMORB
(b)
Iceland array
6
Fig 13 Comparison of age-corrected (230 Ma) Sr^Nd^Hf^Pb isotopic compositions for Karmutsen flood basalts onVancouver Island withage-corrected OIB and MORB and Nb^Zr^Y variation (a) Initial eNd vs 87Sr86Sr (b) NbYand ZrY variation (after Fitton et al 1997)(c) Initial eHf vs eNd (d) Measured and initial 207Pb204Pb vs 206Pb204Pb (e) Measured and initial 208Pb204Pb vs 206Pb204Pb References fordata sources are listed in Electronic Appendix 4 Most of the compiled data were extracted from the GEOROC database (httpgeorocmpch-mainzgwdgdegeoroc) OIB array line in (c) is fromVervoort et al (1999) EPR East Pacific Rise Several high-MgO samples affected bysecondary alteration were not plotted for clarity in Pb isotope plots Estimation of fields for age-corrected Pacific MORB and EPR at 230 Mawere made using parent^daughter concentrations (in ppm) from Salters amp Stracke (2004) of Lufrac14 0063 Hffrac14 0202 Smfrac14 027 Ndfrac14 071Rbfrac14 0086 and Srfrac14 836 In the NbYvs ZrYplot the normal (N)-MORB and OIB fields not including Icelandic OIB are from data com-pilations of Fitton et al (1997 2003) The OIB field is data from 780 basalts (MgO45wt ) from major ocean islands given by Fitton et al(2003) The N-MORB field is based on data from the Reykjanes Ridge EPR and Southwest Indian Ridge from Fitton et al (1997) The twoparallel gray lines in (b) (Iceland array) are the limits of data for Icelandic rocks
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
29
with olivine being produced during the reactioncpxthornopxthorn spfrac14meltthornol for spinel peridotites The mod-eling used coefficients for incongruent melting reactionsbased on experiments on lherzolite melting (eg Kinzleramp Grove 1992 Longhi 2002) and was separated into (1)melting of spinel lherzolite (2) two combined intervals ofmelting for a garnet lherzolite (gt lherz) and spinel lher-zolite (sp lherz) source (3) melting of garnet lherzoliteThe model solutions are non-unique and a range indegree of melting partition coefficients melt reactioncoefficients and proportion of mixed source componentsis possible to achieve acceptable solutions (see Fig 14 cap-tion for description of modeling parameters anduncertainties)The LREE-depleted high-MgO lavas require melting of
a depleted spinel lherzolite source (Fig 14a) The modelingresults indicate a high degree of melting (22^25) similarto the results from PRIMELT1 based on major-elementvariations (discussed above) of a LREE-depleted sourceThe high-MgO lavas lack a residual garnet signature Theenriched tholeiitic basalts involved melting of both garnetand spinel lherzolite and represent aggregate melts pro-duced from continuous melting throughout the depth ofthe melting column (Fig 14b) Trace-element concentra-tions in the tholeiitic and picritic lavas are not consistentwith low-degree melting of garnet lherzolite alone(Fig 14c) The modeling results indicate that the aggregatemelts that formed the tholeiitic basalts would haveinvolved lesser proportions of enriched small-degreemelts (1^6 melting) generated at depth and greateramounts of depleted high-degree melts (12^25) gener-ated at lower pressure (a ratio of garnet lherzolite tospinel lherzolite melt of 14 provides the best fits seeFig 14b and description in figure caption) The totaldegree of melting for the tholeiitic basalts was high(23^27) similar to the degree of partial melting in thespinel lherzolite facies for the primary melts that eruptedas the Keogh Lake picrites of a source that was also prob-ably depleted in LREE
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
(c)
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
Melting of garnet lherzolite
High-MgO lavas
Melting of spinel lherzolite
source (07DM+03PM)
source (07DM+03PM)
6 melting
30 melting
(b)
(a)
30 melting
Sam
ple
Cho
ndrit
eS
ampl
eC
hond
rite 20 melting
1 melting
Melting of garnet lherzoliteand spinel lherzolite
x gt lherz + 4x sp lherz
5 meltingx gt lherz + 4x sp lherz
Sam
ple
Cho
ndrit
e
Tholeiitic lavas
Tholeiitic lavas
High-MgO lavas
source (07DM+03PM)
Fig 14 Trace-element modeling results for incongruent dynamicmantle melting for picritic and tholeiitic lavas from the KarmutsenFormation Three steps of modeling are shown (a) melting of spinellherzolite compared with high-MgO lavas (b) melting of garnet lher-zolite and spinel lherzolite compared with tholeiitic lavas (c) meltingof garnet lherzolite compared with tholeiitic and picritic lavasShaded fields in each panel for tholeiitic basalts and picrites arelabeled Patterns with symbols in each panel are modeling results(patterns are 5 melting increments in (a) and (b) and 1 incre-ments in (c) PM primitive mantle DM depleted MORB mantle gtlherz garnet lherzolite sp lherz spinel lherzolite In (b) the ratios ofper cent melting for garnet and spinel lherzolite are indicated (x gtlherzthorn 4x sp lherz means that for 5 melting 1 melt in garnetfacies and 4 melt in spinel facies where x is per cent melting in theinterval of modeling) The ratio of the respective melt proportions inthe aggregate melt (x gt lherzthorn 4x sp lherz) was kept constant andthis ratio of melt contribution provided the best fit to the data Arange of proportion of source components (07DMthorn 03PM to
09DMthorn 01PM) can reproduce the variation in the high-MgOlavas Melting modeling uses the formulation of Zou amp Reid (2001)an example calculation is shown in their Appendix PM fromMcDonough amp Sun (1995) and DM from Salters amp Stracke (2004)Melt reaction coefficients of spinel lherzolite from Kinzler amp Grove(1992) and garnet lherzolite fromWalter (1998) Partition coefficientsfrom Salters amp Stracke (2004) and Shaw (2000) were kept constantSource mineralogy for spinel lherzolite (018cpx027opx052ol003sp) from Kinzler (1997) and for garnet lherzolite(034cpx008opx053ol005gt) from Salters amp Stracke (2004) Arange of source compositions for melting of spinel lherzolite(07DMthorn 03PM to 03DMthorn 07PM) reproduce REE patternssimilar to those of the high-MgO lavas with best fits for 22ndash25melting and 07DMthorn 03PM source composition Concentrationsof tholeiitic basalts cannot be reproduced from an entirelyprimitive or depleted source
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
30
The uncertainty in the composition of the source in thespinel lherzolite melting model for the high-MgO lavasprecludes determining whether the source was moredepleted in incompatible trace elements than the source ofthe tholeiitic lavas (see Fig 14 caption for description) It ispossible that early high-pressure low-degree melting left aresidue within parts of the plume (possibly the hotter inte-rior portion) that was depleted in trace elements (egElliott et al 1991) further decompression and high-degreemelting of such depleted regions could then have generatedthe depleted high-MgO Karmutsen lavas Shallow high-degree melting would preferentially sample depletedregions whereas deeper low-degree melting may notsample more refractory depleted regions (eg CaribbeanEscuder-Viruete et al 2007) The picrites lie at the highend of the range of Nd isotopic compositions and havelow NbZr similar to depleted Icelandic basalts (Fittonet al 2003) therefore the picrites may represent the partsof the mantle plume that were the most depleted prior tothe initiation of melting As Fitton et al (2003) suggestedthese depleted melts may only rarely be erupted withoutcombining with more voluminous less depleted melts andthey may be detected only as undifferentiated small-volume flows like the Keogh Lake picrites within theKarmutsen Formation on Vancouver IslandDecompression melting within the mantle plume initiatedwithin the garnet stability field (35^25 GPa) at highmantle potential temperatures (Tp414508C) and pro-ceeded beneath oceanic arc lithosphere within the spinelstability field (25^09 GPa) where more extensivedegrees of melting could occur
Magmatic evolution of Karmutsentholeiitic basaltsPrimary magmas in large igneous provinces leave exten-sive coarse-grained cumulate residues within or below thecrust these mafic and ultramafic plutonic sequences repre-sent a significant proportion of the original magma thatpartially crystallized at depth (eg Farnetani et al 1996)For example seismic and petrological studies of OntongJava (eg Farnetani et al 1996) combined with the use ofMELTS (Ghiorso amp Sack 1995) indicate that the volcanicsequence may represent only 40 of the total magmareaching the Moho Neal et al (1997) estimated that30^45 fractional crystallization took place for theOntong Java magmas and produced cumulate thicknessesof 7^14 km corresponding to 11^22 km of flood basaltsdepending on the presence of pre-existing oceanic crustand the total crustal thickness (25^35 km) Interpretationsof seismic velocity measurements suggest the presence ofpyroxene and gabbroic cumulates 9^16 km thick beneaththe Ontong Java Plateau (eg Hussong et al 1979)Wrangellia is characterized by high crustal velocities in
seismic refraction lines which is indicative of mafic
plutonic rocks extending to depth beneath VancouverIsland (Clowes et al 1995)Wrangellia crust is 25^30 kmthick beneath Vancouver Island and is underlain by astrongly reflective zone of high velocity and density thathas been interpreted as a major shear zone where lowerWrangellia lithosphere was detached (Clowes et al 1995)The vast majority of the Karmutsen flows are evolved tho-leiitic lavas (eg low MgO high FeOMgO) indicating animportant role for partial crystallization of melts at lowpressure and a significant portion of the crust beneathVancouver Island has seismic properties that are consistentwith crystalline residues from partially crystallizedKarmutsen magmas To test the proportion of the primarymagma that fractionated within the crust MELTS(Ghiorso amp Sack 1995) was used to simulate fractionalcrystallization using several estimated primary magmacompositions from the PRIMELT1 modeling results Themajor-element composition of the primary magmas couldnot be estimated for the tholeiitic basalts because of exten-sive plagthorn cpxthornol fractionation and thus the MELTSmodeling of the picrites is used as a proxy for the evolutionof major elements in the volcanic sequence A pressure of 1kbar was used with variable water contents (anhydrousand 02wt H2O) calculated at the quartz^fayalite^magnetite (QFM) oxygen buffer Experimental resultsfrom Ontong Java indicate that most crystallizationoccurs in magma chambers shallower than 6 km deep(52 kbar Sano amp Yamashita 2004) however some crys-tallization may take place at greater depths (3^5 kbarFarnetani et al 1996)A selection of the MELTS results are shown in Fig 15
Olivine and spinel crystallize at high temperature until15^20wt of the liquid mass has fractionated and theresidual magma contains 9^10wt MgO (Fig 15f) At1235^12258C plagioclase begins crystallizing and between1235 and 11908C olivine ceases crystallizing and clinopyr-oxene saturates (Fig 15f) As expected the addition of clin-opyroxene and plagioclase to the crystallization sequencecauses substantial changes in the composition of the resid-ual liquid FeO and TiO2 increase and Al2O3 and CaOdecrease while the decrease in MgO lessens considerably(Fig 15) The near-vertical trends for the Karmutsen tho-leiitic basalts especially for CaO do not match the calcu-lated liquid-line-of-descent very well which misses manyof the high-CaO high-Al2O3 low-FeO basalts A positivecorrelation between Al2O3CaO and SrNd indicates thataccumulation of plagioclase played a major role in the evo-lution of the tholeiitic basalts (Fig 15g) Increasing thecrystallization pressure slightly and the water content ofthe melts does not systematically improve the match ofthe MELTS models to the observed dataThe MELTS results indicate that a significant propor-
tion of the crystallization takes place before plagioclase(15^20 crystallization) and clinopyroxene (35^45
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
31
crystallization) join the crystallization sequence Theseresults suggest that445 crystallization of the primarymagmas occurred (estimated from the residual liquid percent in Fig 15f) and if the original flood basalt stratigra-phy on Vancouver Island was 6 km thick at least anequivalent amount of complementary plutonic rocksshould have crystallized The Karmutsen basalts and theirplutonic residues thus represent a significant addition ofcrust (perhaps 412 km thickness) previously thickenedduring the formation of the underlying Paleozoic arc
Growth of the Wrangellia oceanic plateauon Vancouver IslandThe extrusive part of the Wrangellia oceanic plateau onVancouver Island was constructed in a three-layersequence of mostly tholeiitic basalts with a restrictedrange of composition (Fig 16) The plateau overliesDevonian to Mississippian arc volcanic sequences andmarine sediments of Mississippian to Permian ages thatare intruded by gabbroic rocks related to the flood basaltsThe volcanic stratigraphy of theWrangellia plateau formed
00
05
10
15
20
25
30
35
40
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
16
0 2 4 6 8 10 12 14 16 18 20
10
11
12
13
14
15
16
17
18
19
20
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
0 2 4 6 8 10 12 14 16 18 20
MgO (wt)
0 10 20 30 40 50
percent of total mass
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
) Mineral proportions
Sp (e)
Ol
CpxPlag
PicriteHigh-MgO basaltTholeiitic basalt
TiO2 (wt)
MgO (wt)
Al2O3 (wt) CaO (wt)
FeO (wt)
MgO(a) (b)
(c) (d)
MgO (wt)MgO (wt)
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
)
Residual liquid ()
1220
1190
Ol + Sp
Plag+Ol+Sp Plag+Cpx+Sp
02 wt H2O
no H2O
Residual liquid and crystallizing phases
(f)
0102030405060708090100
plagioclaseaccumulation
plagioclaseaccumulation
plagioclase
accumulatio
n
10
12
14
16
0 10 20 30 40
Al2O3CaO
SrNd
tholeiitic basalts
(g)
Fig 15 Forward fractional crystallization modeling results for major elements from MELTS (Ghiorso amp Sack 1995) compared with picriticand tholeiitic lavas from the Karmutsen Formation (a^d) Major-element compositions of residual liquids compared with Karmutsen lavacompositions For clarity the results are shown only for one starting composition (open circle) using the composition of the estimated primarymagma for the high-MgO lava series from the modeling technique of Herzberg et al (2007) for sample 4723A13 as shown in Fig 12 The esti-mated primary magmas for samples 4723A3 and 93G171 yield similar results A pressure of 1 kbar was used with no H2O and 02wt H2Ocalculated at QFM black line is anhydrous and grey line is 02wt H2OThe end-point for the anhydrous result is at 78 fractionation and11308C and the end-point for the result with 02wt H2O is 75 fractionation and11108C (e) Mineral proportions of crystallized phases fromone MELTrun (sample 4723A13) with no H2O (f) Per cent residual liquid vs temperature labelled with intervals of crystallizing phases (g)Al2O3CaO vs SrNd for the tholeiitic basalts
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
32
in a single eruptive phase (c 231^225 Ma) as an emergentbasalt sequence during a deeper-water shallow-water andsubaerial stage and subsided below sea level relativelysoon after its formation (Fig 16) The deeper-water stagecontains pillowed and massive submarine flows (3000mthick) that built up the submarine volcanic edifice from theseafloor The lowest pillow basalts were emplaced onunconsolidated and lithified fine-grained siliciclastic andcarbonaceous sediments containing Daonella beds(c 235^232 Ma) that are intruded by abundantKarmutsen mafic sills (Fig 16a) The different pillowedand massive submarine flows resulted primarily from dif-ferent effusive rates and local topography Some of themassive submarine flows commonly recognized by radiat-ing columnar jointing represent master tubes for the deliv-ery of lava to distal parts of submarine flow fields Thepillowed flows which commonly have uneven topography
in modern sequences may have served as levees to themassive submarine flows rather than allowing unob-structed spreading and inflation typical of the massive sub-aerial flows that form on shallow slopes in continentalflood basalt provinces (eg Self et al 1997) The rapid suc-cession of flows is evident by the near absence of sedimentbetween submarine flows Minor volumes of picritic pillowlavas occur in the upper part of the pillowed flow sequencein areas of the plateau exposed on northern VancouverIsland (Fig 16 Nixon et al 2008)The growth of the plateau in shallow water led to
emplacement of an increasing proportion of volcaniclastics(pillow breccias and hyaloclastites) that accumulated up to1500m thick in some areas (Fig 16b Nixon et al 2008)Horizons of pillowed flows occur locally within the volca-niclastic stratigraphy Sedimentary structures (eg gradedbedding fluidization structures) and petrographic textures
Late Devonian-Mississippian arc
Daonella beds
Middle Triassic
deeper water
Late Devonian-Mississippian arc
pillowed flows massive submarine
flows
pillow brecciahyaloclastiteshallow water
~4 km
~6 km
Late Triassic limestoneinterflow limestonelenses
Karmutsen Formation
mafic intrusions sediment-sill complex
Karmutsen Formation
(a) (b)
(c) (d)
subaerial flows
Late Devonian-Mississippian arc Late Devonian-Mississippian arc
picritic pillow basalts
Miss to Permian marine sediments
sills
Fig 16 Growth of the Wrangellia oceanic plateau onVancouver Island (a) Deep-water eruptive phase of pillowed and massive submarineflows with almost no intervening sediments onto MiddleTriassic shale containing Daonella beds (c 235^232 Ma) The basement of the plateauis composed of Devonian to Mississippian arc sequences overlain by Mississippian to Permian carbonates and siliciclastic sediments(b) Shallow-water flow emplacement with increasing proportion of volcaniclastics resulted primarily from cooling^contraction granulationPicritic pillow basalts occur primarily near the transition between pillowed submarine flows and volcaniclastic units Pillowed and massiveflows occur locally within the volcaniclastic unit (c) Emergent stage of growth consists of overlapping subaerial flow fields similar to continen-tal flood basalts with pahoehoe flow structures Multi-tiered columnar jointing occurrs locally in response to ponding of flows (eg Bondre et al2004) Interflow carbonate deposits and plagioclase megacrystic flows developed locally during the waning stage of volcanism (d) Post-volcanicsubsidence and deposition of Late Carnian to Norian (c 212^223 Ma) platform carbonates on top of the plateau Units are labeled in the dia-grams Thicknesses of the underlying units are not to scale
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
33
within the volcaniclastic rocks indicate formation primar-ily via cooling^contraction granulation magma^water^steam interaction autobrecciation resedimentation andmass-wasting rather than pyroclastic eruption (Nixonet al 2008)A broad subaerial platform was constructed by emplace-
ment of subaerial lava flows during an emergent subaerialstage (Fig 16c) The subaerial unit predominantly com-prises well-layered sheet flows that formed in a similarway to continental flood basalts as hundred(s) of pahoehoeflow fields erupted on shallow slopes (Self et al 1997) how-ever volcaniclastic and pillowed flow units are locally pre-sent mostly in the upper parts of the stratigraphyLocalized ponding of flows led to the formation of multi-tiered columnar jointing in some subaerial flows (egBondre et al 2004) As volcanism waned local interflowcarbonate deposits developed and plagioclase megacrysticflows were erupted Subsidence of the plateau is recordedby the interflow lenses and deposits of platform carbonates(40^1500m thick) on top of the Karmutsen basalts thatcontain Late Carnian to Early Norian fossils (Fig 16d c216^227 Ma Nixon et al 2008)The evolution of the internal architecture of the
Wrangellia oceanic plateau was primarily related to theconditions of emplacement (ie water depth and eruptionenvironment) There does not appear to have been a sys-tematic coincident evolution or shift of lava compositionsonVancouver Island as has been recognized in most con-tinental flood basalts and parts of the Wrangellia oceanicplateau in Alaska and Yukon (Greene et al 2008 2009)In general with the exception of the picritic pillow lavasin the submarine stratigraphy and local variations betweenhigh-MgO and tholeiitic lavas there are no significantcompositional differences between lavas erupted duringthe different stages of growth of the plateau (ie submarineand subaerial flows) nor are substantial compositional dif-ferences generally found between massive submarine flowsin contact with pillowed flows tholeiitic basalt flows nearthe base of the Karmutsen Formation are compositionallysimilar to flows sampled near the top of the volcanicsequence Where it was possible to sample continuous sec-tions of volcanic stratigraphy greater than several hun-dreds of meters (eg Keogh^Maynard Lake area west ofthe Karmutsen Range Fig 2) there is no clear relationshipbetween stratigraphic position and compositionalvariation
CONCLUSIONSThe Karmutsen Formation covers extensive areas ofVancouver Island (20 000 km2) and forms part of a largeoceanic plateau erupted over a geologically short interval(c 225^231 Ma) The tripartite volcanic stratigraphyexposed on Vancouver Island is at least 6 km thick andcomprises a succession of submarine flows volcaniclastic
deposits and massive subaerial flowsThese volcanologicaldifferences are primarily related to the eruption environ-ment (deep-water shallow-water or subaerial) The rapidgrowth of the plateau prevented significant accumulationsof intervening sediment except in the uppermost stratigra-phy where isolated limestone lenses commonly associatedwith pillow basalt and hyaloclastite deposits preserve arecord of the final stages of subsidence of the plateauThe Wrangellia plateau on Vancouver Island was con-
structed dominantly of tholeiitic basalt with restrictedmajor- and trace-element and isotopic compositional var-iations Minor volumes of picritic pillow basalts eruptedlate in the submarine phase The Karmutsen basalts wereemplaced onto older Paleozoic arc volcanic sequences andlow NbLa for the picrites may indicate a contributionfrom the sub-arc lithospheric mantle Modeling of the pet-rogenesis of the Karmutsen picrites indicates melting ofanomalously hot mantle (15008C) and extensive degreesof partial melting (23^27) and is consistent with aplume initiation model Combined Sr^Nd^Hf^Pb isotopicsystematics indicate that the picritic and tholeiitic lavaswere derived from a depleted plume-type mantle sourcedistinct from the source of MORB that has compositionalsimilarities to the source of the lavas forming theCaribbean Plateau Picrites were generated by melting ofspinel lherzolite whereas the tholeiitic basalts formed asaggregate melts of both garnet and spinel lherzolite Thetholeiitic basalts underwent extensive low-pressure fractio-nation (52^3 kbar) and seismic studies indicate that someof theWrangellia crust beneathVancouver Island may cor-respond to the plutonic residues of the Karmutsen floodbasalts
ACKNOWLEDGEMENTSWe are grateful to Nick Arndt for helping us get this pro-ject started and Nick Massey for insights into VancouverIsland geology We greatly appreciate the assistance withPRIMELT modeling from Claude Herzberg We thankMikkel Schau for his insight and enthusiasm during field-work Jane Barling assisted with analyses by MC-ICP-MSThoughtful and thorough reviews of the manuscript byJohn Mahoney Fred Frey Wendy Bohrson DougalJerram and Teal Riley are greatly appreciated Fundingwas generously provided by the Rocks to Riches Programadministered by the BC amp Yukon Chamber of Mines in2004 by the BC Geological Survey and by NSERCDiscovery Grants to JSS and DW ARG was supportedby a University Graduate Fellowship at UBC
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
34
REFERENCESBarker F Brown A S Budahn J R amp Plafker G (1989) Back-arc
with frontal-arc component origin of Triassic Karmutsen basaltBritish Columbia Canada Chemical Geology 75 81^102
Blichert-Toft JWeis D Maerschalk C Agranier A amp Albare de F(2003) Hawaiian hot spot dynamics as inferred from the Hf and Pbisotope evolution of Mauna Kea volcano Geochemistry GeophysicsGeosystems 4(2) 1^27 doi1010292002GC000340
Bondre N R Duraiswami R A amp Dole G (2004) Morphologyand emplacement of flows from the Deccan Volcanic ProvinceIndia Bulletin of Volcanology 66 29^45
Carlisle D (1963) Pillow breccias and their aquagene tuffs QuadraIsland British Columbia Journal of Geology 71 48^71
Carlisle D (1972) Late Paleozoic to Mid-Triassic sedimentary-volcanic sequence on Northeastern Vancouver Island Geological
Society of Canada Paper Report of Activities 72-1 Part B 24^30Carlisle D amp SuzukiT (1974) Emergent basalt and submergent car-
bonate^clastic sequences including the Upper Triassic Dilleri andWelleri zones on Vancouver Island Canadian Journal of Earth
Sciences 11 254^279Chauvel C amp Blichert-Toft J (2001) A hafnium isotope and trace
element perspective on melting of the depleted mantle Earth and
Planetary Science Letters 190 137^151Chayes F (1954)The theory of thin-section analysis Journal of Geology
62 92^101Cho M Liou J G amp Maruyama S (1987) Transition from the zeo-
lite to prehnite^pumpellyite facies in the Karmutsen metabasitesVancouver Island British Columbia Journal of Petrology 27(2)467^494
Clowes R M Zelt C A Amor J R amp Ellis R M (1995)Lithospheric structure in the southern Canadian Cordillera froma network of seismic refraction lines Canadian Journal of Earth
Sciences 32(10) 1485^1513Coffin M F amp Eldholm O (1994) Large igneous provinces Crustal
structure dimensions and external consequences Reviews of
Geophysics 32(1) 1^36Cox K G (1980) A model for flood basalt volcanism Journal of
Petrology 21 629^650Eldholm O amp Coffin M F (2000) Large igneous provinces
and plate tectonics In Richards M A Gordon R G amp vander Hilst R D (eds) The History and Dynamics of Global
Plate Motions American Geophysical Union Geophysical Monograph 121309^326
Elliott T R Hawkesworth C J amp Grolaquo nvold K (1991) Dynamicmelting of the Iceland plume Nature 351 201^206
Escuder-Viruete J Pecurren rez-Estaucurren n A Contreras F Joubert MWeis D Ullrich T D amp Spadea P (2007) Plume mantle sourceheterogeneity through time Insights from the Duarte ComplexHispaniola northeastern Caribbean Journal of Geophysical Research112(B04203) doi1010292006JB004323
Farnetani C G Richards M A amp Ghiorso M S (1996)Petrological models of magma evolution and deep crustal structurebeneath hotspots and flood basalt provinces Earth and Planetary
Science Letters 143 81^94Fitton J G amp Godard M (2004) Origin and evolution of magmas
on the Ontong Java Plateau In Fitton J G Mahoney J JWallace P J amp Saunders A D (eds) Origin and Evolution of the
Ontong Java Plateau Geological Society London Special Publications 229151^178
Fitton J G Saunders A D Norry M J Hardarson B S ampTaylor R N (1997) Thermal and chemical structure of theIceland plume Earth and Planetary Science Letters 153 197^208
Fitton J G Saunders A D Kempton P D amp Hardarson B S(2003) Does depleted mantle form an intrinsic part of the Icelandplume Geochemistry Geophysics Geosystems 4(3) 1032 doi1010292002GC000424
Frey F A Huang S Blichert-Toft J Regelous M amp Boyet M(2005) Origin of depleted components in basalt related to theHawaiian hotspot Evidence from isotopic and incompatible ele-ment ratios Geochemistry Geophysics Geosystems 6(1) 23 doi1010292004GC000757
Galer S J G amp Abouchami W (1998) Practical application of leadtriple spiking for correction of instrumental mass discriminationMineralogical Magazine 62A 491^492
Ghiorso M S amp Sack R O (1995) Chemical mass transfer in mag-matic processes IV A revised and internally consistent thermody-namic model for the interpolation and extrapolation of liquid^solidequilibria in magmatic systems at elevated temperatures and pres-sures Contributions to Mineralogy and Petrology 119 197^212
Greene A R Scoates J S amp Weis D (2005)WrangelliaTerrane onVancouver Island Distribution of flood basalts with implicationsfor potential Ni^Cu^PGE mineralization In Grant B (ed)Geological Fieldwork 2004 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2005-1 209^220Greene A R Scoates J S Nixon G T amp Weis D (2006) Picritic
lavas and basal sills in the Karmutsen flood basalt provinceWrangellia northern Vancouver Island In Grant B (ed)Geological Fieldwork 2005 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2006-1 39^52Greene A R Scoates J S amp Weis D (2008) Wrangellia flood
basalts in Alaska A record of plume^lithosphere interaction in aLate Triassic accreted oceanic plateau Geochemistry Geophysics
Geosystems 9(Q12004) doi1010292008GC002092Greene A R Scoates J S Weis D amp Israel S (2009)
Geochemistry of triassic flood basalts from the Yukon (Canada)segment of the accreted Wrangellia oceanic plateau Lithosdoi101016jlithos200811010
Hauff F Hoernle K Tilton G Graham D W amp Kerr A C(2000) Large volume recycling of oceanic lithosphere over shorttime scales geochemical constraints from the Caribbean LargeIgneous Province Earth and Planetary Science Letters 174 247^263
Hauff F Hoernle K amp Schmidt A (2003) Sr^Nd^Pb compositionof Mesozoic Pacific oceanic crust (Site 1149 and 801 ODP Leg185)Implications for alteration of ocean crust and the input into theIzu^Bonin^Mariana subduction system Geochemistry Geophysics
Geosystems 4(8) doi1010292002GC000421Herzberg C (2004) Partial melting below the Ontong Java Plateau
In Fitton J G Mahoney J J Wallace P J amp Saunders A D(eds) Origin and Evolution of the Ontong Java Plateau Geological SocietyLondon Special Publications 229 179^183
Herzberg C (2006) Petrology and thermal structure of the Hawaiianplume from Mauna Kea volcano Nature 444(7119) 605
Herzberg C amp Asimow P D (2008) Petrology of some oceanicisland basalts PRIMELT2XLS software for primary magma cal-culation Geochemistry Geophysics Geosystems 9(Q09001) doi1010292008GC002057
Herzberg C amp OrsquoHara M J (2002) Plume-associated ultramaficmagmas of Phanerozoic age Journal of Petrology 43(10) 1857^1883
Herzberg C Asimow P D Arndt N Niu Y Lesher C MFitton J G Cheadle M J amp Saunders A D (2007)Temperatures in ambient mantle and plumes Constraints frombasalts picrites and komatiites Geochemistry Geophysics Geosystems8(Q02006) doi1010292006GC001390
Huang S amp Frey F A (2005) Trace element abundances of MaunaKea basalt from phase 2 of the Hawaii Scientific Drilling Project
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
35
Petrogenetic implications of correlations with major element con-tent and isotopic ratios Geochemistry Geophysics Geosystems 4(6)doi1010292002GC000322
Hussong D M Wipperman L K amp Kroenke L W (1979) Thecrustal structure of the Ontong Java and Manihiki OceanicPlateaus Journal of Geophysical Research 84(B11) 6003^6010
Jerram D A amp Widdowson M (2005) The anatomy of ContinentalFlood Basalt Provinces geological constraints on the processes andproducts of flood volcanism Lithos 79(3^4) 385^405
Jones D L Silberling N J amp Hillhouse J (1977)Wrangellia a dis-placed terrane in northwestern North America CanadianJournal ofEarth Sciences 14(11) 2565^2577
Kerr A C (2003) Oceanic Plateaus In Rudnick R (ed) The Crust
Treatise on GeochemistryVol 3 Oxford Elsevier Science pp 537^565Kerr A C amp Mahoney J J (2007) Oceanic plateaus Problematic
plumes potential paradigms Chemical Geology 241 332^353Kerr A CTarney J Marriner G F Klaver GT Saunders A D
amp Thirlwall M F (1996) The geochemistry and petrogenesis ofthe Late-Cretaceous picrites and basalts of Curacao NetherlandsAntilles a remnant of an oceanic plateau Contributions to
Mineralogy and Petrology 124(1) 29^43Kerr A C Marriner G F Tarney J Nivia A Saunders A D
Thirlwall M F amp Sinton A C (1997) Cretaceous basaltic ter-ranes in Western Colombia Elemental chronological and Sr-Ndisotopic constraints on petrogenesis Journal of Petrology 38(6)677^702
Kinzler R J (1997) Melting of mantle peridotite at pressuresapproaching the spinel to garnet transition Application to mid-ocean ridge basalt petrogenesis Journal of Geophysical Research
102(B1) 853^874Kinzler R J amp Grove T L (1992) Primary magmas of mid-ocean
ridge basalts 1 Experiments and methods Journal of Geophysical
Research 97(B5) 6885^6906Korenaga J (2005) Why did not the Ontong Java Plateau form sub-
aerially Earth and Planetary Science Letters 234 385^399Kuniyoshi S (1972) Petrology of the Karmutsen (Volcanic) Group
northeastern Vancouver Island British Columbia PhD disserta-tion University of California Los Angeles 242 pp
Lassiter J C DePaolo D J amp Mahoney J J (1995) Geochemistry ofthe Wrangellia flood basalt province Implications for the role ofcontinental and oceanic lithosphere in flood basalt genesis Journalof Petrology 36(4) 983^1009
LeBas M J (2000) IUGS reclassification of the high-Mg and picriticvolcanic rocks Journal of Petrology 41(10) 1467^1470
Longhi J (2002) Some phase equilibrium systematics of lherzolitemelting I Geochemistry Geophysics Geosystems 3(3) doi1010292001GC000204
MacDonald G A amp Katsura T (1964) Chemical composition ofHawaiian lavas Journal of Petrology 5 82^133
Mahoney J J (1987) An isotopic survey of Pacific oceanic plateausimplications for their nature and origin In Keating B HFryer P Batiza R amp Boehlert GW (eds) Seamounts Islands andAtolls American Geophysical Union Geophysical Monograph 43 207^220
Mahoney J LeRoex A P Peng Z Fisher R L amp Natland J H(1992) Southwestern limits of Indian Ocean Ridge mantle and theorigin of low 206Pb204Pb mid-ocean ridge basalt isotope systema-tics of the central Southwest Indian Ridge (178^508E) Journal ofGeophysical Research 97(B13) 19771^19790
Mahoney J J Sinton J M Kurz M D MacDougall J DSpencer K J amp Lugmair GW (1994) Isotope and trace elementcharacteristics of a super-fast spreading ridge East Pacific rise13^238S Earth and Planetary Science Letters 121 173^193
Massey N W D (1995a) Geology and mineral resources of theAlberni^Nanaimo Lakes sheet Vancouver Island 92F1W 92F2Eand part of 92F7E British Columbia Ministry of Energy Mines and
Petroleum Resources Paper 1992-2 132 ppMassey N W D (1995b) Geology and mineral resources of the
Cowichan Lake sheet Vancouver Island 92C16 British Columbia
Ministry of Energy Mines and Petroleum Resources Paper 1992-3 112 ppMassey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005a) Digital Geology Map of British Columbia Tile NM9 MidCoast BC British Columbia Ministry of Energy and Mines Geofile
2005-2Massey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005b) Digital Geology Map of British Columbia Tile NM10Southwest BC British Columbia Ministry of Energy and Mines Geofile
2005-3McDonough W F amp Sun S (1995) The composition of the Earth
Chemical Geology 120 223^253Monger JW H amp Journeay J M (1994) Basement geology and tec-
tonic evolution of theVancouver region In Monger JW H (ed)Geology and Geological Hazards of the Vancouver Region Southwestern
British Columbia Geological Survey of Canada Bulletin 481 3^25Muller J E (1977) Geology of Vancouver Island Geological Survey of
Canada Open File Map 463Muller J E (1980)The Paleozoic Sicker Group of Vancouver Island British
Columbia Geological Survey of Canada Paper 79-30 22 ppMuller J E Northcote K E amp Carlisle D (1974) Geology and
mineral deposits of Alert Bay^Cape Scott map area VancouverIsland British Columbia Geological Survey of Canada Paper 74-8 77
Neal C R Mahoney J J Kroenke L W Duncan R A ampPetterson M G (1997) The Ontong^Java Plateau InMahoney J J amp Coffin M F (eds) Large Igneous Provinces
Continental Oceanic and Planetary FloodVolcanism American Geophysical
Union Geophysical Monograph 100 183^216Niu Y Collerson K D Batiza R Wendt J I amp Regelous M
(1999) Origin of enriched-typed mid-ocean ridge basalt at ridgesfar from mantle plumes The East Pacific Rise at 11820rsquoN Journalof Geophysical Research 104 7067^7088
Nixon G T amp Orr A J (2007) Recent revisions to the EarlyMesozoic stratigraphy of Northern Vancouver Island (NTS 102I092L) and metallogenic implicatinos British Columbia InGrant B (ed) Geological Fieldwork 2006 British Columbia Ministry of
Energy Mines and Petroleum Resources Paper 2007-1 163^177Nixon GT Hammack J L KoyanagiV M Payie G J Haggart
JW Orchard M JTozerT Friedman R M Archibald D APalfy J amp Cordey F (2006a) Geology of the Quatsino^PortMcNeill area northernVancouver Island British Columbia Ministry
of Energy Mines and Petroleum Resources Geoscience Map 2006-2 scale150 000
Nixon G T Kelman M C Stevenson D Stokes L A ampJohnston K A (2006b) Preliminary geology of the Nimpkishmap area (NTS 092L07) northern Vancouver Island BritishColumbia In Grant B amp Newell J M (eds) Geological Fieldwork2005 British Columbia Ministry of Energy Mines and Petroleum Resources
Paper 2006-1 135^152Nixon G T Laroque J Pals A Styan J Greene A R amp
Scoates J S (2008) High-Mg lavas in the Karmutsen flood basaltsnorthernVancouver Island (NTS 092L) Stratigraphic setting andmetallogenic significance In Grant B (ed) Geological Fieldwork
2007 British Columbia Ministry of Energy Mines and Petroleum
Resources Paper 2008-1 175^190Nowell G M Kempton P D Noble S R Fitton J G
Saunders A Mahoney J J amp Taylor R N (1998) High precisionHf isotope measurements of MORB and OIB by thermal ionisation
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
36
mass spectrometry insights into the depleted mantle Chemical
Geology 149 211^233Parrish R R amp McNicoll V J (1992) U^Pb age determinations
from the southern Vancouver Island area British Columbia InRadiogenic Age and Isotopic Studies Report 5 Geological Survey of
Canada Paper 91-2 79^86Peate DW (1997) The Parana^Etendeka Province In Coffin M F
amp Mahoney J (eds) Large Igneous Provinces Continental Oceanic andPlanetary Flood Volcanism American Geophysical Union Geophysical
Monograph 100 217^245Perfit M R Fornari D J Ridley W I Kirk P D Casey J
Kastens K A Reynolds J R Edwards M Desonie DShuster R amp Paradis S (1996) Recent volcanism in the Siqueirostransform fault picritic basalts and implications for MORBmagma genesis Earth and Planetary Science Letters 141(1^4) 91^108
Pretorius W Weis D Williams G Hanano D Kieffer B ampScoates J S (2006) Complete trace elemental characterization ofgranitoid (USGSG-2GSP-2) reference materials by high resolutioninductively coupled plasma-mass spectrometry Geostandards and
Geoanalytical Research 30(1) 39^54Regelous M Niu Y Wendt J I Batiza R Greig A amp
Collerson K D (1999) Variations in the geochemistry of magma-tism on the East Pacific Rise at 10830rsquoN since 800 ka Earth and
Planetary Science Letters 168 45^63Recurren villon S Chauvel C Arndt N T Pik R Martineau F
Fourcade S amp Marty B (2002) Heterogeneity of the Caribbeanplateau mantle source Sr O and He isotopic compositions of oli-vine and clinopyroxene from Gorgona Island Earth and Planetary
Science Letters 205(1^2) 91Rhodes J M (1996) Geochemical stratigraphy of lava flows samples
by the Hawaii Scientific Drilling Project Journal of Geophysical
Research 101(B5) 11729^11746Rhodes J M amp Vollinger M J (2004) Composition of basaltic lavas
sampled by phase-2 of the Hawaii Scientific Drilling ProjectGeochemical stratigraphy and magma types Geochemistry
Geophysics Geosystems 5(3) doi1010292002GC000434Richards M A Jones D L Duncan R A amp DePaolo D J (1991)
A mantle plume initiation model for the Wrangellia flood basaltand other oceanic plateaus Science 254 263^267
Salters V J M amp Hart S R (1991) The mantle sources of oceanridges islands and arcs the Hf-isotope connection Earth and
Planetary Science Letters 104 364^380Salters V J M amp Stracke A (2004) Composition of the depleted
SaltersV J amp WhiteW (1998) Hf isotope constraints on mantle evo-lution Chemical Geology 145 447^460
SanoT amp Yamashita S (2004) Experimental petrology of basementlavas from Ocean Drilling Program Leg 192 implications for dif-ferentiation processes in Ontong Java Plateau magmas InFitton J G Mahoney J JWallace P J amp Saunders A D (eds)Origin and Evolution of the Ontong Java Plateau Geological Society
London Special Publications 229 185^218Saunders A D (2005) Large igneous provinces Origin and environ-
mental consequences Elements 1 259^263Self S ThordarsonT amp Keszthely L (1997) Emplacement of conti-
nental flood basalt lava flows In Mahoney J J amp Coffin M F(eds) Large Igneous Provinces Continental Oceanic and Planetary Flood
Volcanism American Geophysical Union Geophysical Monograph 100381^410
Shaw D M (2000) Continuous (dynamic) melting theory revisitedCanadian Mineralogist 38(5) 1041^1063
Sluggett C L (2003) Uranium^lead age and geochemical constraintson Paleozoic and Early Mesozoic magmatism in WrangelliaTerrane Saltspring Island British Columbia BSc thesisUniversity of British ColumbiaVancouver 84 pp
Storey M Mahoney J J amp Saunders A D (1997) Cretaceousbasalts in Madagascar and the transition between plume and con-tinental lithospheric mantle sources In Mahoney J J ampCoffin M F (eds) Large Igneous Provinces Continental Oceanic and
Planetary Flood Volcanism Aamerican Geophysical Union Geophysical
Monograph 100 95^122Surdam R C (1967) Low-grade metamorphism of the Karmutsen
Group PhD dissertation University of California Los Angeles288 pp
Sutherland-Brown A Yorath C J Anderson R G amp Dom K(1986) Geological maps of southern Vancouver IslandLITHOPROBE I 92C10 11 14 16 92F1 2 7 8 Geological Survey ofCanada Open File 1272
Tejada M L G Mahoney J J Castillo P R Ingle S P Sheth HC amp Weis D (2004) Pin-pricking the elephant evidence on theorigin of the Ontong Java Plateau from Pb^Sr^Hf^Nd isotopiccharacteristics of ODP Leg 192 basalts In Fitton J GMahoney J J Wallace P J amp Saunders A D (eds) Origin and
Evolution of the Ontong Java Plateau Geological Society London Special
Publications 229 133^150Thompson P M E Kempton P D amp Kerr A C (2008) Evaluation
of the effects of alteration and leaching on Sm^Nd and Lu^Hf sys-tematics in submarine mafic rocks Lithos 104(1^4) 164^176
Thompson P M E Kempton P D White R V Kerr A CTarney J Saunders A D Fitton J G amp McBirney A (2003)Hf-Nd isotope constraints on the origin of the CretaceousCaribbean plateau and its relationship to the Galapagos plumeEarth and Planetary Science Letters 217(1^2) 59^75
Vervoort J D Patchett P Blichert-Toft J amp Albare de F (1999)Relationships between Lu^Hf and Sm^Nd isotopic systems in theglobal sedimentary system Earth and Planetary Science Letters 16879^99
Walter M J (1998) Melting of garnet peridotite and the origin ofkomatiite and depleted lithosphere Journal of Petrology 39(1) 29^60
Weis D Kieffer B Maerschalk C Barling J de Jong JWilliams G A Hanano D Mattielli N Scoates J SGoolaerts A Friedman R A amp Mahoney J B (2006) High-pre-cision isotopic characterization of USGS reference materials byTIMS and MC-ICP-MS Geochemistry Geophysics Geosystems
7(Q08006) doi1010292006GC001283Weis D Kieffer B Hanano D Silva I N Barling J
Pretorius W Maerschalk C amp Mattielli N (2007) Hf isotopecompositions of US Geological Survey reference materialsGeochemistry Geophysics Geosystems 8(Q06006) doi1010292006GC001473
Wheeler J O amp McFeely P (1991) Tectonic assemblage map of theCanadian Cordillera and adjacent part of the United States ofAmerica Geological Survey of Canada Map 1712A
WhiteW M Albare de F amp Tecurren louk P (2000) High-precision analy-sis of Pb isotope ratios by multi-collector ICP-MS Chemical Geology167 257^270
Yorath C J Sutherland Brown A amp Massey N W D (1999)LITHOPROBE southern Vancouver Island British Columbia Geological
Survey of Canada Bulletin 498 145 ppZou H (1998) Trace element fractionation during modal and
non-model dynamic melting and open-system melting Amathematical treatment Geochimica et Cosmochimica Acta 62(11)1937^1945
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
37
Zou H amp Reid M R (2001) Quantitative modeling of trace elementfractionation during incongruent dynamic melting Geochimica et
Cosmochimica Acta 65(1) 153^162
APPENDIX SAMPLEPREPARAT ION ANDANALYT ICAL METHODSOnly the freshest rocks in the field were sampled and onlythe least altered samples were selected for chemical andisotopic analysis based on thorough petrographic inspec-tion Fifty-nine of the 129 collected samples were crushed(400 g) into small pieces 52mm in diameter in aRocklabs hydraulic piston crusher between WC plates tominimize contamination The coarse-crush was mixedand 100 g was powdered in a planetary mill using agatejars and balls cleaned with quartz sand between samples
ActLabs analytical methodsThe major- and trace-element compositions of the whole-rock powders were determined at Activation LaboratoriesLtd (Actlabs) in Ancaster Ontario Analytical techniquesand detection limits are also available from Actlabs (httpwwwactlabscommethsub_code4erehtm) The analyticalmethod for each of the elements analyzed is indicated inTable 2 For the major elements a 02 g sample was mixedwith a lithium metaborate^lithium tetraborate mixtureand fused in a graphite crucible The molten mixture waspoured into a 5 HNO3 solution and shaken until dis-solved (30min) The samples were analyzed for majoroxides and selected trace elements by inductively coupledplasma optical emission spectrometry (ICP-OES) on acombination simultaneous^sequential Thermo Jarrell-AshEnviro II system Internal calibration was achieved usinga variety of international reference materials (eg W-2BIR-1 DNC-1) and independent control samplesAdditional trace elements were analyzed by both instru-mental neutron activation analysis (INAA) and induc-tively couple plasma mass spectrometry (ICP-MS) ForINAA15^25 g of sample was weighed into small polyeth-ylene vials and irradiated with a control international ref-erence material CANMET WMS-1 and NiCr flux wires ata thermal neutron flux of 71012 neutronscm2 per s in theMcMaster Nuclear Reactor Following a 7 day waitingperiod the samples were measured on an Ortec high-purity Ge detector linked to a Canberra Series 95 multi-channel analyzer Activities for each element weredecay- and weight-corrected and compared with a detectorcalibration developed from multiple international certifiedreference materials For ICP-MS 025 g of sample wasdigested in HF followed by HNO3^HClO4 heated andtaken to dryness The samples were brought back into solu-tion with HCl Samples were analyzed using a PerkinElmer Optima 3000 ICP In-lab standards or certified ref-erence materials (eg W-2 BIR-1 DNC-1) were used for
quality control A total of 15 blind duplicates were ana-lyzed from the same sample powders to assessreproducibility
University of Massachusetts analyticalmethodsFourteen sample duplicate powders taken from the same100 g sample aliquot (high-MgO basalts and picrites)were also analyzed at the Ronald B Gilmore X-RayFluorescence (XRF) Laboratory at the University ofMassachusetts Major elements were measured on a fusedLa-bearing lithium borate glass disc using a SiemensMRS-400 spectrometer Trace element concentrations(Rb Sr Ba Ce Nb Zr Y Pb Zn Ga Ni Cr V) weremeasured on a separate powder pellet using a PhilipsPW2400 sequential spectrometer using a Rh tube LOIand ferrous iron measurements were made as describedby Rhodes amp Vollinger (2004) Precision and accuracy esti-mates for the data have been described by Rhodes (1996)and Rhodes amp Vollinger (2004) Results for each sampleare the average of two separate analyses and are shown inElectronic Appendix 2 for a comparison of analyses ofduplicate powders using XRFand ICP-MS
PCIGR analytical methodsA subset of 19 samples was selected for high-precisiontrace-element analysis and Sr Nd Pb and Hf isotopicanalysis at the Pacific Centre for Isotopic andGeochemical Research (PCIGR) at the University ofBritish Columbia (UBC) Samples were selected from the59 samples analyzed for whole-rock chemistry at ActLabsbased on major- and trace-element chemistry alteration(low LOI and petrographic alteration index) and samplelocation Samples were prepared for trace-element analysisat the PCIGR by the technique described by Pretorius et al(2006) on unleached rock powders Sample powders(100mg) were weighed in 7ml screw-top Savillex bea-kers and dissolved in 1ml 14N HNO3 and 5ml 48 HFon a hotplate for 48 h at 1308C with periodic ultrasonica-tion Samples were dried and redissolved in 6ml 6N HClon a hotplate for 24 h and then dried and redissolved in1ml concentrated HNO3 for 24 h before final dryingTrace element abundances were measured by high-resolution inductively coupled plasma-mass spectrometry(HR-ICP-MS) with aThermo Finnigan Element2 systemfollowing the procedures described by Pretorius et al(2006) within 24 h of redissolution HFSE and LILE weremeasured in medium-resolution mode at 2000 dilutionusing a PFATeflon spray chamber washed with aqua regiafor 3min between samples REE were measured in high-resolution mode and U and Pb in low-resolution modeat 2000 dilution using a glass spray chamber washedwith 2 HNO3 between samples Total procedural blanksand reference materials (BCR-2 BHVO-2) were analyzedwith the batch of samples Indium was used as an internal
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
38
standard in all samples and standard solutionsBackground and standard solutions were analyzed afterevery five samples to detect memory effects and massdrift Results for PCIGR trace-element analyses areshown in Electronic Appendix 3 for the full set of analyzedtrace elementsSample digestion for purification of Sr Nd Hf and Pb
for column chemistry began by weighing each samplepowder (400^500mg) prior to leaching All samples wereleached with 6N HCl and placed in an ultrasonic bath for15min Samples were rinsed twice with 18 M-cm H2Obetween each leaching step (15 total) until the supernatantliquid was clear [following the method of Mahoney (1987)]Samples were then dried on a hotplate for 24 h andweighed again Sample solutions were then prepared bydissolving 100^250mg of the leached powder dissolvedin 1ml 14N HNO3 and 10ml 48 HF on a hotplate for48 h at 1308C with periodic ultrasonication Samples weredried and redissolved in 6ml 6N HCl on a hotplate for24 h and then dried Pb was separated using anion exhangecolumns and the discard was used for Sr REE and Hfseparation Nd was separated from the other REE and Hfrequired two additional purification steps Detailed proce-dures for column chemistry for separating Sr Nd and Pbat the PCIGR have been described byWeis et al (2006) andHf purification has been described byWeis et al (2007) Srand Nd isotope ratios were measured by thermal ioniza-tion mass spectrometry (TIMS) on a Thermo FinniganTriton system in static mode with relay matrix rotation ona single Ta and double Re^Ta filament respectively Fourto five filaments per barrel of 21 were occupied by stan-dards (NBS 987 for Sr and LaJolla for Nd) for each barrelwhere samples were run Sample Sr and Nd isotopic com-positions were corrected for mass fractionation using86Sr88Srfrac14 01194 and 146Nd144Ndfrac14 07219 Each samplewas then normalized using the barrel average of the refer-ence material relative to the values of143Nd144Ndfrac14 0511858 and 87Sr86Srfrac14 0710248 (Weiset al 2006) During the course of the Vancouver Islandanalyses the La Jolla Nd standard gave an average valueof 05118566 (nfrac14 7) and NBS987 standard gave an aver-age of 07102408 (nfrac1411) (2 error is reported as times106) 147Sm144Nd errors are 15 or 0006 USGeological Survey (USGS) reference material BHVO-2was processed with the samples and yielded Sr and Nd iso-topic ratios of 07034607 and 05129786 respectivelyThese are in agreement with the published values of070347920 and 051298411 respectively (Weis et al2006)
Pb and Hf isotopic compositions were analyzed by mul-tiple collector-inductively coupled plasma-mass spectrome-try (MC-ICP-MS) on a static multi-collection on a NuPlasma (Nu Instruments) system The detailed analyticalprocedure for Pb isotopic analyses on the Nu Plasma atthe PCIGR has been described by Weis et al (2006) Theconfiguration for Pb analyses allows for collection of PbTl and Hg togetherTl and Hg are used to monitor instru-mental mass discrimination and isobaric overlap repec-tively All sample solutions were analyzed withapproximately the same PbTl ratio (4) as the referencematerial NIST SRM 981 To accomplish this a small ali-quot of each sample solution from the Pb columns was ana-lyzed on the Element2 to determine the precise amount ofPb available for analysis on the Nu Plasma The SRM 981standard was run after every two samples on the NuPlasma During the time samples were run analyses ofthe SRM 981 Pb reference material gave values of206Pb204Pbfrac1416940322 207Pb204Pbfrac1415495823and 208Pb204Pbfrac14 36713164 (nfrac14 61 2 error is reportedas times 104) these values are in excellent agreement withreported TIMS triple-spike values of Galer amp Abouchami(1998) Fractionation-corrected Pb isotopic ratios were fur-ther corrected by the sample-standard bracketing methodor the ln^ln correction method described by White et al(2000) and Blichert-Toft et al (2003) Leached powders ofUSGS reference material BHVO-2 yielded Pb isotopicratios of 206Pb204Pbfrac14186454 8 207Pb204Pbfrac14 1549105 and 208Pb204Pbfrac14 38222514 and BCR-2yielded 206Pb204Pbfrac141880466 207Pb204Pbfrac14 1562518 and 208Pb204Pbfrac14 3883496 (2 error isreported as times 104) These values are in agreement withthose for leached residues of BHVO-2 and BCR-2 fromWeis et al (2006)Hf isotopic compositions were analyzed following the
procedures detailed byWeis et al (2007) The configurationfor Hf analyses monitored Lu mass 175 andYb mass 172 toallow for interference correction to masses 174 and 176 Hfisotopic ratios were normalized internally for mass fractio-nation to a 179Hf177Hf ratio of 07325 using an exponentialcorrection Standards were run after every two samplesand sample results were normalized to the ratio of the in-run daily average and a 176Hf177Hf ratio for JMC-475 of0282160 During the course of analyses the Hf standardJMC-475 gave an average value 02821679 (nfrac14130)USGS reference materials BCR-2 and BHVO-2 were pro-cessed with the samples and yielded Hf isotopic ratios of02828675 and 02831005 respectively Publishedvalues for BCR-2 and BHVO-2 are 02828717 and0283104 8 respectively (Weis et al 2007)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
39
Plagioclase forms most of the phenocrysts and glomero-crysts and the groundmass is fine-grained comprising pla-gioclase microlites clinopyroxene granules small grains ofFe^Ti oxide devitrified glass and secondary minerals
there is no fresh glass preserved The picritic and high-MgO basaltic lavas are characterized by spherulitic plagi-oclase and clinopyroxene with abundant euhedral olivinephenocrysts (Fig 5 Table 1) Olivine is completely replaced
Fig 4 Photographs showing field relations from the Schoen Lake Provincial Park area northernVancouver Island (a) Sediment^sill complexat the base of the Karmutsen Formation on the north side of Mt Schoen 100m below the Daonella locality (see Fig 2 for location) (b)Interbedded mafic sills and deformed finely banded chert and shale with calcareous horizons from location between Mt Adam and MtSchoen
Fig 5 Photomicrographs of picritic pillow basalts Karmutsen Formation northernVancouver Island Scale bars represent 1mm (a) Picritewith abundant euhedral olivine pseudomorphs from Keogh Lake type locality (Fig 2) in cross-polarized light (XPL) (sample 4722A4 19^198wt MgO four analyses) Samples contain dense clusters (24^42 vol ) of olivine pseudomorphs (52mm) in a groundmass of curvedand branching sheaves of acicular plagioclase and intergrown with clinopyroxene and altered glass In many cases plagioclase nucleated on theedges of the olivine phenocrysts (b) Sheaves of intergrown plagioclase and clinopyroxene in aphyric picrite pillow lava from west of MaynardLake (Fig 2) in XPL (sample 4723A2108 wt MgO)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
7
Table 1 Summary of petrographic characteristics and phenocryst proportions of Karmutsen basalts on Vancouver Island
BC
Sample Area Flow Group Texture vol Ol Plag Cpx Ox Alteration Notes
4718A1 MA PIL THOL intersertal 20 5 3 few plag glcr52mm cpx51 mm
4718A2 MA PIL THOL intersertal glomero 15 1 plag glcr52mm very fresh
4718A5 MA FLO THOL intersertal glomero 10 2 mottled few plag glcr54 mm
4723A4 KR PIL PIC intergranular intersertal 0 3 swtl plag51mm no ol phenos
4723A13 KR PIL PIC spherulitic 24 3 swtl plag51 mm
(continued)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
8
by an assemblage of talc^tremolite^serpentine^opaqueoxides there is no fresh olivine present in the Karmutsenlavas onVancouver Island Plagioclase and clinopyroxenephenocrysts are absent in the high-MgO lavas Thecoarse-grained mafic rocks are characterized by sub-ophi-tic textures with an average grain size typically 41mmand are generally not glomeroporphyritic these rocks aremainly from the interiors of massive flows although somemay represent sillsSample preparation and analytical methods for major-
and trace-element and Sr^Nd^Hf^Pb isotopic composi-tions of the Karmutsen volcanic rocks are given in theAppendix
WHOLE -ROCK CHEMISTRYMajor- and trace-element compositionsThe most abundant type of volcanic rock in theKarmutsen Formation is tholeiitic basalt with a restrictedrange of major- and trace-element compositions The tho-leiitic basalts have similar compositions to the coarse-grained mafic rocks and both groups are distinct fromthe picrites and high-MgO basalts [Fig 6 picrites412wt MgO (LeBas 2000) high-MgO basalts8^12wt MgO] The tholeiitic basalts have lower MgO(57^77wt MgO) and higher TiO2 (14^22wt
TiO2) than the picrites (130^198wt MgO05^07wt TiO2) and high-MgO basalts (91^116wt MgO 05^08wt TiO2 Fig 6 Table 2) Almost allcompositions plot within the tholeiitic field in a total alka-lis vs silica plot although there has been substantial K lossin most samples from the Karmutsen Formation (seebelow) and the tholeiitic basalts generally have higherSiO2 Na2OthornK2O CaO and FeOT (total iron expressedas FeO) than the picrites The tholeiitic basalts also havenoticeably lower loss on ignition (LOI mean1713wt ) than the picrites (mean 54 08wt )and high-MgO basalts (mean 3815wt ) (Fig 6)reflecting the presence of abundant altered olivinephenocrysts in the latter groups Ni concentrations are sig-nificantly higher for the picrites (339^755 ppm) and high-MgO basalts (122^551ppm) than the tholeiitic basalts(58^125 ppm except for one anomalous sample) andcoarse-grained rocks (70^213 ppm) (Fig 6 Table 2) Ananomalous tholeiitic pillowed flow (two samples outlierin Tables 1 and 2) from near the picrite type locality atKeogh Lake and a mineralized sill (disseminated sulfide)from the basal sediment^sill complex at Schoen Lake aredistinguished by higher TiO2 and FeOT than the maingroup of tholeiitic basalts The tholeiitic basalts are similarin major-element composition to previously publishedresults for samples from the Karmutsen Formation how-ever the Keogh Lake picrites which encompass the
Table 1 Continued
Sample Area Flow Group Texture vol Ol Plag Cpx Ox Alteration Notes
4722A5 KR FLO OUTLIER intersertal 3 mottled very fg v small pl needles
5615A11 KR PIL OUTLIER intersertal 3 mottled very fg v small pl needles
5617A4 SL SIL MIN intersertal 5 3 fg
Sample number last digit of year month day initial sample station (except 93G171) MA Mount Arrowsmith SLSchoen Lake KR Karmutsen Range PIL pillow BRE breccia FLO flow SIL sill GAB gabbro THOL tholeiiticbasalt PIC picrite HI-MG high-MgO basalt CGcoarse-grained MIN mineralized sill glomero glomeroporphyriticOlivine modes for picrites are based on area calculations from scans (see lsquoOlivine accumulation in high-MgO and picriticlavasrsquo section and Fig 11) Visual alteration index is based primarily on degree of plagioclase alteration and presence ofsecondary minerals (1 least altered 3 most altered) Plagioclase phenocrysts are commonly altered to albite pumpellyiteand chlorite olivine is altered to talc tremolite and clinochlore (determined using the Rietveld method of X-ray powderdiffraction) clinopyroxene is unaltered FendashTi oxide is commonly replaced by sphene glcr glomerocrysts fg fine-grained cg coarse-grained oik oikocryst chad chadacryst enc enclosing swtl swallow-tail ol olivinepseudomorphs plag plagioclase cpx clinopyroxene ox oxides (includes ilmenite thorn titanomagnetite)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
9
00
05
10
15
20
25
30
35
40
0 5 10 15 20 25
10
11
12
13
14
15
16
17
18
19
20
0 5 10 15 20 25
6
7
8
9
10
11
12
13
14
15
0 5 10 15 20 25
6
7
8
9
10
11
12
13
14
15
16
0 5 10 15 20 25
0
1
2
3
4
5
6
40 45 50 55 60
0
200
400
600
800
0 5 10 15 20 25
(wt)(wt)
(wt)
(wt)
+K2ONa2O TiO2
FeO (T)
Ni (ppm)
Al2O3
CaO
alkalic
tholeiitic
Karmutsen Form
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
(compiled data)
Pillowed flowMineralized sill
SiO2 (wt) MgO (wt)
MgO (wt)MgO (wt)
MgO (wt) MgO (wt)
(a) (b)
(g)
(e)
(f)
(c)
LOI (wt )
(d)
MgO (wt )0
1
2
3
4
5
6
7
0 5 10 15 20 25
(wt)
Fig 6 Whole-rock major-element Ni and LOI variation diagrams for the Karmutsen Formation New samples from this study (Table 2) areshown by symbols with black outlines (see legend) Previously published analyses are shown by small gray symbols without black outlines Theboundary of the alkaline and tholeiitic fields is that of MacDonald amp Katsura (1964) Total iron expressed as FeO LOI loss on ignition oxidesare plotted on an anhydrous normalized basis References for the 322 compiled analyses for the Karmutsen Formation are Surdam (1967)Kuniyoshi (1972) Muller et al (1974) Barker et al (1989) Lassiter et al (1995) Massey (1995a1995b)Yorath et al (1999) and G Nixon (unpublisheddata) It should be noted that the compiled dataset has not been filtered many of the samples with high SiO2 (452wt ) are probably notKarmutsen flood basalts but younger Bonanza basalts and andesites Three pillowed flow samples and a mineralized sill are distinguishedseparately because of their anomalous chemistry Coarse-grained samples are indicated separately Three tholeiitic basalt samples (4724A54718A5 4718A6) with416wt Al2O3 and4270 ppm Sr contain 10^25 modal of large plagioclase phenocrysts (7^8mm) or glomerocrysts(44mm)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
10
Table 2 Major element (wt oxide) and trace element (ppm) abundances in whole-rock samples of Karmutsen basalts
1Ni concentrations for these samples were determined by XRFTHOL tholeiitic basalt PIC picrite HI-MG high MgO basalt CG coarse-grained (sill or gabbro) MIN SIL mineralizedsill OUTLIER anomalous pillowed flow in plots MA Mount Arrowsmith SL Schoen Lake KR Karmutsen Range QIQuadra Island Sample locations are given using the Universal Transverse Mercator (UTM) coordinate system (NAD83zones 9 and 10) Analyses were performed at Activation Laboratory (ActLabs) Fe2O3
is total iron expressed as Fe2O3LOI loss on ignition All major elements Sr V and Y were measured by ICP quadrupole OES on solutions of fusedsamples Cu Ni Pb and Zn were measured by total dilution ICP Cs Ga Ge Hf Nb Rb Ta Th U Zr and REE weremeasured by magnetic-sector ICP on solutions of fused samples Co Cr and Sc were measured by INAA Blanks arebelow detection limit See Electronic Appendices 2 and 3 for complete XRF data and PCIGR trace element datarespectively Major elements for sample 93G17 are normalized anhydrous Sample 5615A7 was analyzed by XRFSamples from Quadra Island are not shown in figures
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
17
picritic and high-MgO basalt pillow lavas extend to20wt MgO (Fig 6 and references listed in thecaption)The tholeiitic basalts from the Karmutsen Range dis-
play a tight range of parallel light rare earth element(LREE)-enriched REE patterns (LaYbCNfrac14 21^24mean 2202) whereas the picrites and high-MgObasalts are characterized by sub-parallel LREE-depleted
patterns (LaYbCNfrac14 03^07 mean 06 02) with lowerREE abundances than the tholeiitic basalts (Fig 7)Tholeiitic basalts from the Schoen Lake and MountArrowsmith areas have similar REE patterns tosamples from the Karmutsen Range (Schoen Lake LaYbCNfrac14 21^26 mean 2303 Mount Arrowsmith LaYbCNfrac1419^25 mean 2204) and the coarse-grainedmafic rocks have similar REE patterns to the tholeiitic
Depleted MORB average(Salters amp Stracke 2004)
Schoen LakeSchoen Lake
Mount Arrowsmith
Karmutsen RangeS
ampl
eC
hond
rite
Sam
ple
Cho
ndrit
e
Sam
ple
Prim
itive
Man
tleS
ampl
eP
rimiti
ve M
antle
Sam
ple
Prim
itive
Man
tle
Mount Arrowsmith
Karmutsen Range
(a) (b)
(c) (d)
(e) (f )
Sam
ple
Cho
ndrit
e
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained mafic rock
Pillowed flowMineralized sill
1
10
100
1
10
100
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
01
1
10
100
1
10
100
1
10
100
Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
Fig 7 Whole-rock REE and other incompatible-element concentrations for the Karmutsen Formation (a) (c) and (e) are chondrite-normal-ized REE patterns for samples from each of the three main field areas onVancouver Island (b) (d) and (f) are primitive mantle-normalizedtrace-element patterns for samples from each of the three main field areas All normalization values are from McDonough amp Sun (1995) Theclear distinction between LREE-enriched tholeiitic basalts and LREE-depleted picrites the effects of alteration on the LILE (especially loss ofK and Rb) and the different range for the vertical scale in (b) (extends down to 01) should be noted
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
18
basalts (LaYbCNfrac14 21^29 mean 2508) Two LREE-depleted coarse-grained mafic rocks from the SchoenLake area have similar REE patterns (LaYbCNfrac14 07) tothe picrites from the Keogh Lake area indicating that thisdistinctive suite of rocks is exposed as far south asSchoen Lake that is 75 km away (Fig 7) The anomalouspillowed flow from the Karmutsen Range has lower LaYbCN (13^18) than the main group of tholeiitic basaltsfrom the Karmutsen Range (Fig 7) The mineralized sillfrom the Schoen Lake area is distinctly LREE-enriched(LaYbCNfrac14 33) with the highest REE abundances(LaCNfrac14 940) of all Karmutsen samples this may reflectcontamination by adjacent sediment also evident in thin-section where sulfide blebs are cored by quartz (Greeneet al 2006)The primitive mantle-normalized trace-element pat-
terns (Fig 7) and trace-element variations and ratios(Fig 8) highlight the differences in trace-element
concentrations between the four main groups ofKarmutsen flood basalts onVancouver Island The tholeii-tic basalts from all three areas have relatively smooth par-allel trace-element patterns some samples have smallpositive Sr anomalies relative to Nd and Sm and manysamples have prominent negative K anomalies relative toU and Nb reflecting K loss during alteration (Fig 7) Thelarge ion lithophile elements (LILE) in the tholeiiticbasalts (mainly Rb and K not Cs) are depleted relativeto the high field strength elements (HFSE) and LREELILE segments especially for the Schoen Lake area(Fig 7d) are remarkably parallel The picrites andhigh-MgO basalts form a tight band of parallel trace-element patterns and are depleted in HFSE and LREEwith positive Sr anomalies and relatively low Rb Thecoarse-grained mafic rocks from the Karmutsen Rangehave identical trace-element patterns to the tholeiiticbasalts Samples from the Schoen Lake area have similar
000
025
050
075
100
125
0 1 2 3 4 5
00
04
08
12
16
0 5 10 15 20 25
MgO (wt)
0
5
10
15
0 50 100 150
Hf (ppm)
Th (ppm)
NbLaNb (ppm)
Zr (ppm)
(a)
(c)
(b)
(d)
Th (ppm)
00
02
04
06
08
10
12
00 01 02 03 04 05
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
NbLa=1
4723A4 5615A12
U (ppm)
Fig 8 Whole-rock trace-element concentrations and ratios for the Karmutsen Formation [except (b) which is vs wt MgO] (a) Nb vs Zr(b) NbLa vs MgO (c) Th vs Hf (d) Th vs U There is a clear distinction between the tholeiitic basalts and picrites in Nb and Zr both inconcentration and the slope of each trend
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
19
trace-element patterns to the Karmutsen Range except forthe two LREE-depleted coarse-grained mafic rocks andthe mineralized sill (Fig 7d)
Sr^Nd^Hf^Pb isotopic compositionsA total of 19 samples from the four main groups ofthe Karmutsen Formation were selected for radiogenicisotopic geochemistry on the basis of major- and trace-element variation stratigraphic position and geographicallocation The samples have overlapping age-correctedHf and Nd isotope ratios and distinct ranges of Sr isotopiccompositions (Fig 9) The tholeiitic basalts picriteshigh-MgO basalt and coarse-grained mafic rocks have
initial eHffrac14thorn87 to thorn126 and eNdfrac14thorn66 to thorn88 cor-rected for in situ radioactive decay since 230 Ma (Fig 9Tables 3 and 4) All the Karmutsen samples form a well-defined linear array in a Lu^Hf isochron diagram corre-sponding to an age of 24115 Ma (Fig 9) This is withinerror of the accepted age of the Karmutsen Formationand indicates that the Hf isotope systematics behaved as aclosed system since c 230 Ma The anomalous pillowedflow has similar initial eHf (thorn98) and slightly lower initialeNd (thorn62) compared with the other Karmutsen samplesThe picrites and high-MgO basalt have higher initial87Sr86Sr (070398^070518) than the tholeiitic basalts(initial 87Sr86Srfrac14 070306^070381) and the coarse-grained mafic rocks (initial 87Sr86Srfrac14 070265^070428)
05128
05129
05130
05131
05132
015 017 019 021 023 025
02829
02830
02831
02832
02833
02834
000 002 004 006 008 010
4
6
8
10
0702 0703 0704 0705 07066
8
10
12
14
5 6 7 8 9 10
12
0
1
2
3
4
5
6
7
0702 0703 0704 0705 0706
(d)
LOI (wt )
87Sr 86Sr
4723A2
Nd
143Nd144Nd
147Sm144Nd
87Sr 86Sr (230 Ma)
(230 Ma)
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
(a)
(c)
176Hf177Hf
Age=241 plusmn 15 Ma=+990 plusmn 064
176Lu177Hf
Hf (230 Ma)
Nd (230 Ma)
(e)
Hf (i)
(b)
Fig 9 Whole-rock Sr Nd and Hf isotopic compositions for the Karmutsen Formation (a) 143Nd144Nd vs 147Sm144Nd (b) 176Hf177Hf vs176Lu177Hf The slope of the best-fit line for all samples corresponds to an age of 24115 Ma (c) Initial eNd vs 87Sr86Sr Age-correction to230 Ma (d) LOI vs 87Sr86Sr (e) Initial eHf vs eNd Average 2 error bars are shown in a corner of each panel Complete chemical duplicatesshown inTables 3 and 4 (samples 4720A7 and 4722A4) are circled in each plot
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
overlap the ranges of the picrites and tholeiitic basalts(Fig 9 Table 3)The measured Pb isotopic compositions of the tholeiitic
basalts are more radiogenic than those of the picrites andthe most magnesian Keogh Lake picrites have the leastradiogenic Pb isotopic compositions The range ofinitial Pb isotope ratios for the picrites is206Pb204Pbfrac1417868^18580 207Pb204Pbfrac1415547^15562and 208Pb204Pbfrac14 37873^38257 and the range for thetholeiitic basalts is 206Pb204Pbfrac1418782^19098207Pb204Pbfrac1415570^15584 and 208Pb204Pbfrac14 37312^38587 (Fig 10 Table 5) The coarse-grained mafic rocksoverlap the range of initial Pb isotopic compositions forthe tholeiitic basalts with 206Pb204Pbfrac1418652^19155207Pb204Pbfrac1415568^15588 and 208Pb204Pbfrac14 38153^38735 One high-MgO basalt has the highest initial206Pb204Pb (19242) and 207Pb204Pb (15590) and thelowest initial 208Pb204Pb (37676) The Pb isotopic ratiosfor Karmutsen samples define broadly linearrelationships in Pb isotope plots The age-corrected Pb
isotopic compositions of several picrites have been affectedby U and Pb (andor Th) mobility during alteration(Fig 10)
ALTERAT IONThe Karmutsen basalts have retained most of their origi-nal igneous structures and textures however secondaryalteration and low-grade metamorphism have producedzeolitic- and prehnite^pumpellyite-bearing mineral assem-blages (Table 1 Cho et al 1987) Alteration and metamor-phism have primarily affected the distribution of the LILEand the Sr isotopic systematics The Keogh Lake picriteshave been affected more by alteration than the tholeiiticbasalts and have higher LOI (up to 55wt ) variableLILE and higher measured Sr Nd and Hf and lowermeasured Pb isotope ratios than the tholeiitic basalts Srand Pb isotope compositions for picrites and tholeiiticbasalts show a relationship with LOI (Figs 9 and 10)whereas Nd and Hf isotopic compositions do not correlate
Table 3 Sr and Nd isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Rb Sr 87Sr86Sr 2m87Rb86Sr 87Sr86Srt Sm Nd 143Nd144Nd 2m eNd
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses carried out at the PCIGR thecomplete trace element analyses are given in Electronic Appendix 3
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
21
with LOI (not shown) The relatively high apparent initialSr isotopic compositions for the high-MgO lavas (up to07052) probably resulted from post-magmatic loss of Rbor an increase in 87Sr86Sr through addition of seawater Sr(eg Hauff et al 2003) High-MgO lavas from theCaribbean plateau have similarly high 87Sr86Sr comparedwith basalts (Recurren villon et al 2002)The correction for in situdecay on initial Pb isotopic ratios has been affected bymobilization of U and Th (and Pb) in the whole-rockssince their formation A thorough acid leaching duringsample preparation was used that has been shown to effec-tively remove alteration phases (Weis et al 2006 NobreSilva et al in preparation) Whole-rock SmNd ratiosand age-correction of Nd isotopes may be affected bythe leaching procedure for submarine rocks older than50 Ma (Thompson et al 2008 Fig 9) there ishowever very little variation in Nd isotopic compositionsof the picrites or tholeiitic basalts and this system does notappear to be disturbed by alteration The HFSE abun-dances for tholeiitic and picritic basalts exhibit clear
linear correlations in binary diagrams (Fig 8) whereasplots of LILE vs HFSE are highly scattered (not shown)because of the mobility of some of the alkali and alkalineearth LILE (Cs Rb Ba K) and Sr for most samplesduring alterationThe degree of alteration in the Karmutsen samples does
not appear to be related to eruption environment or depthof burial Pillow rims are compositionally different frompillow cores (Surdam1967 Kuniyoshi1972) and aquagenetuffs are chemically different from isolated pillows withinpillow breccias however submarine and subaerial basaltsexhibit a similar degree of alterationThere is no clear cor-relation between the submarine and subaerial basalts andsome commonly used chemical alteration indices (eg BaRb vs K2OP2O5 Huang amp Frey 2005) There is also nodefinitive correlation between the petrographic alterationindex (Table 1) and chemical alteration indices although10 of 13 tholeiitic basalts with the highest K and LILEabundances have the highest petrographic alterationindex of three (seeTable 1)
Table 4 Hf isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Lu Hf 176Hf177Hf 2m eHf176Lu177Hf 176Hf177Hft eHf(t)
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
OLIV INE ACCUMULAT ION INH IGH-MgO AND PICR IT IC LAVASGeochemical trends and petrographic characteristics indi-cate that accumulation of olivine played an important rolein the formation of the high-MgO basalts and picrites fromthe Keogh Lake area on northern Vancouver Island TheKeogh Lake samples show a strong linear correlation inplots of Al2O3 TiO2 ScYb and Ni vs MgO and many ofthe picrites have abundant clusters of olivine pseudomorphs(Table1 Fig11)There is a strong linear correlationbetween
modal per cent olivine and whole-rock magnesium contentmagnesium contents range between 10 and 19wt MgOand the proportion of olivine phenocrysts in these samplesvaries from 0 to 42 vol (Fig 11)With a few exceptionsboth the size range and median size of the olivine pheno-crysts in most samples are comparable The clear correla-tion between the proportion of olivine phenocrysts andwhole-rock magnesium contents indicates that accumula-tion of olivine was directly responsible for the high MgOcontents (410wt ) in most of the picritic lavas
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
207Pb204Pb
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
208Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
(a) (b)
(c) (e)
4723A24723A2
4723A2
4723A2
4722A4
4723A4
4722A4
4723A4
4723A134723A3
4723A3
4723A13
1556
1558
1560
1562
1564
1566
1568
186 190 194 198 202 206 2101550
1552
1554
1556
1558
1560
178 180 182 184 186 188 190 192 194
380
384
388
392
396
186 190 194 198 202 206 210374
378
382
386
390
178 180 182 184 186 188 190 192 194
206Pb204Pb(d)
LOI (wt )
0
1
2
3
4
5
6
7
186 190 194 198 202 206 210
4723A2
Fig 10 Pb isotopic compositions of leached whole-rock samples measured by MC-ICP-MS for the Karmutsen Formation Error bars are smal-ler than symbols (a) Measured 207Pb204Pb vs 206Pb204Pb (b) Initial 207Pb204Pb vs 206Pb204Pb Age-correction to 230 Ma (c) Measured208Pb204Pb vs 206Pb204Pb (d) LOI vs 206Pb204Pb (e) Initial 208Pb204Pb vs 206Pb204Pb Complete chemical duplicates shown in Table 5(samples 4720A7 and 4722A4) are circled in each plot The dashed lines in (b) and (e) show the differences in age-corrections for two picrites(4723A4 4722A4) using the measured UTh and Pb concentrations for each sample and the age-corrections when concentrations are used fromthe two picrites (4723A3 4723A13) that appear to be least affected by alteration
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
23
DISCUSS IONThe compositional and spatial development of the eruptedlava sequences of the Wrangellia oceanic plateau onVancouver Island provide information about the evolutionof a mantle plume upwelling beneath oceanic lithospherein an analogous way to the interpretation of continentalflood basalts The Caribbean and Ontong Java plateausare the two primary examples of accreted parts of oceanicplateaus that have been studied to date and comparisonscan be made with the Wrangellia oceanic plateauHowever the Wrangellia oceanic plateau is a distinctiveoccurrence of an oceanic plateau because it formed on thecrust of a Devonian to Mississippian intra-oceanic islandarc overlain by Mississippian to Permian limestones andpelagic sediments The nature and thickness of oceaniclithospheric mantle are considered to play a fundamentalrole in the decompression and evolution of a mantleplume and thus the compositions of the erupted lavasequences (Greene et al 2008) The geochemical and
stratigraphic relationships observed in the KarmutsenFormation onVancouver Island and the focus of the subse-quent discussion sections provide constraints on (1) thetemperature and degree of melting required to generatethe primary picritic magmas (2) the composition of themantle source (3) the depth of melting and residual min-eralogy in the source region (4) the low-pressure evolutionof the Karmutsen tholeiitic basalts (5) the growth of theWrangellia oceanic plateau
Melting conditions and major-elementcomposition of the primary magmasThe primary magmas to most flood basalt provinces arebelieved to be picritic (eg Cox 1980) and they have beenfound in many continental and oceanic flood basaltsequences worldwide (eg Siberia Karoo Paranacurren ^Etendeka Caribbean Deccan Saunders 2005) Near-pri-mary picritic lavas with low total alkali abundances
Table 5 Pb isotopic compositions of Karmutsen basaltsVancouver Island BC
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
require high-degree partial melting of the mantle source(eg Herzberg amp OrsquoHara 2002) The Keogh Lake picritesfrom northernVancouver Island are the best candidates foridentifying the least modified partial melts of the mantleplume source despite having accumulated olivine pheno-crysts and can be used to estimate conditions of meltingand the composition of the primary magmas for theKarmutsen FormationPrimary magma compositions mantle potential tem-
peratures and source melt fractions for the picritic lavas ofthe Karmutsen Formation were calculated from primitivewhole-rock compositions using PRIMELT1XLS software(Herzberg et al 2007) and are presented in Fig 12 andTable 6 A detailed discussion of the computationalmethod has been given by Herzberg amp OrsquoHara (2002)and Herzberg et al (2007) key features of the approachare outlined belowPrimary magma composition is calculated for a primi-
tive lava by incremental addition of olivine and compari-son with primary melts of mantle peridotite Lavas thathad experienced plagioclase andor clinopyroxene fractio-nation are excluded from this analysis All calculated pri-mary magma compositions are assumed to be derived byaccumulated fractional melting of fertile peridotite KR-4003 Uncertainties in fertile peridotite composition donot propagate to significant variations in melt fractionand mantle potential temperature melting of depletedperidotite propagates to calculated melt fractions that aretoo high but with a negligible error in mantle potential
temperature PRIMELT1XLS calculates the olivine liqui-dus temperature at 1atm which should be similar to theactual eruption temperature of the primary magmaand is accurate to 308C at the 2 level of confidenceMantle potential temperature is model dependent witha precision that is similar to the accuracy of the olivineliquidus temperature (C Herzberg personal communica-tion 2008)With regard to the evidence for accumulated olivine in
the Keogh Lake picrites it should not matter whether themodeled lava composition is aphyric or laden with accu-mulated olivine phenocrysts PRIMELT1 computes theprimary magma composition by adding olivine to a lavathat has experienced olivine fractionation in this way itinverts the effects of fractional crystallization The onlyrequirement is that the olivine phenocrysts are not acci-dental or exotic to the lava compositions Support for thisprocedure can be seen in the calculated olivine liquid-line-of-descent shown by the curved arrays with 5 olivineaddition increments (Fig 12c) Fractional crystallization ofolivine from model primary magmas having 85^95wt FeO and 153^175wt MgO will yield arrays of deriva-tive liquids and randomly sorted olivines that in most butnot all cases connect the picrites and high-MgO basalts Anewer version of the PRIMELT software (Herzberg ampAsimow 2008) can perform olivine addition to and sub-traction from lavas with highly variably olivine phenocrystcontents and yields results that converge with those shownin Fig 12
Fig 11 Relationship between abundance of olivine phenocrysts and whole-rock MgO contents for the Keogh Lake picrites (a) Tracings ofolivine grains (gray) in six high-MgO samples made from high resolution (2400 dpi) scans of petrographic thin-sections Scale bars represent10mm sample numbers are indicated at the upper left (b) Modal olivine vs whole-rock MgO Continuous line is the best-fit line for 10 high-MgO samplesThe areal proportion of olivine was calculated from raster images of olivine grains using ImageJ image analysis software whichprovides an acceptable estimation of the modal abundance (eg Chayes 1954)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
25
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
Gray lines = final melting pressure
L + Ol + Opx
07
08
09
Pressure (GPa)
10
3
4
5
6
1
SOLIDUS
01
04
0506
L + Ol
2
03
02
PeridotiteKR-4003
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
Thick black lines = initial melting pressure
Dashed lines = melt fraction
Melt Fraction
2
3 4 5 6
L+Ol+Opx
SolidusSpinel
SolidusGarnet Peridotite
7
L+Ol
0 10 20 30 4040
45
50
55
60
MgO (wt)
SiO2
(wt)
PeridotiteKR-4003
Melt Fraction
0 10 20 305
10
15
CaO(wt)
MgO (wt)
3
4 5 6 7
2
46
2
Peridotite Partial Melts
Thick black line = initial melting pressureGray lines = final melting pressure
Peridotite
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
02
Pressure (GPa)
0302
04
Pressure (GPa)
Picrite
High-MgO basalt
Tholeiitic basalt
(c)
(d)
Karmutsen flood basaltsOlivine addition modelOlivine addition (5 increment)Primary magma
PicriteHigh-MgO basaltTholeiitic basalt
(a)
(b)
800 850 900
910
920
930
940
950
Karmutsen flood basalts
Olivine addition modelOlivine addition (5 increment)Primary magma
Gray lines = Fo contents of olivine
Fig 12 Estimated primary magma compositions for three Keogh Lake picrites and high-MgO basalts (samples 4723A4 4723A135616A7) using the forward and inverse modeling technique of Herzberg et al (2007) Karmutsen compositions and modeling results areoverlain on diagrams provided by C Herzberg (a) Whole-rock FeO vs MgO for Karmutsen samples from this study Total iron estimatedto be FeO is 090 Gray lines show olivine compositions that would precipitate from liquid of a given MgO^FeO composition Black lineswith crosses show results of olivine addition (inverse model) using PRIMELT1 (Herzberg et al 2007) Gray field highlights olivinecompositions with Fo contents of 91^92 predicted to precipitate (b) FeO vs MgO showing Karmutsen lava compositions and results ofa forward model for accumulated fractional melting of fertile peridotite KR-4003 (Kettle River Peridotite Walter 1998) (c) SiO2 vsMgO for Karmtusen lavas and model results (d) CaO vs MgO showing Karmtusen lava compositions and model results To brieflysummarize the technique [see Herzberg et al (2007) for complete description] potential parental magma compositions for the high-MgOlava series were selected (highest MgO and appropriate CaO) and using PRIMELT1 software olivine was incrementally added tothe selected compositions to show an array of potential primary magma compositions (inverse model) Then using PRIMELT1 theresults from the inverse model were compared with a range of accumulated fractional melts for fertile peridotite derived fromparameterization of the experimental results for KR-4003 from Walter (1998) forward model Herzberg amp OrsquoHara (2002) A melt fractionwas sought that was unique to both the inverse and forward models (Herzberg et al 2007) A unique solution was found when there was a com-mon melt fraction for both models in FeO^MgO and CaO^MgO^Al2O3^SiO2 (CMAS) projection space This modeling assumes thatolivine was the only phase crystallizing and ignores chromite precipitation and possible augite fractionation in the mantle (Herzbergamp OrsquoHara 2002) Results are best for a residue of spinel lherzolite (not pyroxenite) The presence of accumulated olivine in samples ofthe Keogh Lake picrites used as starting compositions does not significantly affect the results because we are modeling addition of olivine(see text for further description) The tholeiitic basalts cannot be used for modeling because they are all saturated in plagthorn cpxthornolThick black line for initial melting pressure at 4 GPa is not shown in (c) for clarity Gray area in (a) indicates parental magmas thatwill crystallize olivine with Fo 91^92 at the surface Dashed lines in (c) indicate melt fraction Solidi for spinel and garnet peridotite arelabelled in (c)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
The PRIMELT1 results indicate that if the Karmutsenpicritic lavas were derived from accumulated fractionalmelting of fertile peridotite the primary magmas wouldhave contained 15^17wt MgO and 10wt CaO(Fig 12 Table 6) formed from 23^27 partial meltingand would have crystallized olivine with a Fo content of 91 (Fig 12 Table 6) Calculated mantle temperatures forthe Karmutsen picrites (14908C) indicate melting ofanomalously hot mantle (100^2008C above ambient) com-pared with ambient mantle that produces mid-ocean ridgebasalt (MORB 1280^14008C Herzberg et al 2007)which initiated between 3 and 4 GPa Several picritesappear to be near-primary melts and required very littleaddition of olivine (eg sample 93G171 with measured158wt MgO) These temperature and melting esti-mates of the Karmutsen picrites are consistent withdecompression of hot mantle peridotite in an actively con-vecting plume head (ie plume initiation model) Threesamples (picrite 5614A1 and high-MgO basalts 5614A3and 5614A5) are lower in iron than the other Karmutsenlavas with FeO in the 65^80wt range (Fig 12c) andhave MgO and FeO contents that are similar to primitiveMORB from the Sequeiros fracture zone of the EastPacific Rise (EPR Perfit et al 1996) These samples
however differ from EPR MORB in having higher SiO2
and lower Al2O3 and they are restricted geographicallyto one area (Sara Lake Fig 2)We therefore have evidence for primary magmas with
MgO contents that range from a possible low of 110wt to as high as 174wt with inferred mantle potential tem-peratures from 1340 to 15208C This is similar to the rangedisplayed by lavas from the Ontong Java Plateau deter-mined by Herzberg (2004) who found that primarymagma compositions assuming a peridotite sourcewould contain 17wt MgO and 10wt CaO(Table 6) and that these magmas would have formed by27 melting at a mantle potential temperature of15008C (first melting occurring at 36 GPa) Fitton ampGodard (2004) estimated 30 partial melting for theOntong Java lavas based on the Zr contents of primarymagmas of Kroenke-type basalt calculated by incrementaladdition of equilibrium olivine However if a considerableamount of eclogite was involved in the formation of theOntong Java plateau magmas the excess temperatures esti-mated by Herzberg et al (2007) may not be required(Korenaga 2005) The variability in mantle potential tem-perature for the Keogh Lake picrites is also similar to thewide range of temperatures that have been calculated for
Table 6 Estimated primary magma compositions for Karmutsen basalts and other oceanic plateaus or islands
Sample 4723A4 4723A13 5616A7 93G171 Average OJP Mauna Keay Gorgonaz
Ontong Java primary magma composition for accumulated fraction melting (AFM) from Herzberg (2004)yMauna Kea primary magma composition is average of four samples of Herzberg (2006 table 1)zGorgona primary magma composition for AFM for 1F fertile source from Herzberg amp OrsquoHara (2002 table 4)Total iron estimated to be FeO is 090
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
27
lavas from single volcanoes in the Galapagos Hawaii andelsewhere by Herzberg amp Asimow (2008) Those workersinterpreted the range to reflect the transport of magmasfrom both the cool periphery and the hotter axis of aplume A thermally heterogeneous plume will yield pri-mary magmas with different MgO contents and the rangeof primary magma compositions and their inferred mantlepotential temperatures for Karmutsen lavas may reflectmelting from different parts of the plume
Source of the Karmutsen Formation lavasThe Sr^Nd^Hf^Pb isotopic compositions of theKarmutsen volcanic rocks on Vancouver Island indicatethat they formed from plume-type mantle containing adepleted component that is compositionally similar to butdistinct from other oceanic plateaus that formed in thePacific Ocean (eg Ontong Java and Caribbean Fig 13)Trace element and isotopic constraints can be used to testwhether the depleted component of the KarmutsenFormation was intrinsic to the mantle plume and distinctfrom the source of MORB or the result of mixing or invol-vement of ambient depleted MORB mantle with plume-type mantle The Karmutsen tholeiitic basalts have initialeNd and
87Sr86Sr ratios that fall within the range of basaltsfrom the Caribbean Plateau (eg Kerr et al 1996 1997Hauff et al 2000 Kerr 2003) and a range of Hawaiian iso-tope data (eg Frey et al 2005) and lie within and abovethe range for the Ontong Java Plateau (Tejada et al 2004Fig 13a) Karmutsen tholeiitic basalts with the highest ini-tial eNd and lowest initial 87Sr86Sr lie within and justbelow the field for age-corrected northern EPR MORB at230 Ma (eg Niu et al1999 Regelous et al1999) Initial Hfand Nd isotopic compositions for the Karmutsen samplesare noticeably offset to the low side of the ocean islandbasalt (OIB) array (Vervoort et al 1999) and lie belowand slightly overlap the low eHf end of fields for Hawaiiand the Caribbean Plateau (Fig 13c) the Karmutsen sam-ples mostly fall between and slightly overlap a field forage-corrected Pacific MORB at 230 Ma and Ontong Java(Mahoney et al 1992 1994 Nowell et al 1998 Chauvel ampBlichert-Toft 2001) Karmutsen lava Pb isotope ratios over-lap the broad range for the Caribbean Plateau and thelinear trends for the Karmutsen samples intersect thefields for EPR MORB Hawaii and Ontong Java in208Pb^206Pb space but do not intersect these fields in207Pb^206Pb space (Fig 13d and e) The more radiogenic207Pb204Pb for a given 206Pb204Pb of the Karmutsenbasalts requires high UPb ratios at an ancient time when235U was abundant similar to the Caribbean Plateau andenriched in comparison with EPR MORB Hawaii andOntong JavaThe low eHf^low eNd component of the Karmutsen
basalts that places them below the OIB mantle array andpartly within the field for age-corrected Pacific MORBcan form from low ratios of LuHf and high SmNd over
time (eg Salters amp White 1998) MORB will develop aHf^Nd isotopic signature over time that falls below themantle array Low LuHf can also lead to low 176Hf177Hffrom remelting of residues produced by melting in theabsence of garnet (eg Salters amp Hart 1991 Salters ampWhite 1998) The Hf^Nd isotopic compositions of theKarmutsen Formation indicate the possible involvementof depleted MORB mantle however similar to Iceland(Fitton et al 2003) Nb^Zr^Y variations provide a way todistinguish between a depleted component within theplume and a MORB-like component The Karmutsenbasalts and picrites fall within the Iceland array and donot overlap the MORB field in a logarithmic plot of NbYvs ZrY (Fig 13b) using the fields established by Fittonet al (1997) Similar to the depleted component within theCaribbean Plateau (Thompson et al 2003) the trend ofHf^Nd isotopic compositions towards Pacific MORB-likecompositions is not followed by MORB values in Nb^Zr^Y systematics and this suggests that the depleted compo-nent in the source of the Karmutsen basalts is distinctfrom the source of MORB and probably an intrinsic partof the plume The relatively radiogenic Pb isotopic ratiosand REE patterns distinct from MORB also support thisinterpretationTrace-element characteristics suggest the possible invol-
vement of sub-arc lithospheric mantle in the generation ofthe Karmutsen picrites Lassiter et al (1995) suggested thatmixing of a plume-type source with eNdthorn 6 to thorn 7 witharc material with low NbTh could reproduce the varia-tions observed in the Karmutsen basalts however theyconcluded that the absence of low NbLa ratios in theirsample of tholeiitic basalts restricts the amount of arc litho-sphere involved The study of Lassiter et al (1995) wasbased on major and trace element and Sr Nd and Pb iso-topic data for a suite of 29 samples from Buttle Lake in theStrathcona Provincial Park on central Vancouver Island(Fig 1) Whereas there are no clear HFSE depletions inthe Karmutsen tholeiitic basalts the low NbLa (and TaLa) of the Karmutsen picritic lavas in this study indicatethe possible involvement of Paleozoic arc lithosphere onVancouver Island (Fig 8)
REE modeling dynamic melting andsource mineralogyThe combined isotopic and trace-element variations of thepicritic and tholeiitic Karmutsen lavas indicate that differ-ences in trace elements may have originated from differentmelting histories For example the LuHf ratios of theKarmutsen picritic lavas (mean 008) are considerablyhigher than those in the tholeiitic basalts (mean 002)however the small range of overlapping initial eHf valuesfor the two lava suites suggests that these differences devel-oped during melting within the plume (Fig 9b) Below wemodel the effects of source mineralogy and depth of
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
28
melting on differences in trace-element concentrations inthe tholeiitic basalts and picritesDynamic melting models simulate progressive decom-
pression melting where some of the melt fraction isretained in the residue and only when the degree of partialmelting is greater than the critical mass porosity and the
source becomes permeable is excess melt extracted fromthe residue (Zou1998) For the Karmutsen lavas evolutionof trace element concentrations was simulated using theincongruent dynamic melting model developed by Zou ampReid (2001) Melting of the mantle at pressures greater than1 GPa involves incongruent melting (eg Longhi 2002)
OIB array
PacificMORB(230 Ma)
(0 Ma)
EPR(230 Ma)
-4
-2
0
2
4
8
10
12
14
16
18
20
22
-4 -2 0 2 4 6 8 10 12 14
minus2
0
2
4
6
8
10
12
14
0702 0703 0704 0705 0706
IndianMORB
87Sr 86Sr (initial)
Hf (initial)
(a) (c)
Nd (initial)
Karmutsen tholeiitic basalt
HawaiiOntong JavaCaribbean Plateau
Karmutsen high-MgO lava
high-MgO lavasKarmutsen
1540
1545
1550
1555
1560
1565
175 180 185 190 195 200375
380
385
390
395
175 180 185 190 195 200
(d) (e)
EPR
EPR
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb
Karmutsen Formation (initial)
HawaiiOntong JavaCaribbean Plateau
East Pacific Rise
Karmutsen Formation (measured)
Juan de FucaGorda
ε Nd
(initial)
Karmutsen Formation (different age-correction for 3 picrites)
(0 Ma)
NbY
001
01
1
1 10
ZrY
OIB
NMORB
(b)
Iceland array
6
Fig 13 Comparison of age-corrected (230 Ma) Sr^Nd^Hf^Pb isotopic compositions for Karmutsen flood basalts onVancouver Island withage-corrected OIB and MORB and Nb^Zr^Y variation (a) Initial eNd vs 87Sr86Sr (b) NbYand ZrY variation (after Fitton et al 1997)(c) Initial eHf vs eNd (d) Measured and initial 207Pb204Pb vs 206Pb204Pb (e) Measured and initial 208Pb204Pb vs 206Pb204Pb References fordata sources are listed in Electronic Appendix 4 Most of the compiled data were extracted from the GEOROC database (httpgeorocmpch-mainzgwdgdegeoroc) OIB array line in (c) is fromVervoort et al (1999) EPR East Pacific Rise Several high-MgO samples affected bysecondary alteration were not plotted for clarity in Pb isotope plots Estimation of fields for age-corrected Pacific MORB and EPR at 230 Mawere made using parent^daughter concentrations (in ppm) from Salters amp Stracke (2004) of Lufrac14 0063 Hffrac14 0202 Smfrac14 027 Ndfrac14 071Rbfrac14 0086 and Srfrac14 836 In the NbYvs ZrYplot the normal (N)-MORB and OIB fields not including Icelandic OIB are from data com-pilations of Fitton et al (1997 2003) The OIB field is data from 780 basalts (MgO45wt ) from major ocean islands given by Fitton et al(2003) The N-MORB field is based on data from the Reykjanes Ridge EPR and Southwest Indian Ridge from Fitton et al (1997) The twoparallel gray lines in (b) (Iceland array) are the limits of data for Icelandic rocks
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
29
with olivine being produced during the reactioncpxthornopxthorn spfrac14meltthornol for spinel peridotites The mod-eling used coefficients for incongruent melting reactionsbased on experiments on lherzolite melting (eg Kinzleramp Grove 1992 Longhi 2002) and was separated into (1)melting of spinel lherzolite (2) two combined intervals ofmelting for a garnet lherzolite (gt lherz) and spinel lher-zolite (sp lherz) source (3) melting of garnet lherzoliteThe model solutions are non-unique and a range indegree of melting partition coefficients melt reactioncoefficients and proportion of mixed source componentsis possible to achieve acceptable solutions (see Fig 14 cap-tion for description of modeling parameters anduncertainties)The LREE-depleted high-MgO lavas require melting of
a depleted spinel lherzolite source (Fig 14a) The modelingresults indicate a high degree of melting (22^25) similarto the results from PRIMELT1 based on major-elementvariations (discussed above) of a LREE-depleted sourceThe high-MgO lavas lack a residual garnet signature Theenriched tholeiitic basalts involved melting of both garnetand spinel lherzolite and represent aggregate melts pro-duced from continuous melting throughout the depth ofthe melting column (Fig 14b) Trace-element concentra-tions in the tholeiitic and picritic lavas are not consistentwith low-degree melting of garnet lherzolite alone(Fig 14c) The modeling results indicate that the aggregatemelts that formed the tholeiitic basalts would haveinvolved lesser proportions of enriched small-degreemelts (1^6 melting) generated at depth and greateramounts of depleted high-degree melts (12^25) gener-ated at lower pressure (a ratio of garnet lherzolite tospinel lherzolite melt of 14 provides the best fits seeFig 14b and description in figure caption) The totaldegree of melting for the tholeiitic basalts was high(23^27) similar to the degree of partial melting in thespinel lherzolite facies for the primary melts that eruptedas the Keogh Lake picrites of a source that was also prob-ably depleted in LREE
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
(c)
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
Melting of garnet lherzolite
High-MgO lavas
Melting of spinel lherzolite
source (07DM+03PM)
source (07DM+03PM)
6 melting
30 melting
(b)
(a)
30 melting
Sam
ple
Cho
ndrit
eS
ampl
eC
hond
rite 20 melting
1 melting
Melting of garnet lherzoliteand spinel lherzolite
x gt lherz + 4x sp lherz
5 meltingx gt lherz + 4x sp lherz
Sam
ple
Cho
ndrit
e
Tholeiitic lavas
Tholeiitic lavas
High-MgO lavas
source (07DM+03PM)
Fig 14 Trace-element modeling results for incongruent dynamicmantle melting for picritic and tholeiitic lavas from the KarmutsenFormation Three steps of modeling are shown (a) melting of spinellherzolite compared with high-MgO lavas (b) melting of garnet lher-zolite and spinel lherzolite compared with tholeiitic lavas (c) meltingof garnet lherzolite compared with tholeiitic and picritic lavasShaded fields in each panel for tholeiitic basalts and picrites arelabeled Patterns with symbols in each panel are modeling results(patterns are 5 melting increments in (a) and (b) and 1 incre-ments in (c) PM primitive mantle DM depleted MORB mantle gtlherz garnet lherzolite sp lherz spinel lherzolite In (b) the ratios ofper cent melting for garnet and spinel lherzolite are indicated (x gtlherzthorn 4x sp lherz means that for 5 melting 1 melt in garnetfacies and 4 melt in spinel facies where x is per cent melting in theinterval of modeling) The ratio of the respective melt proportions inthe aggregate melt (x gt lherzthorn 4x sp lherz) was kept constant andthis ratio of melt contribution provided the best fit to the data Arange of proportion of source components (07DMthorn 03PM to
09DMthorn 01PM) can reproduce the variation in the high-MgOlavas Melting modeling uses the formulation of Zou amp Reid (2001)an example calculation is shown in their Appendix PM fromMcDonough amp Sun (1995) and DM from Salters amp Stracke (2004)Melt reaction coefficients of spinel lherzolite from Kinzler amp Grove(1992) and garnet lherzolite fromWalter (1998) Partition coefficientsfrom Salters amp Stracke (2004) and Shaw (2000) were kept constantSource mineralogy for spinel lherzolite (018cpx027opx052ol003sp) from Kinzler (1997) and for garnet lherzolite(034cpx008opx053ol005gt) from Salters amp Stracke (2004) Arange of source compositions for melting of spinel lherzolite(07DMthorn 03PM to 03DMthorn 07PM) reproduce REE patternssimilar to those of the high-MgO lavas with best fits for 22ndash25melting and 07DMthorn 03PM source composition Concentrationsof tholeiitic basalts cannot be reproduced from an entirelyprimitive or depleted source
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
30
The uncertainty in the composition of the source in thespinel lherzolite melting model for the high-MgO lavasprecludes determining whether the source was moredepleted in incompatible trace elements than the source ofthe tholeiitic lavas (see Fig 14 caption for description) It ispossible that early high-pressure low-degree melting left aresidue within parts of the plume (possibly the hotter inte-rior portion) that was depleted in trace elements (egElliott et al 1991) further decompression and high-degreemelting of such depleted regions could then have generatedthe depleted high-MgO Karmutsen lavas Shallow high-degree melting would preferentially sample depletedregions whereas deeper low-degree melting may notsample more refractory depleted regions (eg CaribbeanEscuder-Viruete et al 2007) The picrites lie at the highend of the range of Nd isotopic compositions and havelow NbZr similar to depleted Icelandic basalts (Fittonet al 2003) therefore the picrites may represent the partsof the mantle plume that were the most depleted prior tothe initiation of melting As Fitton et al (2003) suggestedthese depleted melts may only rarely be erupted withoutcombining with more voluminous less depleted melts andthey may be detected only as undifferentiated small-volume flows like the Keogh Lake picrites within theKarmutsen Formation on Vancouver IslandDecompression melting within the mantle plume initiatedwithin the garnet stability field (35^25 GPa) at highmantle potential temperatures (Tp414508C) and pro-ceeded beneath oceanic arc lithosphere within the spinelstability field (25^09 GPa) where more extensivedegrees of melting could occur
Magmatic evolution of Karmutsentholeiitic basaltsPrimary magmas in large igneous provinces leave exten-sive coarse-grained cumulate residues within or below thecrust these mafic and ultramafic plutonic sequences repre-sent a significant proportion of the original magma thatpartially crystallized at depth (eg Farnetani et al 1996)For example seismic and petrological studies of OntongJava (eg Farnetani et al 1996) combined with the use ofMELTS (Ghiorso amp Sack 1995) indicate that the volcanicsequence may represent only 40 of the total magmareaching the Moho Neal et al (1997) estimated that30^45 fractional crystallization took place for theOntong Java magmas and produced cumulate thicknessesof 7^14 km corresponding to 11^22 km of flood basaltsdepending on the presence of pre-existing oceanic crustand the total crustal thickness (25^35 km) Interpretationsof seismic velocity measurements suggest the presence ofpyroxene and gabbroic cumulates 9^16 km thick beneaththe Ontong Java Plateau (eg Hussong et al 1979)Wrangellia is characterized by high crustal velocities in
seismic refraction lines which is indicative of mafic
plutonic rocks extending to depth beneath VancouverIsland (Clowes et al 1995)Wrangellia crust is 25^30 kmthick beneath Vancouver Island and is underlain by astrongly reflective zone of high velocity and density thathas been interpreted as a major shear zone where lowerWrangellia lithosphere was detached (Clowes et al 1995)The vast majority of the Karmutsen flows are evolved tho-leiitic lavas (eg low MgO high FeOMgO) indicating animportant role for partial crystallization of melts at lowpressure and a significant portion of the crust beneathVancouver Island has seismic properties that are consistentwith crystalline residues from partially crystallizedKarmutsen magmas To test the proportion of the primarymagma that fractionated within the crust MELTS(Ghiorso amp Sack 1995) was used to simulate fractionalcrystallization using several estimated primary magmacompositions from the PRIMELT1 modeling results Themajor-element composition of the primary magmas couldnot be estimated for the tholeiitic basalts because of exten-sive plagthorn cpxthornol fractionation and thus the MELTSmodeling of the picrites is used as a proxy for the evolutionof major elements in the volcanic sequence A pressure of 1kbar was used with variable water contents (anhydrousand 02wt H2O) calculated at the quartz^fayalite^magnetite (QFM) oxygen buffer Experimental resultsfrom Ontong Java indicate that most crystallizationoccurs in magma chambers shallower than 6 km deep(52 kbar Sano amp Yamashita 2004) however some crys-tallization may take place at greater depths (3^5 kbarFarnetani et al 1996)A selection of the MELTS results are shown in Fig 15
Olivine and spinel crystallize at high temperature until15^20wt of the liquid mass has fractionated and theresidual magma contains 9^10wt MgO (Fig 15f) At1235^12258C plagioclase begins crystallizing and between1235 and 11908C olivine ceases crystallizing and clinopyr-oxene saturates (Fig 15f) As expected the addition of clin-opyroxene and plagioclase to the crystallization sequencecauses substantial changes in the composition of the resid-ual liquid FeO and TiO2 increase and Al2O3 and CaOdecrease while the decrease in MgO lessens considerably(Fig 15) The near-vertical trends for the Karmutsen tho-leiitic basalts especially for CaO do not match the calcu-lated liquid-line-of-descent very well which misses manyof the high-CaO high-Al2O3 low-FeO basalts A positivecorrelation between Al2O3CaO and SrNd indicates thataccumulation of plagioclase played a major role in the evo-lution of the tholeiitic basalts (Fig 15g) Increasing thecrystallization pressure slightly and the water content ofthe melts does not systematically improve the match ofthe MELTS models to the observed dataThe MELTS results indicate that a significant propor-
tion of the crystallization takes place before plagioclase(15^20 crystallization) and clinopyroxene (35^45
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
31
crystallization) join the crystallization sequence Theseresults suggest that445 crystallization of the primarymagmas occurred (estimated from the residual liquid percent in Fig 15f) and if the original flood basalt stratigra-phy on Vancouver Island was 6 km thick at least anequivalent amount of complementary plutonic rocksshould have crystallized The Karmutsen basalts and theirplutonic residues thus represent a significant addition ofcrust (perhaps 412 km thickness) previously thickenedduring the formation of the underlying Paleozoic arc
Growth of the Wrangellia oceanic plateauon Vancouver IslandThe extrusive part of the Wrangellia oceanic plateau onVancouver Island was constructed in a three-layersequence of mostly tholeiitic basalts with a restrictedrange of composition (Fig 16) The plateau overliesDevonian to Mississippian arc volcanic sequences andmarine sediments of Mississippian to Permian ages thatare intruded by gabbroic rocks related to the flood basaltsThe volcanic stratigraphy of theWrangellia plateau formed
00
05
10
15
20
25
30
35
40
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
16
0 2 4 6 8 10 12 14 16 18 20
10
11
12
13
14
15
16
17
18
19
20
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
0 2 4 6 8 10 12 14 16 18 20
MgO (wt)
0 10 20 30 40 50
percent of total mass
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
) Mineral proportions
Sp (e)
Ol
CpxPlag
PicriteHigh-MgO basaltTholeiitic basalt
TiO2 (wt)
MgO (wt)
Al2O3 (wt) CaO (wt)
FeO (wt)
MgO(a) (b)
(c) (d)
MgO (wt)MgO (wt)
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
)
Residual liquid ()
1220
1190
Ol + Sp
Plag+Ol+Sp Plag+Cpx+Sp
02 wt H2O
no H2O
Residual liquid and crystallizing phases
(f)
0102030405060708090100
plagioclaseaccumulation
plagioclaseaccumulation
plagioclase
accumulatio
n
10
12
14
16
0 10 20 30 40
Al2O3CaO
SrNd
tholeiitic basalts
(g)
Fig 15 Forward fractional crystallization modeling results for major elements from MELTS (Ghiorso amp Sack 1995) compared with picriticand tholeiitic lavas from the Karmutsen Formation (a^d) Major-element compositions of residual liquids compared with Karmutsen lavacompositions For clarity the results are shown only for one starting composition (open circle) using the composition of the estimated primarymagma for the high-MgO lava series from the modeling technique of Herzberg et al (2007) for sample 4723A13 as shown in Fig 12 The esti-mated primary magmas for samples 4723A3 and 93G171 yield similar results A pressure of 1 kbar was used with no H2O and 02wt H2Ocalculated at QFM black line is anhydrous and grey line is 02wt H2OThe end-point for the anhydrous result is at 78 fractionation and11308C and the end-point for the result with 02wt H2O is 75 fractionation and11108C (e) Mineral proportions of crystallized phases fromone MELTrun (sample 4723A13) with no H2O (f) Per cent residual liquid vs temperature labelled with intervals of crystallizing phases (g)Al2O3CaO vs SrNd for the tholeiitic basalts
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
32
in a single eruptive phase (c 231^225 Ma) as an emergentbasalt sequence during a deeper-water shallow-water andsubaerial stage and subsided below sea level relativelysoon after its formation (Fig 16) The deeper-water stagecontains pillowed and massive submarine flows (3000mthick) that built up the submarine volcanic edifice from theseafloor The lowest pillow basalts were emplaced onunconsolidated and lithified fine-grained siliciclastic andcarbonaceous sediments containing Daonella beds(c 235^232 Ma) that are intruded by abundantKarmutsen mafic sills (Fig 16a) The different pillowedand massive submarine flows resulted primarily from dif-ferent effusive rates and local topography Some of themassive submarine flows commonly recognized by radiat-ing columnar jointing represent master tubes for the deliv-ery of lava to distal parts of submarine flow fields Thepillowed flows which commonly have uneven topography
in modern sequences may have served as levees to themassive submarine flows rather than allowing unob-structed spreading and inflation typical of the massive sub-aerial flows that form on shallow slopes in continentalflood basalt provinces (eg Self et al 1997) The rapid suc-cession of flows is evident by the near absence of sedimentbetween submarine flows Minor volumes of picritic pillowlavas occur in the upper part of the pillowed flow sequencein areas of the plateau exposed on northern VancouverIsland (Fig 16 Nixon et al 2008)The growth of the plateau in shallow water led to
emplacement of an increasing proportion of volcaniclastics(pillow breccias and hyaloclastites) that accumulated up to1500m thick in some areas (Fig 16b Nixon et al 2008)Horizons of pillowed flows occur locally within the volca-niclastic stratigraphy Sedimentary structures (eg gradedbedding fluidization structures) and petrographic textures
Late Devonian-Mississippian arc
Daonella beds
Middle Triassic
deeper water
Late Devonian-Mississippian arc
pillowed flows massive submarine
flows
pillow brecciahyaloclastiteshallow water
~4 km
~6 km
Late Triassic limestoneinterflow limestonelenses
Karmutsen Formation
mafic intrusions sediment-sill complex
Karmutsen Formation
(a) (b)
(c) (d)
subaerial flows
Late Devonian-Mississippian arc Late Devonian-Mississippian arc
picritic pillow basalts
Miss to Permian marine sediments
sills
Fig 16 Growth of the Wrangellia oceanic plateau onVancouver Island (a) Deep-water eruptive phase of pillowed and massive submarineflows with almost no intervening sediments onto MiddleTriassic shale containing Daonella beds (c 235^232 Ma) The basement of the plateauis composed of Devonian to Mississippian arc sequences overlain by Mississippian to Permian carbonates and siliciclastic sediments(b) Shallow-water flow emplacement with increasing proportion of volcaniclastics resulted primarily from cooling^contraction granulationPicritic pillow basalts occur primarily near the transition between pillowed submarine flows and volcaniclastic units Pillowed and massiveflows occur locally within the volcaniclastic unit (c) Emergent stage of growth consists of overlapping subaerial flow fields similar to continen-tal flood basalts with pahoehoe flow structures Multi-tiered columnar jointing occurrs locally in response to ponding of flows (eg Bondre et al2004) Interflow carbonate deposits and plagioclase megacrystic flows developed locally during the waning stage of volcanism (d) Post-volcanicsubsidence and deposition of Late Carnian to Norian (c 212^223 Ma) platform carbonates on top of the plateau Units are labeled in the dia-grams Thicknesses of the underlying units are not to scale
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
33
within the volcaniclastic rocks indicate formation primar-ily via cooling^contraction granulation magma^water^steam interaction autobrecciation resedimentation andmass-wasting rather than pyroclastic eruption (Nixonet al 2008)A broad subaerial platform was constructed by emplace-
ment of subaerial lava flows during an emergent subaerialstage (Fig 16c) The subaerial unit predominantly com-prises well-layered sheet flows that formed in a similarway to continental flood basalts as hundred(s) of pahoehoeflow fields erupted on shallow slopes (Self et al 1997) how-ever volcaniclastic and pillowed flow units are locally pre-sent mostly in the upper parts of the stratigraphyLocalized ponding of flows led to the formation of multi-tiered columnar jointing in some subaerial flows (egBondre et al 2004) As volcanism waned local interflowcarbonate deposits developed and plagioclase megacrysticflows were erupted Subsidence of the plateau is recordedby the interflow lenses and deposits of platform carbonates(40^1500m thick) on top of the Karmutsen basalts thatcontain Late Carnian to Early Norian fossils (Fig 16d c216^227 Ma Nixon et al 2008)The evolution of the internal architecture of the
Wrangellia oceanic plateau was primarily related to theconditions of emplacement (ie water depth and eruptionenvironment) There does not appear to have been a sys-tematic coincident evolution or shift of lava compositionsonVancouver Island as has been recognized in most con-tinental flood basalts and parts of the Wrangellia oceanicplateau in Alaska and Yukon (Greene et al 2008 2009)In general with the exception of the picritic pillow lavasin the submarine stratigraphy and local variations betweenhigh-MgO and tholeiitic lavas there are no significantcompositional differences between lavas erupted duringthe different stages of growth of the plateau (ie submarineand subaerial flows) nor are substantial compositional dif-ferences generally found between massive submarine flowsin contact with pillowed flows tholeiitic basalt flows nearthe base of the Karmutsen Formation are compositionallysimilar to flows sampled near the top of the volcanicsequence Where it was possible to sample continuous sec-tions of volcanic stratigraphy greater than several hun-dreds of meters (eg Keogh^Maynard Lake area west ofthe Karmutsen Range Fig 2) there is no clear relationshipbetween stratigraphic position and compositionalvariation
CONCLUSIONSThe Karmutsen Formation covers extensive areas ofVancouver Island (20 000 km2) and forms part of a largeoceanic plateau erupted over a geologically short interval(c 225^231 Ma) The tripartite volcanic stratigraphyexposed on Vancouver Island is at least 6 km thick andcomprises a succession of submarine flows volcaniclastic
deposits and massive subaerial flowsThese volcanologicaldifferences are primarily related to the eruption environ-ment (deep-water shallow-water or subaerial) The rapidgrowth of the plateau prevented significant accumulationsof intervening sediment except in the uppermost stratigra-phy where isolated limestone lenses commonly associatedwith pillow basalt and hyaloclastite deposits preserve arecord of the final stages of subsidence of the plateauThe Wrangellia plateau on Vancouver Island was con-
structed dominantly of tholeiitic basalt with restrictedmajor- and trace-element and isotopic compositional var-iations Minor volumes of picritic pillow basalts eruptedlate in the submarine phase The Karmutsen basalts wereemplaced onto older Paleozoic arc volcanic sequences andlow NbLa for the picrites may indicate a contributionfrom the sub-arc lithospheric mantle Modeling of the pet-rogenesis of the Karmutsen picrites indicates melting ofanomalously hot mantle (15008C) and extensive degreesof partial melting (23^27) and is consistent with aplume initiation model Combined Sr^Nd^Hf^Pb isotopicsystematics indicate that the picritic and tholeiitic lavaswere derived from a depleted plume-type mantle sourcedistinct from the source of MORB that has compositionalsimilarities to the source of the lavas forming theCaribbean Plateau Picrites were generated by melting ofspinel lherzolite whereas the tholeiitic basalts formed asaggregate melts of both garnet and spinel lherzolite Thetholeiitic basalts underwent extensive low-pressure fractio-nation (52^3 kbar) and seismic studies indicate that someof theWrangellia crust beneathVancouver Island may cor-respond to the plutonic residues of the Karmutsen floodbasalts
ACKNOWLEDGEMENTSWe are grateful to Nick Arndt for helping us get this pro-ject started and Nick Massey for insights into VancouverIsland geology We greatly appreciate the assistance withPRIMELT modeling from Claude Herzberg We thankMikkel Schau for his insight and enthusiasm during field-work Jane Barling assisted with analyses by MC-ICP-MSThoughtful and thorough reviews of the manuscript byJohn Mahoney Fred Frey Wendy Bohrson DougalJerram and Teal Riley are greatly appreciated Fundingwas generously provided by the Rocks to Riches Programadministered by the BC amp Yukon Chamber of Mines in2004 by the BC Geological Survey and by NSERCDiscovery Grants to JSS and DW ARG was supportedby a University Graduate Fellowship at UBC
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
34
REFERENCESBarker F Brown A S Budahn J R amp Plafker G (1989) Back-arc
with frontal-arc component origin of Triassic Karmutsen basaltBritish Columbia Canada Chemical Geology 75 81^102
Blichert-Toft JWeis D Maerschalk C Agranier A amp Albare de F(2003) Hawaiian hot spot dynamics as inferred from the Hf and Pbisotope evolution of Mauna Kea volcano Geochemistry GeophysicsGeosystems 4(2) 1^27 doi1010292002GC000340
Bondre N R Duraiswami R A amp Dole G (2004) Morphologyand emplacement of flows from the Deccan Volcanic ProvinceIndia Bulletin of Volcanology 66 29^45
Carlisle D (1963) Pillow breccias and their aquagene tuffs QuadraIsland British Columbia Journal of Geology 71 48^71
Carlisle D (1972) Late Paleozoic to Mid-Triassic sedimentary-volcanic sequence on Northeastern Vancouver Island Geological
Society of Canada Paper Report of Activities 72-1 Part B 24^30Carlisle D amp SuzukiT (1974) Emergent basalt and submergent car-
bonate^clastic sequences including the Upper Triassic Dilleri andWelleri zones on Vancouver Island Canadian Journal of Earth
Sciences 11 254^279Chauvel C amp Blichert-Toft J (2001) A hafnium isotope and trace
element perspective on melting of the depleted mantle Earth and
Planetary Science Letters 190 137^151Chayes F (1954)The theory of thin-section analysis Journal of Geology
62 92^101Cho M Liou J G amp Maruyama S (1987) Transition from the zeo-
lite to prehnite^pumpellyite facies in the Karmutsen metabasitesVancouver Island British Columbia Journal of Petrology 27(2)467^494
Clowes R M Zelt C A Amor J R amp Ellis R M (1995)Lithospheric structure in the southern Canadian Cordillera froma network of seismic refraction lines Canadian Journal of Earth
Sciences 32(10) 1485^1513Coffin M F amp Eldholm O (1994) Large igneous provinces Crustal
structure dimensions and external consequences Reviews of
Geophysics 32(1) 1^36Cox K G (1980) A model for flood basalt volcanism Journal of
Petrology 21 629^650Eldholm O amp Coffin M F (2000) Large igneous provinces
and plate tectonics In Richards M A Gordon R G amp vander Hilst R D (eds) The History and Dynamics of Global
Plate Motions American Geophysical Union Geophysical Monograph 121309^326
Elliott T R Hawkesworth C J amp Grolaquo nvold K (1991) Dynamicmelting of the Iceland plume Nature 351 201^206
Escuder-Viruete J Pecurren rez-Estaucurren n A Contreras F Joubert MWeis D Ullrich T D amp Spadea P (2007) Plume mantle sourceheterogeneity through time Insights from the Duarte ComplexHispaniola northeastern Caribbean Journal of Geophysical Research112(B04203) doi1010292006JB004323
Farnetani C G Richards M A amp Ghiorso M S (1996)Petrological models of magma evolution and deep crustal structurebeneath hotspots and flood basalt provinces Earth and Planetary
Science Letters 143 81^94Fitton J G amp Godard M (2004) Origin and evolution of magmas
on the Ontong Java Plateau In Fitton J G Mahoney J JWallace P J amp Saunders A D (eds) Origin and Evolution of the
Ontong Java Plateau Geological Society London Special Publications 229151^178
Fitton J G Saunders A D Norry M J Hardarson B S ampTaylor R N (1997) Thermal and chemical structure of theIceland plume Earth and Planetary Science Letters 153 197^208
Fitton J G Saunders A D Kempton P D amp Hardarson B S(2003) Does depleted mantle form an intrinsic part of the Icelandplume Geochemistry Geophysics Geosystems 4(3) 1032 doi1010292002GC000424
Frey F A Huang S Blichert-Toft J Regelous M amp Boyet M(2005) Origin of depleted components in basalt related to theHawaiian hotspot Evidence from isotopic and incompatible ele-ment ratios Geochemistry Geophysics Geosystems 6(1) 23 doi1010292004GC000757
Galer S J G amp Abouchami W (1998) Practical application of leadtriple spiking for correction of instrumental mass discriminationMineralogical Magazine 62A 491^492
Ghiorso M S amp Sack R O (1995) Chemical mass transfer in mag-matic processes IV A revised and internally consistent thermody-namic model for the interpolation and extrapolation of liquid^solidequilibria in magmatic systems at elevated temperatures and pres-sures Contributions to Mineralogy and Petrology 119 197^212
Greene A R Scoates J S amp Weis D (2005)WrangelliaTerrane onVancouver Island Distribution of flood basalts with implicationsfor potential Ni^Cu^PGE mineralization In Grant B (ed)Geological Fieldwork 2004 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2005-1 209^220Greene A R Scoates J S Nixon G T amp Weis D (2006) Picritic
lavas and basal sills in the Karmutsen flood basalt provinceWrangellia northern Vancouver Island In Grant B (ed)Geological Fieldwork 2005 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2006-1 39^52Greene A R Scoates J S amp Weis D (2008) Wrangellia flood
basalts in Alaska A record of plume^lithosphere interaction in aLate Triassic accreted oceanic plateau Geochemistry Geophysics
Geosystems 9(Q12004) doi1010292008GC002092Greene A R Scoates J S Weis D amp Israel S (2009)
Geochemistry of triassic flood basalts from the Yukon (Canada)segment of the accreted Wrangellia oceanic plateau Lithosdoi101016jlithos200811010
Hauff F Hoernle K Tilton G Graham D W amp Kerr A C(2000) Large volume recycling of oceanic lithosphere over shorttime scales geochemical constraints from the Caribbean LargeIgneous Province Earth and Planetary Science Letters 174 247^263
Hauff F Hoernle K amp Schmidt A (2003) Sr^Nd^Pb compositionof Mesozoic Pacific oceanic crust (Site 1149 and 801 ODP Leg185)Implications for alteration of ocean crust and the input into theIzu^Bonin^Mariana subduction system Geochemistry Geophysics
Geosystems 4(8) doi1010292002GC000421Herzberg C (2004) Partial melting below the Ontong Java Plateau
In Fitton J G Mahoney J J Wallace P J amp Saunders A D(eds) Origin and Evolution of the Ontong Java Plateau Geological SocietyLondon Special Publications 229 179^183
Herzberg C (2006) Petrology and thermal structure of the Hawaiianplume from Mauna Kea volcano Nature 444(7119) 605
Herzberg C amp Asimow P D (2008) Petrology of some oceanicisland basalts PRIMELT2XLS software for primary magma cal-culation Geochemistry Geophysics Geosystems 9(Q09001) doi1010292008GC002057
Herzberg C amp OrsquoHara M J (2002) Plume-associated ultramaficmagmas of Phanerozoic age Journal of Petrology 43(10) 1857^1883
Herzberg C Asimow P D Arndt N Niu Y Lesher C MFitton J G Cheadle M J amp Saunders A D (2007)Temperatures in ambient mantle and plumes Constraints frombasalts picrites and komatiites Geochemistry Geophysics Geosystems8(Q02006) doi1010292006GC001390
Huang S amp Frey F A (2005) Trace element abundances of MaunaKea basalt from phase 2 of the Hawaii Scientific Drilling Project
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
35
Petrogenetic implications of correlations with major element con-tent and isotopic ratios Geochemistry Geophysics Geosystems 4(6)doi1010292002GC000322
Hussong D M Wipperman L K amp Kroenke L W (1979) Thecrustal structure of the Ontong Java and Manihiki OceanicPlateaus Journal of Geophysical Research 84(B11) 6003^6010
Jerram D A amp Widdowson M (2005) The anatomy of ContinentalFlood Basalt Provinces geological constraints on the processes andproducts of flood volcanism Lithos 79(3^4) 385^405
Jones D L Silberling N J amp Hillhouse J (1977)Wrangellia a dis-placed terrane in northwestern North America CanadianJournal ofEarth Sciences 14(11) 2565^2577
Kerr A C (2003) Oceanic Plateaus In Rudnick R (ed) The Crust
Treatise on GeochemistryVol 3 Oxford Elsevier Science pp 537^565Kerr A C amp Mahoney J J (2007) Oceanic plateaus Problematic
plumes potential paradigms Chemical Geology 241 332^353Kerr A CTarney J Marriner G F Klaver GT Saunders A D
amp Thirlwall M F (1996) The geochemistry and petrogenesis ofthe Late-Cretaceous picrites and basalts of Curacao NetherlandsAntilles a remnant of an oceanic plateau Contributions to
Mineralogy and Petrology 124(1) 29^43Kerr A C Marriner G F Tarney J Nivia A Saunders A D
Thirlwall M F amp Sinton A C (1997) Cretaceous basaltic ter-ranes in Western Colombia Elemental chronological and Sr-Ndisotopic constraints on petrogenesis Journal of Petrology 38(6)677^702
Kinzler R J (1997) Melting of mantle peridotite at pressuresapproaching the spinel to garnet transition Application to mid-ocean ridge basalt petrogenesis Journal of Geophysical Research
102(B1) 853^874Kinzler R J amp Grove T L (1992) Primary magmas of mid-ocean
ridge basalts 1 Experiments and methods Journal of Geophysical
Research 97(B5) 6885^6906Korenaga J (2005) Why did not the Ontong Java Plateau form sub-
aerially Earth and Planetary Science Letters 234 385^399Kuniyoshi S (1972) Petrology of the Karmutsen (Volcanic) Group
northeastern Vancouver Island British Columbia PhD disserta-tion University of California Los Angeles 242 pp
Lassiter J C DePaolo D J amp Mahoney J J (1995) Geochemistry ofthe Wrangellia flood basalt province Implications for the role ofcontinental and oceanic lithosphere in flood basalt genesis Journalof Petrology 36(4) 983^1009
LeBas M J (2000) IUGS reclassification of the high-Mg and picriticvolcanic rocks Journal of Petrology 41(10) 1467^1470
Longhi J (2002) Some phase equilibrium systematics of lherzolitemelting I Geochemistry Geophysics Geosystems 3(3) doi1010292001GC000204
MacDonald G A amp Katsura T (1964) Chemical composition ofHawaiian lavas Journal of Petrology 5 82^133
Mahoney J J (1987) An isotopic survey of Pacific oceanic plateausimplications for their nature and origin In Keating B HFryer P Batiza R amp Boehlert GW (eds) Seamounts Islands andAtolls American Geophysical Union Geophysical Monograph 43 207^220
Mahoney J LeRoex A P Peng Z Fisher R L amp Natland J H(1992) Southwestern limits of Indian Ocean Ridge mantle and theorigin of low 206Pb204Pb mid-ocean ridge basalt isotope systema-tics of the central Southwest Indian Ridge (178^508E) Journal ofGeophysical Research 97(B13) 19771^19790
Mahoney J J Sinton J M Kurz M D MacDougall J DSpencer K J amp Lugmair GW (1994) Isotope and trace elementcharacteristics of a super-fast spreading ridge East Pacific rise13^238S Earth and Planetary Science Letters 121 173^193
Massey N W D (1995a) Geology and mineral resources of theAlberni^Nanaimo Lakes sheet Vancouver Island 92F1W 92F2Eand part of 92F7E British Columbia Ministry of Energy Mines and
Petroleum Resources Paper 1992-2 132 ppMassey N W D (1995b) Geology and mineral resources of the
Cowichan Lake sheet Vancouver Island 92C16 British Columbia
Ministry of Energy Mines and Petroleum Resources Paper 1992-3 112 ppMassey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005a) Digital Geology Map of British Columbia Tile NM9 MidCoast BC British Columbia Ministry of Energy and Mines Geofile
2005-2Massey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005b) Digital Geology Map of British Columbia Tile NM10Southwest BC British Columbia Ministry of Energy and Mines Geofile
2005-3McDonough W F amp Sun S (1995) The composition of the Earth
Chemical Geology 120 223^253Monger JW H amp Journeay J M (1994) Basement geology and tec-
tonic evolution of theVancouver region In Monger JW H (ed)Geology and Geological Hazards of the Vancouver Region Southwestern
British Columbia Geological Survey of Canada Bulletin 481 3^25Muller J E (1977) Geology of Vancouver Island Geological Survey of
Canada Open File Map 463Muller J E (1980)The Paleozoic Sicker Group of Vancouver Island British
Columbia Geological Survey of Canada Paper 79-30 22 ppMuller J E Northcote K E amp Carlisle D (1974) Geology and
mineral deposits of Alert Bay^Cape Scott map area VancouverIsland British Columbia Geological Survey of Canada Paper 74-8 77
Neal C R Mahoney J J Kroenke L W Duncan R A ampPetterson M G (1997) The Ontong^Java Plateau InMahoney J J amp Coffin M F (eds) Large Igneous Provinces
Continental Oceanic and Planetary FloodVolcanism American Geophysical
Union Geophysical Monograph 100 183^216Niu Y Collerson K D Batiza R Wendt J I amp Regelous M
(1999) Origin of enriched-typed mid-ocean ridge basalt at ridgesfar from mantle plumes The East Pacific Rise at 11820rsquoN Journalof Geophysical Research 104 7067^7088
Nixon G T amp Orr A J (2007) Recent revisions to the EarlyMesozoic stratigraphy of Northern Vancouver Island (NTS 102I092L) and metallogenic implicatinos British Columbia InGrant B (ed) Geological Fieldwork 2006 British Columbia Ministry of
Energy Mines and Petroleum Resources Paper 2007-1 163^177Nixon GT Hammack J L KoyanagiV M Payie G J Haggart
JW Orchard M JTozerT Friedman R M Archibald D APalfy J amp Cordey F (2006a) Geology of the Quatsino^PortMcNeill area northernVancouver Island British Columbia Ministry
of Energy Mines and Petroleum Resources Geoscience Map 2006-2 scale150 000
Nixon G T Kelman M C Stevenson D Stokes L A ampJohnston K A (2006b) Preliminary geology of the Nimpkishmap area (NTS 092L07) northern Vancouver Island BritishColumbia In Grant B amp Newell J M (eds) Geological Fieldwork2005 British Columbia Ministry of Energy Mines and Petroleum Resources
Paper 2006-1 135^152Nixon G T Laroque J Pals A Styan J Greene A R amp
Scoates J S (2008) High-Mg lavas in the Karmutsen flood basaltsnorthernVancouver Island (NTS 092L) Stratigraphic setting andmetallogenic significance In Grant B (ed) Geological Fieldwork
2007 British Columbia Ministry of Energy Mines and Petroleum
Resources Paper 2008-1 175^190Nowell G M Kempton P D Noble S R Fitton J G
Saunders A Mahoney J J amp Taylor R N (1998) High precisionHf isotope measurements of MORB and OIB by thermal ionisation
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
36
mass spectrometry insights into the depleted mantle Chemical
Geology 149 211^233Parrish R R amp McNicoll V J (1992) U^Pb age determinations
from the southern Vancouver Island area British Columbia InRadiogenic Age and Isotopic Studies Report 5 Geological Survey of
Canada Paper 91-2 79^86Peate DW (1997) The Parana^Etendeka Province In Coffin M F
amp Mahoney J (eds) Large Igneous Provinces Continental Oceanic andPlanetary Flood Volcanism American Geophysical Union Geophysical
Monograph 100 217^245Perfit M R Fornari D J Ridley W I Kirk P D Casey J
Kastens K A Reynolds J R Edwards M Desonie DShuster R amp Paradis S (1996) Recent volcanism in the Siqueirostransform fault picritic basalts and implications for MORBmagma genesis Earth and Planetary Science Letters 141(1^4) 91^108
Pretorius W Weis D Williams G Hanano D Kieffer B ampScoates J S (2006) Complete trace elemental characterization ofgranitoid (USGSG-2GSP-2) reference materials by high resolutioninductively coupled plasma-mass spectrometry Geostandards and
Geoanalytical Research 30(1) 39^54Regelous M Niu Y Wendt J I Batiza R Greig A amp
Collerson K D (1999) Variations in the geochemistry of magma-tism on the East Pacific Rise at 10830rsquoN since 800 ka Earth and
Planetary Science Letters 168 45^63Recurren villon S Chauvel C Arndt N T Pik R Martineau F
Fourcade S amp Marty B (2002) Heterogeneity of the Caribbeanplateau mantle source Sr O and He isotopic compositions of oli-vine and clinopyroxene from Gorgona Island Earth and Planetary
Science Letters 205(1^2) 91Rhodes J M (1996) Geochemical stratigraphy of lava flows samples
by the Hawaii Scientific Drilling Project Journal of Geophysical
Research 101(B5) 11729^11746Rhodes J M amp Vollinger M J (2004) Composition of basaltic lavas
sampled by phase-2 of the Hawaii Scientific Drilling ProjectGeochemical stratigraphy and magma types Geochemistry
Geophysics Geosystems 5(3) doi1010292002GC000434Richards M A Jones D L Duncan R A amp DePaolo D J (1991)
A mantle plume initiation model for the Wrangellia flood basaltand other oceanic plateaus Science 254 263^267
Salters V J M amp Hart S R (1991) The mantle sources of oceanridges islands and arcs the Hf-isotope connection Earth and
Planetary Science Letters 104 364^380Salters V J M amp Stracke A (2004) Composition of the depleted
SaltersV J amp WhiteW (1998) Hf isotope constraints on mantle evo-lution Chemical Geology 145 447^460
SanoT amp Yamashita S (2004) Experimental petrology of basementlavas from Ocean Drilling Program Leg 192 implications for dif-ferentiation processes in Ontong Java Plateau magmas InFitton J G Mahoney J JWallace P J amp Saunders A D (eds)Origin and Evolution of the Ontong Java Plateau Geological Society
London Special Publications 229 185^218Saunders A D (2005) Large igneous provinces Origin and environ-
mental consequences Elements 1 259^263Self S ThordarsonT amp Keszthely L (1997) Emplacement of conti-
nental flood basalt lava flows In Mahoney J J amp Coffin M F(eds) Large Igneous Provinces Continental Oceanic and Planetary Flood
Volcanism American Geophysical Union Geophysical Monograph 100381^410
Shaw D M (2000) Continuous (dynamic) melting theory revisitedCanadian Mineralogist 38(5) 1041^1063
Sluggett C L (2003) Uranium^lead age and geochemical constraintson Paleozoic and Early Mesozoic magmatism in WrangelliaTerrane Saltspring Island British Columbia BSc thesisUniversity of British ColumbiaVancouver 84 pp
Storey M Mahoney J J amp Saunders A D (1997) Cretaceousbasalts in Madagascar and the transition between plume and con-tinental lithospheric mantle sources In Mahoney J J ampCoffin M F (eds) Large Igneous Provinces Continental Oceanic and
Planetary Flood Volcanism Aamerican Geophysical Union Geophysical
Monograph 100 95^122Surdam R C (1967) Low-grade metamorphism of the Karmutsen
Group PhD dissertation University of California Los Angeles288 pp
Sutherland-Brown A Yorath C J Anderson R G amp Dom K(1986) Geological maps of southern Vancouver IslandLITHOPROBE I 92C10 11 14 16 92F1 2 7 8 Geological Survey ofCanada Open File 1272
Tejada M L G Mahoney J J Castillo P R Ingle S P Sheth HC amp Weis D (2004) Pin-pricking the elephant evidence on theorigin of the Ontong Java Plateau from Pb^Sr^Hf^Nd isotopiccharacteristics of ODP Leg 192 basalts In Fitton J GMahoney J J Wallace P J amp Saunders A D (eds) Origin and
Evolution of the Ontong Java Plateau Geological Society London Special
Publications 229 133^150Thompson P M E Kempton P D amp Kerr A C (2008) Evaluation
of the effects of alteration and leaching on Sm^Nd and Lu^Hf sys-tematics in submarine mafic rocks Lithos 104(1^4) 164^176
Thompson P M E Kempton P D White R V Kerr A CTarney J Saunders A D Fitton J G amp McBirney A (2003)Hf-Nd isotope constraints on the origin of the CretaceousCaribbean plateau and its relationship to the Galapagos plumeEarth and Planetary Science Letters 217(1^2) 59^75
Vervoort J D Patchett P Blichert-Toft J amp Albare de F (1999)Relationships between Lu^Hf and Sm^Nd isotopic systems in theglobal sedimentary system Earth and Planetary Science Letters 16879^99
Walter M J (1998) Melting of garnet peridotite and the origin ofkomatiite and depleted lithosphere Journal of Petrology 39(1) 29^60
Weis D Kieffer B Maerschalk C Barling J de Jong JWilliams G A Hanano D Mattielli N Scoates J SGoolaerts A Friedman R A amp Mahoney J B (2006) High-pre-cision isotopic characterization of USGS reference materials byTIMS and MC-ICP-MS Geochemistry Geophysics Geosystems
7(Q08006) doi1010292006GC001283Weis D Kieffer B Hanano D Silva I N Barling J
Pretorius W Maerschalk C amp Mattielli N (2007) Hf isotopecompositions of US Geological Survey reference materialsGeochemistry Geophysics Geosystems 8(Q06006) doi1010292006GC001473
Wheeler J O amp McFeely P (1991) Tectonic assemblage map of theCanadian Cordillera and adjacent part of the United States ofAmerica Geological Survey of Canada Map 1712A
WhiteW M Albare de F amp Tecurren louk P (2000) High-precision analy-sis of Pb isotope ratios by multi-collector ICP-MS Chemical Geology167 257^270
Yorath C J Sutherland Brown A amp Massey N W D (1999)LITHOPROBE southern Vancouver Island British Columbia Geological
Survey of Canada Bulletin 498 145 ppZou H (1998) Trace element fractionation during modal and
non-model dynamic melting and open-system melting Amathematical treatment Geochimica et Cosmochimica Acta 62(11)1937^1945
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
37
Zou H amp Reid M R (2001) Quantitative modeling of trace elementfractionation during incongruent dynamic melting Geochimica et
Cosmochimica Acta 65(1) 153^162
APPENDIX SAMPLEPREPARAT ION ANDANALYT ICAL METHODSOnly the freshest rocks in the field were sampled and onlythe least altered samples were selected for chemical andisotopic analysis based on thorough petrographic inspec-tion Fifty-nine of the 129 collected samples were crushed(400 g) into small pieces 52mm in diameter in aRocklabs hydraulic piston crusher between WC plates tominimize contamination The coarse-crush was mixedand 100 g was powdered in a planetary mill using agatejars and balls cleaned with quartz sand between samples
ActLabs analytical methodsThe major- and trace-element compositions of the whole-rock powders were determined at Activation LaboratoriesLtd (Actlabs) in Ancaster Ontario Analytical techniquesand detection limits are also available from Actlabs (httpwwwactlabscommethsub_code4erehtm) The analyticalmethod for each of the elements analyzed is indicated inTable 2 For the major elements a 02 g sample was mixedwith a lithium metaborate^lithium tetraborate mixtureand fused in a graphite crucible The molten mixture waspoured into a 5 HNO3 solution and shaken until dis-solved (30min) The samples were analyzed for majoroxides and selected trace elements by inductively coupledplasma optical emission spectrometry (ICP-OES) on acombination simultaneous^sequential Thermo Jarrell-AshEnviro II system Internal calibration was achieved usinga variety of international reference materials (eg W-2BIR-1 DNC-1) and independent control samplesAdditional trace elements were analyzed by both instru-mental neutron activation analysis (INAA) and induc-tively couple plasma mass spectrometry (ICP-MS) ForINAA15^25 g of sample was weighed into small polyeth-ylene vials and irradiated with a control international ref-erence material CANMET WMS-1 and NiCr flux wires ata thermal neutron flux of 71012 neutronscm2 per s in theMcMaster Nuclear Reactor Following a 7 day waitingperiod the samples were measured on an Ortec high-purity Ge detector linked to a Canberra Series 95 multi-channel analyzer Activities for each element weredecay- and weight-corrected and compared with a detectorcalibration developed from multiple international certifiedreference materials For ICP-MS 025 g of sample wasdigested in HF followed by HNO3^HClO4 heated andtaken to dryness The samples were brought back into solu-tion with HCl Samples were analyzed using a PerkinElmer Optima 3000 ICP In-lab standards or certified ref-erence materials (eg W-2 BIR-1 DNC-1) were used for
quality control A total of 15 blind duplicates were ana-lyzed from the same sample powders to assessreproducibility
University of Massachusetts analyticalmethodsFourteen sample duplicate powders taken from the same100 g sample aliquot (high-MgO basalts and picrites)were also analyzed at the Ronald B Gilmore X-RayFluorescence (XRF) Laboratory at the University ofMassachusetts Major elements were measured on a fusedLa-bearing lithium borate glass disc using a SiemensMRS-400 spectrometer Trace element concentrations(Rb Sr Ba Ce Nb Zr Y Pb Zn Ga Ni Cr V) weremeasured on a separate powder pellet using a PhilipsPW2400 sequential spectrometer using a Rh tube LOIand ferrous iron measurements were made as describedby Rhodes amp Vollinger (2004) Precision and accuracy esti-mates for the data have been described by Rhodes (1996)and Rhodes amp Vollinger (2004) Results for each sampleare the average of two separate analyses and are shown inElectronic Appendix 2 for a comparison of analyses ofduplicate powders using XRFand ICP-MS
PCIGR analytical methodsA subset of 19 samples was selected for high-precisiontrace-element analysis and Sr Nd Pb and Hf isotopicanalysis at the Pacific Centre for Isotopic andGeochemical Research (PCIGR) at the University ofBritish Columbia (UBC) Samples were selected from the59 samples analyzed for whole-rock chemistry at ActLabsbased on major- and trace-element chemistry alteration(low LOI and petrographic alteration index) and samplelocation Samples were prepared for trace-element analysisat the PCIGR by the technique described by Pretorius et al(2006) on unleached rock powders Sample powders(100mg) were weighed in 7ml screw-top Savillex bea-kers and dissolved in 1ml 14N HNO3 and 5ml 48 HFon a hotplate for 48 h at 1308C with periodic ultrasonica-tion Samples were dried and redissolved in 6ml 6N HClon a hotplate for 24 h and then dried and redissolved in1ml concentrated HNO3 for 24 h before final dryingTrace element abundances were measured by high-resolution inductively coupled plasma-mass spectrometry(HR-ICP-MS) with aThermo Finnigan Element2 systemfollowing the procedures described by Pretorius et al(2006) within 24 h of redissolution HFSE and LILE weremeasured in medium-resolution mode at 2000 dilutionusing a PFATeflon spray chamber washed with aqua regiafor 3min between samples REE were measured in high-resolution mode and U and Pb in low-resolution modeat 2000 dilution using a glass spray chamber washedwith 2 HNO3 between samples Total procedural blanksand reference materials (BCR-2 BHVO-2) were analyzedwith the batch of samples Indium was used as an internal
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
38
standard in all samples and standard solutionsBackground and standard solutions were analyzed afterevery five samples to detect memory effects and massdrift Results for PCIGR trace-element analyses areshown in Electronic Appendix 3 for the full set of analyzedtrace elementsSample digestion for purification of Sr Nd Hf and Pb
for column chemistry began by weighing each samplepowder (400^500mg) prior to leaching All samples wereleached with 6N HCl and placed in an ultrasonic bath for15min Samples were rinsed twice with 18 M-cm H2Obetween each leaching step (15 total) until the supernatantliquid was clear [following the method of Mahoney (1987)]Samples were then dried on a hotplate for 24 h andweighed again Sample solutions were then prepared bydissolving 100^250mg of the leached powder dissolvedin 1ml 14N HNO3 and 10ml 48 HF on a hotplate for48 h at 1308C with periodic ultrasonication Samples weredried and redissolved in 6ml 6N HCl on a hotplate for24 h and then dried Pb was separated using anion exhangecolumns and the discard was used for Sr REE and Hfseparation Nd was separated from the other REE and Hfrequired two additional purification steps Detailed proce-dures for column chemistry for separating Sr Nd and Pbat the PCIGR have been described byWeis et al (2006) andHf purification has been described byWeis et al (2007) Srand Nd isotope ratios were measured by thermal ioniza-tion mass spectrometry (TIMS) on a Thermo FinniganTriton system in static mode with relay matrix rotation ona single Ta and double Re^Ta filament respectively Fourto five filaments per barrel of 21 were occupied by stan-dards (NBS 987 for Sr and LaJolla for Nd) for each barrelwhere samples were run Sample Sr and Nd isotopic com-positions were corrected for mass fractionation using86Sr88Srfrac14 01194 and 146Nd144Ndfrac14 07219 Each samplewas then normalized using the barrel average of the refer-ence material relative to the values of143Nd144Ndfrac14 0511858 and 87Sr86Srfrac14 0710248 (Weiset al 2006) During the course of the Vancouver Islandanalyses the La Jolla Nd standard gave an average valueof 05118566 (nfrac14 7) and NBS987 standard gave an aver-age of 07102408 (nfrac1411) (2 error is reported as times106) 147Sm144Nd errors are 15 or 0006 USGeological Survey (USGS) reference material BHVO-2was processed with the samples and yielded Sr and Nd iso-topic ratios of 07034607 and 05129786 respectivelyThese are in agreement with the published values of070347920 and 051298411 respectively (Weis et al2006)
Pb and Hf isotopic compositions were analyzed by mul-tiple collector-inductively coupled plasma-mass spectrome-try (MC-ICP-MS) on a static multi-collection on a NuPlasma (Nu Instruments) system The detailed analyticalprocedure for Pb isotopic analyses on the Nu Plasma atthe PCIGR has been described by Weis et al (2006) Theconfiguration for Pb analyses allows for collection of PbTl and Hg togetherTl and Hg are used to monitor instru-mental mass discrimination and isobaric overlap repec-tively All sample solutions were analyzed withapproximately the same PbTl ratio (4) as the referencematerial NIST SRM 981 To accomplish this a small ali-quot of each sample solution from the Pb columns was ana-lyzed on the Element2 to determine the precise amount ofPb available for analysis on the Nu Plasma The SRM 981standard was run after every two samples on the NuPlasma During the time samples were run analyses ofthe SRM 981 Pb reference material gave values of206Pb204Pbfrac1416940322 207Pb204Pbfrac1415495823and 208Pb204Pbfrac14 36713164 (nfrac14 61 2 error is reportedas times 104) these values are in excellent agreement withreported TIMS triple-spike values of Galer amp Abouchami(1998) Fractionation-corrected Pb isotopic ratios were fur-ther corrected by the sample-standard bracketing methodor the ln^ln correction method described by White et al(2000) and Blichert-Toft et al (2003) Leached powders ofUSGS reference material BHVO-2 yielded Pb isotopicratios of 206Pb204Pbfrac14186454 8 207Pb204Pbfrac14 1549105 and 208Pb204Pbfrac14 38222514 and BCR-2yielded 206Pb204Pbfrac141880466 207Pb204Pbfrac14 1562518 and 208Pb204Pbfrac14 3883496 (2 error isreported as times 104) These values are in agreement withthose for leached residues of BHVO-2 and BCR-2 fromWeis et al (2006)Hf isotopic compositions were analyzed following the
procedures detailed byWeis et al (2007) The configurationfor Hf analyses monitored Lu mass 175 andYb mass 172 toallow for interference correction to masses 174 and 176 Hfisotopic ratios were normalized internally for mass fractio-nation to a 179Hf177Hf ratio of 07325 using an exponentialcorrection Standards were run after every two samplesand sample results were normalized to the ratio of the in-run daily average and a 176Hf177Hf ratio for JMC-475 of0282160 During the course of analyses the Hf standardJMC-475 gave an average value 02821679 (nfrac14130)USGS reference materials BCR-2 and BHVO-2 were pro-cessed with the samples and yielded Hf isotopic ratios of02828675 and 02831005 respectively Publishedvalues for BCR-2 and BHVO-2 are 02828717 and0283104 8 respectively (Weis et al 2007)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
39
Table 1 Summary of petrographic characteristics and phenocryst proportions of Karmutsen basalts on Vancouver Island
BC
Sample Area Flow Group Texture vol Ol Plag Cpx Ox Alteration Notes
4718A1 MA PIL THOL intersertal 20 5 3 few plag glcr52mm cpx51 mm
4718A2 MA PIL THOL intersertal glomero 15 1 plag glcr52mm very fresh
4718A5 MA FLO THOL intersertal glomero 10 2 mottled few plag glcr54 mm
4723A4 KR PIL PIC intergranular intersertal 0 3 swtl plag51mm no ol phenos
4723A13 KR PIL PIC spherulitic 24 3 swtl plag51 mm
(continued)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
8
by an assemblage of talc^tremolite^serpentine^opaqueoxides there is no fresh olivine present in the Karmutsenlavas onVancouver Island Plagioclase and clinopyroxenephenocrysts are absent in the high-MgO lavas Thecoarse-grained mafic rocks are characterized by sub-ophi-tic textures with an average grain size typically 41mmand are generally not glomeroporphyritic these rocks aremainly from the interiors of massive flows although somemay represent sillsSample preparation and analytical methods for major-
and trace-element and Sr^Nd^Hf^Pb isotopic composi-tions of the Karmutsen volcanic rocks are given in theAppendix
WHOLE -ROCK CHEMISTRYMajor- and trace-element compositionsThe most abundant type of volcanic rock in theKarmutsen Formation is tholeiitic basalt with a restrictedrange of major- and trace-element compositions The tho-leiitic basalts have similar compositions to the coarse-grained mafic rocks and both groups are distinct fromthe picrites and high-MgO basalts [Fig 6 picrites412wt MgO (LeBas 2000) high-MgO basalts8^12wt MgO] The tholeiitic basalts have lower MgO(57^77wt MgO) and higher TiO2 (14^22wt
TiO2) than the picrites (130^198wt MgO05^07wt TiO2) and high-MgO basalts (91^116wt MgO 05^08wt TiO2 Fig 6 Table 2) Almost allcompositions plot within the tholeiitic field in a total alka-lis vs silica plot although there has been substantial K lossin most samples from the Karmutsen Formation (seebelow) and the tholeiitic basalts generally have higherSiO2 Na2OthornK2O CaO and FeOT (total iron expressedas FeO) than the picrites The tholeiitic basalts also havenoticeably lower loss on ignition (LOI mean1713wt ) than the picrites (mean 54 08wt )and high-MgO basalts (mean 3815wt ) (Fig 6)reflecting the presence of abundant altered olivinephenocrysts in the latter groups Ni concentrations are sig-nificantly higher for the picrites (339^755 ppm) and high-MgO basalts (122^551ppm) than the tholeiitic basalts(58^125 ppm except for one anomalous sample) andcoarse-grained rocks (70^213 ppm) (Fig 6 Table 2) Ananomalous tholeiitic pillowed flow (two samples outlierin Tables 1 and 2) from near the picrite type locality atKeogh Lake and a mineralized sill (disseminated sulfide)from the basal sediment^sill complex at Schoen Lake aredistinguished by higher TiO2 and FeOT than the maingroup of tholeiitic basalts The tholeiitic basalts are similarin major-element composition to previously publishedresults for samples from the Karmutsen Formation how-ever the Keogh Lake picrites which encompass the
Table 1 Continued
Sample Area Flow Group Texture vol Ol Plag Cpx Ox Alteration Notes
4722A5 KR FLO OUTLIER intersertal 3 mottled very fg v small pl needles
5615A11 KR PIL OUTLIER intersertal 3 mottled very fg v small pl needles
5617A4 SL SIL MIN intersertal 5 3 fg
Sample number last digit of year month day initial sample station (except 93G171) MA Mount Arrowsmith SLSchoen Lake KR Karmutsen Range PIL pillow BRE breccia FLO flow SIL sill GAB gabbro THOL tholeiiticbasalt PIC picrite HI-MG high-MgO basalt CGcoarse-grained MIN mineralized sill glomero glomeroporphyriticOlivine modes for picrites are based on area calculations from scans (see lsquoOlivine accumulation in high-MgO and picriticlavasrsquo section and Fig 11) Visual alteration index is based primarily on degree of plagioclase alteration and presence ofsecondary minerals (1 least altered 3 most altered) Plagioclase phenocrysts are commonly altered to albite pumpellyiteand chlorite olivine is altered to talc tremolite and clinochlore (determined using the Rietveld method of X-ray powderdiffraction) clinopyroxene is unaltered FendashTi oxide is commonly replaced by sphene glcr glomerocrysts fg fine-grained cg coarse-grained oik oikocryst chad chadacryst enc enclosing swtl swallow-tail ol olivinepseudomorphs plag plagioclase cpx clinopyroxene ox oxides (includes ilmenite thorn titanomagnetite)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
9
00
05
10
15
20
25
30
35
40
0 5 10 15 20 25
10
11
12
13
14
15
16
17
18
19
20
0 5 10 15 20 25
6
7
8
9
10
11
12
13
14
15
0 5 10 15 20 25
6
7
8
9
10
11
12
13
14
15
16
0 5 10 15 20 25
0
1
2
3
4
5
6
40 45 50 55 60
0
200
400
600
800
0 5 10 15 20 25
(wt)(wt)
(wt)
(wt)
+K2ONa2O TiO2
FeO (T)
Ni (ppm)
Al2O3
CaO
alkalic
tholeiitic
Karmutsen Form
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
(compiled data)
Pillowed flowMineralized sill
SiO2 (wt) MgO (wt)
MgO (wt)MgO (wt)
MgO (wt) MgO (wt)
(a) (b)
(g)
(e)
(f)
(c)
LOI (wt )
(d)
MgO (wt )0
1
2
3
4
5
6
7
0 5 10 15 20 25
(wt)
Fig 6 Whole-rock major-element Ni and LOI variation diagrams for the Karmutsen Formation New samples from this study (Table 2) areshown by symbols with black outlines (see legend) Previously published analyses are shown by small gray symbols without black outlines Theboundary of the alkaline and tholeiitic fields is that of MacDonald amp Katsura (1964) Total iron expressed as FeO LOI loss on ignition oxidesare plotted on an anhydrous normalized basis References for the 322 compiled analyses for the Karmutsen Formation are Surdam (1967)Kuniyoshi (1972) Muller et al (1974) Barker et al (1989) Lassiter et al (1995) Massey (1995a1995b)Yorath et al (1999) and G Nixon (unpublisheddata) It should be noted that the compiled dataset has not been filtered many of the samples with high SiO2 (452wt ) are probably notKarmutsen flood basalts but younger Bonanza basalts and andesites Three pillowed flow samples and a mineralized sill are distinguishedseparately because of their anomalous chemistry Coarse-grained samples are indicated separately Three tholeiitic basalt samples (4724A54718A5 4718A6) with416wt Al2O3 and4270 ppm Sr contain 10^25 modal of large plagioclase phenocrysts (7^8mm) or glomerocrysts(44mm)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
10
Table 2 Major element (wt oxide) and trace element (ppm) abundances in whole-rock samples of Karmutsen basalts
1Ni concentrations for these samples were determined by XRFTHOL tholeiitic basalt PIC picrite HI-MG high MgO basalt CG coarse-grained (sill or gabbro) MIN SIL mineralizedsill OUTLIER anomalous pillowed flow in plots MA Mount Arrowsmith SL Schoen Lake KR Karmutsen Range QIQuadra Island Sample locations are given using the Universal Transverse Mercator (UTM) coordinate system (NAD83zones 9 and 10) Analyses were performed at Activation Laboratory (ActLabs) Fe2O3
is total iron expressed as Fe2O3LOI loss on ignition All major elements Sr V and Y were measured by ICP quadrupole OES on solutions of fusedsamples Cu Ni Pb and Zn were measured by total dilution ICP Cs Ga Ge Hf Nb Rb Ta Th U Zr and REE weremeasured by magnetic-sector ICP on solutions of fused samples Co Cr and Sc were measured by INAA Blanks arebelow detection limit See Electronic Appendices 2 and 3 for complete XRF data and PCIGR trace element datarespectively Major elements for sample 93G17 are normalized anhydrous Sample 5615A7 was analyzed by XRFSamples from Quadra Island are not shown in figures
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
17
picritic and high-MgO basalt pillow lavas extend to20wt MgO (Fig 6 and references listed in thecaption)The tholeiitic basalts from the Karmutsen Range dis-
play a tight range of parallel light rare earth element(LREE)-enriched REE patterns (LaYbCNfrac14 21^24mean 2202) whereas the picrites and high-MgObasalts are characterized by sub-parallel LREE-depleted
patterns (LaYbCNfrac14 03^07 mean 06 02) with lowerREE abundances than the tholeiitic basalts (Fig 7)Tholeiitic basalts from the Schoen Lake and MountArrowsmith areas have similar REE patterns tosamples from the Karmutsen Range (Schoen Lake LaYbCNfrac14 21^26 mean 2303 Mount Arrowsmith LaYbCNfrac1419^25 mean 2204) and the coarse-grainedmafic rocks have similar REE patterns to the tholeiitic
Depleted MORB average(Salters amp Stracke 2004)
Schoen LakeSchoen Lake
Mount Arrowsmith
Karmutsen RangeS
ampl
eC
hond
rite
Sam
ple
Cho
ndrit
e
Sam
ple
Prim
itive
Man
tleS
ampl
eP
rimiti
ve M
antle
Sam
ple
Prim
itive
Man
tle
Mount Arrowsmith
Karmutsen Range
(a) (b)
(c) (d)
(e) (f )
Sam
ple
Cho
ndrit
e
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained mafic rock
Pillowed flowMineralized sill
1
10
100
1
10
100
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
01
1
10
100
1
10
100
1
10
100
Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
Fig 7 Whole-rock REE and other incompatible-element concentrations for the Karmutsen Formation (a) (c) and (e) are chondrite-normal-ized REE patterns for samples from each of the three main field areas onVancouver Island (b) (d) and (f) are primitive mantle-normalizedtrace-element patterns for samples from each of the three main field areas All normalization values are from McDonough amp Sun (1995) Theclear distinction between LREE-enriched tholeiitic basalts and LREE-depleted picrites the effects of alteration on the LILE (especially loss ofK and Rb) and the different range for the vertical scale in (b) (extends down to 01) should be noted
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
18
basalts (LaYbCNfrac14 21^29 mean 2508) Two LREE-depleted coarse-grained mafic rocks from the SchoenLake area have similar REE patterns (LaYbCNfrac14 07) tothe picrites from the Keogh Lake area indicating that thisdistinctive suite of rocks is exposed as far south asSchoen Lake that is 75 km away (Fig 7) The anomalouspillowed flow from the Karmutsen Range has lower LaYbCN (13^18) than the main group of tholeiitic basaltsfrom the Karmutsen Range (Fig 7) The mineralized sillfrom the Schoen Lake area is distinctly LREE-enriched(LaYbCNfrac14 33) with the highest REE abundances(LaCNfrac14 940) of all Karmutsen samples this may reflectcontamination by adjacent sediment also evident in thin-section where sulfide blebs are cored by quartz (Greeneet al 2006)The primitive mantle-normalized trace-element pat-
terns (Fig 7) and trace-element variations and ratios(Fig 8) highlight the differences in trace-element
concentrations between the four main groups ofKarmutsen flood basalts onVancouver Island The tholeii-tic basalts from all three areas have relatively smooth par-allel trace-element patterns some samples have smallpositive Sr anomalies relative to Nd and Sm and manysamples have prominent negative K anomalies relative toU and Nb reflecting K loss during alteration (Fig 7) Thelarge ion lithophile elements (LILE) in the tholeiiticbasalts (mainly Rb and K not Cs) are depleted relativeto the high field strength elements (HFSE) and LREELILE segments especially for the Schoen Lake area(Fig 7d) are remarkably parallel The picrites andhigh-MgO basalts form a tight band of parallel trace-element patterns and are depleted in HFSE and LREEwith positive Sr anomalies and relatively low Rb Thecoarse-grained mafic rocks from the Karmutsen Rangehave identical trace-element patterns to the tholeiiticbasalts Samples from the Schoen Lake area have similar
000
025
050
075
100
125
0 1 2 3 4 5
00
04
08
12
16
0 5 10 15 20 25
MgO (wt)
0
5
10
15
0 50 100 150
Hf (ppm)
Th (ppm)
NbLaNb (ppm)
Zr (ppm)
(a)
(c)
(b)
(d)
Th (ppm)
00
02
04
06
08
10
12
00 01 02 03 04 05
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
NbLa=1
4723A4 5615A12
U (ppm)
Fig 8 Whole-rock trace-element concentrations and ratios for the Karmutsen Formation [except (b) which is vs wt MgO] (a) Nb vs Zr(b) NbLa vs MgO (c) Th vs Hf (d) Th vs U There is a clear distinction between the tholeiitic basalts and picrites in Nb and Zr both inconcentration and the slope of each trend
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
19
trace-element patterns to the Karmutsen Range except forthe two LREE-depleted coarse-grained mafic rocks andthe mineralized sill (Fig 7d)
Sr^Nd^Hf^Pb isotopic compositionsA total of 19 samples from the four main groups ofthe Karmutsen Formation were selected for radiogenicisotopic geochemistry on the basis of major- and trace-element variation stratigraphic position and geographicallocation The samples have overlapping age-correctedHf and Nd isotope ratios and distinct ranges of Sr isotopiccompositions (Fig 9) The tholeiitic basalts picriteshigh-MgO basalt and coarse-grained mafic rocks have
initial eHffrac14thorn87 to thorn126 and eNdfrac14thorn66 to thorn88 cor-rected for in situ radioactive decay since 230 Ma (Fig 9Tables 3 and 4) All the Karmutsen samples form a well-defined linear array in a Lu^Hf isochron diagram corre-sponding to an age of 24115 Ma (Fig 9) This is withinerror of the accepted age of the Karmutsen Formationand indicates that the Hf isotope systematics behaved as aclosed system since c 230 Ma The anomalous pillowedflow has similar initial eHf (thorn98) and slightly lower initialeNd (thorn62) compared with the other Karmutsen samplesThe picrites and high-MgO basalt have higher initial87Sr86Sr (070398^070518) than the tholeiitic basalts(initial 87Sr86Srfrac14 070306^070381) and the coarse-grained mafic rocks (initial 87Sr86Srfrac14 070265^070428)
05128
05129
05130
05131
05132
015 017 019 021 023 025
02829
02830
02831
02832
02833
02834
000 002 004 006 008 010
4
6
8
10
0702 0703 0704 0705 07066
8
10
12
14
5 6 7 8 9 10
12
0
1
2
3
4
5
6
7
0702 0703 0704 0705 0706
(d)
LOI (wt )
87Sr 86Sr
4723A2
Nd
143Nd144Nd
147Sm144Nd
87Sr 86Sr (230 Ma)
(230 Ma)
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
(a)
(c)
176Hf177Hf
Age=241 plusmn 15 Ma=+990 plusmn 064
176Lu177Hf
Hf (230 Ma)
Nd (230 Ma)
(e)
Hf (i)
(b)
Fig 9 Whole-rock Sr Nd and Hf isotopic compositions for the Karmutsen Formation (a) 143Nd144Nd vs 147Sm144Nd (b) 176Hf177Hf vs176Lu177Hf The slope of the best-fit line for all samples corresponds to an age of 24115 Ma (c) Initial eNd vs 87Sr86Sr Age-correction to230 Ma (d) LOI vs 87Sr86Sr (e) Initial eHf vs eNd Average 2 error bars are shown in a corner of each panel Complete chemical duplicatesshown inTables 3 and 4 (samples 4720A7 and 4722A4) are circled in each plot
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
overlap the ranges of the picrites and tholeiitic basalts(Fig 9 Table 3)The measured Pb isotopic compositions of the tholeiitic
basalts are more radiogenic than those of the picrites andthe most magnesian Keogh Lake picrites have the leastradiogenic Pb isotopic compositions The range ofinitial Pb isotope ratios for the picrites is206Pb204Pbfrac1417868^18580 207Pb204Pbfrac1415547^15562and 208Pb204Pbfrac14 37873^38257 and the range for thetholeiitic basalts is 206Pb204Pbfrac1418782^19098207Pb204Pbfrac1415570^15584 and 208Pb204Pbfrac14 37312^38587 (Fig 10 Table 5) The coarse-grained mafic rocksoverlap the range of initial Pb isotopic compositions forthe tholeiitic basalts with 206Pb204Pbfrac1418652^19155207Pb204Pbfrac1415568^15588 and 208Pb204Pbfrac14 38153^38735 One high-MgO basalt has the highest initial206Pb204Pb (19242) and 207Pb204Pb (15590) and thelowest initial 208Pb204Pb (37676) The Pb isotopic ratiosfor Karmutsen samples define broadly linearrelationships in Pb isotope plots The age-corrected Pb
isotopic compositions of several picrites have been affectedby U and Pb (andor Th) mobility during alteration(Fig 10)
ALTERAT IONThe Karmutsen basalts have retained most of their origi-nal igneous structures and textures however secondaryalteration and low-grade metamorphism have producedzeolitic- and prehnite^pumpellyite-bearing mineral assem-blages (Table 1 Cho et al 1987) Alteration and metamor-phism have primarily affected the distribution of the LILEand the Sr isotopic systematics The Keogh Lake picriteshave been affected more by alteration than the tholeiiticbasalts and have higher LOI (up to 55wt ) variableLILE and higher measured Sr Nd and Hf and lowermeasured Pb isotope ratios than the tholeiitic basalts Srand Pb isotope compositions for picrites and tholeiiticbasalts show a relationship with LOI (Figs 9 and 10)whereas Nd and Hf isotopic compositions do not correlate
Table 3 Sr and Nd isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Rb Sr 87Sr86Sr 2m87Rb86Sr 87Sr86Srt Sm Nd 143Nd144Nd 2m eNd
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses carried out at the PCIGR thecomplete trace element analyses are given in Electronic Appendix 3
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
21
with LOI (not shown) The relatively high apparent initialSr isotopic compositions for the high-MgO lavas (up to07052) probably resulted from post-magmatic loss of Rbor an increase in 87Sr86Sr through addition of seawater Sr(eg Hauff et al 2003) High-MgO lavas from theCaribbean plateau have similarly high 87Sr86Sr comparedwith basalts (Recurren villon et al 2002)The correction for in situdecay on initial Pb isotopic ratios has been affected bymobilization of U and Th (and Pb) in the whole-rockssince their formation A thorough acid leaching duringsample preparation was used that has been shown to effec-tively remove alteration phases (Weis et al 2006 NobreSilva et al in preparation) Whole-rock SmNd ratiosand age-correction of Nd isotopes may be affected bythe leaching procedure for submarine rocks older than50 Ma (Thompson et al 2008 Fig 9) there ishowever very little variation in Nd isotopic compositionsof the picrites or tholeiitic basalts and this system does notappear to be disturbed by alteration The HFSE abun-dances for tholeiitic and picritic basalts exhibit clear
linear correlations in binary diagrams (Fig 8) whereasplots of LILE vs HFSE are highly scattered (not shown)because of the mobility of some of the alkali and alkalineearth LILE (Cs Rb Ba K) and Sr for most samplesduring alterationThe degree of alteration in the Karmutsen samples does
not appear to be related to eruption environment or depthof burial Pillow rims are compositionally different frompillow cores (Surdam1967 Kuniyoshi1972) and aquagenetuffs are chemically different from isolated pillows withinpillow breccias however submarine and subaerial basaltsexhibit a similar degree of alterationThere is no clear cor-relation between the submarine and subaerial basalts andsome commonly used chemical alteration indices (eg BaRb vs K2OP2O5 Huang amp Frey 2005) There is also nodefinitive correlation between the petrographic alterationindex (Table 1) and chemical alteration indices although10 of 13 tholeiitic basalts with the highest K and LILEabundances have the highest petrographic alterationindex of three (seeTable 1)
Table 4 Hf isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Lu Hf 176Hf177Hf 2m eHf176Lu177Hf 176Hf177Hft eHf(t)
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
OLIV INE ACCUMULAT ION INH IGH-MgO AND PICR IT IC LAVASGeochemical trends and petrographic characteristics indi-cate that accumulation of olivine played an important rolein the formation of the high-MgO basalts and picrites fromthe Keogh Lake area on northern Vancouver Island TheKeogh Lake samples show a strong linear correlation inplots of Al2O3 TiO2 ScYb and Ni vs MgO and many ofthe picrites have abundant clusters of olivine pseudomorphs(Table1 Fig11)There is a strong linear correlationbetween
modal per cent olivine and whole-rock magnesium contentmagnesium contents range between 10 and 19wt MgOand the proportion of olivine phenocrysts in these samplesvaries from 0 to 42 vol (Fig 11)With a few exceptionsboth the size range and median size of the olivine pheno-crysts in most samples are comparable The clear correla-tion between the proportion of olivine phenocrysts andwhole-rock magnesium contents indicates that accumula-tion of olivine was directly responsible for the high MgOcontents (410wt ) in most of the picritic lavas
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
207Pb204Pb
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
208Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
(a) (b)
(c) (e)
4723A24723A2
4723A2
4723A2
4722A4
4723A4
4722A4
4723A4
4723A134723A3
4723A3
4723A13
1556
1558
1560
1562
1564
1566
1568
186 190 194 198 202 206 2101550
1552
1554
1556
1558
1560
178 180 182 184 186 188 190 192 194
380
384
388
392
396
186 190 194 198 202 206 210374
378
382
386
390
178 180 182 184 186 188 190 192 194
206Pb204Pb(d)
LOI (wt )
0
1
2
3
4
5
6
7
186 190 194 198 202 206 210
4723A2
Fig 10 Pb isotopic compositions of leached whole-rock samples measured by MC-ICP-MS for the Karmutsen Formation Error bars are smal-ler than symbols (a) Measured 207Pb204Pb vs 206Pb204Pb (b) Initial 207Pb204Pb vs 206Pb204Pb Age-correction to 230 Ma (c) Measured208Pb204Pb vs 206Pb204Pb (d) LOI vs 206Pb204Pb (e) Initial 208Pb204Pb vs 206Pb204Pb Complete chemical duplicates shown in Table 5(samples 4720A7 and 4722A4) are circled in each plot The dashed lines in (b) and (e) show the differences in age-corrections for two picrites(4723A4 4722A4) using the measured UTh and Pb concentrations for each sample and the age-corrections when concentrations are used fromthe two picrites (4723A3 4723A13) that appear to be least affected by alteration
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
23
DISCUSS IONThe compositional and spatial development of the eruptedlava sequences of the Wrangellia oceanic plateau onVancouver Island provide information about the evolutionof a mantle plume upwelling beneath oceanic lithospherein an analogous way to the interpretation of continentalflood basalts The Caribbean and Ontong Java plateausare the two primary examples of accreted parts of oceanicplateaus that have been studied to date and comparisonscan be made with the Wrangellia oceanic plateauHowever the Wrangellia oceanic plateau is a distinctiveoccurrence of an oceanic plateau because it formed on thecrust of a Devonian to Mississippian intra-oceanic islandarc overlain by Mississippian to Permian limestones andpelagic sediments The nature and thickness of oceaniclithospheric mantle are considered to play a fundamentalrole in the decompression and evolution of a mantleplume and thus the compositions of the erupted lavasequences (Greene et al 2008) The geochemical and
stratigraphic relationships observed in the KarmutsenFormation onVancouver Island and the focus of the subse-quent discussion sections provide constraints on (1) thetemperature and degree of melting required to generatethe primary picritic magmas (2) the composition of themantle source (3) the depth of melting and residual min-eralogy in the source region (4) the low-pressure evolutionof the Karmutsen tholeiitic basalts (5) the growth of theWrangellia oceanic plateau
Melting conditions and major-elementcomposition of the primary magmasThe primary magmas to most flood basalt provinces arebelieved to be picritic (eg Cox 1980) and they have beenfound in many continental and oceanic flood basaltsequences worldwide (eg Siberia Karoo Paranacurren ^Etendeka Caribbean Deccan Saunders 2005) Near-pri-mary picritic lavas with low total alkali abundances
Table 5 Pb isotopic compositions of Karmutsen basaltsVancouver Island BC
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
require high-degree partial melting of the mantle source(eg Herzberg amp OrsquoHara 2002) The Keogh Lake picritesfrom northernVancouver Island are the best candidates foridentifying the least modified partial melts of the mantleplume source despite having accumulated olivine pheno-crysts and can be used to estimate conditions of meltingand the composition of the primary magmas for theKarmutsen FormationPrimary magma compositions mantle potential tem-
peratures and source melt fractions for the picritic lavas ofthe Karmutsen Formation were calculated from primitivewhole-rock compositions using PRIMELT1XLS software(Herzberg et al 2007) and are presented in Fig 12 andTable 6 A detailed discussion of the computationalmethod has been given by Herzberg amp OrsquoHara (2002)and Herzberg et al (2007) key features of the approachare outlined belowPrimary magma composition is calculated for a primi-
tive lava by incremental addition of olivine and compari-son with primary melts of mantle peridotite Lavas thathad experienced plagioclase andor clinopyroxene fractio-nation are excluded from this analysis All calculated pri-mary magma compositions are assumed to be derived byaccumulated fractional melting of fertile peridotite KR-4003 Uncertainties in fertile peridotite composition donot propagate to significant variations in melt fractionand mantle potential temperature melting of depletedperidotite propagates to calculated melt fractions that aretoo high but with a negligible error in mantle potential
temperature PRIMELT1XLS calculates the olivine liqui-dus temperature at 1atm which should be similar to theactual eruption temperature of the primary magmaand is accurate to 308C at the 2 level of confidenceMantle potential temperature is model dependent witha precision that is similar to the accuracy of the olivineliquidus temperature (C Herzberg personal communica-tion 2008)With regard to the evidence for accumulated olivine in
the Keogh Lake picrites it should not matter whether themodeled lava composition is aphyric or laden with accu-mulated olivine phenocrysts PRIMELT1 computes theprimary magma composition by adding olivine to a lavathat has experienced olivine fractionation in this way itinverts the effects of fractional crystallization The onlyrequirement is that the olivine phenocrysts are not acci-dental or exotic to the lava compositions Support for thisprocedure can be seen in the calculated olivine liquid-line-of-descent shown by the curved arrays with 5 olivineaddition increments (Fig 12c) Fractional crystallization ofolivine from model primary magmas having 85^95wt FeO and 153^175wt MgO will yield arrays of deriva-tive liquids and randomly sorted olivines that in most butnot all cases connect the picrites and high-MgO basalts Anewer version of the PRIMELT software (Herzberg ampAsimow 2008) can perform olivine addition to and sub-traction from lavas with highly variably olivine phenocrystcontents and yields results that converge with those shownin Fig 12
Fig 11 Relationship between abundance of olivine phenocrysts and whole-rock MgO contents for the Keogh Lake picrites (a) Tracings ofolivine grains (gray) in six high-MgO samples made from high resolution (2400 dpi) scans of petrographic thin-sections Scale bars represent10mm sample numbers are indicated at the upper left (b) Modal olivine vs whole-rock MgO Continuous line is the best-fit line for 10 high-MgO samplesThe areal proportion of olivine was calculated from raster images of olivine grains using ImageJ image analysis software whichprovides an acceptable estimation of the modal abundance (eg Chayes 1954)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
25
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
Gray lines = final melting pressure
L + Ol + Opx
07
08
09
Pressure (GPa)
10
3
4
5
6
1
SOLIDUS
01
04
0506
L + Ol
2
03
02
PeridotiteKR-4003
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
Thick black lines = initial melting pressure
Dashed lines = melt fraction
Melt Fraction
2
3 4 5 6
L+Ol+Opx
SolidusSpinel
SolidusGarnet Peridotite
7
L+Ol
0 10 20 30 4040
45
50
55
60
MgO (wt)
SiO2
(wt)
PeridotiteKR-4003
Melt Fraction
0 10 20 305
10
15
CaO(wt)
MgO (wt)
3
4 5 6 7
2
46
2
Peridotite Partial Melts
Thick black line = initial melting pressureGray lines = final melting pressure
Peridotite
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
02
Pressure (GPa)
0302
04
Pressure (GPa)
Picrite
High-MgO basalt
Tholeiitic basalt
(c)
(d)
Karmutsen flood basaltsOlivine addition modelOlivine addition (5 increment)Primary magma
PicriteHigh-MgO basaltTholeiitic basalt
(a)
(b)
800 850 900
910
920
930
940
950
Karmutsen flood basalts
Olivine addition modelOlivine addition (5 increment)Primary magma
Gray lines = Fo contents of olivine
Fig 12 Estimated primary magma compositions for three Keogh Lake picrites and high-MgO basalts (samples 4723A4 4723A135616A7) using the forward and inverse modeling technique of Herzberg et al (2007) Karmutsen compositions and modeling results areoverlain on diagrams provided by C Herzberg (a) Whole-rock FeO vs MgO for Karmutsen samples from this study Total iron estimatedto be FeO is 090 Gray lines show olivine compositions that would precipitate from liquid of a given MgO^FeO composition Black lineswith crosses show results of olivine addition (inverse model) using PRIMELT1 (Herzberg et al 2007) Gray field highlights olivinecompositions with Fo contents of 91^92 predicted to precipitate (b) FeO vs MgO showing Karmutsen lava compositions and results ofa forward model for accumulated fractional melting of fertile peridotite KR-4003 (Kettle River Peridotite Walter 1998) (c) SiO2 vsMgO for Karmtusen lavas and model results (d) CaO vs MgO showing Karmtusen lava compositions and model results To brieflysummarize the technique [see Herzberg et al (2007) for complete description] potential parental magma compositions for the high-MgOlava series were selected (highest MgO and appropriate CaO) and using PRIMELT1 software olivine was incrementally added tothe selected compositions to show an array of potential primary magma compositions (inverse model) Then using PRIMELT1 theresults from the inverse model were compared with a range of accumulated fractional melts for fertile peridotite derived fromparameterization of the experimental results for KR-4003 from Walter (1998) forward model Herzberg amp OrsquoHara (2002) A melt fractionwas sought that was unique to both the inverse and forward models (Herzberg et al 2007) A unique solution was found when there was a com-mon melt fraction for both models in FeO^MgO and CaO^MgO^Al2O3^SiO2 (CMAS) projection space This modeling assumes thatolivine was the only phase crystallizing and ignores chromite precipitation and possible augite fractionation in the mantle (Herzbergamp OrsquoHara 2002) Results are best for a residue of spinel lherzolite (not pyroxenite) The presence of accumulated olivine in samples ofthe Keogh Lake picrites used as starting compositions does not significantly affect the results because we are modeling addition of olivine(see text for further description) The tholeiitic basalts cannot be used for modeling because they are all saturated in plagthorn cpxthornolThick black line for initial melting pressure at 4 GPa is not shown in (c) for clarity Gray area in (a) indicates parental magmas thatwill crystallize olivine with Fo 91^92 at the surface Dashed lines in (c) indicate melt fraction Solidi for spinel and garnet peridotite arelabelled in (c)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
The PRIMELT1 results indicate that if the Karmutsenpicritic lavas were derived from accumulated fractionalmelting of fertile peridotite the primary magmas wouldhave contained 15^17wt MgO and 10wt CaO(Fig 12 Table 6) formed from 23^27 partial meltingand would have crystallized olivine with a Fo content of 91 (Fig 12 Table 6) Calculated mantle temperatures forthe Karmutsen picrites (14908C) indicate melting ofanomalously hot mantle (100^2008C above ambient) com-pared with ambient mantle that produces mid-ocean ridgebasalt (MORB 1280^14008C Herzberg et al 2007)which initiated between 3 and 4 GPa Several picritesappear to be near-primary melts and required very littleaddition of olivine (eg sample 93G171 with measured158wt MgO) These temperature and melting esti-mates of the Karmutsen picrites are consistent withdecompression of hot mantle peridotite in an actively con-vecting plume head (ie plume initiation model) Threesamples (picrite 5614A1 and high-MgO basalts 5614A3and 5614A5) are lower in iron than the other Karmutsenlavas with FeO in the 65^80wt range (Fig 12c) andhave MgO and FeO contents that are similar to primitiveMORB from the Sequeiros fracture zone of the EastPacific Rise (EPR Perfit et al 1996) These samples
however differ from EPR MORB in having higher SiO2
and lower Al2O3 and they are restricted geographicallyto one area (Sara Lake Fig 2)We therefore have evidence for primary magmas with
MgO contents that range from a possible low of 110wt to as high as 174wt with inferred mantle potential tem-peratures from 1340 to 15208C This is similar to the rangedisplayed by lavas from the Ontong Java Plateau deter-mined by Herzberg (2004) who found that primarymagma compositions assuming a peridotite sourcewould contain 17wt MgO and 10wt CaO(Table 6) and that these magmas would have formed by27 melting at a mantle potential temperature of15008C (first melting occurring at 36 GPa) Fitton ampGodard (2004) estimated 30 partial melting for theOntong Java lavas based on the Zr contents of primarymagmas of Kroenke-type basalt calculated by incrementaladdition of equilibrium olivine However if a considerableamount of eclogite was involved in the formation of theOntong Java plateau magmas the excess temperatures esti-mated by Herzberg et al (2007) may not be required(Korenaga 2005) The variability in mantle potential tem-perature for the Keogh Lake picrites is also similar to thewide range of temperatures that have been calculated for
Table 6 Estimated primary magma compositions for Karmutsen basalts and other oceanic plateaus or islands
Sample 4723A4 4723A13 5616A7 93G171 Average OJP Mauna Keay Gorgonaz
Ontong Java primary magma composition for accumulated fraction melting (AFM) from Herzberg (2004)yMauna Kea primary magma composition is average of four samples of Herzberg (2006 table 1)zGorgona primary magma composition for AFM for 1F fertile source from Herzberg amp OrsquoHara (2002 table 4)Total iron estimated to be FeO is 090
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
27
lavas from single volcanoes in the Galapagos Hawaii andelsewhere by Herzberg amp Asimow (2008) Those workersinterpreted the range to reflect the transport of magmasfrom both the cool periphery and the hotter axis of aplume A thermally heterogeneous plume will yield pri-mary magmas with different MgO contents and the rangeof primary magma compositions and their inferred mantlepotential temperatures for Karmutsen lavas may reflectmelting from different parts of the plume
Source of the Karmutsen Formation lavasThe Sr^Nd^Hf^Pb isotopic compositions of theKarmutsen volcanic rocks on Vancouver Island indicatethat they formed from plume-type mantle containing adepleted component that is compositionally similar to butdistinct from other oceanic plateaus that formed in thePacific Ocean (eg Ontong Java and Caribbean Fig 13)Trace element and isotopic constraints can be used to testwhether the depleted component of the KarmutsenFormation was intrinsic to the mantle plume and distinctfrom the source of MORB or the result of mixing or invol-vement of ambient depleted MORB mantle with plume-type mantle The Karmutsen tholeiitic basalts have initialeNd and
87Sr86Sr ratios that fall within the range of basaltsfrom the Caribbean Plateau (eg Kerr et al 1996 1997Hauff et al 2000 Kerr 2003) and a range of Hawaiian iso-tope data (eg Frey et al 2005) and lie within and abovethe range for the Ontong Java Plateau (Tejada et al 2004Fig 13a) Karmutsen tholeiitic basalts with the highest ini-tial eNd and lowest initial 87Sr86Sr lie within and justbelow the field for age-corrected northern EPR MORB at230 Ma (eg Niu et al1999 Regelous et al1999) Initial Hfand Nd isotopic compositions for the Karmutsen samplesare noticeably offset to the low side of the ocean islandbasalt (OIB) array (Vervoort et al 1999) and lie belowand slightly overlap the low eHf end of fields for Hawaiiand the Caribbean Plateau (Fig 13c) the Karmutsen sam-ples mostly fall between and slightly overlap a field forage-corrected Pacific MORB at 230 Ma and Ontong Java(Mahoney et al 1992 1994 Nowell et al 1998 Chauvel ampBlichert-Toft 2001) Karmutsen lava Pb isotope ratios over-lap the broad range for the Caribbean Plateau and thelinear trends for the Karmutsen samples intersect thefields for EPR MORB Hawaii and Ontong Java in208Pb^206Pb space but do not intersect these fields in207Pb^206Pb space (Fig 13d and e) The more radiogenic207Pb204Pb for a given 206Pb204Pb of the Karmutsenbasalts requires high UPb ratios at an ancient time when235U was abundant similar to the Caribbean Plateau andenriched in comparison with EPR MORB Hawaii andOntong JavaThe low eHf^low eNd component of the Karmutsen
basalts that places them below the OIB mantle array andpartly within the field for age-corrected Pacific MORBcan form from low ratios of LuHf and high SmNd over
time (eg Salters amp White 1998) MORB will develop aHf^Nd isotopic signature over time that falls below themantle array Low LuHf can also lead to low 176Hf177Hffrom remelting of residues produced by melting in theabsence of garnet (eg Salters amp Hart 1991 Salters ampWhite 1998) The Hf^Nd isotopic compositions of theKarmutsen Formation indicate the possible involvementof depleted MORB mantle however similar to Iceland(Fitton et al 2003) Nb^Zr^Y variations provide a way todistinguish between a depleted component within theplume and a MORB-like component The Karmutsenbasalts and picrites fall within the Iceland array and donot overlap the MORB field in a logarithmic plot of NbYvs ZrY (Fig 13b) using the fields established by Fittonet al (1997) Similar to the depleted component within theCaribbean Plateau (Thompson et al 2003) the trend ofHf^Nd isotopic compositions towards Pacific MORB-likecompositions is not followed by MORB values in Nb^Zr^Y systematics and this suggests that the depleted compo-nent in the source of the Karmutsen basalts is distinctfrom the source of MORB and probably an intrinsic partof the plume The relatively radiogenic Pb isotopic ratiosand REE patterns distinct from MORB also support thisinterpretationTrace-element characteristics suggest the possible invol-
vement of sub-arc lithospheric mantle in the generation ofthe Karmutsen picrites Lassiter et al (1995) suggested thatmixing of a plume-type source with eNdthorn 6 to thorn 7 witharc material with low NbTh could reproduce the varia-tions observed in the Karmutsen basalts however theyconcluded that the absence of low NbLa ratios in theirsample of tholeiitic basalts restricts the amount of arc litho-sphere involved The study of Lassiter et al (1995) wasbased on major and trace element and Sr Nd and Pb iso-topic data for a suite of 29 samples from Buttle Lake in theStrathcona Provincial Park on central Vancouver Island(Fig 1) Whereas there are no clear HFSE depletions inthe Karmutsen tholeiitic basalts the low NbLa (and TaLa) of the Karmutsen picritic lavas in this study indicatethe possible involvement of Paleozoic arc lithosphere onVancouver Island (Fig 8)
REE modeling dynamic melting andsource mineralogyThe combined isotopic and trace-element variations of thepicritic and tholeiitic Karmutsen lavas indicate that differ-ences in trace elements may have originated from differentmelting histories For example the LuHf ratios of theKarmutsen picritic lavas (mean 008) are considerablyhigher than those in the tholeiitic basalts (mean 002)however the small range of overlapping initial eHf valuesfor the two lava suites suggests that these differences devel-oped during melting within the plume (Fig 9b) Below wemodel the effects of source mineralogy and depth of
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
28
melting on differences in trace-element concentrations inthe tholeiitic basalts and picritesDynamic melting models simulate progressive decom-
pression melting where some of the melt fraction isretained in the residue and only when the degree of partialmelting is greater than the critical mass porosity and the
source becomes permeable is excess melt extracted fromthe residue (Zou1998) For the Karmutsen lavas evolutionof trace element concentrations was simulated using theincongruent dynamic melting model developed by Zou ampReid (2001) Melting of the mantle at pressures greater than1 GPa involves incongruent melting (eg Longhi 2002)
OIB array
PacificMORB(230 Ma)
(0 Ma)
EPR(230 Ma)
-4
-2
0
2
4
8
10
12
14
16
18
20
22
-4 -2 0 2 4 6 8 10 12 14
minus2
0
2
4
6
8
10
12
14
0702 0703 0704 0705 0706
IndianMORB
87Sr 86Sr (initial)
Hf (initial)
(a) (c)
Nd (initial)
Karmutsen tholeiitic basalt
HawaiiOntong JavaCaribbean Plateau
Karmutsen high-MgO lava
high-MgO lavasKarmutsen
1540
1545
1550
1555
1560
1565
175 180 185 190 195 200375
380
385
390
395
175 180 185 190 195 200
(d) (e)
EPR
EPR
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb
Karmutsen Formation (initial)
HawaiiOntong JavaCaribbean Plateau
East Pacific Rise
Karmutsen Formation (measured)
Juan de FucaGorda
ε Nd
(initial)
Karmutsen Formation (different age-correction for 3 picrites)
(0 Ma)
NbY
001
01
1
1 10
ZrY
OIB
NMORB
(b)
Iceland array
6
Fig 13 Comparison of age-corrected (230 Ma) Sr^Nd^Hf^Pb isotopic compositions for Karmutsen flood basalts onVancouver Island withage-corrected OIB and MORB and Nb^Zr^Y variation (a) Initial eNd vs 87Sr86Sr (b) NbYand ZrY variation (after Fitton et al 1997)(c) Initial eHf vs eNd (d) Measured and initial 207Pb204Pb vs 206Pb204Pb (e) Measured and initial 208Pb204Pb vs 206Pb204Pb References fordata sources are listed in Electronic Appendix 4 Most of the compiled data were extracted from the GEOROC database (httpgeorocmpch-mainzgwdgdegeoroc) OIB array line in (c) is fromVervoort et al (1999) EPR East Pacific Rise Several high-MgO samples affected bysecondary alteration were not plotted for clarity in Pb isotope plots Estimation of fields for age-corrected Pacific MORB and EPR at 230 Mawere made using parent^daughter concentrations (in ppm) from Salters amp Stracke (2004) of Lufrac14 0063 Hffrac14 0202 Smfrac14 027 Ndfrac14 071Rbfrac14 0086 and Srfrac14 836 In the NbYvs ZrYplot the normal (N)-MORB and OIB fields not including Icelandic OIB are from data com-pilations of Fitton et al (1997 2003) The OIB field is data from 780 basalts (MgO45wt ) from major ocean islands given by Fitton et al(2003) The N-MORB field is based on data from the Reykjanes Ridge EPR and Southwest Indian Ridge from Fitton et al (1997) The twoparallel gray lines in (b) (Iceland array) are the limits of data for Icelandic rocks
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
29
with olivine being produced during the reactioncpxthornopxthorn spfrac14meltthornol for spinel peridotites The mod-eling used coefficients for incongruent melting reactionsbased on experiments on lherzolite melting (eg Kinzleramp Grove 1992 Longhi 2002) and was separated into (1)melting of spinel lherzolite (2) two combined intervals ofmelting for a garnet lherzolite (gt lherz) and spinel lher-zolite (sp lherz) source (3) melting of garnet lherzoliteThe model solutions are non-unique and a range indegree of melting partition coefficients melt reactioncoefficients and proportion of mixed source componentsis possible to achieve acceptable solutions (see Fig 14 cap-tion for description of modeling parameters anduncertainties)The LREE-depleted high-MgO lavas require melting of
a depleted spinel lherzolite source (Fig 14a) The modelingresults indicate a high degree of melting (22^25) similarto the results from PRIMELT1 based on major-elementvariations (discussed above) of a LREE-depleted sourceThe high-MgO lavas lack a residual garnet signature Theenriched tholeiitic basalts involved melting of both garnetand spinel lherzolite and represent aggregate melts pro-duced from continuous melting throughout the depth ofthe melting column (Fig 14b) Trace-element concentra-tions in the tholeiitic and picritic lavas are not consistentwith low-degree melting of garnet lherzolite alone(Fig 14c) The modeling results indicate that the aggregatemelts that formed the tholeiitic basalts would haveinvolved lesser proportions of enriched small-degreemelts (1^6 melting) generated at depth and greateramounts of depleted high-degree melts (12^25) gener-ated at lower pressure (a ratio of garnet lherzolite tospinel lherzolite melt of 14 provides the best fits seeFig 14b and description in figure caption) The totaldegree of melting for the tholeiitic basalts was high(23^27) similar to the degree of partial melting in thespinel lherzolite facies for the primary melts that eruptedas the Keogh Lake picrites of a source that was also prob-ably depleted in LREE
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
(c)
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
Melting of garnet lherzolite
High-MgO lavas
Melting of spinel lherzolite
source (07DM+03PM)
source (07DM+03PM)
6 melting
30 melting
(b)
(a)
30 melting
Sam
ple
Cho
ndrit
eS
ampl
eC
hond
rite 20 melting
1 melting
Melting of garnet lherzoliteand spinel lherzolite
x gt lherz + 4x sp lherz
5 meltingx gt lherz + 4x sp lherz
Sam
ple
Cho
ndrit
e
Tholeiitic lavas
Tholeiitic lavas
High-MgO lavas
source (07DM+03PM)
Fig 14 Trace-element modeling results for incongruent dynamicmantle melting for picritic and tholeiitic lavas from the KarmutsenFormation Three steps of modeling are shown (a) melting of spinellherzolite compared with high-MgO lavas (b) melting of garnet lher-zolite and spinel lherzolite compared with tholeiitic lavas (c) meltingof garnet lherzolite compared with tholeiitic and picritic lavasShaded fields in each panel for tholeiitic basalts and picrites arelabeled Patterns with symbols in each panel are modeling results(patterns are 5 melting increments in (a) and (b) and 1 incre-ments in (c) PM primitive mantle DM depleted MORB mantle gtlherz garnet lherzolite sp lherz spinel lherzolite In (b) the ratios ofper cent melting for garnet and spinel lherzolite are indicated (x gtlherzthorn 4x sp lherz means that for 5 melting 1 melt in garnetfacies and 4 melt in spinel facies where x is per cent melting in theinterval of modeling) The ratio of the respective melt proportions inthe aggregate melt (x gt lherzthorn 4x sp lherz) was kept constant andthis ratio of melt contribution provided the best fit to the data Arange of proportion of source components (07DMthorn 03PM to
09DMthorn 01PM) can reproduce the variation in the high-MgOlavas Melting modeling uses the formulation of Zou amp Reid (2001)an example calculation is shown in their Appendix PM fromMcDonough amp Sun (1995) and DM from Salters amp Stracke (2004)Melt reaction coefficients of spinel lherzolite from Kinzler amp Grove(1992) and garnet lherzolite fromWalter (1998) Partition coefficientsfrom Salters amp Stracke (2004) and Shaw (2000) were kept constantSource mineralogy for spinel lherzolite (018cpx027opx052ol003sp) from Kinzler (1997) and for garnet lherzolite(034cpx008opx053ol005gt) from Salters amp Stracke (2004) Arange of source compositions for melting of spinel lherzolite(07DMthorn 03PM to 03DMthorn 07PM) reproduce REE patternssimilar to those of the high-MgO lavas with best fits for 22ndash25melting and 07DMthorn 03PM source composition Concentrationsof tholeiitic basalts cannot be reproduced from an entirelyprimitive or depleted source
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
30
The uncertainty in the composition of the source in thespinel lherzolite melting model for the high-MgO lavasprecludes determining whether the source was moredepleted in incompatible trace elements than the source ofthe tholeiitic lavas (see Fig 14 caption for description) It ispossible that early high-pressure low-degree melting left aresidue within parts of the plume (possibly the hotter inte-rior portion) that was depleted in trace elements (egElliott et al 1991) further decompression and high-degreemelting of such depleted regions could then have generatedthe depleted high-MgO Karmutsen lavas Shallow high-degree melting would preferentially sample depletedregions whereas deeper low-degree melting may notsample more refractory depleted regions (eg CaribbeanEscuder-Viruete et al 2007) The picrites lie at the highend of the range of Nd isotopic compositions and havelow NbZr similar to depleted Icelandic basalts (Fittonet al 2003) therefore the picrites may represent the partsof the mantle plume that were the most depleted prior tothe initiation of melting As Fitton et al (2003) suggestedthese depleted melts may only rarely be erupted withoutcombining with more voluminous less depleted melts andthey may be detected only as undifferentiated small-volume flows like the Keogh Lake picrites within theKarmutsen Formation on Vancouver IslandDecompression melting within the mantle plume initiatedwithin the garnet stability field (35^25 GPa) at highmantle potential temperatures (Tp414508C) and pro-ceeded beneath oceanic arc lithosphere within the spinelstability field (25^09 GPa) where more extensivedegrees of melting could occur
Magmatic evolution of Karmutsentholeiitic basaltsPrimary magmas in large igneous provinces leave exten-sive coarse-grained cumulate residues within or below thecrust these mafic and ultramafic plutonic sequences repre-sent a significant proportion of the original magma thatpartially crystallized at depth (eg Farnetani et al 1996)For example seismic and petrological studies of OntongJava (eg Farnetani et al 1996) combined with the use ofMELTS (Ghiorso amp Sack 1995) indicate that the volcanicsequence may represent only 40 of the total magmareaching the Moho Neal et al (1997) estimated that30^45 fractional crystallization took place for theOntong Java magmas and produced cumulate thicknessesof 7^14 km corresponding to 11^22 km of flood basaltsdepending on the presence of pre-existing oceanic crustand the total crustal thickness (25^35 km) Interpretationsof seismic velocity measurements suggest the presence ofpyroxene and gabbroic cumulates 9^16 km thick beneaththe Ontong Java Plateau (eg Hussong et al 1979)Wrangellia is characterized by high crustal velocities in
seismic refraction lines which is indicative of mafic
plutonic rocks extending to depth beneath VancouverIsland (Clowes et al 1995)Wrangellia crust is 25^30 kmthick beneath Vancouver Island and is underlain by astrongly reflective zone of high velocity and density thathas been interpreted as a major shear zone where lowerWrangellia lithosphere was detached (Clowes et al 1995)The vast majority of the Karmutsen flows are evolved tho-leiitic lavas (eg low MgO high FeOMgO) indicating animportant role for partial crystallization of melts at lowpressure and a significant portion of the crust beneathVancouver Island has seismic properties that are consistentwith crystalline residues from partially crystallizedKarmutsen magmas To test the proportion of the primarymagma that fractionated within the crust MELTS(Ghiorso amp Sack 1995) was used to simulate fractionalcrystallization using several estimated primary magmacompositions from the PRIMELT1 modeling results Themajor-element composition of the primary magmas couldnot be estimated for the tholeiitic basalts because of exten-sive plagthorn cpxthornol fractionation and thus the MELTSmodeling of the picrites is used as a proxy for the evolutionof major elements in the volcanic sequence A pressure of 1kbar was used with variable water contents (anhydrousand 02wt H2O) calculated at the quartz^fayalite^magnetite (QFM) oxygen buffer Experimental resultsfrom Ontong Java indicate that most crystallizationoccurs in magma chambers shallower than 6 km deep(52 kbar Sano amp Yamashita 2004) however some crys-tallization may take place at greater depths (3^5 kbarFarnetani et al 1996)A selection of the MELTS results are shown in Fig 15
Olivine and spinel crystallize at high temperature until15^20wt of the liquid mass has fractionated and theresidual magma contains 9^10wt MgO (Fig 15f) At1235^12258C plagioclase begins crystallizing and between1235 and 11908C olivine ceases crystallizing and clinopyr-oxene saturates (Fig 15f) As expected the addition of clin-opyroxene and plagioclase to the crystallization sequencecauses substantial changes in the composition of the resid-ual liquid FeO and TiO2 increase and Al2O3 and CaOdecrease while the decrease in MgO lessens considerably(Fig 15) The near-vertical trends for the Karmutsen tho-leiitic basalts especially for CaO do not match the calcu-lated liquid-line-of-descent very well which misses manyof the high-CaO high-Al2O3 low-FeO basalts A positivecorrelation between Al2O3CaO and SrNd indicates thataccumulation of plagioclase played a major role in the evo-lution of the tholeiitic basalts (Fig 15g) Increasing thecrystallization pressure slightly and the water content ofthe melts does not systematically improve the match ofthe MELTS models to the observed dataThe MELTS results indicate that a significant propor-
tion of the crystallization takes place before plagioclase(15^20 crystallization) and clinopyroxene (35^45
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
31
crystallization) join the crystallization sequence Theseresults suggest that445 crystallization of the primarymagmas occurred (estimated from the residual liquid percent in Fig 15f) and if the original flood basalt stratigra-phy on Vancouver Island was 6 km thick at least anequivalent amount of complementary plutonic rocksshould have crystallized The Karmutsen basalts and theirplutonic residues thus represent a significant addition ofcrust (perhaps 412 km thickness) previously thickenedduring the formation of the underlying Paleozoic arc
Growth of the Wrangellia oceanic plateauon Vancouver IslandThe extrusive part of the Wrangellia oceanic plateau onVancouver Island was constructed in a three-layersequence of mostly tholeiitic basalts with a restrictedrange of composition (Fig 16) The plateau overliesDevonian to Mississippian arc volcanic sequences andmarine sediments of Mississippian to Permian ages thatare intruded by gabbroic rocks related to the flood basaltsThe volcanic stratigraphy of theWrangellia plateau formed
00
05
10
15
20
25
30
35
40
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
16
0 2 4 6 8 10 12 14 16 18 20
10
11
12
13
14
15
16
17
18
19
20
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
0 2 4 6 8 10 12 14 16 18 20
MgO (wt)
0 10 20 30 40 50
percent of total mass
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
) Mineral proportions
Sp (e)
Ol
CpxPlag
PicriteHigh-MgO basaltTholeiitic basalt
TiO2 (wt)
MgO (wt)
Al2O3 (wt) CaO (wt)
FeO (wt)
MgO(a) (b)
(c) (d)
MgO (wt)MgO (wt)
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
)
Residual liquid ()
1220
1190
Ol + Sp
Plag+Ol+Sp Plag+Cpx+Sp
02 wt H2O
no H2O
Residual liquid and crystallizing phases
(f)
0102030405060708090100
plagioclaseaccumulation
plagioclaseaccumulation
plagioclase
accumulatio
n
10
12
14
16
0 10 20 30 40
Al2O3CaO
SrNd
tholeiitic basalts
(g)
Fig 15 Forward fractional crystallization modeling results for major elements from MELTS (Ghiorso amp Sack 1995) compared with picriticand tholeiitic lavas from the Karmutsen Formation (a^d) Major-element compositions of residual liquids compared with Karmutsen lavacompositions For clarity the results are shown only for one starting composition (open circle) using the composition of the estimated primarymagma for the high-MgO lava series from the modeling technique of Herzberg et al (2007) for sample 4723A13 as shown in Fig 12 The esti-mated primary magmas for samples 4723A3 and 93G171 yield similar results A pressure of 1 kbar was used with no H2O and 02wt H2Ocalculated at QFM black line is anhydrous and grey line is 02wt H2OThe end-point for the anhydrous result is at 78 fractionation and11308C and the end-point for the result with 02wt H2O is 75 fractionation and11108C (e) Mineral proportions of crystallized phases fromone MELTrun (sample 4723A13) with no H2O (f) Per cent residual liquid vs temperature labelled with intervals of crystallizing phases (g)Al2O3CaO vs SrNd for the tholeiitic basalts
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
32
in a single eruptive phase (c 231^225 Ma) as an emergentbasalt sequence during a deeper-water shallow-water andsubaerial stage and subsided below sea level relativelysoon after its formation (Fig 16) The deeper-water stagecontains pillowed and massive submarine flows (3000mthick) that built up the submarine volcanic edifice from theseafloor The lowest pillow basalts were emplaced onunconsolidated and lithified fine-grained siliciclastic andcarbonaceous sediments containing Daonella beds(c 235^232 Ma) that are intruded by abundantKarmutsen mafic sills (Fig 16a) The different pillowedand massive submarine flows resulted primarily from dif-ferent effusive rates and local topography Some of themassive submarine flows commonly recognized by radiat-ing columnar jointing represent master tubes for the deliv-ery of lava to distal parts of submarine flow fields Thepillowed flows which commonly have uneven topography
in modern sequences may have served as levees to themassive submarine flows rather than allowing unob-structed spreading and inflation typical of the massive sub-aerial flows that form on shallow slopes in continentalflood basalt provinces (eg Self et al 1997) The rapid suc-cession of flows is evident by the near absence of sedimentbetween submarine flows Minor volumes of picritic pillowlavas occur in the upper part of the pillowed flow sequencein areas of the plateau exposed on northern VancouverIsland (Fig 16 Nixon et al 2008)The growth of the plateau in shallow water led to
emplacement of an increasing proportion of volcaniclastics(pillow breccias and hyaloclastites) that accumulated up to1500m thick in some areas (Fig 16b Nixon et al 2008)Horizons of pillowed flows occur locally within the volca-niclastic stratigraphy Sedimentary structures (eg gradedbedding fluidization structures) and petrographic textures
Late Devonian-Mississippian arc
Daonella beds
Middle Triassic
deeper water
Late Devonian-Mississippian arc
pillowed flows massive submarine
flows
pillow brecciahyaloclastiteshallow water
~4 km
~6 km
Late Triassic limestoneinterflow limestonelenses
Karmutsen Formation
mafic intrusions sediment-sill complex
Karmutsen Formation
(a) (b)
(c) (d)
subaerial flows
Late Devonian-Mississippian arc Late Devonian-Mississippian arc
picritic pillow basalts
Miss to Permian marine sediments
sills
Fig 16 Growth of the Wrangellia oceanic plateau onVancouver Island (a) Deep-water eruptive phase of pillowed and massive submarineflows with almost no intervening sediments onto MiddleTriassic shale containing Daonella beds (c 235^232 Ma) The basement of the plateauis composed of Devonian to Mississippian arc sequences overlain by Mississippian to Permian carbonates and siliciclastic sediments(b) Shallow-water flow emplacement with increasing proportion of volcaniclastics resulted primarily from cooling^contraction granulationPicritic pillow basalts occur primarily near the transition between pillowed submarine flows and volcaniclastic units Pillowed and massiveflows occur locally within the volcaniclastic unit (c) Emergent stage of growth consists of overlapping subaerial flow fields similar to continen-tal flood basalts with pahoehoe flow structures Multi-tiered columnar jointing occurrs locally in response to ponding of flows (eg Bondre et al2004) Interflow carbonate deposits and plagioclase megacrystic flows developed locally during the waning stage of volcanism (d) Post-volcanicsubsidence and deposition of Late Carnian to Norian (c 212^223 Ma) platform carbonates on top of the plateau Units are labeled in the dia-grams Thicknesses of the underlying units are not to scale
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
33
within the volcaniclastic rocks indicate formation primar-ily via cooling^contraction granulation magma^water^steam interaction autobrecciation resedimentation andmass-wasting rather than pyroclastic eruption (Nixonet al 2008)A broad subaerial platform was constructed by emplace-
ment of subaerial lava flows during an emergent subaerialstage (Fig 16c) The subaerial unit predominantly com-prises well-layered sheet flows that formed in a similarway to continental flood basalts as hundred(s) of pahoehoeflow fields erupted on shallow slopes (Self et al 1997) how-ever volcaniclastic and pillowed flow units are locally pre-sent mostly in the upper parts of the stratigraphyLocalized ponding of flows led to the formation of multi-tiered columnar jointing in some subaerial flows (egBondre et al 2004) As volcanism waned local interflowcarbonate deposits developed and plagioclase megacrysticflows were erupted Subsidence of the plateau is recordedby the interflow lenses and deposits of platform carbonates(40^1500m thick) on top of the Karmutsen basalts thatcontain Late Carnian to Early Norian fossils (Fig 16d c216^227 Ma Nixon et al 2008)The evolution of the internal architecture of the
Wrangellia oceanic plateau was primarily related to theconditions of emplacement (ie water depth and eruptionenvironment) There does not appear to have been a sys-tematic coincident evolution or shift of lava compositionsonVancouver Island as has been recognized in most con-tinental flood basalts and parts of the Wrangellia oceanicplateau in Alaska and Yukon (Greene et al 2008 2009)In general with the exception of the picritic pillow lavasin the submarine stratigraphy and local variations betweenhigh-MgO and tholeiitic lavas there are no significantcompositional differences between lavas erupted duringthe different stages of growth of the plateau (ie submarineand subaerial flows) nor are substantial compositional dif-ferences generally found between massive submarine flowsin contact with pillowed flows tholeiitic basalt flows nearthe base of the Karmutsen Formation are compositionallysimilar to flows sampled near the top of the volcanicsequence Where it was possible to sample continuous sec-tions of volcanic stratigraphy greater than several hun-dreds of meters (eg Keogh^Maynard Lake area west ofthe Karmutsen Range Fig 2) there is no clear relationshipbetween stratigraphic position and compositionalvariation
CONCLUSIONSThe Karmutsen Formation covers extensive areas ofVancouver Island (20 000 km2) and forms part of a largeoceanic plateau erupted over a geologically short interval(c 225^231 Ma) The tripartite volcanic stratigraphyexposed on Vancouver Island is at least 6 km thick andcomprises a succession of submarine flows volcaniclastic
deposits and massive subaerial flowsThese volcanologicaldifferences are primarily related to the eruption environ-ment (deep-water shallow-water or subaerial) The rapidgrowth of the plateau prevented significant accumulationsof intervening sediment except in the uppermost stratigra-phy where isolated limestone lenses commonly associatedwith pillow basalt and hyaloclastite deposits preserve arecord of the final stages of subsidence of the plateauThe Wrangellia plateau on Vancouver Island was con-
structed dominantly of tholeiitic basalt with restrictedmajor- and trace-element and isotopic compositional var-iations Minor volumes of picritic pillow basalts eruptedlate in the submarine phase The Karmutsen basalts wereemplaced onto older Paleozoic arc volcanic sequences andlow NbLa for the picrites may indicate a contributionfrom the sub-arc lithospheric mantle Modeling of the pet-rogenesis of the Karmutsen picrites indicates melting ofanomalously hot mantle (15008C) and extensive degreesof partial melting (23^27) and is consistent with aplume initiation model Combined Sr^Nd^Hf^Pb isotopicsystematics indicate that the picritic and tholeiitic lavaswere derived from a depleted plume-type mantle sourcedistinct from the source of MORB that has compositionalsimilarities to the source of the lavas forming theCaribbean Plateau Picrites were generated by melting ofspinel lherzolite whereas the tholeiitic basalts formed asaggregate melts of both garnet and spinel lherzolite Thetholeiitic basalts underwent extensive low-pressure fractio-nation (52^3 kbar) and seismic studies indicate that someof theWrangellia crust beneathVancouver Island may cor-respond to the plutonic residues of the Karmutsen floodbasalts
ACKNOWLEDGEMENTSWe are grateful to Nick Arndt for helping us get this pro-ject started and Nick Massey for insights into VancouverIsland geology We greatly appreciate the assistance withPRIMELT modeling from Claude Herzberg We thankMikkel Schau for his insight and enthusiasm during field-work Jane Barling assisted with analyses by MC-ICP-MSThoughtful and thorough reviews of the manuscript byJohn Mahoney Fred Frey Wendy Bohrson DougalJerram and Teal Riley are greatly appreciated Fundingwas generously provided by the Rocks to Riches Programadministered by the BC amp Yukon Chamber of Mines in2004 by the BC Geological Survey and by NSERCDiscovery Grants to JSS and DW ARG was supportedby a University Graduate Fellowship at UBC
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
34
REFERENCESBarker F Brown A S Budahn J R amp Plafker G (1989) Back-arc
with frontal-arc component origin of Triassic Karmutsen basaltBritish Columbia Canada Chemical Geology 75 81^102
Blichert-Toft JWeis D Maerschalk C Agranier A amp Albare de F(2003) Hawaiian hot spot dynamics as inferred from the Hf and Pbisotope evolution of Mauna Kea volcano Geochemistry GeophysicsGeosystems 4(2) 1^27 doi1010292002GC000340
Bondre N R Duraiswami R A amp Dole G (2004) Morphologyand emplacement of flows from the Deccan Volcanic ProvinceIndia Bulletin of Volcanology 66 29^45
Carlisle D (1963) Pillow breccias and their aquagene tuffs QuadraIsland British Columbia Journal of Geology 71 48^71
Carlisle D (1972) Late Paleozoic to Mid-Triassic sedimentary-volcanic sequence on Northeastern Vancouver Island Geological
Society of Canada Paper Report of Activities 72-1 Part B 24^30Carlisle D amp SuzukiT (1974) Emergent basalt and submergent car-
bonate^clastic sequences including the Upper Triassic Dilleri andWelleri zones on Vancouver Island Canadian Journal of Earth
Sciences 11 254^279Chauvel C amp Blichert-Toft J (2001) A hafnium isotope and trace
element perspective on melting of the depleted mantle Earth and
Planetary Science Letters 190 137^151Chayes F (1954)The theory of thin-section analysis Journal of Geology
62 92^101Cho M Liou J G amp Maruyama S (1987) Transition from the zeo-
lite to prehnite^pumpellyite facies in the Karmutsen metabasitesVancouver Island British Columbia Journal of Petrology 27(2)467^494
Clowes R M Zelt C A Amor J R amp Ellis R M (1995)Lithospheric structure in the southern Canadian Cordillera froma network of seismic refraction lines Canadian Journal of Earth
Sciences 32(10) 1485^1513Coffin M F amp Eldholm O (1994) Large igneous provinces Crustal
structure dimensions and external consequences Reviews of
Geophysics 32(1) 1^36Cox K G (1980) A model for flood basalt volcanism Journal of
Petrology 21 629^650Eldholm O amp Coffin M F (2000) Large igneous provinces
and plate tectonics In Richards M A Gordon R G amp vander Hilst R D (eds) The History and Dynamics of Global
Plate Motions American Geophysical Union Geophysical Monograph 121309^326
Elliott T R Hawkesworth C J amp Grolaquo nvold K (1991) Dynamicmelting of the Iceland plume Nature 351 201^206
Escuder-Viruete J Pecurren rez-Estaucurren n A Contreras F Joubert MWeis D Ullrich T D amp Spadea P (2007) Plume mantle sourceheterogeneity through time Insights from the Duarte ComplexHispaniola northeastern Caribbean Journal of Geophysical Research112(B04203) doi1010292006JB004323
Farnetani C G Richards M A amp Ghiorso M S (1996)Petrological models of magma evolution and deep crustal structurebeneath hotspots and flood basalt provinces Earth and Planetary
Science Letters 143 81^94Fitton J G amp Godard M (2004) Origin and evolution of magmas
on the Ontong Java Plateau In Fitton J G Mahoney J JWallace P J amp Saunders A D (eds) Origin and Evolution of the
Ontong Java Plateau Geological Society London Special Publications 229151^178
Fitton J G Saunders A D Norry M J Hardarson B S ampTaylor R N (1997) Thermal and chemical structure of theIceland plume Earth and Planetary Science Letters 153 197^208
Fitton J G Saunders A D Kempton P D amp Hardarson B S(2003) Does depleted mantle form an intrinsic part of the Icelandplume Geochemistry Geophysics Geosystems 4(3) 1032 doi1010292002GC000424
Frey F A Huang S Blichert-Toft J Regelous M amp Boyet M(2005) Origin of depleted components in basalt related to theHawaiian hotspot Evidence from isotopic and incompatible ele-ment ratios Geochemistry Geophysics Geosystems 6(1) 23 doi1010292004GC000757
Galer S J G amp Abouchami W (1998) Practical application of leadtriple spiking for correction of instrumental mass discriminationMineralogical Magazine 62A 491^492
Ghiorso M S amp Sack R O (1995) Chemical mass transfer in mag-matic processes IV A revised and internally consistent thermody-namic model for the interpolation and extrapolation of liquid^solidequilibria in magmatic systems at elevated temperatures and pres-sures Contributions to Mineralogy and Petrology 119 197^212
Greene A R Scoates J S amp Weis D (2005)WrangelliaTerrane onVancouver Island Distribution of flood basalts with implicationsfor potential Ni^Cu^PGE mineralization In Grant B (ed)Geological Fieldwork 2004 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2005-1 209^220Greene A R Scoates J S Nixon G T amp Weis D (2006) Picritic
lavas and basal sills in the Karmutsen flood basalt provinceWrangellia northern Vancouver Island In Grant B (ed)Geological Fieldwork 2005 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2006-1 39^52Greene A R Scoates J S amp Weis D (2008) Wrangellia flood
basalts in Alaska A record of plume^lithosphere interaction in aLate Triassic accreted oceanic plateau Geochemistry Geophysics
Geosystems 9(Q12004) doi1010292008GC002092Greene A R Scoates J S Weis D amp Israel S (2009)
Geochemistry of triassic flood basalts from the Yukon (Canada)segment of the accreted Wrangellia oceanic plateau Lithosdoi101016jlithos200811010
Hauff F Hoernle K Tilton G Graham D W amp Kerr A C(2000) Large volume recycling of oceanic lithosphere over shorttime scales geochemical constraints from the Caribbean LargeIgneous Province Earth and Planetary Science Letters 174 247^263
Hauff F Hoernle K amp Schmidt A (2003) Sr^Nd^Pb compositionof Mesozoic Pacific oceanic crust (Site 1149 and 801 ODP Leg185)Implications for alteration of ocean crust and the input into theIzu^Bonin^Mariana subduction system Geochemistry Geophysics
Geosystems 4(8) doi1010292002GC000421Herzberg C (2004) Partial melting below the Ontong Java Plateau
In Fitton J G Mahoney J J Wallace P J amp Saunders A D(eds) Origin and Evolution of the Ontong Java Plateau Geological SocietyLondon Special Publications 229 179^183
Herzberg C (2006) Petrology and thermal structure of the Hawaiianplume from Mauna Kea volcano Nature 444(7119) 605
Herzberg C amp Asimow P D (2008) Petrology of some oceanicisland basalts PRIMELT2XLS software for primary magma cal-culation Geochemistry Geophysics Geosystems 9(Q09001) doi1010292008GC002057
Herzberg C amp OrsquoHara M J (2002) Plume-associated ultramaficmagmas of Phanerozoic age Journal of Petrology 43(10) 1857^1883
Herzberg C Asimow P D Arndt N Niu Y Lesher C MFitton J G Cheadle M J amp Saunders A D (2007)Temperatures in ambient mantle and plumes Constraints frombasalts picrites and komatiites Geochemistry Geophysics Geosystems8(Q02006) doi1010292006GC001390
Huang S amp Frey F A (2005) Trace element abundances of MaunaKea basalt from phase 2 of the Hawaii Scientific Drilling Project
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
35
Petrogenetic implications of correlations with major element con-tent and isotopic ratios Geochemistry Geophysics Geosystems 4(6)doi1010292002GC000322
Hussong D M Wipperman L K amp Kroenke L W (1979) Thecrustal structure of the Ontong Java and Manihiki OceanicPlateaus Journal of Geophysical Research 84(B11) 6003^6010
Jerram D A amp Widdowson M (2005) The anatomy of ContinentalFlood Basalt Provinces geological constraints on the processes andproducts of flood volcanism Lithos 79(3^4) 385^405
Jones D L Silberling N J amp Hillhouse J (1977)Wrangellia a dis-placed terrane in northwestern North America CanadianJournal ofEarth Sciences 14(11) 2565^2577
Kerr A C (2003) Oceanic Plateaus In Rudnick R (ed) The Crust
Treatise on GeochemistryVol 3 Oxford Elsevier Science pp 537^565Kerr A C amp Mahoney J J (2007) Oceanic plateaus Problematic
plumes potential paradigms Chemical Geology 241 332^353Kerr A CTarney J Marriner G F Klaver GT Saunders A D
amp Thirlwall M F (1996) The geochemistry and petrogenesis ofthe Late-Cretaceous picrites and basalts of Curacao NetherlandsAntilles a remnant of an oceanic plateau Contributions to
Mineralogy and Petrology 124(1) 29^43Kerr A C Marriner G F Tarney J Nivia A Saunders A D
Thirlwall M F amp Sinton A C (1997) Cretaceous basaltic ter-ranes in Western Colombia Elemental chronological and Sr-Ndisotopic constraints on petrogenesis Journal of Petrology 38(6)677^702
Kinzler R J (1997) Melting of mantle peridotite at pressuresapproaching the spinel to garnet transition Application to mid-ocean ridge basalt petrogenesis Journal of Geophysical Research
102(B1) 853^874Kinzler R J amp Grove T L (1992) Primary magmas of mid-ocean
ridge basalts 1 Experiments and methods Journal of Geophysical
Research 97(B5) 6885^6906Korenaga J (2005) Why did not the Ontong Java Plateau form sub-
aerially Earth and Planetary Science Letters 234 385^399Kuniyoshi S (1972) Petrology of the Karmutsen (Volcanic) Group
northeastern Vancouver Island British Columbia PhD disserta-tion University of California Los Angeles 242 pp
Lassiter J C DePaolo D J amp Mahoney J J (1995) Geochemistry ofthe Wrangellia flood basalt province Implications for the role ofcontinental and oceanic lithosphere in flood basalt genesis Journalof Petrology 36(4) 983^1009
LeBas M J (2000) IUGS reclassification of the high-Mg and picriticvolcanic rocks Journal of Petrology 41(10) 1467^1470
Longhi J (2002) Some phase equilibrium systematics of lherzolitemelting I Geochemistry Geophysics Geosystems 3(3) doi1010292001GC000204
MacDonald G A amp Katsura T (1964) Chemical composition ofHawaiian lavas Journal of Petrology 5 82^133
Mahoney J J (1987) An isotopic survey of Pacific oceanic plateausimplications for their nature and origin In Keating B HFryer P Batiza R amp Boehlert GW (eds) Seamounts Islands andAtolls American Geophysical Union Geophysical Monograph 43 207^220
Mahoney J LeRoex A P Peng Z Fisher R L amp Natland J H(1992) Southwestern limits of Indian Ocean Ridge mantle and theorigin of low 206Pb204Pb mid-ocean ridge basalt isotope systema-tics of the central Southwest Indian Ridge (178^508E) Journal ofGeophysical Research 97(B13) 19771^19790
Mahoney J J Sinton J M Kurz M D MacDougall J DSpencer K J amp Lugmair GW (1994) Isotope and trace elementcharacteristics of a super-fast spreading ridge East Pacific rise13^238S Earth and Planetary Science Letters 121 173^193
Massey N W D (1995a) Geology and mineral resources of theAlberni^Nanaimo Lakes sheet Vancouver Island 92F1W 92F2Eand part of 92F7E British Columbia Ministry of Energy Mines and
Petroleum Resources Paper 1992-2 132 ppMassey N W D (1995b) Geology and mineral resources of the
Cowichan Lake sheet Vancouver Island 92C16 British Columbia
Ministry of Energy Mines and Petroleum Resources Paper 1992-3 112 ppMassey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005a) Digital Geology Map of British Columbia Tile NM9 MidCoast BC British Columbia Ministry of Energy and Mines Geofile
2005-2Massey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005b) Digital Geology Map of British Columbia Tile NM10Southwest BC British Columbia Ministry of Energy and Mines Geofile
2005-3McDonough W F amp Sun S (1995) The composition of the Earth
Chemical Geology 120 223^253Monger JW H amp Journeay J M (1994) Basement geology and tec-
tonic evolution of theVancouver region In Monger JW H (ed)Geology and Geological Hazards of the Vancouver Region Southwestern
British Columbia Geological Survey of Canada Bulletin 481 3^25Muller J E (1977) Geology of Vancouver Island Geological Survey of
Canada Open File Map 463Muller J E (1980)The Paleozoic Sicker Group of Vancouver Island British
Columbia Geological Survey of Canada Paper 79-30 22 ppMuller J E Northcote K E amp Carlisle D (1974) Geology and
mineral deposits of Alert Bay^Cape Scott map area VancouverIsland British Columbia Geological Survey of Canada Paper 74-8 77
Neal C R Mahoney J J Kroenke L W Duncan R A ampPetterson M G (1997) The Ontong^Java Plateau InMahoney J J amp Coffin M F (eds) Large Igneous Provinces
Continental Oceanic and Planetary FloodVolcanism American Geophysical
Union Geophysical Monograph 100 183^216Niu Y Collerson K D Batiza R Wendt J I amp Regelous M
(1999) Origin of enriched-typed mid-ocean ridge basalt at ridgesfar from mantle plumes The East Pacific Rise at 11820rsquoN Journalof Geophysical Research 104 7067^7088
Nixon G T amp Orr A J (2007) Recent revisions to the EarlyMesozoic stratigraphy of Northern Vancouver Island (NTS 102I092L) and metallogenic implicatinos British Columbia InGrant B (ed) Geological Fieldwork 2006 British Columbia Ministry of
Energy Mines and Petroleum Resources Paper 2007-1 163^177Nixon GT Hammack J L KoyanagiV M Payie G J Haggart
JW Orchard M JTozerT Friedman R M Archibald D APalfy J amp Cordey F (2006a) Geology of the Quatsino^PortMcNeill area northernVancouver Island British Columbia Ministry
of Energy Mines and Petroleum Resources Geoscience Map 2006-2 scale150 000
Nixon G T Kelman M C Stevenson D Stokes L A ampJohnston K A (2006b) Preliminary geology of the Nimpkishmap area (NTS 092L07) northern Vancouver Island BritishColumbia In Grant B amp Newell J M (eds) Geological Fieldwork2005 British Columbia Ministry of Energy Mines and Petroleum Resources
Paper 2006-1 135^152Nixon G T Laroque J Pals A Styan J Greene A R amp
Scoates J S (2008) High-Mg lavas in the Karmutsen flood basaltsnorthernVancouver Island (NTS 092L) Stratigraphic setting andmetallogenic significance In Grant B (ed) Geological Fieldwork
2007 British Columbia Ministry of Energy Mines and Petroleum
Resources Paper 2008-1 175^190Nowell G M Kempton P D Noble S R Fitton J G
Saunders A Mahoney J J amp Taylor R N (1998) High precisionHf isotope measurements of MORB and OIB by thermal ionisation
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
36
mass spectrometry insights into the depleted mantle Chemical
Geology 149 211^233Parrish R R amp McNicoll V J (1992) U^Pb age determinations
from the southern Vancouver Island area British Columbia InRadiogenic Age and Isotopic Studies Report 5 Geological Survey of
Canada Paper 91-2 79^86Peate DW (1997) The Parana^Etendeka Province In Coffin M F
amp Mahoney J (eds) Large Igneous Provinces Continental Oceanic andPlanetary Flood Volcanism American Geophysical Union Geophysical
Monograph 100 217^245Perfit M R Fornari D J Ridley W I Kirk P D Casey J
Kastens K A Reynolds J R Edwards M Desonie DShuster R amp Paradis S (1996) Recent volcanism in the Siqueirostransform fault picritic basalts and implications for MORBmagma genesis Earth and Planetary Science Letters 141(1^4) 91^108
Pretorius W Weis D Williams G Hanano D Kieffer B ampScoates J S (2006) Complete trace elemental characterization ofgranitoid (USGSG-2GSP-2) reference materials by high resolutioninductively coupled plasma-mass spectrometry Geostandards and
Geoanalytical Research 30(1) 39^54Regelous M Niu Y Wendt J I Batiza R Greig A amp
Collerson K D (1999) Variations in the geochemistry of magma-tism on the East Pacific Rise at 10830rsquoN since 800 ka Earth and
Planetary Science Letters 168 45^63Recurren villon S Chauvel C Arndt N T Pik R Martineau F
Fourcade S amp Marty B (2002) Heterogeneity of the Caribbeanplateau mantle source Sr O and He isotopic compositions of oli-vine and clinopyroxene from Gorgona Island Earth and Planetary
Science Letters 205(1^2) 91Rhodes J M (1996) Geochemical stratigraphy of lava flows samples
by the Hawaii Scientific Drilling Project Journal of Geophysical
Research 101(B5) 11729^11746Rhodes J M amp Vollinger M J (2004) Composition of basaltic lavas
sampled by phase-2 of the Hawaii Scientific Drilling ProjectGeochemical stratigraphy and magma types Geochemistry
Geophysics Geosystems 5(3) doi1010292002GC000434Richards M A Jones D L Duncan R A amp DePaolo D J (1991)
A mantle plume initiation model for the Wrangellia flood basaltand other oceanic plateaus Science 254 263^267
Salters V J M amp Hart S R (1991) The mantle sources of oceanridges islands and arcs the Hf-isotope connection Earth and
Planetary Science Letters 104 364^380Salters V J M amp Stracke A (2004) Composition of the depleted
SaltersV J amp WhiteW (1998) Hf isotope constraints on mantle evo-lution Chemical Geology 145 447^460
SanoT amp Yamashita S (2004) Experimental petrology of basementlavas from Ocean Drilling Program Leg 192 implications for dif-ferentiation processes in Ontong Java Plateau magmas InFitton J G Mahoney J JWallace P J amp Saunders A D (eds)Origin and Evolution of the Ontong Java Plateau Geological Society
London Special Publications 229 185^218Saunders A D (2005) Large igneous provinces Origin and environ-
mental consequences Elements 1 259^263Self S ThordarsonT amp Keszthely L (1997) Emplacement of conti-
nental flood basalt lava flows In Mahoney J J amp Coffin M F(eds) Large Igneous Provinces Continental Oceanic and Planetary Flood
Volcanism American Geophysical Union Geophysical Monograph 100381^410
Shaw D M (2000) Continuous (dynamic) melting theory revisitedCanadian Mineralogist 38(5) 1041^1063
Sluggett C L (2003) Uranium^lead age and geochemical constraintson Paleozoic and Early Mesozoic magmatism in WrangelliaTerrane Saltspring Island British Columbia BSc thesisUniversity of British ColumbiaVancouver 84 pp
Storey M Mahoney J J amp Saunders A D (1997) Cretaceousbasalts in Madagascar and the transition between plume and con-tinental lithospheric mantle sources In Mahoney J J ampCoffin M F (eds) Large Igneous Provinces Continental Oceanic and
Planetary Flood Volcanism Aamerican Geophysical Union Geophysical
Monograph 100 95^122Surdam R C (1967) Low-grade metamorphism of the Karmutsen
Group PhD dissertation University of California Los Angeles288 pp
Sutherland-Brown A Yorath C J Anderson R G amp Dom K(1986) Geological maps of southern Vancouver IslandLITHOPROBE I 92C10 11 14 16 92F1 2 7 8 Geological Survey ofCanada Open File 1272
Tejada M L G Mahoney J J Castillo P R Ingle S P Sheth HC amp Weis D (2004) Pin-pricking the elephant evidence on theorigin of the Ontong Java Plateau from Pb^Sr^Hf^Nd isotopiccharacteristics of ODP Leg 192 basalts In Fitton J GMahoney J J Wallace P J amp Saunders A D (eds) Origin and
Evolution of the Ontong Java Plateau Geological Society London Special
Publications 229 133^150Thompson P M E Kempton P D amp Kerr A C (2008) Evaluation
of the effects of alteration and leaching on Sm^Nd and Lu^Hf sys-tematics in submarine mafic rocks Lithos 104(1^4) 164^176
Thompson P M E Kempton P D White R V Kerr A CTarney J Saunders A D Fitton J G amp McBirney A (2003)Hf-Nd isotope constraints on the origin of the CretaceousCaribbean plateau and its relationship to the Galapagos plumeEarth and Planetary Science Letters 217(1^2) 59^75
Vervoort J D Patchett P Blichert-Toft J amp Albare de F (1999)Relationships between Lu^Hf and Sm^Nd isotopic systems in theglobal sedimentary system Earth and Planetary Science Letters 16879^99
Walter M J (1998) Melting of garnet peridotite and the origin ofkomatiite and depleted lithosphere Journal of Petrology 39(1) 29^60
Weis D Kieffer B Maerschalk C Barling J de Jong JWilliams G A Hanano D Mattielli N Scoates J SGoolaerts A Friedman R A amp Mahoney J B (2006) High-pre-cision isotopic characterization of USGS reference materials byTIMS and MC-ICP-MS Geochemistry Geophysics Geosystems
7(Q08006) doi1010292006GC001283Weis D Kieffer B Hanano D Silva I N Barling J
Pretorius W Maerschalk C amp Mattielli N (2007) Hf isotopecompositions of US Geological Survey reference materialsGeochemistry Geophysics Geosystems 8(Q06006) doi1010292006GC001473
Wheeler J O amp McFeely P (1991) Tectonic assemblage map of theCanadian Cordillera and adjacent part of the United States ofAmerica Geological Survey of Canada Map 1712A
WhiteW M Albare de F amp Tecurren louk P (2000) High-precision analy-sis of Pb isotope ratios by multi-collector ICP-MS Chemical Geology167 257^270
Yorath C J Sutherland Brown A amp Massey N W D (1999)LITHOPROBE southern Vancouver Island British Columbia Geological
Survey of Canada Bulletin 498 145 ppZou H (1998) Trace element fractionation during modal and
non-model dynamic melting and open-system melting Amathematical treatment Geochimica et Cosmochimica Acta 62(11)1937^1945
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
37
Zou H amp Reid M R (2001) Quantitative modeling of trace elementfractionation during incongruent dynamic melting Geochimica et
Cosmochimica Acta 65(1) 153^162
APPENDIX SAMPLEPREPARAT ION ANDANALYT ICAL METHODSOnly the freshest rocks in the field were sampled and onlythe least altered samples were selected for chemical andisotopic analysis based on thorough petrographic inspec-tion Fifty-nine of the 129 collected samples were crushed(400 g) into small pieces 52mm in diameter in aRocklabs hydraulic piston crusher between WC plates tominimize contamination The coarse-crush was mixedand 100 g was powdered in a planetary mill using agatejars and balls cleaned with quartz sand between samples
ActLabs analytical methodsThe major- and trace-element compositions of the whole-rock powders were determined at Activation LaboratoriesLtd (Actlabs) in Ancaster Ontario Analytical techniquesand detection limits are also available from Actlabs (httpwwwactlabscommethsub_code4erehtm) The analyticalmethod for each of the elements analyzed is indicated inTable 2 For the major elements a 02 g sample was mixedwith a lithium metaborate^lithium tetraborate mixtureand fused in a graphite crucible The molten mixture waspoured into a 5 HNO3 solution and shaken until dis-solved (30min) The samples were analyzed for majoroxides and selected trace elements by inductively coupledplasma optical emission spectrometry (ICP-OES) on acombination simultaneous^sequential Thermo Jarrell-AshEnviro II system Internal calibration was achieved usinga variety of international reference materials (eg W-2BIR-1 DNC-1) and independent control samplesAdditional trace elements were analyzed by both instru-mental neutron activation analysis (INAA) and induc-tively couple plasma mass spectrometry (ICP-MS) ForINAA15^25 g of sample was weighed into small polyeth-ylene vials and irradiated with a control international ref-erence material CANMET WMS-1 and NiCr flux wires ata thermal neutron flux of 71012 neutronscm2 per s in theMcMaster Nuclear Reactor Following a 7 day waitingperiod the samples were measured on an Ortec high-purity Ge detector linked to a Canberra Series 95 multi-channel analyzer Activities for each element weredecay- and weight-corrected and compared with a detectorcalibration developed from multiple international certifiedreference materials For ICP-MS 025 g of sample wasdigested in HF followed by HNO3^HClO4 heated andtaken to dryness The samples were brought back into solu-tion with HCl Samples were analyzed using a PerkinElmer Optima 3000 ICP In-lab standards or certified ref-erence materials (eg W-2 BIR-1 DNC-1) were used for
quality control A total of 15 blind duplicates were ana-lyzed from the same sample powders to assessreproducibility
University of Massachusetts analyticalmethodsFourteen sample duplicate powders taken from the same100 g sample aliquot (high-MgO basalts and picrites)were also analyzed at the Ronald B Gilmore X-RayFluorescence (XRF) Laboratory at the University ofMassachusetts Major elements were measured on a fusedLa-bearing lithium borate glass disc using a SiemensMRS-400 spectrometer Trace element concentrations(Rb Sr Ba Ce Nb Zr Y Pb Zn Ga Ni Cr V) weremeasured on a separate powder pellet using a PhilipsPW2400 sequential spectrometer using a Rh tube LOIand ferrous iron measurements were made as describedby Rhodes amp Vollinger (2004) Precision and accuracy esti-mates for the data have been described by Rhodes (1996)and Rhodes amp Vollinger (2004) Results for each sampleare the average of two separate analyses and are shown inElectronic Appendix 2 for a comparison of analyses ofduplicate powders using XRFand ICP-MS
PCIGR analytical methodsA subset of 19 samples was selected for high-precisiontrace-element analysis and Sr Nd Pb and Hf isotopicanalysis at the Pacific Centre for Isotopic andGeochemical Research (PCIGR) at the University ofBritish Columbia (UBC) Samples were selected from the59 samples analyzed for whole-rock chemistry at ActLabsbased on major- and trace-element chemistry alteration(low LOI and petrographic alteration index) and samplelocation Samples were prepared for trace-element analysisat the PCIGR by the technique described by Pretorius et al(2006) on unleached rock powders Sample powders(100mg) were weighed in 7ml screw-top Savillex bea-kers and dissolved in 1ml 14N HNO3 and 5ml 48 HFon a hotplate for 48 h at 1308C with periodic ultrasonica-tion Samples were dried and redissolved in 6ml 6N HClon a hotplate for 24 h and then dried and redissolved in1ml concentrated HNO3 for 24 h before final dryingTrace element abundances were measured by high-resolution inductively coupled plasma-mass spectrometry(HR-ICP-MS) with aThermo Finnigan Element2 systemfollowing the procedures described by Pretorius et al(2006) within 24 h of redissolution HFSE and LILE weremeasured in medium-resolution mode at 2000 dilutionusing a PFATeflon spray chamber washed with aqua regiafor 3min between samples REE were measured in high-resolution mode and U and Pb in low-resolution modeat 2000 dilution using a glass spray chamber washedwith 2 HNO3 between samples Total procedural blanksand reference materials (BCR-2 BHVO-2) were analyzedwith the batch of samples Indium was used as an internal
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
38
standard in all samples and standard solutionsBackground and standard solutions were analyzed afterevery five samples to detect memory effects and massdrift Results for PCIGR trace-element analyses areshown in Electronic Appendix 3 for the full set of analyzedtrace elementsSample digestion for purification of Sr Nd Hf and Pb
for column chemistry began by weighing each samplepowder (400^500mg) prior to leaching All samples wereleached with 6N HCl and placed in an ultrasonic bath for15min Samples were rinsed twice with 18 M-cm H2Obetween each leaching step (15 total) until the supernatantliquid was clear [following the method of Mahoney (1987)]Samples were then dried on a hotplate for 24 h andweighed again Sample solutions were then prepared bydissolving 100^250mg of the leached powder dissolvedin 1ml 14N HNO3 and 10ml 48 HF on a hotplate for48 h at 1308C with periodic ultrasonication Samples weredried and redissolved in 6ml 6N HCl on a hotplate for24 h and then dried Pb was separated using anion exhangecolumns and the discard was used for Sr REE and Hfseparation Nd was separated from the other REE and Hfrequired two additional purification steps Detailed proce-dures for column chemistry for separating Sr Nd and Pbat the PCIGR have been described byWeis et al (2006) andHf purification has been described byWeis et al (2007) Srand Nd isotope ratios were measured by thermal ioniza-tion mass spectrometry (TIMS) on a Thermo FinniganTriton system in static mode with relay matrix rotation ona single Ta and double Re^Ta filament respectively Fourto five filaments per barrel of 21 were occupied by stan-dards (NBS 987 for Sr and LaJolla for Nd) for each barrelwhere samples were run Sample Sr and Nd isotopic com-positions were corrected for mass fractionation using86Sr88Srfrac14 01194 and 146Nd144Ndfrac14 07219 Each samplewas then normalized using the barrel average of the refer-ence material relative to the values of143Nd144Ndfrac14 0511858 and 87Sr86Srfrac14 0710248 (Weiset al 2006) During the course of the Vancouver Islandanalyses the La Jolla Nd standard gave an average valueof 05118566 (nfrac14 7) and NBS987 standard gave an aver-age of 07102408 (nfrac1411) (2 error is reported as times106) 147Sm144Nd errors are 15 or 0006 USGeological Survey (USGS) reference material BHVO-2was processed with the samples and yielded Sr and Nd iso-topic ratios of 07034607 and 05129786 respectivelyThese are in agreement with the published values of070347920 and 051298411 respectively (Weis et al2006)
Pb and Hf isotopic compositions were analyzed by mul-tiple collector-inductively coupled plasma-mass spectrome-try (MC-ICP-MS) on a static multi-collection on a NuPlasma (Nu Instruments) system The detailed analyticalprocedure for Pb isotopic analyses on the Nu Plasma atthe PCIGR has been described by Weis et al (2006) Theconfiguration for Pb analyses allows for collection of PbTl and Hg togetherTl and Hg are used to monitor instru-mental mass discrimination and isobaric overlap repec-tively All sample solutions were analyzed withapproximately the same PbTl ratio (4) as the referencematerial NIST SRM 981 To accomplish this a small ali-quot of each sample solution from the Pb columns was ana-lyzed on the Element2 to determine the precise amount ofPb available for analysis on the Nu Plasma The SRM 981standard was run after every two samples on the NuPlasma During the time samples were run analyses ofthe SRM 981 Pb reference material gave values of206Pb204Pbfrac1416940322 207Pb204Pbfrac1415495823and 208Pb204Pbfrac14 36713164 (nfrac14 61 2 error is reportedas times 104) these values are in excellent agreement withreported TIMS triple-spike values of Galer amp Abouchami(1998) Fractionation-corrected Pb isotopic ratios were fur-ther corrected by the sample-standard bracketing methodor the ln^ln correction method described by White et al(2000) and Blichert-Toft et al (2003) Leached powders ofUSGS reference material BHVO-2 yielded Pb isotopicratios of 206Pb204Pbfrac14186454 8 207Pb204Pbfrac14 1549105 and 208Pb204Pbfrac14 38222514 and BCR-2yielded 206Pb204Pbfrac141880466 207Pb204Pbfrac14 1562518 and 208Pb204Pbfrac14 3883496 (2 error isreported as times 104) These values are in agreement withthose for leached residues of BHVO-2 and BCR-2 fromWeis et al (2006)Hf isotopic compositions were analyzed following the
procedures detailed byWeis et al (2007) The configurationfor Hf analyses monitored Lu mass 175 andYb mass 172 toallow for interference correction to masses 174 and 176 Hfisotopic ratios were normalized internally for mass fractio-nation to a 179Hf177Hf ratio of 07325 using an exponentialcorrection Standards were run after every two samplesand sample results were normalized to the ratio of the in-run daily average and a 176Hf177Hf ratio for JMC-475 of0282160 During the course of analyses the Hf standardJMC-475 gave an average value 02821679 (nfrac14130)USGS reference materials BCR-2 and BHVO-2 were pro-cessed with the samples and yielded Hf isotopic ratios of02828675 and 02831005 respectively Publishedvalues for BCR-2 and BHVO-2 are 02828717 and0283104 8 respectively (Weis et al 2007)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
39
by an assemblage of talc^tremolite^serpentine^opaqueoxides there is no fresh olivine present in the Karmutsenlavas onVancouver Island Plagioclase and clinopyroxenephenocrysts are absent in the high-MgO lavas Thecoarse-grained mafic rocks are characterized by sub-ophi-tic textures with an average grain size typically 41mmand are generally not glomeroporphyritic these rocks aremainly from the interiors of massive flows although somemay represent sillsSample preparation and analytical methods for major-
and trace-element and Sr^Nd^Hf^Pb isotopic composi-tions of the Karmutsen volcanic rocks are given in theAppendix
WHOLE -ROCK CHEMISTRYMajor- and trace-element compositionsThe most abundant type of volcanic rock in theKarmutsen Formation is tholeiitic basalt with a restrictedrange of major- and trace-element compositions The tho-leiitic basalts have similar compositions to the coarse-grained mafic rocks and both groups are distinct fromthe picrites and high-MgO basalts [Fig 6 picrites412wt MgO (LeBas 2000) high-MgO basalts8^12wt MgO] The tholeiitic basalts have lower MgO(57^77wt MgO) and higher TiO2 (14^22wt
TiO2) than the picrites (130^198wt MgO05^07wt TiO2) and high-MgO basalts (91^116wt MgO 05^08wt TiO2 Fig 6 Table 2) Almost allcompositions plot within the tholeiitic field in a total alka-lis vs silica plot although there has been substantial K lossin most samples from the Karmutsen Formation (seebelow) and the tholeiitic basalts generally have higherSiO2 Na2OthornK2O CaO and FeOT (total iron expressedas FeO) than the picrites The tholeiitic basalts also havenoticeably lower loss on ignition (LOI mean1713wt ) than the picrites (mean 54 08wt )and high-MgO basalts (mean 3815wt ) (Fig 6)reflecting the presence of abundant altered olivinephenocrysts in the latter groups Ni concentrations are sig-nificantly higher for the picrites (339^755 ppm) and high-MgO basalts (122^551ppm) than the tholeiitic basalts(58^125 ppm except for one anomalous sample) andcoarse-grained rocks (70^213 ppm) (Fig 6 Table 2) Ananomalous tholeiitic pillowed flow (two samples outlierin Tables 1 and 2) from near the picrite type locality atKeogh Lake and a mineralized sill (disseminated sulfide)from the basal sediment^sill complex at Schoen Lake aredistinguished by higher TiO2 and FeOT than the maingroup of tholeiitic basalts The tholeiitic basalts are similarin major-element composition to previously publishedresults for samples from the Karmutsen Formation how-ever the Keogh Lake picrites which encompass the
Table 1 Continued
Sample Area Flow Group Texture vol Ol Plag Cpx Ox Alteration Notes
4722A5 KR FLO OUTLIER intersertal 3 mottled very fg v small pl needles
5615A11 KR PIL OUTLIER intersertal 3 mottled very fg v small pl needles
5617A4 SL SIL MIN intersertal 5 3 fg
Sample number last digit of year month day initial sample station (except 93G171) MA Mount Arrowsmith SLSchoen Lake KR Karmutsen Range PIL pillow BRE breccia FLO flow SIL sill GAB gabbro THOL tholeiiticbasalt PIC picrite HI-MG high-MgO basalt CGcoarse-grained MIN mineralized sill glomero glomeroporphyriticOlivine modes for picrites are based on area calculations from scans (see lsquoOlivine accumulation in high-MgO and picriticlavasrsquo section and Fig 11) Visual alteration index is based primarily on degree of plagioclase alteration and presence ofsecondary minerals (1 least altered 3 most altered) Plagioclase phenocrysts are commonly altered to albite pumpellyiteand chlorite olivine is altered to talc tremolite and clinochlore (determined using the Rietveld method of X-ray powderdiffraction) clinopyroxene is unaltered FendashTi oxide is commonly replaced by sphene glcr glomerocrysts fg fine-grained cg coarse-grained oik oikocryst chad chadacryst enc enclosing swtl swallow-tail ol olivinepseudomorphs plag plagioclase cpx clinopyroxene ox oxides (includes ilmenite thorn titanomagnetite)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
9
00
05
10
15
20
25
30
35
40
0 5 10 15 20 25
10
11
12
13
14
15
16
17
18
19
20
0 5 10 15 20 25
6
7
8
9
10
11
12
13
14
15
0 5 10 15 20 25
6
7
8
9
10
11
12
13
14
15
16
0 5 10 15 20 25
0
1
2
3
4
5
6
40 45 50 55 60
0
200
400
600
800
0 5 10 15 20 25
(wt)(wt)
(wt)
(wt)
+K2ONa2O TiO2
FeO (T)
Ni (ppm)
Al2O3
CaO
alkalic
tholeiitic
Karmutsen Form
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
(compiled data)
Pillowed flowMineralized sill
SiO2 (wt) MgO (wt)
MgO (wt)MgO (wt)
MgO (wt) MgO (wt)
(a) (b)
(g)
(e)
(f)
(c)
LOI (wt )
(d)
MgO (wt )0
1
2
3
4
5
6
7
0 5 10 15 20 25
(wt)
Fig 6 Whole-rock major-element Ni and LOI variation diagrams for the Karmutsen Formation New samples from this study (Table 2) areshown by symbols with black outlines (see legend) Previously published analyses are shown by small gray symbols without black outlines Theboundary of the alkaline and tholeiitic fields is that of MacDonald amp Katsura (1964) Total iron expressed as FeO LOI loss on ignition oxidesare plotted on an anhydrous normalized basis References for the 322 compiled analyses for the Karmutsen Formation are Surdam (1967)Kuniyoshi (1972) Muller et al (1974) Barker et al (1989) Lassiter et al (1995) Massey (1995a1995b)Yorath et al (1999) and G Nixon (unpublisheddata) It should be noted that the compiled dataset has not been filtered many of the samples with high SiO2 (452wt ) are probably notKarmutsen flood basalts but younger Bonanza basalts and andesites Three pillowed flow samples and a mineralized sill are distinguishedseparately because of their anomalous chemistry Coarse-grained samples are indicated separately Three tholeiitic basalt samples (4724A54718A5 4718A6) with416wt Al2O3 and4270 ppm Sr contain 10^25 modal of large plagioclase phenocrysts (7^8mm) or glomerocrysts(44mm)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
10
Table 2 Major element (wt oxide) and trace element (ppm) abundances in whole-rock samples of Karmutsen basalts
1Ni concentrations for these samples were determined by XRFTHOL tholeiitic basalt PIC picrite HI-MG high MgO basalt CG coarse-grained (sill or gabbro) MIN SIL mineralizedsill OUTLIER anomalous pillowed flow in plots MA Mount Arrowsmith SL Schoen Lake KR Karmutsen Range QIQuadra Island Sample locations are given using the Universal Transverse Mercator (UTM) coordinate system (NAD83zones 9 and 10) Analyses were performed at Activation Laboratory (ActLabs) Fe2O3
is total iron expressed as Fe2O3LOI loss on ignition All major elements Sr V and Y were measured by ICP quadrupole OES on solutions of fusedsamples Cu Ni Pb and Zn were measured by total dilution ICP Cs Ga Ge Hf Nb Rb Ta Th U Zr and REE weremeasured by magnetic-sector ICP on solutions of fused samples Co Cr and Sc were measured by INAA Blanks arebelow detection limit See Electronic Appendices 2 and 3 for complete XRF data and PCIGR trace element datarespectively Major elements for sample 93G17 are normalized anhydrous Sample 5615A7 was analyzed by XRFSamples from Quadra Island are not shown in figures
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
17
picritic and high-MgO basalt pillow lavas extend to20wt MgO (Fig 6 and references listed in thecaption)The tholeiitic basalts from the Karmutsen Range dis-
play a tight range of parallel light rare earth element(LREE)-enriched REE patterns (LaYbCNfrac14 21^24mean 2202) whereas the picrites and high-MgObasalts are characterized by sub-parallel LREE-depleted
patterns (LaYbCNfrac14 03^07 mean 06 02) with lowerREE abundances than the tholeiitic basalts (Fig 7)Tholeiitic basalts from the Schoen Lake and MountArrowsmith areas have similar REE patterns tosamples from the Karmutsen Range (Schoen Lake LaYbCNfrac14 21^26 mean 2303 Mount Arrowsmith LaYbCNfrac1419^25 mean 2204) and the coarse-grainedmafic rocks have similar REE patterns to the tholeiitic
Depleted MORB average(Salters amp Stracke 2004)
Schoen LakeSchoen Lake
Mount Arrowsmith
Karmutsen RangeS
ampl
eC
hond
rite
Sam
ple
Cho
ndrit
e
Sam
ple
Prim
itive
Man
tleS
ampl
eP
rimiti
ve M
antle
Sam
ple
Prim
itive
Man
tle
Mount Arrowsmith
Karmutsen Range
(a) (b)
(c) (d)
(e) (f )
Sam
ple
Cho
ndrit
e
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained mafic rock
Pillowed flowMineralized sill
1
10
100
1
10
100
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
01
1
10
100
1
10
100
1
10
100
Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
Fig 7 Whole-rock REE and other incompatible-element concentrations for the Karmutsen Formation (a) (c) and (e) are chondrite-normal-ized REE patterns for samples from each of the three main field areas onVancouver Island (b) (d) and (f) are primitive mantle-normalizedtrace-element patterns for samples from each of the three main field areas All normalization values are from McDonough amp Sun (1995) Theclear distinction between LREE-enriched tholeiitic basalts and LREE-depleted picrites the effects of alteration on the LILE (especially loss ofK and Rb) and the different range for the vertical scale in (b) (extends down to 01) should be noted
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
18
basalts (LaYbCNfrac14 21^29 mean 2508) Two LREE-depleted coarse-grained mafic rocks from the SchoenLake area have similar REE patterns (LaYbCNfrac14 07) tothe picrites from the Keogh Lake area indicating that thisdistinctive suite of rocks is exposed as far south asSchoen Lake that is 75 km away (Fig 7) The anomalouspillowed flow from the Karmutsen Range has lower LaYbCN (13^18) than the main group of tholeiitic basaltsfrom the Karmutsen Range (Fig 7) The mineralized sillfrom the Schoen Lake area is distinctly LREE-enriched(LaYbCNfrac14 33) with the highest REE abundances(LaCNfrac14 940) of all Karmutsen samples this may reflectcontamination by adjacent sediment also evident in thin-section where sulfide blebs are cored by quartz (Greeneet al 2006)The primitive mantle-normalized trace-element pat-
terns (Fig 7) and trace-element variations and ratios(Fig 8) highlight the differences in trace-element
concentrations between the four main groups ofKarmutsen flood basalts onVancouver Island The tholeii-tic basalts from all three areas have relatively smooth par-allel trace-element patterns some samples have smallpositive Sr anomalies relative to Nd and Sm and manysamples have prominent negative K anomalies relative toU and Nb reflecting K loss during alteration (Fig 7) Thelarge ion lithophile elements (LILE) in the tholeiiticbasalts (mainly Rb and K not Cs) are depleted relativeto the high field strength elements (HFSE) and LREELILE segments especially for the Schoen Lake area(Fig 7d) are remarkably parallel The picrites andhigh-MgO basalts form a tight band of parallel trace-element patterns and are depleted in HFSE and LREEwith positive Sr anomalies and relatively low Rb Thecoarse-grained mafic rocks from the Karmutsen Rangehave identical trace-element patterns to the tholeiiticbasalts Samples from the Schoen Lake area have similar
000
025
050
075
100
125
0 1 2 3 4 5
00
04
08
12
16
0 5 10 15 20 25
MgO (wt)
0
5
10
15
0 50 100 150
Hf (ppm)
Th (ppm)
NbLaNb (ppm)
Zr (ppm)
(a)
(c)
(b)
(d)
Th (ppm)
00
02
04
06
08
10
12
00 01 02 03 04 05
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
NbLa=1
4723A4 5615A12
U (ppm)
Fig 8 Whole-rock trace-element concentrations and ratios for the Karmutsen Formation [except (b) which is vs wt MgO] (a) Nb vs Zr(b) NbLa vs MgO (c) Th vs Hf (d) Th vs U There is a clear distinction between the tholeiitic basalts and picrites in Nb and Zr both inconcentration and the slope of each trend
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
19
trace-element patterns to the Karmutsen Range except forthe two LREE-depleted coarse-grained mafic rocks andthe mineralized sill (Fig 7d)
Sr^Nd^Hf^Pb isotopic compositionsA total of 19 samples from the four main groups ofthe Karmutsen Formation were selected for radiogenicisotopic geochemistry on the basis of major- and trace-element variation stratigraphic position and geographicallocation The samples have overlapping age-correctedHf and Nd isotope ratios and distinct ranges of Sr isotopiccompositions (Fig 9) The tholeiitic basalts picriteshigh-MgO basalt and coarse-grained mafic rocks have
initial eHffrac14thorn87 to thorn126 and eNdfrac14thorn66 to thorn88 cor-rected for in situ radioactive decay since 230 Ma (Fig 9Tables 3 and 4) All the Karmutsen samples form a well-defined linear array in a Lu^Hf isochron diagram corre-sponding to an age of 24115 Ma (Fig 9) This is withinerror of the accepted age of the Karmutsen Formationand indicates that the Hf isotope systematics behaved as aclosed system since c 230 Ma The anomalous pillowedflow has similar initial eHf (thorn98) and slightly lower initialeNd (thorn62) compared with the other Karmutsen samplesThe picrites and high-MgO basalt have higher initial87Sr86Sr (070398^070518) than the tholeiitic basalts(initial 87Sr86Srfrac14 070306^070381) and the coarse-grained mafic rocks (initial 87Sr86Srfrac14 070265^070428)
05128
05129
05130
05131
05132
015 017 019 021 023 025
02829
02830
02831
02832
02833
02834
000 002 004 006 008 010
4
6
8
10
0702 0703 0704 0705 07066
8
10
12
14
5 6 7 8 9 10
12
0
1
2
3
4
5
6
7
0702 0703 0704 0705 0706
(d)
LOI (wt )
87Sr 86Sr
4723A2
Nd
143Nd144Nd
147Sm144Nd
87Sr 86Sr (230 Ma)
(230 Ma)
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
(a)
(c)
176Hf177Hf
Age=241 plusmn 15 Ma=+990 plusmn 064
176Lu177Hf
Hf (230 Ma)
Nd (230 Ma)
(e)
Hf (i)
(b)
Fig 9 Whole-rock Sr Nd and Hf isotopic compositions for the Karmutsen Formation (a) 143Nd144Nd vs 147Sm144Nd (b) 176Hf177Hf vs176Lu177Hf The slope of the best-fit line for all samples corresponds to an age of 24115 Ma (c) Initial eNd vs 87Sr86Sr Age-correction to230 Ma (d) LOI vs 87Sr86Sr (e) Initial eHf vs eNd Average 2 error bars are shown in a corner of each panel Complete chemical duplicatesshown inTables 3 and 4 (samples 4720A7 and 4722A4) are circled in each plot
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
overlap the ranges of the picrites and tholeiitic basalts(Fig 9 Table 3)The measured Pb isotopic compositions of the tholeiitic
basalts are more radiogenic than those of the picrites andthe most magnesian Keogh Lake picrites have the leastradiogenic Pb isotopic compositions The range ofinitial Pb isotope ratios for the picrites is206Pb204Pbfrac1417868^18580 207Pb204Pbfrac1415547^15562and 208Pb204Pbfrac14 37873^38257 and the range for thetholeiitic basalts is 206Pb204Pbfrac1418782^19098207Pb204Pbfrac1415570^15584 and 208Pb204Pbfrac14 37312^38587 (Fig 10 Table 5) The coarse-grained mafic rocksoverlap the range of initial Pb isotopic compositions forthe tholeiitic basalts with 206Pb204Pbfrac1418652^19155207Pb204Pbfrac1415568^15588 and 208Pb204Pbfrac14 38153^38735 One high-MgO basalt has the highest initial206Pb204Pb (19242) and 207Pb204Pb (15590) and thelowest initial 208Pb204Pb (37676) The Pb isotopic ratiosfor Karmutsen samples define broadly linearrelationships in Pb isotope plots The age-corrected Pb
isotopic compositions of several picrites have been affectedby U and Pb (andor Th) mobility during alteration(Fig 10)
ALTERAT IONThe Karmutsen basalts have retained most of their origi-nal igneous structures and textures however secondaryalteration and low-grade metamorphism have producedzeolitic- and prehnite^pumpellyite-bearing mineral assem-blages (Table 1 Cho et al 1987) Alteration and metamor-phism have primarily affected the distribution of the LILEand the Sr isotopic systematics The Keogh Lake picriteshave been affected more by alteration than the tholeiiticbasalts and have higher LOI (up to 55wt ) variableLILE and higher measured Sr Nd and Hf and lowermeasured Pb isotope ratios than the tholeiitic basalts Srand Pb isotope compositions for picrites and tholeiiticbasalts show a relationship with LOI (Figs 9 and 10)whereas Nd and Hf isotopic compositions do not correlate
Table 3 Sr and Nd isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Rb Sr 87Sr86Sr 2m87Rb86Sr 87Sr86Srt Sm Nd 143Nd144Nd 2m eNd
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses carried out at the PCIGR thecomplete trace element analyses are given in Electronic Appendix 3
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
21
with LOI (not shown) The relatively high apparent initialSr isotopic compositions for the high-MgO lavas (up to07052) probably resulted from post-magmatic loss of Rbor an increase in 87Sr86Sr through addition of seawater Sr(eg Hauff et al 2003) High-MgO lavas from theCaribbean plateau have similarly high 87Sr86Sr comparedwith basalts (Recurren villon et al 2002)The correction for in situdecay on initial Pb isotopic ratios has been affected bymobilization of U and Th (and Pb) in the whole-rockssince their formation A thorough acid leaching duringsample preparation was used that has been shown to effec-tively remove alteration phases (Weis et al 2006 NobreSilva et al in preparation) Whole-rock SmNd ratiosand age-correction of Nd isotopes may be affected bythe leaching procedure for submarine rocks older than50 Ma (Thompson et al 2008 Fig 9) there ishowever very little variation in Nd isotopic compositionsof the picrites or tholeiitic basalts and this system does notappear to be disturbed by alteration The HFSE abun-dances for tholeiitic and picritic basalts exhibit clear
linear correlations in binary diagrams (Fig 8) whereasplots of LILE vs HFSE are highly scattered (not shown)because of the mobility of some of the alkali and alkalineearth LILE (Cs Rb Ba K) and Sr for most samplesduring alterationThe degree of alteration in the Karmutsen samples does
not appear to be related to eruption environment or depthof burial Pillow rims are compositionally different frompillow cores (Surdam1967 Kuniyoshi1972) and aquagenetuffs are chemically different from isolated pillows withinpillow breccias however submarine and subaerial basaltsexhibit a similar degree of alterationThere is no clear cor-relation between the submarine and subaerial basalts andsome commonly used chemical alteration indices (eg BaRb vs K2OP2O5 Huang amp Frey 2005) There is also nodefinitive correlation between the petrographic alterationindex (Table 1) and chemical alteration indices although10 of 13 tholeiitic basalts with the highest K and LILEabundances have the highest petrographic alterationindex of three (seeTable 1)
Table 4 Hf isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Lu Hf 176Hf177Hf 2m eHf176Lu177Hf 176Hf177Hft eHf(t)
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
OLIV INE ACCUMULAT ION INH IGH-MgO AND PICR IT IC LAVASGeochemical trends and petrographic characteristics indi-cate that accumulation of olivine played an important rolein the formation of the high-MgO basalts and picrites fromthe Keogh Lake area on northern Vancouver Island TheKeogh Lake samples show a strong linear correlation inplots of Al2O3 TiO2 ScYb and Ni vs MgO and many ofthe picrites have abundant clusters of olivine pseudomorphs(Table1 Fig11)There is a strong linear correlationbetween
modal per cent olivine and whole-rock magnesium contentmagnesium contents range between 10 and 19wt MgOand the proportion of olivine phenocrysts in these samplesvaries from 0 to 42 vol (Fig 11)With a few exceptionsboth the size range and median size of the olivine pheno-crysts in most samples are comparable The clear correla-tion between the proportion of olivine phenocrysts andwhole-rock magnesium contents indicates that accumula-tion of olivine was directly responsible for the high MgOcontents (410wt ) in most of the picritic lavas
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
207Pb204Pb
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
208Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
(a) (b)
(c) (e)
4723A24723A2
4723A2
4723A2
4722A4
4723A4
4722A4
4723A4
4723A134723A3
4723A3
4723A13
1556
1558
1560
1562
1564
1566
1568
186 190 194 198 202 206 2101550
1552
1554
1556
1558
1560
178 180 182 184 186 188 190 192 194
380
384
388
392
396
186 190 194 198 202 206 210374
378
382
386
390
178 180 182 184 186 188 190 192 194
206Pb204Pb(d)
LOI (wt )
0
1
2
3
4
5
6
7
186 190 194 198 202 206 210
4723A2
Fig 10 Pb isotopic compositions of leached whole-rock samples measured by MC-ICP-MS for the Karmutsen Formation Error bars are smal-ler than symbols (a) Measured 207Pb204Pb vs 206Pb204Pb (b) Initial 207Pb204Pb vs 206Pb204Pb Age-correction to 230 Ma (c) Measured208Pb204Pb vs 206Pb204Pb (d) LOI vs 206Pb204Pb (e) Initial 208Pb204Pb vs 206Pb204Pb Complete chemical duplicates shown in Table 5(samples 4720A7 and 4722A4) are circled in each plot The dashed lines in (b) and (e) show the differences in age-corrections for two picrites(4723A4 4722A4) using the measured UTh and Pb concentrations for each sample and the age-corrections when concentrations are used fromthe two picrites (4723A3 4723A13) that appear to be least affected by alteration
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
23
DISCUSS IONThe compositional and spatial development of the eruptedlava sequences of the Wrangellia oceanic plateau onVancouver Island provide information about the evolutionof a mantle plume upwelling beneath oceanic lithospherein an analogous way to the interpretation of continentalflood basalts The Caribbean and Ontong Java plateausare the two primary examples of accreted parts of oceanicplateaus that have been studied to date and comparisonscan be made with the Wrangellia oceanic plateauHowever the Wrangellia oceanic plateau is a distinctiveoccurrence of an oceanic plateau because it formed on thecrust of a Devonian to Mississippian intra-oceanic islandarc overlain by Mississippian to Permian limestones andpelagic sediments The nature and thickness of oceaniclithospheric mantle are considered to play a fundamentalrole in the decompression and evolution of a mantleplume and thus the compositions of the erupted lavasequences (Greene et al 2008) The geochemical and
stratigraphic relationships observed in the KarmutsenFormation onVancouver Island and the focus of the subse-quent discussion sections provide constraints on (1) thetemperature and degree of melting required to generatethe primary picritic magmas (2) the composition of themantle source (3) the depth of melting and residual min-eralogy in the source region (4) the low-pressure evolutionof the Karmutsen tholeiitic basalts (5) the growth of theWrangellia oceanic plateau
Melting conditions and major-elementcomposition of the primary magmasThe primary magmas to most flood basalt provinces arebelieved to be picritic (eg Cox 1980) and they have beenfound in many continental and oceanic flood basaltsequences worldwide (eg Siberia Karoo Paranacurren ^Etendeka Caribbean Deccan Saunders 2005) Near-pri-mary picritic lavas with low total alkali abundances
Table 5 Pb isotopic compositions of Karmutsen basaltsVancouver Island BC
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
require high-degree partial melting of the mantle source(eg Herzberg amp OrsquoHara 2002) The Keogh Lake picritesfrom northernVancouver Island are the best candidates foridentifying the least modified partial melts of the mantleplume source despite having accumulated olivine pheno-crysts and can be used to estimate conditions of meltingand the composition of the primary magmas for theKarmutsen FormationPrimary magma compositions mantle potential tem-
peratures and source melt fractions for the picritic lavas ofthe Karmutsen Formation were calculated from primitivewhole-rock compositions using PRIMELT1XLS software(Herzberg et al 2007) and are presented in Fig 12 andTable 6 A detailed discussion of the computationalmethod has been given by Herzberg amp OrsquoHara (2002)and Herzberg et al (2007) key features of the approachare outlined belowPrimary magma composition is calculated for a primi-
tive lava by incremental addition of olivine and compari-son with primary melts of mantle peridotite Lavas thathad experienced plagioclase andor clinopyroxene fractio-nation are excluded from this analysis All calculated pri-mary magma compositions are assumed to be derived byaccumulated fractional melting of fertile peridotite KR-4003 Uncertainties in fertile peridotite composition donot propagate to significant variations in melt fractionand mantle potential temperature melting of depletedperidotite propagates to calculated melt fractions that aretoo high but with a negligible error in mantle potential
temperature PRIMELT1XLS calculates the olivine liqui-dus temperature at 1atm which should be similar to theactual eruption temperature of the primary magmaand is accurate to 308C at the 2 level of confidenceMantle potential temperature is model dependent witha precision that is similar to the accuracy of the olivineliquidus temperature (C Herzberg personal communica-tion 2008)With regard to the evidence for accumulated olivine in
the Keogh Lake picrites it should not matter whether themodeled lava composition is aphyric or laden with accu-mulated olivine phenocrysts PRIMELT1 computes theprimary magma composition by adding olivine to a lavathat has experienced olivine fractionation in this way itinverts the effects of fractional crystallization The onlyrequirement is that the olivine phenocrysts are not acci-dental or exotic to the lava compositions Support for thisprocedure can be seen in the calculated olivine liquid-line-of-descent shown by the curved arrays with 5 olivineaddition increments (Fig 12c) Fractional crystallization ofolivine from model primary magmas having 85^95wt FeO and 153^175wt MgO will yield arrays of deriva-tive liquids and randomly sorted olivines that in most butnot all cases connect the picrites and high-MgO basalts Anewer version of the PRIMELT software (Herzberg ampAsimow 2008) can perform olivine addition to and sub-traction from lavas with highly variably olivine phenocrystcontents and yields results that converge with those shownin Fig 12
Fig 11 Relationship between abundance of olivine phenocrysts and whole-rock MgO contents for the Keogh Lake picrites (a) Tracings ofolivine grains (gray) in six high-MgO samples made from high resolution (2400 dpi) scans of petrographic thin-sections Scale bars represent10mm sample numbers are indicated at the upper left (b) Modal olivine vs whole-rock MgO Continuous line is the best-fit line for 10 high-MgO samplesThe areal proportion of olivine was calculated from raster images of olivine grains using ImageJ image analysis software whichprovides an acceptable estimation of the modal abundance (eg Chayes 1954)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
25
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
Gray lines = final melting pressure
L + Ol + Opx
07
08
09
Pressure (GPa)
10
3
4
5
6
1
SOLIDUS
01
04
0506
L + Ol
2
03
02
PeridotiteKR-4003
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
Thick black lines = initial melting pressure
Dashed lines = melt fraction
Melt Fraction
2
3 4 5 6
L+Ol+Opx
SolidusSpinel
SolidusGarnet Peridotite
7
L+Ol
0 10 20 30 4040
45
50
55
60
MgO (wt)
SiO2
(wt)
PeridotiteKR-4003
Melt Fraction
0 10 20 305
10
15
CaO(wt)
MgO (wt)
3
4 5 6 7
2
46
2
Peridotite Partial Melts
Thick black line = initial melting pressureGray lines = final melting pressure
Peridotite
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
02
Pressure (GPa)
0302
04
Pressure (GPa)
Picrite
High-MgO basalt
Tholeiitic basalt
(c)
(d)
Karmutsen flood basaltsOlivine addition modelOlivine addition (5 increment)Primary magma
PicriteHigh-MgO basaltTholeiitic basalt
(a)
(b)
800 850 900
910
920
930
940
950
Karmutsen flood basalts
Olivine addition modelOlivine addition (5 increment)Primary magma
Gray lines = Fo contents of olivine
Fig 12 Estimated primary magma compositions for three Keogh Lake picrites and high-MgO basalts (samples 4723A4 4723A135616A7) using the forward and inverse modeling technique of Herzberg et al (2007) Karmutsen compositions and modeling results areoverlain on diagrams provided by C Herzberg (a) Whole-rock FeO vs MgO for Karmutsen samples from this study Total iron estimatedto be FeO is 090 Gray lines show olivine compositions that would precipitate from liquid of a given MgO^FeO composition Black lineswith crosses show results of olivine addition (inverse model) using PRIMELT1 (Herzberg et al 2007) Gray field highlights olivinecompositions with Fo contents of 91^92 predicted to precipitate (b) FeO vs MgO showing Karmutsen lava compositions and results ofa forward model for accumulated fractional melting of fertile peridotite KR-4003 (Kettle River Peridotite Walter 1998) (c) SiO2 vsMgO for Karmtusen lavas and model results (d) CaO vs MgO showing Karmtusen lava compositions and model results To brieflysummarize the technique [see Herzberg et al (2007) for complete description] potential parental magma compositions for the high-MgOlava series were selected (highest MgO and appropriate CaO) and using PRIMELT1 software olivine was incrementally added tothe selected compositions to show an array of potential primary magma compositions (inverse model) Then using PRIMELT1 theresults from the inverse model were compared with a range of accumulated fractional melts for fertile peridotite derived fromparameterization of the experimental results for KR-4003 from Walter (1998) forward model Herzberg amp OrsquoHara (2002) A melt fractionwas sought that was unique to both the inverse and forward models (Herzberg et al 2007) A unique solution was found when there was a com-mon melt fraction for both models in FeO^MgO and CaO^MgO^Al2O3^SiO2 (CMAS) projection space This modeling assumes thatolivine was the only phase crystallizing and ignores chromite precipitation and possible augite fractionation in the mantle (Herzbergamp OrsquoHara 2002) Results are best for a residue of spinel lherzolite (not pyroxenite) The presence of accumulated olivine in samples ofthe Keogh Lake picrites used as starting compositions does not significantly affect the results because we are modeling addition of olivine(see text for further description) The tholeiitic basalts cannot be used for modeling because they are all saturated in plagthorn cpxthornolThick black line for initial melting pressure at 4 GPa is not shown in (c) for clarity Gray area in (a) indicates parental magmas thatwill crystallize olivine with Fo 91^92 at the surface Dashed lines in (c) indicate melt fraction Solidi for spinel and garnet peridotite arelabelled in (c)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
The PRIMELT1 results indicate that if the Karmutsenpicritic lavas were derived from accumulated fractionalmelting of fertile peridotite the primary magmas wouldhave contained 15^17wt MgO and 10wt CaO(Fig 12 Table 6) formed from 23^27 partial meltingand would have crystallized olivine with a Fo content of 91 (Fig 12 Table 6) Calculated mantle temperatures forthe Karmutsen picrites (14908C) indicate melting ofanomalously hot mantle (100^2008C above ambient) com-pared with ambient mantle that produces mid-ocean ridgebasalt (MORB 1280^14008C Herzberg et al 2007)which initiated between 3 and 4 GPa Several picritesappear to be near-primary melts and required very littleaddition of olivine (eg sample 93G171 with measured158wt MgO) These temperature and melting esti-mates of the Karmutsen picrites are consistent withdecompression of hot mantle peridotite in an actively con-vecting plume head (ie plume initiation model) Threesamples (picrite 5614A1 and high-MgO basalts 5614A3and 5614A5) are lower in iron than the other Karmutsenlavas with FeO in the 65^80wt range (Fig 12c) andhave MgO and FeO contents that are similar to primitiveMORB from the Sequeiros fracture zone of the EastPacific Rise (EPR Perfit et al 1996) These samples
however differ from EPR MORB in having higher SiO2
and lower Al2O3 and they are restricted geographicallyto one area (Sara Lake Fig 2)We therefore have evidence for primary magmas with
MgO contents that range from a possible low of 110wt to as high as 174wt with inferred mantle potential tem-peratures from 1340 to 15208C This is similar to the rangedisplayed by lavas from the Ontong Java Plateau deter-mined by Herzberg (2004) who found that primarymagma compositions assuming a peridotite sourcewould contain 17wt MgO and 10wt CaO(Table 6) and that these magmas would have formed by27 melting at a mantle potential temperature of15008C (first melting occurring at 36 GPa) Fitton ampGodard (2004) estimated 30 partial melting for theOntong Java lavas based on the Zr contents of primarymagmas of Kroenke-type basalt calculated by incrementaladdition of equilibrium olivine However if a considerableamount of eclogite was involved in the formation of theOntong Java plateau magmas the excess temperatures esti-mated by Herzberg et al (2007) may not be required(Korenaga 2005) The variability in mantle potential tem-perature for the Keogh Lake picrites is also similar to thewide range of temperatures that have been calculated for
Table 6 Estimated primary magma compositions for Karmutsen basalts and other oceanic plateaus or islands
Sample 4723A4 4723A13 5616A7 93G171 Average OJP Mauna Keay Gorgonaz
Ontong Java primary magma composition for accumulated fraction melting (AFM) from Herzberg (2004)yMauna Kea primary magma composition is average of four samples of Herzberg (2006 table 1)zGorgona primary magma composition for AFM for 1F fertile source from Herzberg amp OrsquoHara (2002 table 4)Total iron estimated to be FeO is 090
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
27
lavas from single volcanoes in the Galapagos Hawaii andelsewhere by Herzberg amp Asimow (2008) Those workersinterpreted the range to reflect the transport of magmasfrom both the cool periphery and the hotter axis of aplume A thermally heterogeneous plume will yield pri-mary magmas with different MgO contents and the rangeof primary magma compositions and their inferred mantlepotential temperatures for Karmutsen lavas may reflectmelting from different parts of the plume
Source of the Karmutsen Formation lavasThe Sr^Nd^Hf^Pb isotopic compositions of theKarmutsen volcanic rocks on Vancouver Island indicatethat they formed from plume-type mantle containing adepleted component that is compositionally similar to butdistinct from other oceanic plateaus that formed in thePacific Ocean (eg Ontong Java and Caribbean Fig 13)Trace element and isotopic constraints can be used to testwhether the depleted component of the KarmutsenFormation was intrinsic to the mantle plume and distinctfrom the source of MORB or the result of mixing or invol-vement of ambient depleted MORB mantle with plume-type mantle The Karmutsen tholeiitic basalts have initialeNd and
87Sr86Sr ratios that fall within the range of basaltsfrom the Caribbean Plateau (eg Kerr et al 1996 1997Hauff et al 2000 Kerr 2003) and a range of Hawaiian iso-tope data (eg Frey et al 2005) and lie within and abovethe range for the Ontong Java Plateau (Tejada et al 2004Fig 13a) Karmutsen tholeiitic basalts with the highest ini-tial eNd and lowest initial 87Sr86Sr lie within and justbelow the field for age-corrected northern EPR MORB at230 Ma (eg Niu et al1999 Regelous et al1999) Initial Hfand Nd isotopic compositions for the Karmutsen samplesare noticeably offset to the low side of the ocean islandbasalt (OIB) array (Vervoort et al 1999) and lie belowand slightly overlap the low eHf end of fields for Hawaiiand the Caribbean Plateau (Fig 13c) the Karmutsen sam-ples mostly fall between and slightly overlap a field forage-corrected Pacific MORB at 230 Ma and Ontong Java(Mahoney et al 1992 1994 Nowell et al 1998 Chauvel ampBlichert-Toft 2001) Karmutsen lava Pb isotope ratios over-lap the broad range for the Caribbean Plateau and thelinear trends for the Karmutsen samples intersect thefields for EPR MORB Hawaii and Ontong Java in208Pb^206Pb space but do not intersect these fields in207Pb^206Pb space (Fig 13d and e) The more radiogenic207Pb204Pb for a given 206Pb204Pb of the Karmutsenbasalts requires high UPb ratios at an ancient time when235U was abundant similar to the Caribbean Plateau andenriched in comparison with EPR MORB Hawaii andOntong JavaThe low eHf^low eNd component of the Karmutsen
basalts that places them below the OIB mantle array andpartly within the field for age-corrected Pacific MORBcan form from low ratios of LuHf and high SmNd over
time (eg Salters amp White 1998) MORB will develop aHf^Nd isotopic signature over time that falls below themantle array Low LuHf can also lead to low 176Hf177Hffrom remelting of residues produced by melting in theabsence of garnet (eg Salters amp Hart 1991 Salters ampWhite 1998) The Hf^Nd isotopic compositions of theKarmutsen Formation indicate the possible involvementof depleted MORB mantle however similar to Iceland(Fitton et al 2003) Nb^Zr^Y variations provide a way todistinguish between a depleted component within theplume and a MORB-like component The Karmutsenbasalts and picrites fall within the Iceland array and donot overlap the MORB field in a logarithmic plot of NbYvs ZrY (Fig 13b) using the fields established by Fittonet al (1997) Similar to the depleted component within theCaribbean Plateau (Thompson et al 2003) the trend ofHf^Nd isotopic compositions towards Pacific MORB-likecompositions is not followed by MORB values in Nb^Zr^Y systematics and this suggests that the depleted compo-nent in the source of the Karmutsen basalts is distinctfrom the source of MORB and probably an intrinsic partof the plume The relatively radiogenic Pb isotopic ratiosand REE patterns distinct from MORB also support thisinterpretationTrace-element characteristics suggest the possible invol-
vement of sub-arc lithospheric mantle in the generation ofthe Karmutsen picrites Lassiter et al (1995) suggested thatmixing of a plume-type source with eNdthorn 6 to thorn 7 witharc material with low NbTh could reproduce the varia-tions observed in the Karmutsen basalts however theyconcluded that the absence of low NbLa ratios in theirsample of tholeiitic basalts restricts the amount of arc litho-sphere involved The study of Lassiter et al (1995) wasbased on major and trace element and Sr Nd and Pb iso-topic data for a suite of 29 samples from Buttle Lake in theStrathcona Provincial Park on central Vancouver Island(Fig 1) Whereas there are no clear HFSE depletions inthe Karmutsen tholeiitic basalts the low NbLa (and TaLa) of the Karmutsen picritic lavas in this study indicatethe possible involvement of Paleozoic arc lithosphere onVancouver Island (Fig 8)
REE modeling dynamic melting andsource mineralogyThe combined isotopic and trace-element variations of thepicritic and tholeiitic Karmutsen lavas indicate that differ-ences in trace elements may have originated from differentmelting histories For example the LuHf ratios of theKarmutsen picritic lavas (mean 008) are considerablyhigher than those in the tholeiitic basalts (mean 002)however the small range of overlapping initial eHf valuesfor the two lava suites suggests that these differences devel-oped during melting within the plume (Fig 9b) Below wemodel the effects of source mineralogy and depth of
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
28
melting on differences in trace-element concentrations inthe tholeiitic basalts and picritesDynamic melting models simulate progressive decom-
pression melting where some of the melt fraction isretained in the residue and only when the degree of partialmelting is greater than the critical mass porosity and the
source becomes permeable is excess melt extracted fromthe residue (Zou1998) For the Karmutsen lavas evolutionof trace element concentrations was simulated using theincongruent dynamic melting model developed by Zou ampReid (2001) Melting of the mantle at pressures greater than1 GPa involves incongruent melting (eg Longhi 2002)
OIB array
PacificMORB(230 Ma)
(0 Ma)
EPR(230 Ma)
-4
-2
0
2
4
8
10
12
14
16
18
20
22
-4 -2 0 2 4 6 8 10 12 14
minus2
0
2
4
6
8
10
12
14
0702 0703 0704 0705 0706
IndianMORB
87Sr 86Sr (initial)
Hf (initial)
(a) (c)
Nd (initial)
Karmutsen tholeiitic basalt
HawaiiOntong JavaCaribbean Plateau
Karmutsen high-MgO lava
high-MgO lavasKarmutsen
1540
1545
1550
1555
1560
1565
175 180 185 190 195 200375
380
385
390
395
175 180 185 190 195 200
(d) (e)
EPR
EPR
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb
Karmutsen Formation (initial)
HawaiiOntong JavaCaribbean Plateau
East Pacific Rise
Karmutsen Formation (measured)
Juan de FucaGorda
ε Nd
(initial)
Karmutsen Formation (different age-correction for 3 picrites)
(0 Ma)
NbY
001
01
1
1 10
ZrY
OIB
NMORB
(b)
Iceland array
6
Fig 13 Comparison of age-corrected (230 Ma) Sr^Nd^Hf^Pb isotopic compositions for Karmutsen flood basalts onVancouver Island withage-corrected OIB and MORB and Nb^Zr^Y variation (a) Initial eNd vs 87Sr86Sr (b) NbYand ZrY variation (after Fitton et al 1997)(c) Initial eHf vs eNd (d) Measured and initial 207Pb204Pb vs 206Pb204Pb (e) Measured and initial 208Pb204Pb vs 206Pb204Pb References fordata sources are listed in Electronic Appendix 4 Most of the compiled data were extracted from the GEOROC database (httpgeorocmpch-mainzgwdgdegeoroc) OIB array line in (c) is fromVervoort et al (1999) EPR East Pacific Rise Several high-MgO samples affected bysecondary alteration were not plotted for clarity in Pb isotope plots Estimation of fields for age-corrected Pacific MORB and EPR at 230 Mawere made using parent^daughter concentrations (in ppm) from Salters amp Stracke (2004) of Lufrac14 0063 Hffrac14 0202 Smfrac14 027 Ndfrac14 071Rbfrac14 0086 and Srfrac14 836 In the NbYvs ZrYplot the normal (N)-MORB and OIB fields not including Icelandic OIB are from data com-pilations of Fitton et al (1997 2003) The OIB field is data from 780 basalts (MgO45wt ) from major ocean islands given by Fitton et al(2003) The N-MORB field is based on data from the Reykjanes Ridge EPR and Southwest Indian Ridge from Fitton et al (1997) The twoparallel gray lines in (b) (Iceland array) are the limits of data for Icelandic rocks
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
29
with olivine being produced during the reactioncpxthornopxthorn spfrac14meltthornol for spinel peridotites The mod-eling used coefficients for incongruent melting reactionsbased on experiments on lherzolite melting (eg Kinzleramp Grove 1992 Longhi 2002) and was separated into (1)melting of spinel lherzolite (2) two combined intervals ofmelting for a garnet lherzolite (gt lherz) and spinel lher-zolite (sp lherz) source (3) melting of garnet lherzoliteThe model solutions are non-unique and a range indegree of melting partition coefficients melt reactioncoefficients and proportion of mixed source componentsis possible to achieve acceptable solutions (see Fig 14 cap-tion for description of modeling parameters anduncertainties)The LREE-depleted high-MgO lavas require melting of
a depleted spinel lherzolite source (Fig 14a) The modelingresults indicate a high degree of melting (22^25) similarto the results from PRIMELT1 based on major-elementvariations (discussed above) of a LREE-depleted sourceThe high-MgO lavas lack a residual garnet signature Theenriched tholeiitic basalts involved melting of both garnetand spinel lherzolite and represent aggregate melts pro-duced from continuous melting throughout the depth ofthe melting column (Fig 14b) Trace-element concentra-tions in the tholeiitic and picritic lavas are not consistentwith low-degree melting of garnet lherzolite alone(Fig 14c) The modeling results indicate that the aggregatemelts that formed the tholeiitic basalts would haveinvolved lesser proportions of enriched small-degreemelts (1^6 melting) generated at depth and greateramounts of depleted high-degree melts (12^25) gener-ated at lower pressure (a ratio of garnet lherzolite tospinel lherzolite melt of 14 provides the best fits seeFig 14b and description in figure caption) The totaldegree of melting for the tholeiitic basalts was high(23^27) similar to the degree of partial melting in thespinel lherzolite facies for the primary melts that eruptedas the Keogh Lake picrites of a source that was also prob-ably depleted in LREE
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
(c)
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
Melting of garnet lherzolite
High-MgO lavas
Melting of spinel lherzolite
source (07DM+03PM)
source (07DM+03PM)
6 melting
30 melting
(b)
(a)
30 melting
Sam
ple
Cho
ndrit
eS
ampl
eC
hond
rite 20 melting
1 melting
Melting of garnet lherzoliteand spinel lherzolite
x gt lherz + 4x sp lherz
5 meltingx gt lherz + 4x sp lherz
Sam
ple
Cho
ndrit
e
Tholeiitic lavas
Tholeiitic lavas
High-MgO lavas
source (07DM+03PM)
Fig 14 Trace-element modeling results for incongruent dynamicmantle melting for picritic and tholeiitic lavas from the KarmutsenFormation Three steps of modeling are shown (a) melting of spinellherzolite compared with high-MgO lavas (b) melting of garnet lher-zolite and spinel lherzolite compared with tholeiitic lavas (c) meltingof garnet lherzolite compared with tholeiitic and picritic lavasShaded fields in each panel for tholeiitic basalts and picrites arelabeled Patterns with symbols in each panel are modeling results(patterns are 5 melting increments in (a) and (b) and 1 incre-ments in (c) PM primitive mantle DM depleted MORB mantle gtlherz garnet lherzolite sp lherz spinel lherzolite In (b) the ratios ofper cent melting for garnet and spinel lherzolite are indicated (x gtlherzthorn 4x sp lherz means that for 5 melting 1 melt in garnetfacies and 4 melt in spinel facies where x is per cent melting in theinterval of modeling) The ratio of the respective melt proportions inthe aggregate melt (x gt lherzthorn 4x sp lherz) was kept constant andthis ratio of melt contribution provided the best fit to the data Arange of proportion of source components (07DMthorn 03PM to
09DMthorn 01PM) can reproduce the variation in the high-MgOlavas Melting modeling uses the formulation of Zou amp Reid (2001)an example calculation is shown in their Appendix PM fromMcDonough amp Sun (1995) and DM from Salters amp Stracke (2004)Melt reaction coefficients of spinel lherzolite from Kinzler amp Grove(1992) and garnet lherzolite fromWalter (1998) Partition coefficientsfrom Salters amp Stracke (2004) and Shaw (2000) were kept constantSource mineralogy for spinel lherzolite (018cpx027opx052ol003sp) from Kinzler (1997) and for garnet lherzolite(034cpx008opx053ol005gt) from Salters amp Stracke (2004) Arange of source compositions for melting of spinel lherzolite(07DMthorn 03PM to 03DMthorn 07PM) reproduce REE patternssimilar to those of the high-MgO lavas with best fits for 22ndash25melting and 07DMthorn 03PM source composition Concentrationsof tholeiitic basalts cannot be reproduced from an entirelyprimitive or depleted source
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
30
The uncertainty in the composition of the source in thespinel lherzolite melting model for the high-MgO lavasprecludes determining whether the source was moredepleted in incompatible trace elements than the source ofthe tholeiitic lavas (see Fig 14 caption for description) It ispossible that early high-pressure low-degree melting left aresidue within parts of the plume (possibly the hotter inte-rior portion) that was depleted in trace elements (egElliott et al 1991) further decompression and high-degreemelting of such depleted regions could then have generatedthe depleted high-MgO Karmutsen lavas Shallow high-degree melting would preferentially sample depletedregions whereas deeper low-degree melting may notsample more refractory depleted regions (eg CaribbeanEscuder-Viruete et al 2007) The picrites lie at the highend of the range of Nd isotopic compositions and havelow NbZr similar to depleted Icelandic basalts (Fittonet al 2003) therefore the picrites may represent the partsof the mantle plume that were the most depleted prior tothe initiation of melting As Fitton et al (2003) suggestedthese depleted melts may only rarely be erupted withoutcombining with more voluminous less depleted melts andthey may be detected only as undifferentiated small-volume flows like the Keogh Lake picrites within theKarmutsen Formation on Vancouver IslandDecompression melting within the mantle plume initiatedwithin the garnet stability field (35^25 GPa) at highmantle potential temperatures (Tp414508C) and pro-ceeded beneath oceanic arc lithosphere within the spinelstability field (25^09 GPa) where more extensivedegrees of melting could occur
Magmatic evolution of Karmutsentholeiitic basaltsPrimary magmas in large igneous provinces leave exten-sive coarse-grained cumulate residues within or below thecrust these mafic and ultramafic plutonic sequences repre-sent a significant proportion of the original magma thatpartially crystallized at depth (eg Farnetani et al 1996)For example seismic and petrological studies of OntongJava (eg Farnetani et al 1996) combined with the use ofMELTS (Ghiorso amp Sack 1995) indicate that the volcanicsequence may represent only 40 of the total magmareaching the Moho Neal et al (1997) estimated that30^45 fractional crystallization took place for theOntong Java magmas and produced cumulate thicknessesof 7^14 km corresponding to 11^22 km of flood basaltsdepending on the presence of pre-existing oceanic crustand the total crustal thickness (25^35 km) Interpretationsof seismic velocity measurements suggest the presence ofpyroxene and gabbroic cumulates 9^16 km thick beneaththe Ontong Java Plateau (eg Hussong et al 1979)Wrangellia is characterized by high crustal velocities in
seismic refraction lines which is indicative of mafic
plutonic rocks extending to depth beneath VancouverIsland (Clowes et al 1995)Wrangellia crust is 25^30 kmthick beneath Vancouver Island and is underlain by astrongly reflective zone of high velocity and density thathas been interpreted as a major shear zone where lowerWrangellia lithosphere was detached (Clowes et al 1995)The vast majority of the Karmutsen flows are evolved tho-leiitic lavas (eg low MgO high FeOMgO) indicating animportant role for partial crystallization of melts at lowpressure and a significant portion of the crust beneathVancouver Island has seismic properties that are consistentwith crystalline residues from partially crystallizedKarmutsen magmas To test the proportion of the primarymagma that fractionated within the crust MELTS(Ghiorso amp Sack 1995) was used to simulate fractionalcrystallization using several estimated primary magmacompositions from the PRIMELT1 modeling results Themajor-element composition of the primary magmas couldnot be estimated for the tholeiitic basalts because of exten-sive plagthorn cpxthornol fractionation and thus the MELTSmodeling of the picrites is used as a proxy for the evolutionof major elements in the volcanic sequence A pressure of 1kbar was used with variable water contents (anhydrousand 02wt H2O) calculated at the quartz^fayalite^magnetite (QFM) oxygen buffer Experimental resultsfrom Ontong Java indicate that most crystallizationoccurs in magma chambers shallower than 6 km deep(52 kbar Sano amp Yamashita 2004) however some crys-tallization may take place at greater depths (3^5 kbarFarnetani et al 1996)A selection of the MELTS results are shown in Fig 15
Olivine and spinel crystallize at high temperature until15^20wt of the liquid mass has fractionated and theresidual magma contains 9^10wt MgO (Fig 15f) At1235^12258C plagioclase begins crystallizing and between1235 and 11908C olivine ceases crystallizing and clinopyr-oxene saturates (Fig 15f) As expected the addition of clin-opyroxene and plagioclase to the crystallization sequencecauses substantial changes in the composition of the resid-ual liquid FeO and TiO2 increase and Al2O3 and CaOdecrease while the decrease in MgO lessens considerably(Fig 15) The near-vertical trends for the Karmutsen tho-leiitic basalts especially for CaO do not match the calcu-lated liquid-line-of-descent very well which misses manyof the high-CaO high-Al2O3 low-FeO basalts A positivecorrelation between Al2O3CaO and SrNd indicates thataccumulation of plagioclase played a major role in the evo-lution of the tholeiitic basalts (Fig 15g) Increasing thecrystallization pressure slightly and the water content ofthe melts does not systematically improve the match ofthe MELTS models to the observed dataThe MELTS results indicate that a significant propor-
tion of the crystallization takes place before plagioclase(15^20 crystallization) and clinopyroxene (35^45
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
31
crystallization) join the crystallization sequence Theseresults suggest that445 crystallization of the primarymagmas occurred (estimated from the residual liquid percent in Fig 15f) and if the original flood basalt stratigra-phy on Vancouver Island was 6 km thick at least anequivalent amount of complementary plutonic rocksshould have crystallized The Karmutsen basalts and theirplutonic residues thus represent a significant addition ofcrust (perhaps 412 km thickness) previously thickenedduring the formation of the underlying Paleozoic arc
Growth of the Wrangellia oceanic plateauon Vancouver IslandThe extrusive part of the Wrangellia oceanic plateau onVancouver Island was constructed in a three-layersequence of mostly tholeiitic basalts with a restrictedrange of composition (Fig 16) The plateau overliesDevonian to Mississippian arc volcanic sequences andmarine sediments of Mississippian to Permian ages thatare intruded by gabbroic rocks related to the flood basaltsThe volcanic stratigraphy of theWrangellia plateau formed
00
05
10
15
20
25
30
35
40
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
16
0 2 4 6 8 10 12 14 16 18 20
10
11
12
13
14
15
16
17
18
19
20
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
0 2 4 6 8 10 12 14 16 18 20
MgO (wt)
0 10 20 30 40 50
percent of total mass
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
) Mineral proportions
Sp (e)
Ol
CpxPlag
PicriteHigh-MgO basaltTholeiitic basalt
TiO2 (wt)
MgO (wt)
Al2O3 (wt) CaO (wt)
FeO (wt)
MgO(a) (b)
(c) (d)
MgO (wt)MgO (wt)
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
)
Residual liquid ()
1220
1190
Ol + Sp
Plag+Ol+Sp Plag+Cpx+Sp
02 wt H2O
no H2O
Residual liquid and crystallizing phases
(f)
0102030405060708090100
plagioclaseaccumulation
plagioclaseaccumulation
plagioclase
accumulatio
n
10
12
14
16
0 10 20 30 40
Al2O3CaO
SrNd
tholeiitic basalts
(g)
Fig 15 Forward fractional crystallization modeling results for major elements from MELTS (Ghiorso amp Sack 1995) compared with picriticand tholeiitic lavas from the Karmutsen Formation (a^d) Major-element compositions of residual liquids compared with Karmutsen lavacompositions For clarity the results are shown only for one starting composition (open circle) using the composition of the estimated primarymagma for the high-MgO lava series from the modeling technique of Herzberg et al (2007) for sample 4723A13 as shown in Fig 12 The esti-mated primary magmas for samples 4723A3 and 93G171 yield similar results A pressure of 1 kbar was used with no H2O and 02wt H2Ocalculated at QFM black line is anhydrous and grey line is 02wt H2OThe end-point for the anhydrous result is at 78 fractionation and11308C and the end-point for the result with 02wt H2O is 75 fractionation and11108C (e) Mineral proportions of crystallized phases fromone MELTrun (sample 4723A13) with no H2O (f) Per cent residual liquid vs temperature labelled with intervals of crystallizing phases (g)Al2O3CaO vs SrNd for the tholeiitic basalts
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
32
in a single eruptive phase (c 231^225 Ma) as an emergentbasalt sequence during a deeper-water shallow-water andsubaerial stage and subsided below sea level relativelysoon after its formation (Fig 16) The deeper-water stagecontains pillowed and massive submarine flows (3000mthick) that built up the submarine volcanic edifice from theseafloor The lowest pillow basalts were emplaced onunconsolidated and lithified fine-grained siliciclastic andcarbonaceous sediments containing Daonella beds(c 235^232 Ma) that are intruded by abundantKarmutsen mafic sills (Fig 16a) The different pillowedand massive submarine flows resulted primarily from dif-ferent effusive rates and local topography Some of themassive submarine flows commonly recognized by radiat-ing columnar jointing represent master tubes for the deliv-ery of lava to distal parts of submarine flow fields Thepillowed flows which commonly have uneven topography
in modern sequences may have served as levees to themassive submarine flows rather than allowing unob-structed spreading and inflation typical of the massive sub-aerial flows that form on shallow slopes in continentalflood basalt provinces (eg Self et al 1997) The rapid suc-cession of flows is evident by the near absence of sedimentbetween submarine flows Minor volumes of picritic pillowlavas occur in the upper part of the pillowed flow sequencein areas of the plateau exposed on northern VancouverIsland (Fig 16 Nixon et al 2008)The growth of the plateau in shallow water led to
emplacement of an increasing proportion of volcaniclastics(pillow breccias and hyaloclastites) that accumulated up to1500m thick in some areas (Fig 16b Nixon et al 2008)Horizons of pillowed flows occur locally within the volca-niclastic stratigraphy Sedimentary structures (eg gradedbedding fluidization structures) and petrographic textures
Late Devonian-Mississippian arc
Daonella beds
Middle Triassic
deeper water
Late Devonian-Mississippian arc
pillowed flows massive submarine
flows
pillow brecciahyaloclastiteshallow water
~4 km
~6 km
Late Triassic limestoneinterflow limestonelenses
Karmutsen Formation
mafic intrusions sediment-sill complex
Karmutsen Formation
(a) (b)
(c) (d)
subaerial flows
Late Devonian-Mississippian arc Late Devonian-Mississippian arc
picritic pillow basalts
Miss to Permian marine sediments
sills
Fig 16 Growth of the Wrangellia oceanic plateau onVancouver Island (a) Deep-water eruptive phase of pillowed and massive submarineflows with almost no intervening sediments onto MiddleTriassic shale containing Daonella beds (c 235^232 Ma) The basement of the plateauis composed of Devonian to Mississippian arc sequences overlain by Mississippian to Permian carbonates and siliciclastic sediments(b) Shallow-water flow emplacement with increasing proportion of volcaniclastics resulted primarily from cooling^contraction granulationPicritic pillow basalts occur primarily near the transition between pillowed submarine flows and volcaniclastic units Pillowed and massiveflows occur locally within the volcaniclastic unit (c) Emergent stage of growth consists of overlapping subaerial flow fields similar to continen-tal flood basalts with pahoehoe flow structures Multi-tiered columnar jointing occurrs locally in response to ponding of flows (eg Bondre et al2004) Interflow carbonate deposits and plagioclase megacrystic flows developed locally during the waning stage of volcanism (d) Post-volcanicsubsidence and deposition of Late Carnian to Norian (c 212^223 Ma) platform carbonates on top of the plateau Units are labeled in the dia-grams Thicknesses of the underlying units are not to scale
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
33
within the volcaniclastic rocks indicate formation primar-ily via cooling^contraction granulation magma^water^steam interaction autobrecciation resedimentation andmass-wasting rather than pyroclastic eruption (Nixonet al 2008)A broad subaerial platform was constructed by emplace-
ment of subaerial lava flows during an emergent subaerialstage (Fig 16c) The subaerial unit predominantly com-prises well-layered sheet flows that formed in a similarway to continental flood basalts as hundred(s) of pahoehoeflow fields erupted on shallow slopes (Self et al 1997) how-ever volcaniclastic and pillowed flow units are locally pre-sent mostly in the upper parts of the stratigraphyLocalized ponding of flows led to the formation of multi-tiered columnar jointing in some subaerial flows (egBondre et al 2004) As volcanism waned local interflowcarbonate deposits developed and plagioclase megacrysticflows were erupted Subsidence of the plateau is recordedby the interflow lenses and deposits of platform carbonates(40^1500m thick) on top of the Karmutsen basalts thatcontain Late Carnian to Early Norian fossils (Fig 16d c216^227 Ma Nixon et al 2008)The evolution of the internal architecture of the
Wrangellia oceanic plateau was primarily related to theconditions of emplacement (ie water depth and eruptionenvironment) There does not appear to have been a sys-tematic coincident evolution or shift of lava compositionsonVancouver Island as has been recognized in most con-tinental flood basalts and parts of the Wrangellia oceanicplateau in Alaska and Yukon (Greene et al 2008 2009)In general with the exception of the picritic pillow lavasin the submarine stratigraphy and local variations betweenhigh-MgO and tholeiitic lavas there are no significantcompositional differences between lavas erupted duringthe different stages of growth of the plateau (ie submarineand subaerial flows) nor are substantial compositional dif-ferences generally found between massive submarine flowsin contact with pillowed flows tholeiitic basalt flows nearthe base of the Karmutsen Formation are compositionallysimilar to flows sampled near the top of the volcanicsequence Where it was possible to sample continuous sec-tions of volcanic stratigraphy greater than several hun-dreds of meters (eg Keogh^Maynard Lake area west ofthe Karmutsen Range Fig 2) there is no clear relationshipbetween stratigraphic position and compositionalvariation
CONCLUSIONSThe Karmutsen Formation covers extensive areas ofVancouver Island (20 000 km2) and forms part of a largeoceanic plateau erupted over a geologically short interval(c 225^231 Ma) The tripartite volcanic stratigraphyexposed on Vancouver Island is at least 6 km thick andcomprises a succession of submarine flows volcaniclastic
deposits and massive subaerial flowsThese volcanologicaldifferences are primarily related to the eruption environ-ment (deep-water shallow-water or subaerial) The rapidgrowth of the plateau prevented significant accumulationsof intervening sediment except in the uppermost stratigra-phy where isolated limestone lenses commonly associatedwith pillow basalt and hyaloclastite deposits preserve arecord of the final stages of subsidence of the plateauThe Wrangellia plateau on Vancouver Island was con-
structed dominantly of tholeiitic basalt with restrictedmajor- and trace-element and isotopic compositional var-iations Minor volumes of picritic pillow basalts eruptedlate in the submarine phase The Karmutsen basalts wereemplaced onto older Paleozoic arc volcanic sequences andlow NbLa for the picrites may indicate a contributionfrom the sub-arc lithospheric mantle Modeling of the pet-rogenesis of the Karmutsen picrites indicates melting ofanomalously hot mantle (15008C) and extensive degreesof partial melting (23^27) and is consistent with aplume initiation model Combined Sr^Nd^Hf^Pb isotopicsystematics indicate that the picritic and tholeiitic lavaswere derived from a depleted plume-type mantle sourcedistinct from the source of MORB that has compositionalsimilarities to the source of the lavas forming theCaribbean Plateau Picrites were generated by melting ofspinel lherzolite whereas the tholeiitic basalts formed asaggregate melts of both garnet and spinel lherzolite Thetholeiitic basalts underwent extensive low-pressure fractio-nation (52^3 kbar) and seismic studies indicate that someof theWrangellia crust beneathVancouver Island may cor-respond to the plutonic residues of the Karmutsen floodbasalts
ACKNOWLEDGEMENTSWe are grateful to Nick Arndt for helping us get this pro-ject started and Nick Massey for insights into VancouverIsland geology We greatly appreciate the assistance withPRIMELT modeling from Claude Herzberg We thankMikkel Schau for his insight and enthusiasm during field-work Jane Barling assisted with analyses by MC-ICP-MSThoughtful and thorough reviews of the manuscript byJohn Mahoney Fred Frey Wendy Bohrson DougalJerram and Teal Riley are greatly appreciated Fundingwas generously provided by the Rocks to Riches Programadministered by the BC amp Yukon Chamber of Mines in2004 by the BC Geological Survey and by NSERCDiscovery Grants to JSS and DW ARG was supportedby a University Graduate Fellowship at UBC
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
34
REFERENCESBarker F Brown A S Budahn J R amp Plafker G (1989) Back-arc
with frontal-arc component origin of Triassic Karmutsen basaltBritish Columbia Canada Chemical Geology 75 81^102
Blichert-Toft JWeis D Maerschalk C Agranier A amp Albare de F(2003) Hawaiian hot spot dynamics as inferred from the Hf and Pbisotope evolution of Mauna Kea volcano Geochemistry GeophysicsGeosystems 4(2) 1^27 doi1010292002GC000340
Bondre N R Duraiswami R A amp Dole G (2004) Morphologyand emplacement of flows from the Deccan Volcanic ProvinceIndia Bulletin of Volcanology 66 29^45
Carlisle D (1963) Pillow breccias and their aquagene tuffs QuadraIsland British Columbia Journal of Geology 71 48^71
Carlisle D (1972) Late Paleozoic to Mid-Triassic sedimentary-volcanic sequence on Northeastern Vancouver Island Geological
Society of Canada Paper Report of Activities 72-1 Part B 24^30Carlisle D amp SuzukiT (1974) Emergent basalt and submergent car-
bonate^clastic sequences including the Upper Triassic Dilleri andWelleri zones on Vancouver Island Canadian Journal of Earth
Sciences 11 254^279Chauvel C amp Blichert-Toft J (2001) A hafnium isotope and trace
element perspective on melting of the depleted mantle Earth and
Planetary Science Letters 190 137^151Chayes F (1954)The theory of thin-section analysis Journal of Geology
62 92^101Cho M Liou J G amp Maruyama S (1987) Transition from the zeo-
lite to prehnite^pumpellyite facies in the Karmutsen metabasitesVancouver Island British Columbia Journal of Petrology 27(2)467^494
Clowes R M Zelt C A Amor J R amp Ellis R M (1995)Lithospheric structure in the southern Canadian Cordillera froma network of seismic refraction lines Canadian Journal of Earth
Sciences 32(10) 1485^1513Coffin M F amp Eldholm O (1994) Large igneous provinces Crustal
structure dimensions and external consequences Reviews of
Geophysics 32(1) 1^36Cox K G (1980) A model for flood basalt volcanism Journal of
Petrology 21 629^650Eldholm O amp Coffin M F (2000) Large igneous provinces
and plate tectonics In Richards M A Gordon R G amp vander Hilst R D (eds) The History and Dynamics of Global
Plate Motions American Geophysical Union Geophysical Monograph 121309^326
Elliott T R Hawkesworth C J amp Grolaquo nvold K (1991) Dynamicmelting of the Iceland plume Nature 351 201^206
Escuder-Viruete J Pecurren rez-Estaucurren n A Contreras F Joubert MWeis D Ullrich T D amp Spadea P (2007) Plume mantle sourceheterogeneity through time Insights from the Duarte ComplexHispaniola northeastern Caribbean Journal of Geophysical Research112(B04203) doi1010292006JB004323
Farnetani C G Richards M A amp Ghiorso M S (1996)Petrological models of magma evolution and deep crustal structurebeneath hotspots and flood basalt provinces Earth and Planetary
Science Letters 143 81^94Fitton J G amp Godard M (2004) Origin and evolution of magmas
on the Ontong Java Plateau In Fitton J G Mahoney J JWallace P J amp Saunders A D (eds) Origin and Evolution of the
Ontong Java Plateau Geological Society London Special Publications 229151^178
Fitton J G Saunders A D Norry M J Hardarson B S ampTaylor R N (1997) Thermal and chemical structure of theIceland plume Earth and Planetary Science Letters 153 197^208
Fitton J G Saunders A D Kempton P D amp Hardarson B S(2003) Does depleted mantle form an intrinsic part of the Icelandplume Geochemistry Geophysics Geosystems 4(3) 1032 doi1010292002GC000424
Frey F A Huang S Blichert-Toft J Regelous M amp Boyet M(2005) Origin of depleted components in basalt related to theHawaiian hotspot Evidence from isotopic and incompatible ele-ment ratios Geochemistry Geophysics Geosystems 6(1) 23 doi1010292004GC000757
Galer S J G amp Abouchami W (1998) Practical application of leadtriple spiking for correction of instrumental mass discriminationMineralogical Magazine 62A 491^492
Ghiorso M S amp Sack R O (1995) Chemical mass transfer in mag-matic processes IV A revised and internally consistent thermody-namic model for the interpolation and extrapolation of liquid^solidequilibria in magmatic systems at elevated temperatures and pres-sures Contributions to Mineralogy and Petrology 119 197^212
Greene A R Scoates J S amp Weis D (2005)WrangelliaTerrane onVancouver Island Distribution of flood basalts with implicationsfor potential Ni^Cu^PGE mineralization In Grant B (ed)Geological Fieldwork 2004 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2005-1 209^220Greene A R Scoates J S Nixon G T amp Weis D (2006) Picritic
lavas and basal sills in the Karmutsen flood basalt provinceWrangellia northern Vancouver Island In Grant B (ed)Geological Fieldwork 2005 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2006-1 39^52Greene A R Scoates J S amp Weis D (2008) Wrangellia flood
basalts in Alaska A record of plume^lithosphere interaction in aLate Triassic accreted oceanic plateau Geochemistry Geophysics
Geosystems 9(Q12004) doi1010292008GC002092Greene A R Scoates J S Weis D amp Israel S (2009)
Geochemistry of triassic flood basalts from the Yukon (Canada)segment of the accreted Wrangellia oceanic plateau Lithosdoi101016jlithos200811010
Hauff F Hoernle K Tilton G Graham D W amp Kerr A C(2000) Large volume recycling of oceanic lithosphere over shorttime scales geochemical constraints from the Caribbean LargeIgneous Province Earth and Planetary Science Letters 174 247^263
Hauff F Hoernle K amp Schmidt A (2003) Sr^Nd^Pb compositionof Mesozoic Pacific oceanic crust (Site 1149 and 801 ODP Leg185)Implications for alteration of ocean crust and the input into theIzu^Bonin^Mariana subduction system Geochemistry Geophysics
Geosystems 4(8) doi1010292002GC000421Herzberg C (2004) Partial melting below the Ontong Java Plateau
In Fitton J G Mahoney J J Wallace P J amp Saunders A D(eds) Origin and Evolution of the Ontong Java Plateau Geological SocietyLondon Special Publications 229 179^183
Herzberg C (2006) Petrology and thermal structure of the Hawaiianplume from Mauna Kea volcano Nature 444(7119) 605
Herzberg C amp Asimow P D (2008) Petrology of some oceanicisland basalts PRIMELT2XLS software for primary magma cal-culation Geochemistry Geophysics Geosystems 9(Q09001) doi1010292008GC002057
Herzberg C amp OrsquoHara M J (2002) Plume-associated ultramaficmagmas of Phanerozoic age Journal of Petrology 43(10) 1857^1883
Herzberg C Asimow P D Arndt N Niu Y Lesher C MFitton J G Cheadle M J amp Saunders A D (2007)Temperatures in ambient mantle and plumes Constraints frombasalts picrites and komatiites Geochemistry Geophysics Geosystems8(Q02006) doi1010292006GC001390
Huang S amp Frey F A (2005) Trace element abundances of MaunaKea basalt from phase 2 of the Hawaii Scientific Drilling Project
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
35
Petrogenetic implications of correlations with major element con-tent and isotopic ratios Geochemistry Geophysics Geosystems 4(6)doi1010292002GC000322
Hussong D M Wipperman L K amp Kroenke L W (1979) Thecrustal structure of the Ontong Java and Manihiki OceanicPlateaus Journal of Geophysical Research 84(B11) 6003^6010
Jerram D A amp Widdowson M (2005) The anatomy of ContinentalFlood Basalt Provinces geological constraints on the processes andproducts of flood volcanism Lithos 79(3^4) 385^405
Jones D L Silberling N J amp Hillhouse J (1977)Wrangellia a dis-placed terrane in northwestern North America CanadianJournal ofEarth Sciences 14(11) 2565^2577
Kerr A C (2003) Oceanic Plateaus In Rudnick R (ed) The Crust
Treatise on GeochemistryVol 3 Oxford Elsevier Science pp 537^565Kerr A C amp Mahoney J J (2007) Oceanic plateaus Problematic
plumes potential paradigms Chemical Geology 241 332^353Kerr A CTarney J Marriner G F Klaver GT Saunders A D
amp Thirlwall M F (1996) The geochemistry and petrogenesis ofthe Late-Cretaceous picrites and basalts of Curacao NetherlandsAntilles a remnant of an oceanic plateau Contributions to
Mineralogy and Petrology 124(1) 29^43Kerr A C Marriner G F Tarney J Nivia A Saunders A D
Thirlwall M F amp Sinton A C (1997) Cretaceous basaltic ter-ranes in Western Colombia Elemental chronological and Sr-Ndisotopic constraints on petrogenesis Journal of Petrology 38(6)677^702
Kinzler R J (1997) Melting of mantle peridotite at pressuresapproaching the spinel to garnet transition Application to mid-ocean ridge basalt petrogenesis Journal of Geophysical Research
102(B1) 853^874Kinzler R J amp Grove T L (1992) Primary magmas of mid-ocean
ridge basalts 1 Experiments and methods Journal of Geophysical
Research 97(B5) 6885^6906Korenaga J (2005) Why did not the Ontong Java Plateau form sub-
aerially Earth and Planetary Science Letters 234 385^399Kuniyoshi S (1972) Petrology of the Karmutsen (Volcanic) Group
northeastern Vancouver Island British Columbia PhD disserta-tion University of California Los Angeles 242 pp
Lassiter J C DePaolo D J amp Mahoney J J (1995) Geochemistry ofthe Wrangellia flood basalt province Implications for the role ofcontinental and oceanic lithosphere in flood basalt genesis Journalof Petrology 36(4) 983^1009
LeBas M J (2000) IUGS reclassification of the high-Mg and picriticvolcanic rocks Journal of Petrology 41(10) 1467^1470
Longhi J (2002) Some phase equilibrium systematics of lherzolitemelting I Geochemistry Geophysics Geosystems 3(3) doi1010292001GC000204
MacDonald G A amp Katsura T (1964) Chemical composition ofHawaiian lavas Journal of Petrology 5 82^133
Mahoney J J (1987) An isotopic survey of Pacific oceanic plateausimplications for their nature and origin In Keating B HFryer P Batiza R amp Boehlert GW (eds) Seamounts Islands andAtolls American Geophysical Union Geophysical Monograph 43 207^220
Mahoney J LeRoex A P Peng Z Fisher R L amp Natland J H(1992) Southwestern limits of Indian Ocean Ridge mantle and theorigin of low 206Pb204Pb mid-ocean ridge basalt isotope systema-tics of the central Southwest Indian Ridge (178^508E) Journal ofGeophysical Research 97(B13) 19771^19790
Mahoney J J Sinton J M Kurz M D MacDougall J DSpencer K J amp Lugmair GW (1994) Isotope and trace elementcharacteristics of a super-fast spreading ridge East Pacific rise13^238S Earth and Planetary Science Letters 121 173^193
Massey N W D (1995a) Geology and mineral resources of theAlberni^Nanaimo Lakes sheet Vancouver Island 92F1W 92F2Eand part of 92F7E British Columbia Ministry of Energy Mines and
Petroleum Resources Paper 1992-2 132 ppMassey N W D (1995b) Geology and mineral resources of the
Cowichan Lake sheet Vancouver Island 92C16 British Columbia
Ministry of Energy Mines and Petroleum Resources Paper 1992-3 112 ppMassey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005a) Digital Geology Map of British Columbia Tile NM9 MidCoast BC British Columbia Ministry of Energy and Mines Geofile
2005-2Massey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005b) Digital Geology Map of British Columbia Tile NM10Southwest BC British Columbia Ministry of Energy and Mines Geofile
2005-3McDonough W F amp Sun S (1995) The composition of the Earth
Chemical Geology 120 223^253Monger JW H amp Journeay J M (1994) Basement geology and tec-
tonic evolution of theVancouver region In Monger JW H (ed)Geology and Geological Hazards of the Vancouver Region Southwestern
British Columbia Geological Survey of Canada Bulletin 481 3^25Muller J E (1977) Geology of Vancouver Island Geological Survey of
Canada Open File Map 463Muller J E (1980)The Paleozoic Sicker Group of Vancouver Island British
Columbia Geological Survey of Canada Paper 79-30 22 ppMuller J E Northcote K E amp Carlisle D (1974) Geology and
mineral deposits of Alert Bay^Cape Scott map area VancouverIsland British Columbia Geological Survey of Canada Paper 74-8 77
Neal C R Mahoney J J Kroenke L W Duncan R A ampPetterson M G (1997) The Ontong^Java Plateau InMahoney J J amp Coffin M F (eds) Large Igneous Provinces
Continental Oceanic and Planetary FloodVolcanism American Geophysical
Union Geophysical Monograph 100 183^216Niu Y Collerson K D Batiza R Wendt J I amp Regelous M
(1999) Origin of enriched-typed mid-ocean ridge basalt at ridgesfar from mantle plumes The East Pacific Rise at 11820rsquoN Journalof Geophysical Research 104 7067^7088
Nixon G T amp Orr A J (2007) Recent revisions to the EarlyMesozoic stratigraphy of Northern Vancouver Island (NTS 102I092L) and metallogenic implicatinos British Columbia InGrant B (ed) Geological Fieldwork 2006 British Columbia Ministry of
Energy Mines and Petroleum Resources Paper 2007-1 163^177Nixon GT Hammack J L KoyanagiV M Payie G J Haggart
JW Orchard M JTozerT Friedman R M Archibald D APalfy J amp Cordey F (2006a) Geology of the Quatsino^PortMcNeill area northernVancouver Island British Columbia Ministry
of Energy Mines and Petroleum Resources Geoscience Map 2006-2 scale150 000
Nixon G T Kelman M C Stevenson D Stokes L A ampJohnston K A (2006b) Preliminary geology of the Nimpkishmap area (NTS 092L07) northern Vancouver Island BritishColumbia In Grant B amp Newell J M (eds) Geological Fieldwork2005 British Columbia Ministry of Energy Mines and Petroleum Resources
Paper 2006-1 135^152Nixon G T Laroque J Pals A Styan J Greene A R amp
Scoates J S (2008) High-Mg lavas in the Karmutsen flood basaltsnorthernVancouver Island (NTS 092L) Stratigraphic setting andmetallogenic significance In Grant B (ed) Geological Fieldwork
2007 British Columbia Ministry of Energy Mines and Petroleum
Resources Paper 2008-1 175^190Nowell G M Kempton P D Noble S R Fitton J G
Saunders A Mahoney J J amp Taylor R N (1998) High precisionHf isotope measurements of MORB and OIB by thermal ionisation
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
36
mass spectrometry insights into the depleted mantle Chemical
Geology 149 211^233Parrish R R amp McNicoll V J (1992) U^Pb age determinations
from the southern Vancouver Island area British Columbia InRadiogenic Age and Isotopic Studies Report 5 Geological Survey of
Canada Paper 91-2 79^86Peate DW (1997) The Parana^Etendeka Province In Coffin M F
amp Mahoney J (eds) Large Igneous Provinces Continental Oceanic andPlanetary Flood Volcanism American Geophysical Union Geophysical
Monograph 100 217^245Perfit M R Fornari D J Ridley W I Kirk P D Casey J
Kastens K A Reynolds J R Edwards M Desonie DShuster R amp Paradis S (1996) Recent volcanism in the Siqueirostransform fault picritic basalts and implications for MORBmagma genesis Earth and Planetary Science Letters 141(1^4) 91^108
Pretorius W Weis D Williams G Hanano D Kieffer B ampScoates J S (2006) Complete trace elemental characterization ofgranitoid (USGSG-2GSP-2) reference materials by high resolutioninductively coupled plasma-mass spectrometry Geostandards and
Geoanalytical Research 30(1) 39^54Regelous M Niu Y Wendt J I Batiza R Greig A amp
Collerson K D (1999) Variations in the geochemistry of magma-tism on the East Pacific Rise at 10830rsquoN since 800 ka Earth and
Planetary Science Letters 168 45^63Recurren villon S Chauvel C Arndt N T Pik R Martineau F
Fourcade S amp Marty B (2002) Heterogeneity of the Caribbeanplateau mantle source Sr O and He isotopic compositions of oli-vine and clinopyroxene from Gorgona Island Earth and Planetary
Science Letters 205(1^2) 91Rhodes J M (1996) Geochemical stratigraphy of lava flows samples
by the Hawaii Scientific Drilling Project Journal of Geophysical
Research 101(B5) 11729^11746Rhodes J M amp Vollinger M J (2004) Composition of basaltic lavas
sampled by phase-2 of the Hawaii Scientific Drilling ProjectGeochemical stratigraphy and magma types Geochemistry
Geophysics Geosystems 5(3) doi1010292002GC000434Richards M A Jones D L Duncan R A amp DePaolo D J (1991)
A mantle plume initiation model for the Wrangellia flood basaltand other oceanic plateaus Science 254 263^267
Salters V J M amp Hart S R (1991) The mantle sources of oceanridges islands and arcs the Hf-isotope connection Earth and
Planetary Science Letters 104 364^380Salters V J M amp Stracke A (2004) Composition of the depleted
SaltersV J amp WhiteW (1998) Hf isotope constraints on mantle evo-lution Chemical Geology 145 447^460
SanoT amp Yamashita S (2004) Experimental petrology of basementlavas from Ocean Drilling Program Leg 192 implications for dif-ferentiation processes in Ontong Java Plateau magmas InFitton J G Mahoney J JWallace P J amp Saunders A D (eds)Origin and Evolution of the Ontong Java Plateau Geological Society
London Special Publications 229 185^218Saunders A D (2005) Large igneous provinces Origin and environ-
mental consequences Elements 1 259^263Self S ThordarsonT amp Keszthely L (1997) Emplacement of conti-
nental flood basalt lava flows In Mahoney J J amp Coffin M F(eds) Large Igneous Provinces Continental Oceanic and Planetary Flood
Volcanism American Geophysical Union Geophysical Monograph 100381^410
Shaw D M (2000) Continuous (dynamic) melting theory revisitedCanadian Mineralogist 38(5) 1041^1063
Sluggett C L (2003) Uranium^lead age and geochemical constraintson Paleozoic and Early Mesozoic magmatism in WrangelliaTerrane Saltspring Island British Columbia BSc thesisUniversity of British ColumbiaVancouver 84 pp
Storey M Mahoney J J amp Saunders A D (1997) Cretaceousbasalts in Madagascar and the transition between plume and con-tinental lithospheric mantle sources In Mahoney J J ampCoffin M F (eds) Large Igneous Provinces Continental Oceanic and
Planetary Flood Volcanism Aamerican Geophysical Union Geophysical
Monograph 100 95^122Surdam R C (1967) Low-grade metamorphism of the Karmutsen
Group PhD dissertation University of California Los Angeles288 pp
Sutherland-Brown A Yorath C J Anderson R G amp Dom K(1986) Geological maps of southern Vancouver IslandLITHOPROBE I 92C10 11 14 16 92F1 2 7 8 Geological Survey ofCanada Open File 1272
Tejada M L G Mahoney J J Castillo P R Ingle S P Sheth HC amp Weis D (2004) Pin-pricking the elephant evidence on theorigin of the Ontong Java Plateau from Pb^Sr^Hf^Nd isotopiccharacteristics of ODP Leg 192 basalts In Fitton J GMahoney J J Wallace P J amp Saunders A D (eds) Origin and
Evolution of the Ontong Java Plateau Geological Society London Special
Publications 229 133^150Thompson P M E Kempton P D amp Kerr A C (2008) Evaluation
of the effects of alteration and leaching on Sm^Nd and Lu^Hf sys-tematics in submarine mafic rocks Lithos 104(1^4) 164^176
Thompson P M E Kempton P D White R V Kerr A CTarney J Saunders A D Fitton J G amp McBirney A (2003)Hf-Nd isotope constraints on the origin of the CretaceousCaribbean plateau and its relationship to the Galapagos plumeEarth and Planetary Science Letters 217(1^2) 59^75
Vervoort J D Patchett P Blichert-Toft J amp Albare de F (1999)Relationships between Lu^Hf and Sm^Nd isotopic systems in theglobal sedimentary system Earth and Planetary Science Letters 16879^99
Walter M J (1998) Melting of garnet peridotite and the origin ofkomatiite and depleted lithosphere Journal of Petrology 39(1) 29^60
Weis D Kieffer B Maerschalk C Barling J de Jong JWilliams G A Hanano D Mattielli N Scoates J SGoolaerts A Friedman R A amp Mahoney J B (2006) High-pre-cision isotopic characterization of USGS reference materials byTIMS and MC-ICP-MS Geochemistry Geophysics Geosystems
7(Q08006) doi1010292006GC001283Weis D Kieffer B Hanano D Silva I N Barling J
Pretorius W Maerschalk C amp Mattielli N (2007) Hf isotopecompositions of US Geological Survey reference materialsGeochemistry Geophysics Geosystems 8(Q06006) doi1010292006GC001473
Wheeler J O amp McFeely P (1991) Tectonic assemblage map of theCanadian Cordillera and adjacent part of the United States ofAmerica Geological Survey of Canada Map 1712A
WhiteW M Albare de F amp Tecurren louk P (2000) High-precision analy-sis of Pb isotope ratios by multi-collector ICP-MS Chemical Geology167 257^270
Yorath C J Sutherland Brown A amp Massey N W D (1999)LITHOPROBE southern Vancouver Island British Columbia Geological
Survey of Canada Bulletin 498 145 ppZou H (1998) Trace element fractionation during modal and
non-model dynamic melting and open-system melting Amathematical treatment Geochimica et Cosmochimica Acta 62(11)1937^1945
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
37
Zou H amp Reid M R (2001) Quantitative modeling of trace elementfractionation during incongruent dynamic melting Geochimica et
Cosmochimica Acta 65(1) 153^162
APPENDIX SAMPLEPREPARAT ION ANDANALYT ICAL METHODSOnly the freshest rocks in the field were sampled and onlythe least altered samples were selected for chemical andisotopic analysis based on thorough petrographic inspec-tion Fifty-nine of the 129 collected samples were crushed(400 g) into small pieces 52mm in diameter in aRocklabs hydraulic piston crusher between WC plates tominimize contamination The coarse-crush was mixedand 100 g was powdered in a planetary mill using agatejars and balls cleaned with quartz sand between samples
ActLabs analytical methodsThe major- and trace-element compositions of the whole-rock powders were determined at Activation LaboratoriesLtd (Actlabs) in Ancaster Ontario Analytical techniquesand detection limits are also available from Actlabs (httpwwwactlabscommethsub_code4erehtm) The analyticalmethod for each of the elements analyzed is indicated inTable 2 For the major elements a 02 g sample was mixedwith a lithium metaborate^lithium tetraborate mixtureand fused in a graphite crucible The molten mixture waspoured into a 5 HNO3 solution and shaken until dis-solved (30min) The samples were analyzed for majoroxides and selected trace elements by inductively coupledplasma optical emission spectrometry (ICP-OES) on acombination simultaneous^sequential Thermo Jarrell-AshEnviro II system Internal calibration was achieved usinga variety of international reference materials (eg W-2BIR-1 DNC-1) and independent control samplesAdditional trace elements were analyzed by both instru-mental neutron activation analysis (INAA) and induc-tively couple plasma mass spectrometry (ICP-MS) ForINAA15^25 g of sample was weighed into small polyeth-ylene vials and irradiated with a control international ref-erence material CANMET WMS-1 and NiCr flux wires ata thermal neutron flux of 71012 neutronscm2 per s in theMcMaster Nuclear Reactor Following a 7 day waitingperiod the samples were measured on an Ortec high-purity Ge detector linked to a Canberra Series 95 multi-channel analyzer Activities for each element weredecay- and weight-corrected and compared with a detectorcalibration developed from multiple international certifiedreference materials For ICP-MS 025 g of sample wasdigested in HF followed by HNO3^HClO4 heated andtaken to dryness The samples were brought back into solu-tion with HCl Samples were analyzed using a PerkinElmer Optima 3000 ICP In-lab standards or certified ref-erence materials (eg W-2 BIR-1 DNC-1) were used for
quality control A total of 15 blind duplicates were ana-lyzed from the same sample powders to assessreproducibility
University of Massachusetts analyticalmethodsFourteen sample duplicate powders taken from the same100 g sample aliquot (high-MgO basalts and picrites)were also analyzed at the Ronald B Gilmore X-RayFluorescence (XRF) Laboratory at the University ofMassachusetts Major elements were measured on a fusedLa-bearing lithium borate glass disc using a SiemensMRS-400 spectrometer Trace element concentrations(Rb Sr Ba Ce Nb Zr Y Pb Zn Ga Ni Cr V) weremeasured on a separate powder pellet using a PhilipsPW2400 sequential spectrometer using a Rh tube LOIand ferrous iron measurements were made as describedby Rhodes amp Vollinger (2004) Precision and accuracy esti-mates for the data have been described by Rhodes (1996)and Rhodes amp Vollinger (2004) Results for each sampleare the average of two separate analyses and are shown inElectronic Appendix 2 for a comparison of analyses ofduplicate powders using XRFand ICP-MS
PCIGR analytical methodsA subset of 19 samples was selected for high-precisiontrace-element analysis and Sr Nd Pb and Hf isotopicanalysis at the Pacific Centre for Isotopic andGeochemical Research (PCIGR) at the University ofBritish Columbia (UBC) Samples were selected from the59 samples analyzed for whole-rock chemistry at ActLabsbased on major- and trace-element chemistry alteration(low LOI and petrographic alteration index) and samplelocation Samples were prepared for trace-element analysisat the PCIGR by the technique described by Pretorius et al(2006) on unleached rock powders Sample powders(100mg) were weighed in 7ml screw-top Savillex bea-kers and dissolved in 1ml 14N HNO3 and 5ml 48 HFon a hotplate for 48 h at 1308C with periodic ultrasonica-tion Samples were dried and redissolved in 6ml 6N HClon a hotplate for 24 h and then dried and redissolved in1ml concentrated HNO3 for 24 h before final dryingTrace element abundances were measured by high-resolution inductively coupled plasma-mass spectrometry(HR-ICP-MS) with aThermo Finnigan Element2 systemfollowing the procedures described by Pretorius et al(2006) within 24 h of redissolution HFSE and LILE weremeasured in medium-resolution mode at 2000 dilutionusing a PFATeflon spray chamber washed with aqua regiafor 3min between samples REE were measured in high-resolution mode and U and Pb in low-resolution modeat 2000 dilution using a glass spray chamber washedwith 2 HNO3 between samples Total procedural blanksand reference materials (BCR-2 BHVO-2) were analyzedwith the batch of samples Indium was used as an internal
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
38
standard in all samples and standard solutionsBackground and standard solutions were analyzed afterevery five samples to detect memory effects and massdrift Results for PCIGR trace-element analyses areshown in Electronic Appendix 3 for the full set of analyzedtrace elementsSample digestion for purification of Sr Nd Hf and Pb
for column chemistry began by weighing each samplepowder (400^500mg) prior to leaching All samples wereleached with 6N HCl and placed in an ultrasonic bath for15min Samples were rinsed twice with 18 M-cm H2Obetween each leaching step (15 total) until the supernatantliquid was clear [following the method of Mahoney (1987)]Samples were then dried on a hotplate for 24 h andweighed again Sample solutions were then prepared bydissolving 100^250mg of the leached powder dissolvedin 1ml 14N HNO3 and 10ml 48 HF on a hotplate for48 h at 1308C with periodic ultrasonication Samples weredried and redissolved in 6ml 6N HCl on a hotplate for24 h and then dried Pb was separated using anion exhangecolumns and the discard was used for Sr REE and Hfseparation Nd was separated from the other REE and Hfrequired two additional purification steps Detailed proce-dures for column chemistry for separating Sr Nd and Pbat the PCIGR have been described byWeis et al (2006) andHf purification has been described byWeis et al (2007) Srand Nd isotope ratios were measured by thermal ioniza-tion mass spectrometry (TIMS) on a Thermo FinniganTriton system in static mode with relay matrix rotation ona single Ta and double Re^Ta filament respectively Fourto five filaments per barrel of 21 were occupied by stan-dards (NBS 987 for Sr and LaJolla for Nd) for each barrelwhere samples were run Sample Sr and Nd isotopic com-positions were corrected for mass fractionation using86Sr88Srfrac14 01194 and 146Nd144Ndfrac14 07219 Each samplewas then normalized using the barrel average of the refer-ence material relative to the values of143Nd144Ndfrac14 0511858 and 87Sr86Srfrac14 0710248 (Weiset al 2006) During the course of the Vancouver Islandanalyses the La Jolla Nd standard gave an average valueof 05118566 (nfrac14 7) and NBS987 standard gave an aver-age of 07102408 (nfrac1411) (2 error is reported as times106) 147Sm144Nd errors are 15 or 0006 USGeological Survey (USGS) reference material BHVO-2was processed with the samples and yielded Sr and Nd iso-topic ratios of 07034607 and 05129786 respectivelyThese are in agreement with the published values of070347920 and 051298411 respectively (Weis et al2006)
Pb and Hf isotopic compositions were analyzed by mul-tiple collector-inductively coupled plasma-mass spectrome-try (MC-ICP-MS) on a static multi-collection on a NuPlasma (Nu Instruments) system The detailed analyticalprocedure for Pb isotopic analyses on the Nu Plasma atthe PCIGR has been described by Weis et al (2006) Theconfiguration for Pb analyses allows for collection of PbTl and Hg togetherTl and Hg are used to monitor instru-mental mass discrimination and isobaric overlap repec-tively All sample solutions were analyzed withapproximately the same PbTl ratio (4) as the referencematerial NIST SRM 981 To accomplish this a small ali-quot of each sample solution from the Pb columns was ana-lyzed on the Element2 to determine the precise amount ofPb available for analysis on the Nu Plasma The SRM 981standard was run after every two samples on the NuPlasma During the time samples were run analyses ofthe SRM 981 Pb reference material gave values of206Pb204Pbfrac1416940322 207Pb204Pbfrac1415495823and 208Pb204Pbfrac14 36713164 (nfrac14 61 2 error is reportedas times 104) these values are in excellent agreement withreported TIMS triple-spike values of Galer amp Abouchami(1998) Fractionation-corrected Pb isotopic ratios were fur-ther corrected by the sample-standard bracketing methodor the ln^ln correction method described by White et al(2000) and Blichert-Toft et al (2003) Leached powders ofUSGS reference material BHVO-2 yielded Pb isotopicratios of 206Pb204Pbfrac14186454 8 207Pb204Pbfrac14 1549105 and 208Pb204Pbfrac14 38222514 and BCR-2yielded 206Pb204Pbfrac141880466 207Pb204Pbfrac14 1562518 and 208Pb204Pbfrac14 3883496 (2 error isreported as times 104) These values are in agreement withthose for leached residues of BHVO-2 and BCR-2 fromWeis et al (2006)Hf isotopic compositions were analyzed following the
procedures detailed byWeis et al (2007) The configurationfor Hf analyses monitored Lu mass 175 andYb mass 172 toallow for interference correction to masses 174 and 176 Hfisotopic ratios were normalized internally for mass fractio-nation to a 179Hf177Hf ratio of 07325 using an exponentialcorrection Standards were run after every two samplesand sample results were normalized to the ratio of the in-run daily average and a 176Hf177Hf ratio for JMC-475 of0282160 During the course of analyses the Hf standardJMC-475 gave an average value 02821679 (nfrac14130)USGS reference materials BCR-2 and BHVO-2 were pro-cessed with the samples and yielded Hf isotopic ratios of02828675 and 02831005 respectively Publishedvalues for BCR-2 and BHVO-2 are 02828717 and0283104 8 respectively (Weis et al 2007)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
39
00
05
10
15
20
25
30
35
40
0 5 10 15 20 25
10
11
12
13
14
15
16
17
18
19
20
0 5 10 15 20 25
6
7
8
9
10
11
12
13
14
15
0 5 10 15 20 25
6
7
8
9
10
11
12
13
14
15
16
0 5 10 15 20 25
0
1
2
3
4
5
6
40 45 50 55 60
0
200
400
600
800
0 5 10 15 20 25
(wt)(wt)
(wt)
(wt)
+K2ONa2O TiO2
FeO (T)
Ni (ppm)
Al2O3
CaO
alkalic
tholeiitic
Karmutsen Form
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
(compiled data)
Pillowed flowMineralized sill
SiO2 (wt) MgO (wt)
MgO (wt)MgO (wt)
MgO (wt) MgO (wt)
(a) (b)
(g)
(e)
(f)
(c)
LOI (wt )
(d)
MgO (wt )0
1
2
3
4
5
6
7
0 5 10 15 20 25
(wt)
Fig 6 Whole-rock major-element Ni and LOI variation diagrams for the Karmutsen Formation New samples from this study (Table 2) areshown by symbols with black outlines (see legend) Previously published analyses are shown by small gray symbols without black outlines Theboundary of the alkaline and tholeiitic fields is that of MacDonald amp Katsura (1964) Total iron expressed as FeO LOI loss on ignition oxidesare plotted on an anhydrous normalized basis References for the 322 compiled analyses for the Karmutsen Formation are Surdam (1967)Kuniyoshi (1972) Muller et al (1974) Barker et al (1989) Lassiter et al (1995) Massey (1995a1995b)Yorath et al (1999) and G Nixon (unpublisheddata) It should be noted that the compiled dataset has not been filtered many of the samples with high SiO2 (452wt ) are probably notKarmutsen flood basalts but younger Bonanza basalts and andesites Three pillowed flow samples and a mineralized sill are distinguishedseparately because of their anomalous chemistry Coarse-grained samples are indicated separately Three tholeiitic basalt samples (4724A54718A5 4718A6) with416wt Al2O3 and4270 ppm Sr contain 10^25 modal of large plagioclase phenocrysts (7^8mm) or glomerocrysts(44mm)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
10
Table 2 Major element (wt oxide) and trace element (ppm) abundances in whole-rock samples of Karmutsen basalts
1Ni concentrations for these samples were determined by XRFTHOL tholeiitic basalt PIC picrite HI-MG high MgO basalt CG coarse-grained (sill or gabbro) MIN SIL mineralizedsill OUTLIER anomalous pillowed flow in plots MA Mount Arrowsmith SL Schoen Lake KR Karmutsen Range QIQuadra Island Sample locations are given using the Universal Transverse Mercator (UTM) coordinate system (NAD83zones 9 and 10) Analyses were performed at Activation Laboratory (ActLabs) Fe2O3
is total iron expressed as Fe2O3LOI loss on ignition All major elements Sr V and Y were measured by ICP quadrupole OES on solutions of fusedsamples Cu Ni Pb and Zn were measured by total dilution ICP Cs Ga Ge Hf Nb Rb Ta Th U Zr and REE weremeasured by magnetic-sector ICP on solutions of fused samples Co Cr and Sc were measured by INAA Blanks arebelow detection limit See Electronic Appendices 2 and 3 for complete XRF data and PCIGR trace element datarespectively Major elements for sample 93G17 are normalized anhydrous Sample 5615A7 was analyzed by XRFSamples from Quadra Island are not shown in figures
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
17
picritic and high-MgO basalt pillow lavas extend to20wt MgO (Fig 6 and references listed in thecaption)The tholeiitic basalts from the Karmutsen Range dis-
play a tight range of parallel light rare earth element(LREE)-enriched REE patterns (LaYbCNfrac14 21^24mean 2202) whereas the picrites and high-MgObasalts are characterized by sub-parallel LREE-depleted
patterns (LaYbCNfrac14 03^07 mean 06 02) with lowerREE abundances than the tholeiitic basalts (Fig 7)Tholeiitic basalts from the Schoen Lake and MountArrowsmith areas have similar REE patterns tosamples from the Karmutsen Range (Schoen Lake LaYbCNfrac14 21^26 mean 2303 Mount Arrowsmith LaYbCNfrac1419^25 mean 2204) and the coarse-grainedmafic rocks have similar REE patterns to the tholeiitic
Depleted MORB average(Salters amp Stracke 2004)
Schoen LakeSchoen Lake
Mount Arrowsmith
Karmutsen RangeS
ampl
eC
hond
rite
Sam
ple
Cho
ndrit
e
Sam
ple
Prim
itive
Man
tleS
ampl
eP
rimiti
ve M
antle
Sam
ple
Prim
itive
Man
tle
Mount Arrowsmith
Karmutsen Range
(a) (b)
(c) (d)
(e) (f )
Sam
ple
Cho
ndrit
e
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained mafic rock
Pillowed flowMineralized sill
1
10
100
1
10
100
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
01
1
10
100
1
10
100
1
10
100
Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
Fig 7 Whole-rock REE and other incompatible-element concentrations for the Karmutsen Formation (a) (c) and (e) are chondrite-normal-ized REE patterns for samples from each of the three main field areas onVancouver Island (b) (d) and (f) are primitive mantle-normalizedtrace-element patterns for samples from each of the three main field areas All normalization values are from McDonough amp Sun (1995) Theclear distinction between LREE-enriched tholeiitic basalts and LREE-depleted picrites the effects of alteration on the LILE (especially loss ofK and Rb) and the different range for the vertical scale in (b) (extends down to 01) should be noted
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
18
basalts (LaYbCNfrac14 21^29 mean 2508) Two LREE-depleted coarse-grained mafic rocks from the SchoenLake area have similar REE patterns (LaYbCNfrac14 07) tothe picrites from the Keogh Lake area indicating that thisdistinctive suite of rocks is exposed as far south asSchoen Lake that is 75 km away (Fig 7) The anomalouspillowed flow from the Karmutsen Range has lower LaYbCN (13^18) than the main group of tholeiitic basaltsfrom the Karmutsen Range (Fig 7) The mineralized sillfrom the Schoen Lake area is distinctly LREE-enriched(LaYbCNfrac14 33) with the highest REE abundances(LaCNfrac14 940) of all Karmutsen samples this may reflectcontamination by adjacent sediment also evident in thin-section where sulfide blebs are cored by quartz (Greeneet al 2006)The primitive mantle-normalized trace-element pat-
terns (Fig 7) and trace-element variations and ratios(Fig 8) highlight the differences in trace-element
concentrations between the four main groups ofKarmutsen flood basalts onVancouver Island The tholeii-tic basalts from all three areas have relatively smooth par-allel trace-element patterns some samples have smallpositive Sr anomalies relative to Nd and Sm and manysamples have prominent negative K anomalies relative toU and Nb reflecting K loss during alteration (Fig 7) Thelarge ion lithophile elements (LILE) in the tholeiiticbasalts (mainly Rb and K not Cs) are depleted relativeto the high field strength elements (HFSE) and LREELILE segments especially for the Schoen Lake area(Fig 7d) are remarkably parallel The picrites andhigh-MgO basalts form a tight band of parallel trace-element patterns and are depleted in HFSE and LREEwith positive Sr anomalies and relatively low Rb Thecoarse-grained mafic rocks from the Karmutsen Rangehave identical trace-element patterns to the tholeiiticbasalts Samples from the Schoen Lake area have similar
000
025
050
075
100
125
0 1 2 3 4 5
00
04
08
12
16
0 5 10 15 20 25
MgO (wt)
0
5
10
15
0 50 100 150
Hf (ppm)
Th (ppm)
NbLaNb (ppm)
Zr (ppm)
(a)
(c)
(b)
(d)
Th (ppm)
00
02
04
06
08
10
12
00 01 02 03 04 05
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
NbLa=1
4723A4 5615A12
U (ppm)
Fig 8 Whole-rock trace-element concentrations and ratios for the Karmutsen Formation [except (b) which is vs wt MgO] (a) Nb vs Zr(b) NbLa vs MgO (c) Th vs Hf (d) Th vs U There is a clear distinction between the tholeiitic basalts and picrites in Nb and Zr both inconcentration and the slope of each trend
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
19
trace-element patterns to the Karmutsen Range except forthe two LREE-depleted coarse-grained mafic rocks andthe mineralized sill (Fig 7d)
Sr^Nd^Hf^Pb isotopic compositionsA total of 19 samples from the four main groups ofthe Karmutsen Formation were selected for radiogenicisotopic geochemistry on the basis of major- and trace-element variation stratigraphic position and geographicallocation The samples have overlapping age-correctedHf and Nd isotope ratios and distinct ranges of Sr isotopiccompositions (Fig 9) The tholeiitic basalts picriteshigh-MgO basalt and coarse-grained mafic rocks have
initial eHffrac14thorn87 to thorn126 and eNdfrac14thorn66 to thorn88 cor-rected for in situ radioactive decay since 230 Ma (Fig 9Tables 3 and 4) All the Karmutsen samples form a well-defined linear array in a Lu^Hf isochron diagram corre-sponding to an age of 24115 Ma (Fig 9) This is withinerror of the accepted age of the Karmutsen Formationand indicates that the Hf isotope systematics behaved as aclosed system since c 230 Ma The anomalous pillowedflow has similar initial eHf (thorn98) and slightly lower initialeNd (thorn62) compared with the other Karmutsen samplesThe picrites and high-MgO basalt have higher initial87Sr86Sr (070398^070518) than the tholeiitic basalts(initial 87Sr86Srfrac14 070306^070381) and the coarse-grained mafic rocks (initial 87Sr86Srfrac14 070265^070428)
05128
05129
05130
05131
05132
015 017 019 021 023 025
02829
02830
02831
02832
02833
02834
000 002 004 006 008 010
4
6
8
10
0702 0703 0704 0705 07066
8
10
12
14
5 6 7 8 9 10
12
0
1
2
3
4
5
6
7
0702 0703 0704 0705 0706
(d)
LOI (wt )
87Sr 86Sr
4723A2
Nd
143Nd144Nd
147Sm144Nd
87Sr 86Sr (230 Ma)
(230 Ma)
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
(a)
(c)
176Hf177Hf
Age=241 plusmn 15 Ma=+990 plusmn 064
176Lu177Hf
Hf (230 Ma)
Nd (230 Ma)
(e)
Hf (i)
(b)
Fig 9 Whole-rock Sr Nd and Hf isotopic compositions for the Karmutsen Formation (a) 143Nd144Nd vs 147Sm144Nd (b) 176Hf177Hf vs176Lu177Hf The slope of the best-fit line for all samples corresponds to an age of 24115 Ma (c) Initial eNd vs 87Sr86Sr Age-correction to230 Ma (d) LOI vs 87Sr86Sr (e) Initial eHf vs eNd Average 2 error bars are shown in a corner of each panel Complete chemical duplicatesshown inTables 3 and 4 (samples 4720A7 and 4722A4) are circled in each plot
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
overlap the ranges of the picrites and tholeiitic basalts(Fig 9 Table 3)The measured Pb isotopic compositions of the tholeiitic
basalts are more radiogenic than those of the picrites andthe most magnesian Keogh Lake picrites have the leastradiogenic Pb isotopic compositions The range ofinitial Pb isotope ratios for the picrites is206Pb204Pbfrac1417868^18580 207Pb204Pbfrac1415547^15562and 208Pb204Pbfrac14 37873^38257 and the range for thetholeiitic basalts is 206Pb204Pbfrac1418782^19098207Pb204Pbfrac1415570^15584 and 208Pb204Pbfrac14 37312^38587 (Fig 10 Table 5) The coarse-grained mafic rocksoverlap the range of initial Pb isotopic compositions forthe tholeiitic basalts with 206Pb204Pbfrac1418652^19155207Pb204Pbfrac1415568^15588 and 208Pb204Pbfrac14 38153^38735 One high-MgO basalt has the highest initial206Pb204Pb (19242) and 207Pb204Pb (15590) and thelowest initial 208Pb204Pb (37676) The Pb isotopic ratiosfor Karmutsen samples define broadly linearrelationships in Pb isotope plots The age-corrected Pb
isotopic compositions of several picrites have been affectedby U and Pb (andor Th) mobility during alteration(Fig 10)
ALTERAT IONThe Karmutsen basalts have retained most of their origi-nal igneous structures and textures however secondaryalteration and low-grade metamorphism have producedzeolitic- and prehnite^pumpellyite-bearing mineral assem-blages (Table 1 Cho et al 1987) Alteration and metamor-phism have primarily affected the distribution of the LILEand the Sr isotopic systematics The Keogh Lake picriteshave been affected more by alteration than the tholeiiticbasalts and have higher LOI (up to 55wt ) variableLILE and higher measured Sr Nd and Hf and lowermeasured Pb isotope ratios than the tholeiitic basalts Srand Pb isotope compositions for picrites and tholeiiticbasalts show a relationship with LOI (Figs 9 and 10)whereas Nd and Hf isotopic compositions do not correlate
Table 3 Sr and Nd isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Rb Sr 87Sr86Sr 2m87Rb86Sr 87Sr86Srt Sm Nd 143Nd144Nd 2m eNd
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses carried out at the PCIGR thecomplete trace element analyses are given in Electronic Appendix 3
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
21
with LOI (not shown) The relatively high apparent initialSr isotopic compositions for the high-MgO lavas (up to07052) probably resulted from post-magmatic loss of Rbor an increase in 87Sr86Sr through addition of seawater Sr(eg Hauff et al 2003) High-MgO lavas from theCaribbean plateau have similarly high 87Sr86Sr comparedwith basalts (Recurren villon et al 2002)The correction for in situdecay on initial Pb isotopic ratios has been affected bymobilization of U and Th (and Pb) in the whole-rockssince their formation A thorough acid leaching duringsample preparation was used that has been shown to effec-tively remove alteration phases (Weis et al 2006 NobreSilva et al in preparation) Whole-rock SmNd ratiosand age-correction of Nd isotopes may be affected bythe leaching procedure for submarine rocks older than50 Ma (Thompson et al 2008 Fig 9) there ishowever very little variation in Nd isotopic compositionsof the picrites or tholeiitic basalts and this system does notappear to be disturbed by alteration The HFSE abun-dances for tholeiitic and picritic basalts exhibit clear
linear correlations in binary diagrams (Fig 8) whereasplots of LILE vs HFSE are highly scattered (not shown)because of the mobility of some of the alkali and alkalineearth LILE (Cs Rb Ba K) and Sr for most samplesduring alterationThe degree of alteration in the Karmutsen samples does
not appear to be related to eruption environment or depthof burial Pillow rims are compositionally different frompillow cores (Surdam1967 Kuniyoshi1972) and aquagenetuffs are chemically different from isolated pillows withinpillow breccias however submarine and subaerial basaltsexhibit a similar degree of alterationThere is no clear cor-relation between the submarine and subaerial basalts andsome commonly used chemical alteration indices (eg BaRb vs K2OP2O5 Huang amp Frey 2005) There is also nodefinitive correlation between the petrographic alterationindex (Table 1) and chemical alteration indices although10 of 13 tholeiitic basalts with the highest K and LILEabundances have the highest petrographic alterationindex of three (seeTable 1)
Table 4 Hf isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Lu Hf 176Hf177Hf 2m eHf176Lu177Hf 176Hf177Hft eHf(t)
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
OLIV INE ACCUMULAT ION INH IGH-MgO AND PICR IT IC LAVASGeochemical trends and petrographic characteristics indi-cate that accumulation of olivine played an important rolein the formation of the high-MgO basalts and picrites fromthe Keogh Lake area on northern Vancouver Island TheKeogh Lake samples show a strong linear correlation inplots of Al2O3 TiO2 ScYb and Ni vs MgO and many ofthe picrites have abundant clusters of olivine pseudomorphs(Table1 Fig11)There is a strong linear correlationbetween
modal per cent olivine and whole-rock magnesium contentmagnesium contents range between 10 and 19wt MgOand the proportion of olivine phenocrysts in these samplesvaries from 0 to 42 vol (Fig 11)With a few exceptionsboth the size range and median size of the olivine pheno-crysts in most samples are comparable The clear correla-tion between the proportion of olivine phenocrysts andwhole-rock magnesium contents indicates that accumula-tion of olivine was directly responsible for the high MgOcontents (410wt ) in most of the picritic lavas
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
207Pb204Pb
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
208Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
(a) (b)
(c) (e)
4723A24723A2
4723A2
4723A2
4722A4
4723A4
4722A4
4723A4
4723A134723A3
4723A3
4723A13
1556
1558
1560
1562
1564
1566
1568
186 190 194 198 202 206 2101550
1552
1554
1556
1558
1560
178 180 182 184 186 188 190 192 194
380
384
388
392
396
186 190 194 198 202 206 210374
378
382
386
390
178 180 182 184 186 188 190 192 194
206Pb204Pb(d)
LOI (wt )
0
1
2
3
4
5
6
7
186 190 194 198 202 206 210
4723A2
Fig 10 Pb isotopic compositions of leached whole-rock samples measured by MC-ICP-MS for the Karmutsen Formation Error bars are smal-ler than symbols (a) Measured 207Pb204Pb vs 206Pb204Pb (b) Initial 207Pb204Pb vs 206Pb204Pb Age-correction to 230 Ma (c) Measured208Pb204Pb vs 206Pb204Pb (d) LOI vs 206Pb204Pb (e) Initial 208Pb204Pb vs 206Pb204Pb Complete chemical duplicates shown in Table 5(samples 4720A7 and 4722A4) are circled in each plot The dashed lines in (b) and (e) show the differences in age-corrections for two picrites(4723A4 4722A4) using the measured UTh and Pb concentrations for each sample and the age-corrections when concentrations are used fromthe two picrites (4723A3 4723A13) that appear to be least affected by alteration
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
23
DISCUSS IONThe compositional and spatial development of the eruptedlava sequences of the Wrangellia oceanic plateau onVancouver Island provide information about the evolutionof a mantle plume upwelling beneath oceanic lithospherein an analogous way to the interpretation of continentalflood basalts The Caribbean and Ontong Java plateausare the two primary examples of accreted parts of oceanicplateaus that have been studied to date and comparisonscan be made with the Wrangellia oceanic plateauHowever the Wrangellia oceanic plateau is a distinctiveoccurrence of an oceanic plateau because it formed on thecrust of a Devonian to Mississippian intra-oceanic islandarc overlain by Mississippian to Permian limestones andpelagic sediments The nature and thickness of oceaniclithospheric mantle are considered to play a fundamentalrole in the decompression and evolution of a mantleplume and thus the compositions of the erupted lavasequences (Greene et al 2008) The geochemical and
stratigraphic relationships observed in the KarmutsenFormation onVancouver Island and the focus of the subse-quent discussion sections provide constraints on (1) thetemperature and degree of melting required to generatethe primary picritic magmas (2) the composition of themantle source (3) the depth of melting and residual min-eralogy in the source region (4) the low-pressure evolutionof the Karmutsen tholeiitic basalts (5) the growth of theWrangellia oceanic plateau
Melting conditions and major-elementcomposition of the primary magmasThe primary magmas to most flood basalt provinces arebelieved to be picritic (eg Cox 1980) and they have beenfound in many continental and oceanic flood basaltsequences worldwide (eg Siberia Karoo Paranacurren ^Etendeka Caribbean Deccan Saunders 2005) Near-pri-mary picritic lavas with low total alkali abundances
Table 5 Pb isotopic compositions of Karmutsen basaltsVancouver Island BC
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
require high-degree partial melting of the mantle source(eg Herzberg amp OrsquoHara 2002) The Keogh Lake picritesfrom northernVancouver Island are the best candidates foridentifying the least modified partial melts of the mantleplume source despite having accumulated olivine pheno-crysts and can be used to estimate conditions of meltingand the composition of the primary magmas for theKarmutsen FormationPrimary magma compositions mantle potential tem-
peratures and source melt fractions for the picritic lavas ofthe Karmutsen Formation were calculated from primitivewhole-rock compositions using PRIMELT1XLS software(Herzberg et al 2007) and are presented in Fig 12 andTable 6 A detailed discussion of the computationalmethod has been given by Herzberg amp OrsquoHara (2002)and Herzberg et al (2007) key features of the approachare outlined belowPrimary magma composition is calculated for a primi-
tive lava by incremental addition of olivine and compari-son with primary melts of mantle peridotite Lavas thathad experienced plagioclase andor clinopyroxene fractio-nation are excluded from this analysis All calculated pri-mary magma compositions are assumed to be derived byaccumulated fractional melting of fertile peridotite KR-4003 Uncertainties in fertile peridotite composition donot propagate to significant variations in melt fractionand mantle potential temperature melting of depletedperidotite propagates to calculated melt fractions that aretoo high but with a negligible error in mantle potential
temperature PRIMELT1XLS calculates the olivine liqui-dus temperature at 1atm which should be similar to theactual eruption temperature of the primary magmaand is accurate to 308C at the 2 level of confidenceMantle potential temperature is model dependent witha precision that is similar to the accuracy of the olivineliquidus temperature (C Herzberg personal communica-tion 2008)With regard to the evidence for accumulated olivine in
the Keogh Lake picrites it should not matter whether themodeled lava composition is aphyric or laden with accu-mulated olivine phenocrysts PRIMELT1 computes theprimary magma composition by adding olivine to a lavathat has experienced olivine fractionation in this way itinverts the effects of fractional crystallization The onlyrequirement is that the olivine phenocrysts are not acci-dental or exotic to the lava compositions Support for thisprocedure can be seen in the calculated olivine liquid-line-of-descent shown by the curved arrays with 5 olivineaddition increments (Fig 12c) Fractional crystallization ofolivine from model primary magmas having 85^95wt FeO and 153^175wt MgO will yield arrays of deriva-tive liquids and randomly sorted olivines that in most butnot all cases connect the picrites and high-MgO basalts Anewer version of the PRIMELT software (Herzberg ampAsimow 2008) can perform olivine addition to and sub-traction from lavas with highly variably olivine phenocrystcontents and yields results that converge with those shownin Fig 12
Fig 11 Relationship between abundance of olivine phenocrysts and whole-rock MgO contents for the Keogh Lake picrites (a) Tracings ofolivine grains (gray) in six high-MgO samples made from high resolution (2400 dpi) scans of petrographic thin-sections Scale bars represent10mm sample numbers are indicated at the upper left (b) Modal olivine vs whole-rock MgO Continuous line is the best-fit line for 10 high-MgO samplesThe areal proportion of olivine was calculated from raster images of olivine grains using ImageJ image analysis software whichprovides an acceptable estimation of the modal abundance (eg Chayes 1954)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
25
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
Gray lines = final melting pressure
L + Ol + Opx
07
08
09
Pressure (GPa)
10
3
4
5
6
1
SOLIDUS
01
04
0506
L + Ol
2
03
02
PeridotiteKR-4003
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
Thick black lines = initial melting pressure
Dashed lines = melt fraction
Melt Fraction
2
3 4 5 6
L+Ol+Opx
SolidusSpinel
SolidusGarnet Peridotite
7
L+Ol
0 10 20 30 4040
45
50
55
60
MgO (wt)
SiO2
(wt)
PeridotiteKR-4003
Melt Fraction
0 10 20 305
10
15
CaO(wt)
MgO (wt)
3
4 5 6 7
2
46
2
Peridotite Partial Melts
Thick black line = initial melting pressureGray lines = final melting pressure
Peridotite
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
02
Pressure (GPa)
0302
04
Pressure (GPa)
Picrite
High-MgO basalt
Tholeiitic basalt
(c)
(d)
Karmutsen flood basaltsOlivine addition modelOlivine addition (5 increment)Primary magma
PicriteHigh-MgO basaltTholeiitic basalt
(a)
(b)
800 850 900
910
920
930
940
950
Karmutsen flood basalts
Olivine addition modelOlivine addition (5 increment)Primary magma
Gray lines = Fo contents of olivine
Fig 12 Estimated primary magma compositions for three Keogh Lake picrites and high-MgO basalts (samples 4723A4 4723A135616A7) using the forward and inverse modeling technique of Herzberg et al (2007) Karmutsen compositions and modeling results areoverlain on diagrams provided by C Herzberg (a) Whole-rock FeO vs MgO for Karmutsen samples from this study Total iron estimatedto be FeO is 090 Gray lines show olivine compositions that would precipitate from liquid of a given MgO^FeO composition Black lineswith crosses show results of olivine addition (inverse model) using PRIMELT1 (Herzberg et al 2007) Gray field highlights olivinecompositions with Fo contents of 91^92 predicted to precipitate (b) FeO vs MgO showing Karmutsen lava compositions and results ofa forward model for accumulated fractional melting of fertile peridotite KR-4003 (Kettle River Peridotite Walter 1998) (c) SiO2 vsMgO for Karmtusen lavas and model results (d) CaO vs MgO showing Karmtusen lava compositions and model results To brieflysummarize the technique [see Herzberg et al (2007) for complete description] potential parental magma compositions for the high-MgOlava series were selected (highest MgO and appropriate CaO) and using PRIMELT1 software olivine was incrementally added tothe selected compositions to show an array of potential primary magma compositions (inverse model) Then using PRIMELT1 theresults from the inverse model were compared with a range of accumulated fractional melts for fertile peridotite derived fromparameterization of the experimental results for KR-4003 from Walter (1998) forward model Herzberg amp OrsquoHara (2002) A melt fractionwas sought that was unique to both the inverse and forward models (Herzberg et al 2007) A unique solution was found when there was a com-mon melt fraction for both models in FeO^MgO and CaO^MgO^Al2O3^SiO2 (CMAS) projection space This modeling assumes thatolivine was the only phase crystallizing and ignores chromite precipitation and possible augite fractionation in the mantle (Herzbergamp OrsquoHara 2002) Results are best for a residue of spinel lherzolite (not pyroxenite) The presence of accumulated olivine in samples ofthe Keogh Lake picrites used as starting compositions does not significantly affect the results because we are modeling addition of olivine(see text for further description) The tholeiitic basalts cannot be used for modeling because they are all saturated in plagthorn cpxthornolThick black line for initial melting pressure at 4 GPa is not shown in (c) for clarity Gray area in (a) indicates parental magmas thatwill crystallize olivine with Fo 91^92 at the surface Dashed lines in (c) indicate melt fraction Solidi for spinel and garnet peridotite arelabelled in (c)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
The PRIMELT1 results indicate that if the Karmutsenpicritic lavas were derived from accumulated fractionalmelting of fertile peridotite the primary magmas wouldhave contained 15^17wt MgO and 10wt CaO(Fig 12 Table 6) formed from 23^27 partial meltingand would have crystallized olivine with a Fo content of 91 (Fig 12 Table 6) Calculated mantle temperatures forthe Karmutsen picrites (14908C) indicate melting ofanomalously hot mantle (100^2008C above ambient) com-pared with ambient mantle that produces mid-ocean ridgebasalt (MORB 1280^14008C Herzberg et al 2007)which initiated between 3 and 4 GPa Several picritesappear to be near-primary melts and required very littleaddition of olivine (eg sample 93G171 with measured158wt MgO) These temperature and melting esti-mates of the Karmutsen picrites are consistent withdecompression of hot mantle peridotite in an actively con-vecting plume head (ie plume initiation model) Threesamples (picrite 5614A1 and high-MgO basalts 5614A3and 5614A5) are lower in iron than the other Karmutsenlavas with FeO in the 65^80wt range (Fig 12c) andhave MgO and FeO contents that are similar to primitiveMORB from the Sequeiros fracture zone of the EastPacific Rise (EPR Perfit et al 1996) These samples
however differ from EPR MORB in having higher SiO2
and lower Al2O3 and they are restricted geographicallyto one area (Sara Lake Fig 2)We therefore have evidence for primary magmas with
MgO contents that range from a possible low of 110wt to as high as 174wt with inferred mantle potential tem-peratures from 1340 to 15208C This is similar to the rangedisplayed by lavas from the Ontong Java Plateau deter-mined by Herzberg (2004) who found that primarymagma compositions assuming a peridotite sourcewould contain 17wt MgO and 10wt CaO(Table 6) and that these magmas would have formed by27 melting at a mantle potential temperature of15008C (first melting occurring at 36 GPa) Fitton ampGodard (2004) estimated 30 partial melting for theOntong Java lavas based on the Zr contents of primarymagmas of Kroenke-type basalt calculated by incrementaladdition of equilibrium olivine However if a considerableamount of eclogite was involved in the formation of theOntong Java plateau magmas the excess temperatures esti-mated by Herzberg et al (2007) may not be required(Korenaga 2005) The variability in mantle potential tem-perature for the Keogh Lake picrites is also similar to thewide range of temperatures that have been calculated for
Table 6 Estimated primary magma compositions for Karmutsen basalts and other oceanic plateaus or islands
Sample 4723A4 4723A13 5616A7 93G171 Average OJP Mauna Keay Gorgonaz
Ontong Java primary magma composition for accumulated fraction melting (AFM) from Herzberg (2004)yMauna Kea primary magma composition is average of four samples of Herzberg (2006 table 1)zGorgona primary magma composition for AFM for 1F fertile source from Herzberg amp OrsquoHara (2002 table 4)Total iron estimated to be FeO is 090
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
27
lavas from single volcanoes in the Galapagos Hawaii andelsewhere by Herzberg amp Asimow (2008) Those workersinterpreted the range to reflect the transport of magmasfrom both the cool periphery and the hotter axis of aplume A thermally heterogeneous plume will yield pri-mary magmas with different MgO contents and the rangeof primary magma compositions and their inferred mantlepotential temperatures for Karmutsen lavas may reflectmelting from different parts of the plume
Source of the Karmutsen Formation lavasThe Sr^Nd^Hf^Pb isotopic compositions of theKarmutsen volcanic rocks on Vancouver Island indicatethat they formed from plume-type mantle containing adepleted component that is compositionally similar to butdistinct from other oceanic plateaus that formed in thePacific Ocean (eg Ontong Java and Caribbean Fig 13)Trace element and isotopic constraints can be used to testwhether the depleted component of the KarmutsenFormation was intrinsic to the mantle plume and distinctfrom the source of MORB or the result of mixing or invol-vement of ambient depleted MORB mantle with plume-type mantle The Karmutsen tholeiitic basalts have initialeNd and
87Sr86Sr ratios that fall within the range of basaltsfrom the Caribbean Plateau (eg Kerr et al 1996 1997Hauff et al 2000 Kerr 2003) and a range of Hawaiian iso-tope data (eg Frey et al 2005) and lie within and abovethe range for the Ontong Java Plateau (Tejada et al 2004Fig 13a) Karmutsen tholeiitic basalts with the highest ini-tial eNd and lowest initial 87Sr86Sr lie within and justbelow the field for age-corrected northern EPR MORB at230 Ma (eg Niu et al1999 Regelous et al1999) Initial Hfand Nd isotopic compositions for the Karmutsen samplesare noticeably offset to the low side of the ocean islandbasalt (OIB) array (Vervoort et al 1999) and lie belowand slightly overlap the low eHf end of fields for Hawaiiand the Caribbean Plateau (Fig 13c) the Karmutsen sam-ples mostly fall between and slightly overlap a field forage-corrected Pacific MORB at 230 Ma and Ontong Java(Mahoney et al 1992 1994 Nowell et al 1998 Chauvel ampBlichert-Toft 2001) Karmutsen lava Pb isotope ratios over-lap the broad range for the Caribbean Plateau and thelinear trends for the Karmutsen samples intersect thefields for EPR MORB Hawaii and Ontong Java in208Pb^206Pb space but do not intersect these fields in207Pb^206Pb space (Fig 13d and e) The more radiogenic207Pb204Pb for a given 206Pb204Pb of the Karmutsenbasalts requires high UPb ratios at an ancient time when235U was abundant similar to the Caribbean Plateau andenriched in comparison with EPR MORB Hawaii andOntong JavaThe low eHf^low eNd component of the Karmutsen
basalts that places them below the OIB mantle array andpartly within the field for age-corrected Pacific MORBcan form from low ratios of LuHf and high SmNd over
time (eg Salters amp White 1998) MORB will develop aHf^Nd isotopic signature over time that falls below themantle array Low LuHf can also lead to low 176Hf177Hffrom remelting of residues produced by melting in theabsence of garnet (eg Salters amp Hart 1991 Salters ampWhite 1998) The Hf^Nd isotopic compositions of theKarmutsen Formation indicate the possible involvementof depleted MORB mantle however similar to Iceland(Fitton et al 2003) Nb^Zr^Y variations provide a way todistinguish between a depleted component within theplume and a MORB-like component The Karmutsenbasalts and picrites fall within the Iceland array and donot overlap the MORB field in a logarithmic plot of NbYvs ZrY (Fig 13b) using the fields established by Fittonet al (1997) Similar to the depleted component within theCaribbean Plateau (Thompson et al 2003) the trend ofHf^Nd isotopic compositions towards Pacific MORB-likecompositions is not followed by MORB values in Nb^Zr^Y systematics and this suggests that the depleted compo-nent in the source of the Karmutsen basalts is distinctfrom the source of MORB and probably an intrinsic partof the plume The relatively radiogenic Pb isotopic ratiosand REE patterns distinct from MORB also support thisinterpretationTrace-element characteristics suggest the possible invol-
vement of sub-arc lithospheric mantle in the generation ofthe Karmutsen picrites Lassiter et al (1995) suggested thatmixing of a plume-type source with eNdthorn 6 to thorn 7 witharc material with low NbTh could reproduce the varia-tions observed in the Karmutsen basalts however theyconcluded that the absence of low NbLa ratios in theirsample of tholeiitic basalts restricts the amount of arc litho-sphere involved The study of Lassiter et al (1995) wasbased on major and trace element and Sr Nd and Pb iso-topic data for a suite of 29 samples from Buttle Lake in theStrathcona Provincial Park on central Vancouver Island(Fig 1) Whereas there are no clear HFSE depletions inthe Karmutsen tholeiitic basalts the low NbLa (and TaLa) of the Karmutsen picritic lavas in this study indicatethe possible involvement of Paleozoic arc lithosphere onVancouver Island (Fig 8)
REE modeling dynamic melting andsource mineralogyThe combined isotopic and trace-element variations of thepicritic and tholeiitic Karmutsen lavas indicate that differ-ences in trace elements may have originated from differentmelting histories For example the LuHf ratios of theKarmutsen picritic lavas (mean 008) are considerablyhigher than those in the tholeiitic basalts (mean 002)however the small range of overlapping initial eHf valuesfor the two lava suites suggests that these differences devel-oped during melting within the plume (Fig 9b) Below wemodel the effects of source mineralogy and depth of
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
28
melting on differences in trace-element concentrations inthe tholeiitic basalts and picritesDynamic melting models simulate progressive decom-
pression melting where some of the melt fraction isretained in the residue and only when the degree of partialmelting is greater than the critical mass porosity and the
source becomes permeable is excess melt extracted fromthe residue (Zou1998) For the Karmutsen lavas evolutionof trace element concentrations was simulated using theincongruent dynamic melting model developed by Zou ampReid (2001) Melting of the mantle at pressures greater than1 GPa involves incongruent melting (eg Longhi 2002)
OIB array
PacificMORB(230 Ma)
(0 Ma)
EPR(230 Ma)
-4
-2
0
2
4
8
10
12
14
16
18
20
22
-4 -2 0 2 4 6 8 10 12 14
minus2
0
2
4
6
8
10
12
14
0702 0703 0704 0705 0706
IndianMORB
87Sr 86Sr (initial)
Hf (initial)
(a) (c)
Nd (initial)
Karmutsen tholeiitic basalt
HawaiiOntong JavaCaribbean Plateau
Karmutsen high-MgO lava
high-MgO lavasKarmutsen
1540
1545
1550
1555
1560
1565
175 180 185 190 195 200375
380
385
390
395
175 180 185 190 195 200
(d) (e)
EPR
EPR
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb
Karmutsen Formation (initial)
HawaiiOntong JavaCaribbean Plateau
East Pacific Rise
Karmutsen Formation (measured)
Juan de FucaGorda
ε Nd
(initial)
Karmutsen Formation (different age-correction for 3 picrites)
(0 Ma)
NbY
001
01
1
1 10
ZrY
OIB
NMORB
(b)
Iceland array
6
Fig 13 Comparison of age-corrected (230 Ma) Sr^Nd^Hf^Pb isotopic compositions for Karmutsen flood basalts onVancouver Island withage-corrected OIB and MORB and Nb^Zr^Y variation (a) Initial eNd vs 87Sr86Sr (b) NbYand ZrY variation (after Fitton et al 1997)(c) Initial eHf vs eNd (d) Measured and initial 207Pb204Pb vs 206Pb204Pb (e) Measured and initial 208Pb204Pb vs 206Pb204Pb References fordata sources are listed in Electronic Appendix 4 Most of the compiled data were extracted from the GEOROC database (httpgeorocmpch-mainzgwdgdegeoroc) OIB array line in (c) is fromVervoort et al (1999) EPR East Pacific Rise Several high-MgO samples affected bysecondary alteration were not plotted for clarity in Pb isotope plots Estimation of fields for age-corrected Pacific MORB and EPR at 230 Mawere made using parent^daughter concentrations (in ppm) from Salters amp Stracke (2004) of Lufrac14 0063 Hffrac14 0202 Smfrac14 027 Ndfrac14 071Rbfrac14 0086 and Srfrac14 836 In the NbYvs ZrYplot the normal (N)-MORB and OIB fields not including Icelandic OIB are from data com-pilations of Fitton et al (1997 2003) The OIB field is data from 780 basalts (MgO45wt ) from major ocean islands given by Fitton et al(2003) The N-MORB field is based on data from the Reykjanes Ridge EPR and Southwest Indian Ridge from Fitton et al (1997) The twoparallel gray lines in (b) (Iceland array) are the limits of data for Icelandic rocks
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
29
with olivine being produced during the reactioncpxthornopxthorn spfrac14meltthornol for spinel peridotites The mod-eling used coefficients for incongruent melting reactionsbased on experiments on lherzolite melting (eg Kinzleramp Grove 1992 Longhi 2002) and was separated into (1)melting of spinel lherzolite (2) two combined intervals ofmelting for a garnet lherzolite (gt lherz) and spinel lher-zolite (sp lherz) source (3) melting of garnet lherzoliteThe model solutions are non-unique and a range indegree of melting partition coefficients melt reactioncoefficients and proportion of mixed source componentsis possible to achieve acceptable solutions (see Fig 14 cap-tion for description of modeling parameters anduncertainties)The LREE-depleted high-MgO lavas require melting of
a depleted spinel lherzolite source (Fig 14a) The modelingresults indicate a high degree of melting (22^25) similarto the results from PRIMELT1 based on major-elementvariations (discussed above) of a LREE-depleted sourceThe high-MgO lavas lack a residual garnet signature Theenriched tholeiitic basalts involved melting of both garnetand spinel lherzolite and represent aggregate melts pro-duced from continuous melting throughout the depth ofthe melting column (Fig 14b) Trace-element concentra-tions in the tholeiitic and picritic lavas are not consistentwith low-degree melting of garnet lherzolite alone(Fig 14c) The modeling results indicate that the aggregatemelts that formed the tholeiitic basalts would haveinvolved lesser proportions of enriched small-degreemelts (1^6 melting) generated at depth and greateramounts of depleted high-degree melts (12^25) gener-ated at lower pressure (a ratio of garnet lherzolite tospinel lherzolite melt of 14 provides the best fits seeFig 14b and description in figure caption) The totaldegree of melting for the tholeiitic basalts was high(23^27) similar to the degree of partial melting in thespinel lherzolite facies for the primary melts that eruptedas the Keogh Lake picrites of a source that was also prob-ably depleted in LREE
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
(c)
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
Melting of garnet lherzolite
High-MgO lavas
Melting of spinel lherzolite
source (07DM+03PM)
source (07DM+03PM)
6 melting
30 melting
(b)
(a)
30 melting
Sam
ple
Cho
ndrit
eS
ampl
eC
hond
rite 20 melting
1 melting
Melting of garnet lherzoliteand spinel lherzolite
x gt lherz + 4x sp lherz
5 meltingx gt lherz + 4x sp lherz
Sam
ple
Cho
ndrit
e
Tholeiitic lavas
Tholeiitic lavas
High-MgO lavas
source (07DM+03PM)
Fig 14 Trace-element modeling results for incongruent dynamicmantle melting for picritic and tholeiitic lavas from the KarmutsenFormation Three steps of modeling are shown (a) melting of spinellherzolite compared with high-MgO lavas (b) melting of garnet lher-zolite and spinel lherzolite compared with tholeiitic lavas (c) meltingof garnet lherzolite compared with tholeiitic and picritic lavasShaded fields in each panel for tholeiitic basalts and picrites arelabeled Patterns with symbols in each panel are modeling results(patterns are 5 melting increments in (a) and (b) and 1 incre-ments in (c) PM primitive mantle DM depleted MORB mantle gtlherz garnet lherzolite sp lherz spinel lherzolite In (b) the ratios ofper cent melting for garnet and spinel lherzolite are indicated (x gtlherzthorn 4x sp lherz means that for 5 melting 1 melt in garnetfacies and 4 melt in spinel facies where x is per cent melting in theinterval of modeling) The ratio of the respective melt proportions inthe aggregate melt (x gt lherzthorn 4x sp lherz) was kept constant andthis ratio of melt contribution provided the best fit to the data Arange of proportion of source components (07DMthorn 03PM to
09DMthorn 01PM) can reproduce the variation in the high-MgOlavas Melting modeling uses the formulation of Zou amp Reid (2001)an example calculation is shown in their Appendix PM fromMcDonough amp Sun (1995) and DM from Salters amp Stracke (2004)Melt reaction coefficients of spinel lherzolite from Kinzler amp Grove(1992) and garnet lherzolite fromWalter (1998) Partition coefficientsfrom Salters amp Stracke (2004) and Shaw (2000) were kept constantSource mineralogy for spinel lherzolite (018cpx027opx052ol003sp) from Kinzler (1997) and for garnet lherzolite(034cpx008opx053ol005gt) from Salters amp Stracke (2004) Arange of source compositions for melting of spinel lherzolite(07DMthorn 03PM to 03DMthorn 07PM) reproduce REE patternssimilar to those of the high-MgO lavas with best fits for 22ndash25melting and 07DMthorn 03PM source composition Concentrationsof tholeiitic basalts cannot be reproduced from an entirelyprimitive or depleted source
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
30
The uncertainty in the composition of the source in thespinel lherzolite melting model for the high-MgO lavasprecludes determining whether the source was moredepleted in incompatible trace elements than the source ofthe tholeiitic lavas (see Fig 14 caption for description) It ispossible that early high-pressure low-degree melting left aresidue within parts of the plume (possibly the hotter inte-rior portion) that was depleted in trace elements (egElliott et al 1991) further decompression and high-degreemelting of such depleted regions could then have generatedthe depleted high-MgO Karmutsen lavas Shallow high-degree melting would preferentially sample depletedregions whereas deeper low-degree melting may notsample more refractory depleted regions (eg CaribbeanEscuder-Viruete et al 2007) The picrites lie at the highend of the range of Nd isotopic compositions and havelow NbZr similar to depleted Icelandic basalts (Fittonet al 2003) therefore the picrites may represent the partsof the mantle plume that were the most depleted prior tothe initiation of melting As Fitton et al (2003) suggestedthese depleted melts may only rarely be erupted withoutcombining with more voluminous less depleted melts andthey may be detected only as undifferentiated small-volume flows like the Keogh Lake picrites within theKarmutsen Formation on Vancouver IslandDecompression melting within the mantle plume initiatedwithin the garnet stability field (35^25 GPa) at highmantle potential temperatures (Tp414508C) and pro-ceeded beneath oceanic arc lithosphere within the spinelstability field (25^09 GPa) where more extensivedegrees of melting could occur
Magmatic evolution of Karmutsentholeiitic basaltsPrimary magmas in large igneous provinces leave exten-sive coarse-grained cumulate residues within or below thecrust these mafic and ultramafic plutonic sequences repre-sent a significant proportion of the original magma thatpartially crystallized at depth (eg Farnetani et al 1996)For example seismic and petrological studies of OntongJava (eg Farnetani et al 1996) combined with the use ofMELTS (Ghiorso amp Sack 1995) indicate that the volcanicsequence may represent only 40 of the total magmareaching the Moho Neal et al (1997) estimated that30^45 fractional crystallization took place for theOntong Java magmas and produced cumulate thicknessesof 7^14 km corresponding to 11^22 km of flood basaltsdepending on the presence of pre-existing oceanic crustand the total crustal thickness (25^35 km) Interpretationsof seismic velocity measurements suggest the presence ofpyroxene and gabbroic cumulates 9^16 km thick beneaththe Ontong Java Plateau (eg Hussong et al 1979)Wrangellia is characterized by high crustal velocities in
seismic refraction lines which is indicative of mafic
plutonic rocks extending to depth beneath VancouverIsland (Clowes et al 1995)Wrangellia crust is 25^30 kmthick beneath Vancouver Island and is underlain by astrongly reflective zone of high velocity and density thathas been interpreted as a major shear zone where lowerWrangellia lithosphere was detached (Clowes et al 1995)The vast majority of the Karmutsen flows are evolved tho-leiitic lavas (eg low MgO high FeOMgO) indicating animportant role for partial crystallization of melts at lowpressure and a significant portion of the crust beneathVancouver Island has seismic properties that are consistentwith crystalline residues from partially crystallizedKarmutsen magmas To test the proportion of the primarymagma that fractionated within the crust MELTS(Ghiorso amp Sack 1995) was used to simulate fractionalcrystallization using several estimated primary magmacompositions from the PRIMELT1 modeling results Themajor-element composition of the primary magmas couldnot be estimated for the tholeiitic basalts because of exten-sive plagthorn cpxthornol fractionation and thus the MELTSmodeling of the picrites is used as a proxy for the evolutionof major elements in the volcanic sequence A pressure of 1kbar was used with variable water contents (anhydrousand 02wt H2O) calculated at the quartz^fayalite^magnetite (QFM) oxygen buffer Experimental resultsfrom Ontong Java indicate that most crystallizationoccurs in magma chambers shallower than 6 km deep(52 kbar Sano amp Yamashita 2004) however some crys-tallization may take place at greater depths (3^5 kbarFarnetani et al 1996)A selection of the MELTS results are shown in Fig 15
Olivine and spinel crystallize at high temperature until15^20wt of the liquid mass has fractionated and theresidual magma contains 9^10wt MgO (Fig 15f) At1235^12258C plagioclase begins crystallizing and between1235 and 11908C olivine ceases crystallizing and clinopyr-oxene saturates (Fig 15f) As expected the addition of clin-opyroxene and plagioclase to the crystallization sequencecauses substantial changes in the composition of the resid-ual liquid FeO and TiO2 increase and Al2O3 and CaOdecrease while the decrease in MgO lessens considerably(Fig 15) The near-vertical trends for the Karmutsen tho-leiitic basalts especially for CaO do not match the calcu-lated liquid-line-of-descent very well which misses manyof the high-CaO high-Al2O3 low-FeO basalts A positivecorrelation between Al2O3CaO and SrNd indicates thataccumulation of plagioclase played a major role in the evo-lution of the tholeiitic basalts (Fig 15g) Increasing thecrystallization pressure slightly and the water content ofthe melts does not systematically improve the match ofthe MELTS models to the observed dataThe MELTS results indicate that a significant propor-
tion of the crystallization takes place before plagioclase(15^20 crystallization) and clinopyroxene (35^45
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
31
crystallization) join the crystallization sequence Theseresults suggest that445 crystallization of the primarymagmas occurred (estimated from the residual liquid percent in Fig 15f) and if the original flood basalt stratigra-phy on Vancouver Island was 6 km thick at least anequivalent amount of complementary plutonic rocksshould have crystallized The Karmutsen basalts and theirplutonic residues thus represent a significant addition ofcrust (perhaps 412 km thickness) previously thickenedduring the formation of the underlying Paleozoic arc
Growth of the Wrangellia oceanic plateauon Vancouver IslandThe extrusive part of the Wrangellia oceanic plateau onVancouver Island was constructed in a three-layersequence of mostly tholeiitic basalts with a restrictedrange of composition (Fig 16) The plateau overliesDevonian to Mississippian arc volcanic sequences andmarine sediments of Mississippian to Permian ages thatare intruded by gabbroic rocks related to the flood basaltsThe volcanic stratigraphy of theWrangellia plateau formed
00
05
10
15
20
25
30
35
40
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
16
0 2 4 6 8 10 12 14 16 18 20
10
11
12
13
14
15
16
17
18
19
20
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
0 2 4 6 8 10 12 14 16 18 20
MgO (wt)
0 10 20 30 40 50
percent of total mass
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
) Mineral proportions
Sp (e)
Ol
CpxPlag
PicriteHigh-MgO basaltTholeiitic basalt
TiO2 (wt)
MgO (wt)
Al2O3 (wt) CaO (wt)
FeO (wt)
MgO(a) (b)
(c) (d)
MgO (wt)MgO (wt)
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
)
Residual liquid ()
1220
1190
Ol + Sp
Plag+Ol+Sp Plag+Cpx+Sp
02 wt H2O
no H2O
Residual liquid and crystallizing phases
(f)
0102030405060708090100
plagioclaseaccumulation
plagioclaseaccumulation
plagioclase
accumulatio
n
10
12
14
16
0 10 20 30 40
Al2O3CaO
SrNd
tholeiitic basalts
(g)
Fig 15 Forward fractional crystallization modeling results for major elements from MELTS (Ghiorso amp Sack 1995) compared with picriticand tholeiitic lavas from the Karmutsen Formation (a^d) Major-element compositions of residual liquids compared with Karmutsen lavacompositions For clarity the results are shown only for one starting composition (open circle) using the composition of the estimated primarymagma for the high-MgO lava series from the modeling technique of Herzberg et al (2007) for sample 4723A13 as shown in Fig 12 The esti-mated primary magmas for samples 4723A3 and 93G171 yield similar results A pressure of 1 kbar was used with no H2O and 02wt H2Ocalculated at QFM black line is anhydrous and grey line is 02wt H2OThe end-point for the anhydrous result is at 78 fractionation and11308C and the end-point for the result with 02wt H2O is 75 fractionation and11108C (e) Mineral proportions of crystallized phases fromone MELTrun (sample 4723A13) with no H2O (f) Per cent residual liquid vs temperature labelled with intervals of crystallizing phases (g)Al2O3CaO vs SrNd for the tholeiitic basalts
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
32
in a single eruptive phase (c 231^225 Ma) as an emergentbasalt sequence during a deeper-water shallow-water andsubaerial stage and subsided below sea level relativelysoon after its formation (Fig 16) The deeper-water stagecontains pillowed and massive submarine flows (3000mthick) that built up the submarine volcanic edifice from theseafloor The lowest pillow basalts were emplaced onunconsolidated and lithified fine-grained siliciclastic andcarbonaceous sediments containing Daonella beds(c 235^232 Ma) that are intruded by abundantKarmutsen mafic sills (Fig 16a) The different pillowedand massive submarine flows resulted primarily from dif-ferent effusive rates and local topography Some of themassive submarine flows commonly recognized by radiat-ing columnar jointing represent master tubes for the deliv-ery of lava to distal parts of submarine flow fields Thepillowed flows which commonly have uneven topography
in modern sequences may have served as levees to themassive submarine flows rather than allowing unob-structed spreading and inflation typical of the massive sub-aerial flows that form on shallow slopes in continentalflood basalt provinces (eg Self et al 1997) The rapid suc-cession of flows is evident by the near absence of sedimentbetween submarine flows Minor volumes of picritic pillowlavas occur in the upper part of the pillowed flow sequencein areas of the plateau exposed on northern VancouverIsland (Fig 16 Nixon et al 2008)The growth of the plateau in shallow water led to
emplacement of an increasing proportion of volcaniclastics(pillow breccias and hyaloclastites) that accumulated up to1500m thick in some areas (Fig 16b Nixon et al 2008)Horizons of pillowed flows occur locally within the volca-niclastic stratigraphy Sedimentary structures (eg gradedbedding fluidization structures) and petrographic textures
Late Devonian-Mississippian arc
Daonella beds
Middle Triassic
deeper water
Late Devonian-Mississippian arc
pillowed flows massive submarine
flows
pillow brecciahyaloclastiteshallow water
~4 km
~6 km
Late Triassic limestoneinterflow limestonelenses
Karmutsen Formation
mafic intrusions sediment-sill complex
Karmutsen Formation
(a) (b)
(c) (d)
subaerial flows
Late Devonian-Mississippian arc Late Devonian-Mississippian arc
picritic pillow basalts
Miss to Permian marine sediments
sills
Fig 16 Growth of the Wrangellia oceanic plateau onVancouver Island (a) Deep-water eruptive phase of pillowed and massive submarineflows with almost no intervening sediments onto MiddleTriassic shale containing Daonella beds (c 235^232 Ma) The basement of the plateauis composed of Devonian to Mississippian arc sequences overlain by Mississippian to Permian carbonates and siliciclastic sediments(b) Shallow-water flow emplacement with increasing proportion of volcaniclastics resulted primarily from cooling^contraction granulationPicritic pillow basalts occur primarily near the transition between pillowed submarine flows and volcaniclastic units Pillowed and massiveflows occur locally within the volcaniclastic unit (c) Emergent stage of growth consists of overlapping subaerial flow fields similar to continen-tal flood basalts with pahoehoe flow structures Multi-tiered columnar jointing occurrs locally in response to ponding of flows (eg Bondre et al2004) Interflow carbonate deposits and plagioclase megacrystic flows developed locally during the waning stage of volcanism (d) Post-volcanicsubsidence and deposition of Late Carnian to Norian (c 212^223 Ma) platform carbonates on top of the plateau Units are labeled in the dia-grams Thicknesses of the underlying units are not to scale
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
33
within the volcaniclastic rocks indicate formation primar-ily via cooling^contraction granulation magma^water^steam interaction autobrecciation resedimentation andmass-wasting rather than pyroclastic eruption (Nixonet al 2008)A broad subaerial platform was constructed by emplace-
ment of subaerial lava flows during an emergent subaerialstage (Fig 16c) The subaerial unit predominantly com-prises well-layered sheet flows that formed in a similarway to continental flood basalts as hundred(s) of pahoehoeflow fields erupted on shallow slopes (Self et al 1997) how-ever volcaniclastic and pillowed flow units are locally pre-sent mostly in the upper parts of the stratigraphyLocalized ponding of flows led to the formation of multi-tiered columnar jointing in some subaerial flows (egBondre et al 2004) As volcanism waned local interflowcarbonate deposits developed and plagioclase megacrysticflows were erupted Subsidence of the plateau is recordedby the interflow lenses and deposits of platform carbonates(40^1500m thick) on top of the Karmutsen basalts thatcontain Late Carnian to Early Norian fossils (Fig 16d c216^227 Ma Nixon et al 2008)The evolution of the internal architecture of the
Wrangellia oceanic plateau was primarily related to theconditions of emplacement (ie water depth and eruptionenvironment) There does not appear to have been a sys-tematic coincident evolution or shift of lava compositionsonVancouver Island as has been recognized in most con-tinental flood basalts and parts of the Wrangellia oceanicplateau in Alaska and Yukon (Greene et al 2008 2009)In general with the exception of the picritic pillow lavasin the submarine stratigraphy and local variations betweenhigh-MgO and tholeiitic lavas there are no significantcompositional differences between lavas erupted duringthe different stages of growth of the plateau (ie submarineand subaerial flows) nor are substantial compositional dif-ferences generally found between massive submarine flowsin contact with pillowed flows tholeiitic basalt flows nearthe base of the Karmutsen Formation are compositionallysimilar to flows sampled near the top of the volcanicsequence Where it was possible to sample continuous sec-tions of volcanic stratigraphy greater than several hun-dreds of meters (eg Keogh^Maynard Lake area west ofthe Karmutsen Range Fig 2) there is no clear relationshipbetween stratigraphic position and compositionalvariation
CONCLUSIONSThe Karmutsen Formation covers extensive areas ofVancouver Island (20 000 km2) and forms part of a largeoceanic plateau erupted over a geologically short interval(c 225^231 Ma) The tripartite volcanic stratigraphyexposed on Vancouver Island is at least 6 km thick andcomprises a succession of submarine flows volcaniclastic
deposits and massive subaerial flowsThese volcanologicaldifferences are primarily related to the eruption environ-ment (deep-water shallow-water or subaerial) The rapidgrowth of the plateau prevented significant accumulationsof intervening sediment except in the uppermost stratigra-phy where isolated limestone lenses commonly associatedwith pillow basalt and hyaloclastite deposits preserve arecord of the final stages of subsidence of the plateauThe Wrangellia plateau on Vancouver Island was con-
structed dominantly of tholeiitic basalt with restrictedmajor- and trace-element and isotopic compositional var-iations Minor volumes of picritic pillow basalts eruptedlate in the submarine phase The Karmutsen basalts wereemplaced onto older Paleozoic arc volcanic sequences andlow NbLa for the picrites may indicate a contributionfrom the sub-arc lithospheric mantle Modeling of the pet-rogenesis of the Karmutsen picrites indicates melting ofanomalously hot mantle (15008C) and extensive degreesof partial melting (23^27) and is consistent with aplume initiation model Combined Sr^Nd^Hf^Pb isotopicsystematics indicate that the picritic and tholeiitic lavaswere derived from a depleted plume-type mantle sourcedistinct from the source of MORB that has compositionalsimilarities to the source of the lavas forming theCaribbean Plateau Picrites were generated by melting ofspinel lherzolite whereas the tholeiitic basalts formed asaggregate melts of both garnet and spinel lherzolite Thetholeiitic basalts underwent extensive low-pressure fractio-nation (52^3 kbar) and seismic studies indicate that someof theWrangellia crust beneathVancouver Island may cor-respond to the plutonic residues of the Karmutsen floodbasalts
ACKNOWLEDGEMENTSWe are grateful to Nick Arndt for helping us get this pro-ject started and Nick Massey for insights into VancouverIsland geology We greatly appreciate the assistance withPRIMELT modeling from Claude Herzberg We thankMikkel Schau for his insight and enthusiasm during field-work Jane Barling assisted with analyses by MC-ICP-MSThoughtful and thorough reviews of the manuscript byJohn Mahoney Fred Frey Wendy Bohrson DougalJerram and Teal Riley are greatly appreciated Fundingwas generously provided by the Rocks to Riches Programadministered by the BC amp Yukon Chamber of Mines in2004 by the BC Geological Survey and by NSERCDiscovery Grants to JSS and DW ARG was supportedby a University Graduate Fellowship at UBC
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
34
REFERENCESBarker F Brown A S Budahn J R amp Plafker G (1989) Back-arc
with frontal-arc component origin of Triassic Karmutsen basaltBritish Columbia Canada Chemical Geology 75 81^102
Blichert-Toft JWeis D Maerschalk C Agranier A amp Albare de F(2003) Hawaiian hot spot dynamics as inferred from the Hf and Pbisotope evolution of Mauna Kea volcano Geochemistry GeophysicsGeosystems 4(2) 1^27 doi1010292002GC000340
Bondre N R Duraiswami R A amp Dole G (2004) Morphologyand emplacement of flows from the Deccan Volcanic ProvinceIndia Bulletin of Volcanology 66 29^45
Carlisle D (1963) Pillow breccias and their aquagene tuffs QuadraIsland British Columbia Journal of Geology 71 48^71
Carlisle D (1972) Late Paleozoic to Mid-Triassic sedimentary-volcanic sequence on Northeastern Vancouver Island Geological
Society of Canada Paper Report of Activities 72-1 Part B 24^30Carlisle D amp SuzukiT (1974) Emergent basalt and submergent car-
bonate^clastic sequences including the Upper Triassic Dilleri andWelleri zones on Vancouver Island Canadian Journal of Earth
Sciences 11 254^279Chauvel C amp Blichert-Toft J (2001) A hafnium isotope and trace
element perspective on melting of the depleted mantle Earth and
Planetary Science Letters 190 137^151Chayes F (1954)The theory of thin-section analysis Journal of Geology
62 92^101Cho M Liou J G amp Maruyama S (1987) Transition from the zeo-
lite to prehnite^pumpellyite facies in the Karmutsen metabasitesVancouver Island British Columbia Journal of Petrology 27(2)467^494
Clowes R M Zelt C A Amor J R amp Ellis R M (1995)Lithospheric structure in the southern Canadian Cordillera froma network of seismic refraction lines Canadian Journal of Earth
Sciences 32(10) 1485^1513Coffin M F amp Eldholm O (1994) Large igneous provinces Crustal
structure dimensions and external consequences Reviews of
Geophysics 32(1) 1^36Cox K G (1980) A model for flood basalt volcanism Journal of
Petrology 21 629^650Eldholm O amp Coffin M F (2000) Large igneous provinces
and plate tectonics In Richards M A Gordon R G amp vander Hilst R D (eds) The History and Dynamics of Global
Plate Motions American Geophysical Union Geophysical Monograph 121309^326
Elliott T R Hawkesworth C J amp Grolaquo nvold K (1991) Dynamicmelting of the Iceland plume Nature 351 201^206
Escuder-Viruete J Pecurren rez-Estaucurren n A Contreras F Joubert MWeis D Ullrich T D amp Spadea P (2007) Plume mantle sourceheterogeneity through time Insights from the Duarte ComplexHispaniola northeastern Caribbean Journal of Geophysical Research112(B04203) doi1010292006JB004323
Farnetani C G Richards M A amp Ghiorso M S (1996)Petrological models of magma evolution and deep crustal structurebeneath hotspots and flood basalt provinces Earth and Planetary
Science Letters 143 81^94Fitton J G amp Godard M (2004) Origin and evolution of magmas
on the Ontong Java Plateau In Fitton J G Mahoney J JWallace P J amp Saunders A D (eds) Origin and Evolution of the
Ontong Java Plateau Geological Society London Special Publications 229151^178
Fitton J G Saunders A D Norry M J Hardarson B S ampTaylor R N (1997) Thermal and chemical structure of theIceland plume Earth and Planetary Science Letters 153 197^208
Fitton J G Saunders A D Kempton P D amp Hardarson B S(2003) Does depleted mantle form an intrinsic part of the Icelandplume Geochemistry Geophysics Geosystems 4(3) 1032 doi1010292002GC000424
Frey F A Huang S Blichert-Toft J Regelous M amp Boyet M(2005) Origin of depleted components in basalt related to theHawaiian hotspot Evidence from isotopic and incompatible ele-ment ratios Geochemistry Geophysics Geosystems 6(1) 23 doi1010292004GC000757
Galer S J G amp Abouchami W (1998) Practical application of leadtriple spiking for correction of instrumental mass discriminationMineralogical Magazine 62A 491^492
Ghiorso M S amp Sack R O (1995) Chemical mass transfer in mag-matic processes IV A revised and internally consistent thermody-namic model for the interpolation and extrapolation of liquid^solidequilibria in magmatic systems at elevated temperatures and pres-sures Contributions to Mineralogy and Petrology 119 197^212
Greene A R Scoates J S amp Weis D (2005)WrangelliaTerrane onVancouver Island Distribution of flood basalts with implicationsfor potential Ni^Cu^PGE mineralization In Grant B (ed)Geological Fieldwork 2004 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2005-1 209^220Greene A R Scoates J S Nixon G T amp Weis D (2006) Picritic
lavas and basal sills in the Karmutsen flood basalt provinceWrangellia northern Vancouver Island In Grant B (ed)Geological Fieldwork 2005 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2006-1 39^52Greene A R Scoates J S amp Weis D (2008) Wrangellia flood
basalts in Alaska A record of plume^lithosphere interaction in aLate Triassic accreted oceanic plateau Geochemistry Geophysics
Geosystems 9(Q12004) doi1010292008GC002092Greene A R Scoates J S Weis D amp Israel S (2009)
Geochemistry of triassic flood basalts from the Yukon (Canada)segment of the accreted Wrangellia oceanic plateau Lithosdoi101016jlithos200811010
Hauff F Hoernle K Tilton G Graham D W amp Kerr A C(2000) Large volume recycling of oceanic lithosphere over shorttime scales geochemical constraints from the Caribbean LargeIgneous Province Earth and Planetary Science Letters 174 247^263
Hauff F Hoernle K amp Schmidt A (2003) Sr^Nd^Pb compositionof Mesozoic Pacific oceanic crust (Site 1149 and 801 ODP Leg185)Implications for alteration of ocean crust and the input into theIzu^Bonin^Mariana subduction system Geochemistry Geophysics
Geosystems 4(8) doi1010292002GC000421Herzberg C (2004) Partial melting below the Ontong Java Plateau
In Fitton J G Mahoney J J Wallace P J amp Saunders A D(eds) Origin and Evolution of the Ontong Java Plateau Geological SocietyLondon Special Publications 229 179^183
Herzberg C (2006) Petrology and thermal structure of the Hawaiianplume from Mauna Kea volcano Nature 444(7119) 605
Herzberg C amp Asimow P D (2008) Petrology of some oceanicisland basalts PRIMELT2XLS software for primary magma cal-culation Geochemistry Geophysics Geosystems 9(Q09001) doi1010292008GC002057
Herzberg C amp OrsquoHara M J (2002) Plume-associated ultramaficmagmas of Phanerozoic age Journal of Petrology 43(10) 1857^1883
Herzberg C Asimow P D Arndt N Niu Y Lesher C MFitton J G Cheadle M J amp Saunders A D (2007)Temperatures in ambient mantle and plumes Constraints frombasalts picrites and komatiites Geochemistry Geophysics Geosystems8(Q02006) doi1010292006GC001390
Huang S amp Frey F A (2005) Trace element abundances of MaunaKea basalt from phase 2 of the Hawaii Scientific Drilling Project
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
35
Petrogenetic implications of correlations with major element con-tent and isotopic ratios Geochemistry Geophysics Geosystems 4(6)doi1010292002GC000322
Hussong D M Wipperman L K amp Kroenke L W (1979) Thecrustal structure of the Ontong Java and Manihiki OceanicPlateaus Journal of Geophysical Research 84(B11) 6003^6010
Jerram D A amp Widdowson M (2005) The anatomy of ContinentalFlood Basalt Provinces geological constraints on the processes andproducts of flood volcanism Lithos 79(3^4) 385^405
Jones D L Silberling N J amp Hillhouse J (1977)Wrangellia a dis-placed terrane in northwestern North America CanadianJournal ofEarth Sciences 14(11) 2565^2577
Kerr A C (2003) Oceanic Plateaus In Rudnick R (ed) The Crust
Treatise on GeochemistryVol 3 Oxford Elsevier Science pp 537^565Kerr A C amp Mahoney J J (2007) Oceanic plateaus Problematic
plumes potential paradigms Chemical Geology 241 332^353Kerr A CTarney J Marriner G F Klaver GT Saunders A D
amp Thirlwall M F (1996) The geochemistry and petrogenesis ofthe Late-Cretaceous picrites and basalts of Curacao NetherlandsAntilles a remnant of an oceanic plateau Contributions to
Mineralogy and Petrology 124(1) 29^43Kerr A C Marriner G F Tarney J Nivia A Saunders A D
Thirlwall M F amp Sinton A C (1997) Cretaceous basaltic ter-ranes in Western Colombia Elemental chronological and Sr-Ndisotopic constraints on petrogenesis Journal of Petrology 38(6)677^702
Kinzler R J (1997) Melting of mantle peridotite at pressuresapproaching the spinel to garnet transition Application to mid-ocean ridge basalt petrogenesis Journal of Geophysical Research
102(B1) 853^874Kinzler R J amp Grove T L (1992) Primary magmas of mid-ocean
ridge basalts 1 Experiments and methods Journal of Geophysical
Research 97(B5) 6885^6906Korenaga J (2005) Why did not the Ontong Java Plateau form sub-
aerially Earth and Planetary Science Letters 234 385^399Kuniyoshi S (1972) Petrology of the Karmutsen (Volcanic) Group
northeastern Vancouver Island British Columbia PhD disserta-tion University of California Los Angeles 242 pp
Lassiter J C DePaolo D J amp Mahoney J J (1995) Geochemistry ofthe Wrangellia flood basalt province Implications for the role ofcontinental and oceanic lithosphere in flood basalt genesis Journalof Petrology 36(4) 983^1009
LeBas M J (2000) IUGS reclassification of the high-Mg and picriticvolcanic rocks Journal of Petrology 41(10) 1467^1470
Longhi J (2002) Some phase equilibrium systematics of lherzolitemelting I Geochemistry Geophysics Geosystems 3(3) doi1010292001GC000204
MacDonald G A amp Katsura T (1964) Chemical composition ofHawaiian lavas Journal of Petrology 5 82^133
Mahoney J J (1987) An isotopic survey of Pacific oceanic plateausimplications for their nature and origin In Keating B HFryer P Batiza R amp Boehlert GW (eds) Seamounts Islands andAtolls American Geophysical Union Geophysical Monograph 43 207^220
Mahoney J LeRoex A P Peng Z Fisher R L amp Natland J H(1992) Southwestern limits of Indian Ocean Ridge mantle and theorigin of low 206Pb204Pb mid-ocean ridge basalt isotope systema-tics of the central Southwest Indian Ridge (178^508E) Journal ofGeophysical Research 97(B13) 19771^19790
Mahoney J J Sinton J M Kurz M D MacDougall J DSpencer K J amp Lugmair GW (1994) Isotope and trace elementcharacteristics of a super-fast spreading ridge East Pacific rise13^238S Earth and Planetary Science Letters 121 173^193
Massey N W D (1995a) Geology and mineral resources of theAlberni^Nanaimo Lakes sheet Vancouver Island 92F1W 92F2Eand part of 92F7E British Columbia Ministry of Energy Mines and
Petroleum Resources Paper 1992-2 132 ppMassey N W D (1995b) Geology and mineral resources of the
Cowichan Lake sheet Vancouver Island 92C16 British Columbia
Ministry of Energy Mines and Petroleum Resources Paper 1992-3 112 ppMassey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005a) Digital Geology Map of British Columbia Tile NM9 MidCoast BC British Columbia Ministry of Energy and Mines Geofile
2005-2Massey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005b) Digital Geology Map of British Columbia Tile NM10Southwest BC British Columbia Ministry of Energy and Mines Geofile
2005-3McDonough W F amp Sun S (1995) The composition of the Earth
Chemical Geology 120 223^253Monger JW H amp Journeay J M (1994) Basement geology and tec-
tonic evolution of theVancouver region In Monger JW H (ed)Geology and Geological Hazards of the Vancouver Region Southwestern
British Columbia Geological Survey of Canada Bulletin 481 3^25Muller J E (1977) Geology of Vancouver Island Geological Survey of
Canada Open File Map 463Muller J E (1980)The Paleozoic Sicker Group of Vancouver Island British
Columbia Geological Survey of Canada Paper 79-30 22 ppMuller J E Northcote K E amp Carlisle D (1974) Geology and
mineral deposits of Alert Bay^Cape Scott map area VancouverIsland British Columbia Geological Survey of Canada Paper 74-8 77
Neal C R Mahoney J J Kroenke L W Duncan R A ampPetterson M G (1997) The Ontong^Java Plateau InMahoney J J amp Coffin M F (eds) Large Igneous Provinces
Continental Oceanic and Planetary FloodVolcanism American Geophysical
Union Geophysical Monograph 100 183^216Niu Y Collerson K D Batiza R Wendt J I amp Regelous M
(1999) Origin of enriched-typed mid-ocean ridge basalt at ridgesfar from mantle plumes The East Pacific Rise at 11820rsquoN Journalof Geophysical Research 104 7067^7088
Nixon G T amp Orr A J (2007) Recent revisions to the EarlyMesozoic stratigraphy of Northern Vancouver Island (NTS 102I092L) and metallogenic implicatinos British Columbia InGrant B (ed) Geological Fieldwork 2006 British Columbia Ministry of
Energy Mines and Petroleum Resources Paper 2007-1 163^177Nixon GT Hammack J L KoyanagiV M Payie G J Haggart
JW Orchard M JTozerT Friedman R M Archibald D APalfy J amp Cordey F (2006a) Geology of the Quatsino^PortMcNeill area northernVancouver Island British Columbia Ministry
of Energy Mines and Petroleum Resources Geoscience Map 2006-2 scale150 000
Nixon G T Kelman M C Stevenson D Stokes L A ampJohnston K A (2006b) Preliminary geology of the Nimpkishmap area (NTS 092L07) northern Vancouver Island BritishColumbia In Grant B amp Newell J M (eds) Geological Fieldwork2005 British Columbia Ministry of Energy Mines and Petroleum Resources
Paper 2006-1 135^152Nixon G T Laroque J Pals A Styan J Greene A R amp
Scoates J S (2008) High-Mg lavas in the Karmutsen flood basaltsnorthernVancouver Island (NTS 092L) Stratigraphic setting andmetallogenic significance In Grant B (ed) Geological Fieldwork
2007 British Columbia Ministry of Energy Mines and Petroleum
Resources Paper 2008-1 175^190Nowell G M Kempton P D Noble S R Fitton J G
Saunders A Mahoney J J amp Taylor R N (1998) High precisionHf isotope measurements of MORB and OIB by thermal ionisation
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
36
mass spectrometry insights into the depleted mantle Chemical
Geology 149 211^233Parrish R R amp McNicoll V J (1992) U^Pb age determinations
from the southern Vancouver Island area British Columbia InRadiogenic Age and Isotopic Studies Report 5 Geological Survey of
Canada Paper 91-2 79^86Peate DW (1997) The Parana^Etendeka Province In Coffin M F
amp Mahoney J (eds) Large Igneous Provinces Continental Oceanic andPlanetary Flood Volcanism American Geophysical Union Geophysical
Monograph 100 217^245Perfit M R Fornari D J Ridley W I Kirk P D Casey J
Kastens K A Reynolds J R Edwards M Desonie DShuster R amp Paradis S (1996) Recent volcanism in the Siqueirostransform fault picritic basalts and implications for MORBmagma genesis Earth and Planetary Science Letters 141(1^4) 91^108
Pretorius W Weis D Williams G Hanano D Kieffer B ampScoates J S (2006) Complete trace elemental characterization ofgranitoid (USGSG-2GSP-2) reference materials by high resolutioninductively coupled plasma-mass spectrometry Geostandards and
Geoanalytical Research 30(1) 39^54Regelous M Niu Y Wendt J I Batiza R Greig A amp
Collerson K D (1999) Variations in the geochemistry of magma-tism on the East Pacific Rise at 10830rsquoN since 800 ka Earth and
Planetary Science Letters 168 45^63Recurren villon S Chauvel C Arndt N T Pik R Martineau F
Fourcade S amp Marty B (2002) Heterogeneity of the Caribbeanplateau mantle source Sr O and He isotopic compositions of oli-vine and clinopyroxene from Gorgona Island Earth and Planetary
Science Letters 205(1^2) 91Rhodes J M (1996) Geochemical stratigraphy of lava flows samples
by the Hawaii Scientific Drilling Project Journal of Geophysical
Research 101(B5) 11729^11746Rhodes J M amp Vollinger M J (2004) Composition of basaltic lavas
sampled by phase-2 of the Hawaii Scientific Drilling ProjectGeochemical stratigraphy and magma types Geochemistry
Geophysics Geosystems 5(3) doi1010292002GC000434Richards M A Jones D L Duncan R A amp DePaolo D J (1991)
A mantle plume initiation model for the Wrangellia flood basaltand other oceanic plateaus Science 254 263^267
Salters V J M amp Hart S R (1991) The mantle sources of oceanridges islands and arcs the Hf-isotope connection Earth and
Planetary Science Letters 104 364^380Salters V J M amp Stracke A (2004) Composition of the depleted
SaltersV J amp WhiteW (1998) Hf isotope constraints on mantle evo-lution Chemical Geology 145 447^460
SanoT amp Yamashita S (2004) Experimental petrology of basementlavas from Ocean Drilling Program Leg 192 implications for dif-ferentiation processes in Ontong Java Plateau magmas InFitton J G Mahoney J JWallace P J amp Saunders A D (eds)Origin and Evolution of the Ontong Java Plateau Geological Society
London Special Publications 229 185^218Saunders A D (2005) Large igneous provinces Origin and environ-
mental consequences Elements 1 259^263Self S ThordarsonT amp Keszthely L (1997) Emplacement of conti-
nental flood basalt lava flows In Mahoney J J amp Coffin M F(eds) Large Igneous Provinces Continental Oceanic and Planetary Flood
Volcanism American Geophysical Union Geophysical Monograph 100381^410
Shaw D M (2000) Continuous (dynamic) melting theory revisitedCanadian Mineralogist 38(5) 1041^1063
Sluggett C L (2003) Uranium^lead age and geochemical constraintson Paleozoic and Early Mesozoic magmatism in WrangelliaTerrane Saltspring Island British Columbia BSc thesisUniversity of British ColumbiaVancouver 84 pp
Storey M Mahoney J J amp Saunders A D (1997) Cretaceousbasalts in Madagascar and the transition between plume and con-tinental lithospheric mantle sources In Mahoney J J ampCoffin M F (eds) Large Igneous Provinces Continental Oceanic and
Planetary Flood Volcanism Aamerican Geophysical Union Geophysical
Monograph 100 95^122Surdam R C (1967) Low-grade metamorphism of the Karmutsen
Group PhD dissertation University of California Los Angeles288 pp
Sutherland-Brown A Yorath C J Anderson R G amp Dom K(1986) Geological maps of southern Vancouver IslandLITHOPROBE I 92C10 11 14 16 92F1 2 7 8 Geological Survey ofCanada Open File 1272
Tejada M L G Mahoney J J Castillo P R Ingle S P Sheth HC amp Weis D (2004) Pin-pricking the elephant evidence on theorigin of the Ontong Java Plateau from Pb^Sr^Hf^Nd isotopiccharacteristics of ODP Leg 192 basalts In Fitton J GMahoney J J Wallace P J amp Saunders A D (eds) Origin and
Evolution of the Ontong Java Plateau Geological Society London Special
Publications 229 133^150Thompson P M E Kempton P D amp Kerr A C (2008) Evaluation
of the effects of alteration and leaching on Sm^Nd and Lu^Hf sys-tematics in submarine mafic rocks Lithos 104(1^4) 164^176
Thompson P M E Kempton P D White R V Kerr A CTarney J Saunders A D Fitton J G amp McBirney A (2003)Hf-Nd isotope constraints on the origin of the CretaceousCaribbean plateau and its relationship to the Galapagos plumeEarth and Planetary Science Letters 217(1^2) 59^75
Vervoort J D Patchett P Blichert-Toft J amp Albare de F (1999)Relationships between Lu^Hf and Sm^Nd isotopic systems in theglobal sedimentary system Earth and Planetary Science Letters 16879^99
Walter M J (1998) Melting of garnet peridotite and the origin ofkomatiite and depleted lithosphere Journal of Petrology 39(1) 29^60
Weis D Kieffer B Maerschalk C Barling J de Jong JWilliams G A Hanano D Mattielli N Scoates J SGoolaerts A Friedman R A amp Mahoney J B (2006) High-pre-cision isotopic characterization of USGS reference materials byTIMS and MC-ICP-MS Geochemistry Geophysics Geosystems
7(Q08006) doi1010292006GC001283Weis D Kieffer B Hanano D Silva I N Barling J
Pretorius W Maerschalk C amp Mattielli N (2007) Hf isotopecompositions of US Geological Survey reference materialsGeochemistry Geophysics Geosystems 8(Q06006) doi1010292006GC001473
Wheeler J O amp McFeely P (1991) Tectonic assemblage map of theCanadian Cordillera and adjacent part of the United States ofAmerica Geological Survey of Canada Map 1712A
WhiteW M Albare de F amp Tecurren louk P (2000) High-precision analy-sis of Pb isotope ratios by multi-collector ICP-MS Chemical Geology167 257^270
Yorath C J Sutherland Brown A amp Massey N W D (1999)LITHOPROBE southern Vancouver Island British Columbia Geological
Survey of Canada Bulletin 498 145 ppZou H (1998) Trace element fractionation during modal and
non-model dynamic melting and open-system melting Amathematical treatment Geochimica et Cosmochimica Acta 62(11)1937^1945
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
37
Zou H amp Reid M R (2001) Quantitative modeling of trace elementfractionation during incongruent dynamic melting Geochimica et
Cosmochimica Acta 65(1) 153^162
APPENDIX SAMPLEPREPARAT ION ANDANALYT ICAL METHODSOnly the freshest rocks in the field were sampled and onlythe least altered samples were selected for chemical andisotopic analysis based on thorough petrographic inspec-tion Fifty-nine of the 129 collected samples were crushed(400 g) into small pieces 52mm in diameter in aRocklabs hydraulic piston crusher between WC plates tominimize contamination The coarse-crush was mixedand 100 g was powdered in a planetary mill using agatejars and balls cleaned with quartz sand between samples
ActLabs analytical methodsThe major- and trace-element compositions of the whole-rock powders were determined at Activation LaboratoriesLtd (Actlabs) in Ancaster Ontario Analytical techniquesand detection limits are also available from Actlabs (httpwwwactlabscommethsub_code4erehtm) The analyticalmethod for each of the elements analyzed is indicated inTable 2 For the major elements a 02 g sample was mixedwith a lithium metaborate^lithium tetraborate mixtureand fused in a graphite crucible The molten mixture waspoured into a 5 HNO3 solution and shaken until dis-solved (30min) The samples were analyzed for majoroxides and selected trace elements by inductively coupledplasma optical emission spectrometry (ICP-OES) on acombination simultaneous^sequential Thermo Jarrell-AshEnviro II system Internal calibration was achieved usinga variety of international reference materials (eg W-2BIR-1 DNC-1) and independent control samplesAdditional trace elements were analyzed by both instru-mental neutron activation analysis (INAA) and induc-tively couple plasma mass spectrometry (ICP-MS) ForINAA15^25 g of sample was weighed into small polyeth-ylene vials and irradiated with a control international ref-erence material CANMET WMS-1 and NiCr flux wires ata thermal neutron flux of 71012 neutronscm2 per s in theMcMaster Nuclear Reactor Following a 7 day waitingperiod the samples were measured on an Ortec high-purity Ge detector linked to a Canberra Series 95 multi-channel analyzer Activities for each element weredecay- and weight-corrected and compared with a detectorcalibration developed from multiple international certifiedreference materials For ICP-MS 025 g of sample wasdigested in HF followed by HNO3^HClO4 heated andtaken to dryness The samples were brought back into solu-tion with HCl Samples were analyzed using a PerkinElmer Optima 3000 ICP In-lab standards or certified ref-erence materials (eg W-2 BIR-1 DNC-1) were used for
quality control A total of 15 blind duplicates were ana-lyzed from the same sample powders to assessreproducibility
University of Massachusetts analyticalmethodsFourteen sample duplicate powders taken from the same100 g sample aliquot (high-MgO basalts and picrites)were also analyzed at the Ronald B Gilmore X-RayFluorescence (XRF) Laboratory at the University ofMassachusetts Major elements were measured on a fusedLa-bearing lithium borate glass disc using a SiemensMRS-400 spectrometer Trace element concentrations(Rb Sr Ba Ce Nb Zr Y Pb Zn Ga Ni Cr V) weremeasured on a separate powder pellet using a PhilipsPW2400 sequential spectrometer using a Rh tube LOIand ferrous iron measurements were made as describedby Rhodes amp Vollinger (2004) Precision and accuracy esti-mates for the data have been described by Rhodes (1996)and Rhodes amp Vollinger (2004) Results for each sampleare the average of two separate analyses and are shown inElectronic Appendix 2 for a comparison of analyses ofduplicate powders using XRFand ICP-MS
PCIGR analytical methodsA subset of 19 samples was selected for high-precisiontrace-element analysis and Sr Nd Pb and Hf isotopicanalysis at the Pacific Centre for Isotopic andGeochemical Research (PCIGR) at the University ofBritish Columbia (UBC) Samples were selected from the59 samples analyzed for whole-rock chemistry at ActLabsbased on major- and trace-element chemistry alteration(low LOI and petrographic alteration index) and samplelocation Samples were prepared for trace-element analysisat the PCIGR by the technique described by Pretorius et al(2006) on unleached rock powders Sample powders(100mg) were weighed in 7ml screw-top Savillex bea-kers and dissolved in 1ml 14N HNO3 and 5ml 48 HFon a hotplate for 48 h at 1308C with periodic ultrasonica-tion Samples were dried and redissolved in 6ml 6N HClon a hotplate for 24 h and then dried and redissolved in1ml concentrated HNO3 for 24 h before final dryingTrace element abundances were measured by high-resolution inductively coupled plasma-mass spectrometry(HR-ICP-MS) with aThermo Finnigan Element2 systemfollowing the procedures described by Pretorius et al(2006) within 24 h of redissolution HFSE and LILE weremeasured in medium-resolution mode at 2000 dilutionusing a PFATeflon spray chamber washed with aqua regiafor 3min between samples REE were measured in high-resolution mode and U and Pb in low-resolution modeat 2000 dilution using a glass spray chamber washedwith 2 HNO3 between samples Total procedural blanksand reference materials (BCR-2 BHVO-2) were analyzedwith the batch of samples Indium was used as an internal
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
38
standard in all samples and standard solutionsBackground and standard solutions were analyzed afterevery five samples to detect memory effects and massdrift Results for PCIGR trace-element analyses areshown in Electronic Appendix 3 for the full set of analyzedtrace elementsSample digestion for purification of Sr Nd Hf and Pb
for column chemistry began by weighing each samplepowder (400^500mg) prior to leaching All samples wereleached with 6N HCl and placed in an ultrasonic bath for15min Samples were rinsed twice with 18 M-cm H2Obetween each leaching step (15 total) until the supernatantliquid was clear [following the method of Mahoney (1987)]Samples were then dried on a hotplate for 24 h andweighed again Sample solutions were then prepared bydissolving 100^250mg of the leached powder dissolvedin 1ml 14N HNO3 and 10ml 48 HF on a hotplate for48 h at 1308C with periodic ultrasonication Samples weredried and redissolved in 6ml 6N HCl on a hotplate for24 h and then dried Pb was separated using anion exhangecolumns and the discard was used for Sr REE and Hfseparation Nd was separated from the other REE and Hfrequired two additional purification steps Detailed proce-dures for column chemistry for separating Sr Nd and Pbat the PCIGR have been described byWeis et al (2006) andHf purification has been described byWeis et al (2007) Srand Nd isotope ratios were measured by thermal ioniza-tion mass spectrometry (TIMS) on a Thermo FinniganTriton system in static mode with relay matrix rotation ona single Ta and double Re^Ta filament respectively Fourto five filaments per barrel of 21 were occupied by stan-dards (NBS 987 for Sr and LaJolla for Nd) for each barrelwhere samples were run Sample Sr and Nd isotopic com-positions were corrected for mass fractionation using86Sr88Srfrac14 01194 and 146Nd144Ndfrac14 07219 Each samplewas then normalized using the barrel average of the refer-ence material relative to the values of143Nd144Ndfrac14 0511858 and 87Sr86Srfrac14 0710248 (Weiset al 2006) During the course of the Vancouver Islandanalyses the La Jolla Nd standard gave an average valueof 05118566 (nfrac14 7) and NBS987 standard gave an aver-age of 07102408 (nfrac1411) (2 error is reported as times106) 147Sm144Nd errors are 15 or 0006 USGeological Survey (USGS) reference material BHVO-2was processed with the samples and yielded Sr and Nd iso-topic ratios of 07034607 and 05129786 respectivelyThese are in agreement with the published values of070347920 and 051298411 respectively (Weis et al2006)
Pb and Hf isotopic compositions were analyzed by mul-tiple collector-inductively coupled plasma-mass spectrome-try (MC-ICP-MS) on a static multi-collection on a NuPlasma (Nu Instruments) system The detailed analyticalprocedure for Pb isotopic analyses on the Nu Plasma atthe PCIGR has been described by Weis et al (2006) Theconfiguration for Pb analyses allows for collection of PbTl and Hg togetherTl and Hg are used to monitor instru-mental mass discrimination and isobaric overlap repec-tively All sample solutions were analyzed withapproximately the same PbTl ratio (4) as the referencematerial NIST SRM 981 To accomplish this a small ali-quot of each sample solution from the Pb columns was ana-lyzed on the Element2 to determine the precise amount ofPb available for analysis on the Nu Plasma The SRM 981standard was run after every two samples on the NuPlasma During the time samples were run analyses ofthe SRM 981 Pb reference material gave values of206Pb204Pbfrac1416940322 207Pb204Pbfrac1415495823and 208Pb204Pbfrac14 36713164 (nfrac14 61 2 error is reportedas times 104) these values are in excellent agreement withreported TIMS triple-spike values of Galer amp Abouchami(1998) Fractionation-corrected Pb isotopic ratios were fur-ther corrected by the sample-standard bracketing methodor the ln^ln correction method described by White et al(2000) and Blichert-Toft et al (2003) Leached powders ofUSGS reference material BHVO-2 yielded Pb isotopicratios of 206Pb204Pbfrac14186454 8 207Pb204Pbfrac14 1549105 and 208Pb204Pbfrac14 38222514 and BCR-2yielded 206Pb204Pbfrac141880466 207Pb204Pbfrac14 1562518 and 208Pb204Pbfrac14 3883496 (2 error isreported as times 104) These values are in agreement withthose for leached residues of BHVO-2 and BCR-2 fromWeis et al (2006)Hf isotopic compositions were analyzed following the
procedures detailed byWeis et al (2007) The configurationfor Hf analyses monitored Lu mass 175 andYb mass 172 toallow for interference correction to masses 174 and 176 Hfisotopic ratios were normalized internally for mass fractio-nation to a 179Hf177Hf ratio of 07325 using an exponentialcorrection Standards were run after every two samplesand sample results were normalized to the ratio of the in-run daily average and a 176Hf177Hf ratio for JMC-475 of0282160 During the course of analyses the Hf standardJMC-475 gave an average value 02821679 (nfrac14130)USGS reference materials BCR-2 and BHVO-2 were pro-cessed with the samples and yielded Hf isotopic ratios of02828675 and 02831005 respectively Publishedvalues for BCR-2 and BHVO-2 are 02828717 and0283104 8 respectively (Weis et al 2007)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
39
Table 2 Major element (wt oxide) and trace element (ppm) abundances in whole-rock samples of Karmutsen basalts
1Ni concentrations for these samples were determined by XRFTHOL tholeiitic basalt PIC picrite HI-MG high MgO basalt CG coarse-grained (sill or gabbro) MIN SIL mineralizedsill OUTLIER anomalous pillowed flow in plots MA Mount Arrowsmith SL Schoen Lake KR Karmutsen Range QIQuadra Island Sample locations are given using the Universal Transverse Mercator (UTM) coordinate system (NAD83zones 9 and 10) Analyses were performed at Activation Laboratory (ActLabs) Fe2O3
is total iron expressed as Fe2O3LOI loss on ignition All major elements Sr V and Y were measured by ICP quadrupole OES on solutions of fusedsamples Cu Ni Pb and Zn were measured by total dilution ICP Cs Ga Ge Hf Nb Rb Ta Th U Zr and REE weremeasured by magnetic-sector ICP on solutions of fused samples Co Cr and Sc were measured by INAA Blanks arebelow detection limit See Electronic Appendices 2 and 3 for complete XRF data and PCIGR trace element datarespectively Major elements for sample 93G17 are normalized anhydrous Sample 5615A7 was analyzed by XRFSamples from Quadra Island are not shown in figures
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
17
picritic and high-MgO basalt pillow lavas extend to20wt MgO (Fig 6 and references listed in thecaption)The tholeiitic basalts from the Karmutsen Range dis-
play a tight range of parallel light rare earth element(LREE)-enriched REE patterns (LaYbCNfrac14 21^24mean 2202) whereas the picrites and high-MgObasalts are characterized by sub-parallel LREE-depleted
patterns (LaYbCNfrac14 03^07 mean 06 02) with lowerREE abundances than the tholeiitic basalts (Fig 7)Tholeiitic basalts from the Schoen Lake and MountArrowsmith areas have similar REE patterns tosamples from the Karmutsen Range (Schoen Lake LaYbCNfrac14 21^26 mean 2303 Mount Arrowsmith LaYbCNfrac1419^25 mean 2204) and the coarse-grainedmafic rocks have similar REE patterns to the tholeiitic
Depleted MORB average(Salters amp Stracke 2004)
Schoen LakeSchoen Lake
Mount Arrowsmith
Karmutsen RangeS
ampl
eC
hond
rite
Sam
ple
Cho
ndrit
e
Sam
ple
Prim
itive
Man
tleS
ampl
eP
rimiti
ve M
antle
Sam
ple
Prim
itive
Man
tle
Mount Arrowsmith
Karmutsen Range
(a) (b)
(c) (d)
(e) (f )
Sam
ple
Cho
ndrit
e
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained mafic rock
Pillowed flowMineralized sill
1
10
100
1
10
100
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
01
1
10
100
1
10
100
1
10
100
Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
Fig 7 Whole-rock REE and other incompatible-element concentrations for the Karmutsen Formation (a) (c) and (e) are chondrite-normal-ized REE patterns for samples from each of the three main field areas onVancouver Island (b) (d) and (f) are primitive mantle-normalizedtrace-element patterns for samples from each of the three main field areas All normalization values are from McDonough amp Sun (1995) Theclear distinction between LREE-enriched tholeiitic basalts and LREE-depleted picrites the effects of alteration on the LILE (especially loss ofK and Rb) and the different range for the vertical scale in (b) (extends down to 01) should be noted
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
18
basalts (LaYbCNfrac14 21^29 mean 2508) Two LREE-depleted coarse-grained mafic rocks from the SchoenLake area have similar REE patterns (LaYbCNfrac14 07) tothe picrites from the Keogh Lake area indicating that thisdistinctive suite of rocks is exposed as far south asSchoen Lake that is 75 km away (Fig 7) The anomalouspillowed flow from the Karmutsen Range has lower LaYbCN (13^18) than the main group of tholeiitic basaltsfrom the Karmutsen Range (Fig 7) The mineralized sillfrom the Schoen Lake area is distinctly LREE-enriched(LaYbCNfrac14 33) with the highest REE abundances(LaCNfrac14 940) of all Karmutsen samples this may reflectcontamination by adjacent sediment also evident in thin-section where sulfide blebs are cored by quartz (Greeneet al 2006)The primitive mantle-normalized trace-element pat-
terns (Fig 7) and trace-element variations and ratios(Fig 8) highlight the differences in trace-element
concentrations between the four main groups ofKarmutsen flood basalts onVancouver Island The tholeii-tic basalts from all three areas have relatively smooth par-allel trace-element patterns some samples have smallpositive Sr anomalies relative to Nd and Sm and manysamples have prominent negative K anomalies relative toU and Nb reflecting K loss during alteration (Fig 7) Thelarge ion lithophile elements (LILE) in the tholeiiticbasalts (mainly Rb and K not Cs) are depleted relativeto the high field strength elements (HFSE) and LREELILE segments especially for the Schoen Lake area(Fig 7d) are remarkably parallel The picrites andhigh-MgO basalts form a tight band of parallel trace-element patterns and are depleted in HFSE and LREEwith positive Sr anomalies and relatively low Rb Thecoarse-grained mafic rocks from the Karmutsen Rangehave identical trace-element patterns to the tholeiiticbasalts Samples from the Schoen Lake area have similar
000
025
050
075
100
125
0 1 2 3 4 5
00
04
08
12
16
0 5 10 15 20 25
MgO (wt)
0
5
10
15
0 50 100 150
Hf (ppm)
Th (ppm)
NbLaNb (ppm)
Zr (ppm)
(a)
(c)
(b)
(d)
Th (ppm)
00
02
04
06
08
10
12
00 01 02 03 04 05
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
NbLa=1
4723A4 5615A12
U (ppm)
Fig 8 Whole-rock trace-element concentrations and ratios for the Karmutsen Formation [except (b) which is vs wt MgO] (a) Nb vs Zr(b) NbLa vs MgO (c) Th vs Hf (d) Th vs U There is a clear distinction between the tholeiitic basalts and picrites in Nb and Zr both inconcentration and the slope of each trend
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
19
trace-element patterns to the Karmutsen Range except forthe two LREE-depleted coarse-grained mafic rocks andthe mineralized sill (Fig 7d)
Sr^Nd^Hf^Pb isotopic compositionsA total of 19 samples from the four main groups ofthe Karmutsen Formation were selected for radiogenicisotopic geochemistry on the basis of major- and trace-element variation stratigraphic position and geographicallocation The samples have overlapping age-correctedHf and Nd isotope ratios and distinct ranges of Sr isotopiccompositions (Fig 9) The tholeiitic basalts picriteshigh-MgO basalt and coarse-grained mafic rocks have
initial eHffrac14thorn87 to thorn126 and eNdfrac14thorn66 to thorn88 cor-rected for in situ radioactive decay since 230 Ma (Fig 9Tables 3 and 4) All the Karmutsen samples form a well-defined linear array in a Lu^Hf isochron diagram corre-sponding to an age of 24115 Ma (Fig 9) This is withinerror of the accepted age of the Karmutsen Formationand indicates that the Hf isotope systematics behaved as aclosed system since c 230 Ma The anomalous pillowedflow has similar initial eHf (thorn98) and slightly lower initialeNd (thorn62) compared with the other Karmutsen samplesThe picrites and high-MgO basalt have higher initial87Sr86Sr (070398^070518) than the tholeiitic basalts(initial 87Sr86Srfrac14 070306^070381) and the coarse-grained mafic rocks (initial 87Sr86Srfrac14 070265^070428)
05128
05129
05130
05131
05132
015 017 019 021 023 025
02829
02830
02831
02832
02833
02834
000 002 004 006 008 010
4
6
8
10
0702 0703 0704 0705 07066
8
10
12
14
5 6 7 8 9 10
12
0
1
2
3
4
5
6
7
0702 0703 0704 0705 0706
(d)
LOI (wt )
87Sr 86Sr
4723A2
Nd
143Nd144Nd
147Sm144Nd
87Sr 86Sr (230 Ma)
(230 Ma)
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
(a)
(c)
176Hf177Hf
Age=241 plusmn 15 Ma=+990 plusmn 064
176Lu177Hf
Hf (230 Ma)
Nd (230 Ma)
(e)
Hf (i)
(b)
Fig 9 Whole-rock Sr Nd and Hf isotopic compositions for the Karmutsen Formation (a) 143Nd144Nd vs 147Sm144Nd (b) 176Hf177Hf vs176Lu177Hf The slope of the best-fit line for all samples corresponds to an age of 24115 Ma (c) Initial eNd vs 87Sr86Sr Age-correction to230 Ma (d) LOI vs 87Sr86Sr (e) Initial eHf vs eNd Average 2 error bars are shown in a corner of each panel Complete chemical duplicatesshown inTables 3 and 4 (samples 4720A7 and 4722A4) are circled in each plot
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
overlap the ranges of the picrites and tholeiitic basalts(Fig 9 Table 3)The measured Pb isotopic compositions of the tholeiitic
basalts are more radiogenic than those of the picrites andthe most magnesian Keogh Lake picrites have the leastradiogenic Pb isotopic compositions The range ofinitial Pb isotope ratios for the picrites is206Pb204Pbfrac1417868^18580 207Pb204Pbfrac1415547^15562and 208Pb204Pbfrac14 37873^38257 and the range for thetholeiitic basalts is 206Pb204Pbfrac1418782^19098207Pb204Pbfrac1415570^15584 and 208Pb204Pbfrac14 37312^38587 (Fig 10 Table 5) The coarse-grained mafic rocksoverlap the range of initial Pb isotopic compositions forthe tholeiitic basalts with 206Pb204Pbfrac1418652^19155207Pb204Pbfrac1415568^15588 and 208Pb204Pbfrac14 38153^38735 One high-MgO basalt has the highest initial206Pb204Pb (19242) and 207Pb204Pb (15590) and thelowest initial 208Pb204Pb (37676) The Pb isotopic ratiosfor Karmutsen samples define broadly linearrelationships in Pb isotope plots The age-corrected Pb
isotopic compositions of several picrites have been affectedby U and Pb (andor Th) mobility during alteration(Fig 10)
ALTERAT IONThe Karmutsen basalts have retained most of their origi-nal igneous structures and textures however secondaryalteration and low-grade metamorphism have producedzeolitic- and prehnite^pumpellyite-bearing mineral assem-blages (Table 1 Cho et al 1987) Alteration and metamor-phism have primarily affected the distribution of the LILEand the Sr isotopic systematics The Keogh Lake picriteshave been affected more by alteration than the tholeiiticbasalts and have higher LOI (up to 55wt ) variableLILE and higher measured Sr Nd and Hf and lowermeasured Pb isotope ratios than the tholeiitic basalts Srand Pb isotope compositions for picrites and tholeiiticbasalts show a relationship with LOI (Figs 9 and 10)whereas Nd and Hf isotopic compositions do not correlate
Table 3 Sr and Nd isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Rb Sr 87Sr86Sr 2m87Rb86Sr 87Sr86Srt Sm Nd 143Nd144Nd 2m eNd
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses carried out at the PCIGR thecomplete trace element analyses are given in Electronic Appendix 3
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
21
with LOI (not shown) The relatively high apparent initialSr isotopic compositions for the high-MgO lavas (up to07052) probably resulted from post-magmatic loss of Rbor an increase in 87Sr86Sr through addition of seawater Sr(eg Hauff et al 2003) High-MgO lavas from theCaribbean plateau have similarly high 87Sr86Sr comparedwith basalts (Recurren villon et al 2002)The correction for in situdecay on initial Pb isotopic ratios has been affected bymobilization of U and Th (and Pb) in the whole-rockssince their formation A thorough acid leaching duringsample preparation was used that has been shown to effec-tively remove alteration phases (Weis et al 2006 NobreSilva et al in preparation) Whole-rock SmNd ratiosand age-correction of Nd isotopes may be affected bythe leaching procedure for submarine rocks older than50 Ma (Thompson et al 2008 Fig 9) there ishowever very little variation in Nd isotopic compositionsof the picrites or tholeiitic basalts and this system does notappear to be disturbed by alteration The HFSE abun-dances for tholeiitic and picritic basalts exhibit clear
linear correlations in binary diagrams (Fig 8) whereasplots of LILE vs HFSE are highly scattered (not shown)because of the mobility of some of the alkali and alkalineearth LILE (Cs Rb Ba K) and Sr for most samplesduring alterationThe degree of alteration in the Karmutsen samples does
not appear to be related to eruption environment or depthof burial Pillow rims are compositionally different frompillow cores (Surdam1967 Kuniyoshi1972) and aquagenetuffs are chemically different from isolated pillows withinpillow breccias however submarine and subaerial basaltsexhibit a similar degree of alterationThere is no clear cor-relation between the submarine and subaerial basalts andsome commonly used chemical alteration indices (eg BaRb vs K2OP2O5 Huang amp Frey 2005) There is also nodefinitive correlation between the petrographic alterationindex (Table 1) and chemical alteration indices although10 of 13 tholeiitic basalts with the highest K and LILEabundances have the highest petrographic alterationindex of three (seeTable 1)
Table 4 Hf isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Lu Hf 176Hf177Hf 2m eHf176Lu177Hf 176Hf177Hft eHf(t)
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
OLIV INE ACCUMULAT ION INH IGH-MgO AND PICR IT IC LAVASGeochemical trends and petrographic characteristics indi-cate that accumulation of olivine played an important rolein the formation of the high-MgO basalts and picrites fromthe Keogh Lake area on northern Vancouver Island TheKeogh Lake samples show a strong linear correlation inplots of Al2O3 TiO2 ScYb and Ni vs MgO and many ofthe picrites have abundant clusters of olivine pseudomorphs(Table1 Fig11)There is a strong linear correlationbetween
modal per cent olivine and whole-rock magnesium contentmagnesium contents range between 10 and 19wt MgOand the proportion of olivine phenocrysts in these samplesvaries from 0 to 42 vol (Fig 11)With a few exceptionsboth the size range and median size of the olivine pheno-crysts in most samples are comparable The clear correla-tion between the proportion of olivine phenocrysts andwhole-rock magnesium contents indicates that accumula-tion of olivine was directly responsible for the high MgOcontents (410wt ) in most of the picritic lavas
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
207Pb204Pb
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
208Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
(a) (b)
(c) (e)
4723A24723A2
4723A2
4723A2
4722A4
4723A4
4722A4
4723A4
4723A134723A3
4723A3
4723A13
1556
1558
1560
1562
1564
1566
1568
186 190 194 198 202 206 2101550
1552
1554
1556
1558
1560
178 180 182 184 186 188 190 192 194
380
384
388
392
396
186 190 194 198 202 206 210374
378
382
386
390
178 180 182 184 186 188 190 192 194
206Pb204Pb(d)
LOI (wt )
0
1
2
3
4
5
6
7
186 190 194 198 202 206 210
4723A2
Fig 10 Pb isotopic compositions of leached whole-rock samples measured by MC-ICP-MS for the Karmutsen Formation Error bars are smal-ler than symbols (a) Measured 207Pb204Pb vs 206Pb204Pb (b) Initial 207Pb204Pb vs 206Pb204Pb Age-correction to 230 Ma (c) Measured208Pb204Pb vs 206Pb204Pb (d) LOI vs 206Pb204Pb (e) Initial 208Pb204Pb vs 206Pb204Pb Complete chemical duplicates shown in Table 5(samples 4720A7 and 4722A4) are circled in each plot The dashed lines in (b) and (e) show the differences in age-corrections for two picrites(4723A4 4722A4) using the measured UTh and Pb concentrations for each sample and the age-corrections when concentrations are used fromthe two picrites (4723A3 4723A13) that appear to be least affected by alteration
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
23
DISCUSS IONThe compositional and spatial development of the eruptedlava sequences of the Wrangellia oceanic plateau onVancouver Island provide information about the evolutionof a mantle plume upwelling beneath oceanic lithospherein an analogous way to the interpretation of continentalflood basalts The Caribbean and Ontong Java plateausare the two primary examples of accreted parts of oceanicplateaus that have been studied to date and comparisonscan be made with the Wrangellia oceanic plateauHowever the Wrangellia oceanic plateau is a distinctiveoccurrence of an oceanic plateau because it formed on thecrust of a Devonian to Mississippian intra-oceanic islandarc overlain by Mississippian to Permian limestones andpelagic sediments The nature and thickness of oceaniclithospheric mantle are considered to play a fundamentalrole in the decompression and evolution of a mantleplume and thus the compositions of the erupted lavasequences (Greene et al 2008) The geochemical and
stratigraphic relationships observed in the KarmutsenFormation onVancouver Island and the focus of the subse-quent discussion sections provide constraints on (1) thetemperature and degree of melting required to generatethe primary picritic magmas (2) the composition of themantle source (3) the depth of melting and residual min-eralogy in the source region (4) the low-pressure evolutionof the Karmutsen tholeiitic basalts (5) the growth of theWrangellia oceanic plateau
Melting conditions and major-elementcomposition of the primary magmasThe primary magmas to most flood basalt provinces arebelieved to be picritic (eg Cox 1980) and they have beenfound in many continental and oceanic flood basaltsequences worldwide (eg Siberia Karoo Paranacurren ^Etendeka Caribbean Deccan Saunders 2005) Near-pri-mary picritic lavas with low total alkali abundances
Table 5 Pb isotopic compositions of Karmutsen basaltsVancouver Island BC
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
require high-degree partial melting of the mantle source(eg Herzberg amp OrsquoHara 2002) The Keogh Lake picritesfrom northernVancouver Island are the best candidates foridentifying the least modified partial melts of the mantleplume source despite having accumulated olivine pheno-crysts and can be used to estimate conditions of meltingand the composition of the primary magmas for theKarmutsen FormationPrimary magma compositions mantle potential tem-
peratures and source melt fractions for the picritic lavas ofthe Karmutsen Formation were calculated from primitivewhole-rock compositions using PRIMELT1XLS software(Herzberg et al 2007) and are presented in Fig 12 andTable 6 A detailed discussion of the computationalmethod has been given by Herzberg amp OrsquoHara (2002)and Herzberg et al (2007) key features of the approachare outlined belowPrimary magma composition is calculated for a primi-
tive lava by incremental addition of olivine and compari-son with primary melts of mantle peridotite Lavas thathad experienced plagioclase andor clinopyroxene fractio-nation are excluded from this analysis All calculated pri-mary magma compositions are assumed to be derived byaccumulated fractional melting of fertile peridotite KR-4003 Uncertainties in fertile peridotite composition donot propagate to significant variations in melt fractionand mantle potential temperature melting of depletedperidotite propagates to calculated melt fractions that aretoo high but with a negligible error in mantle potential
temperature PRIMELT1XLS calculates the olivine liqui-dus temperature at 1atm which should be similar to theactual eruption temperature of the primary magmaand is accurate to 308C at the 2 level of confidenceMantle potential temperature is model dependent witha precision that is similar to the accuracy of the olivineliquidus temperature (C Herzberg personal communica-tion 2008)With regard to the evidence for accumulated olivine in
the Keogh Lake picrites it should not matter whether themodeled lava composition is aphyric or laden with accu-mulated olivine phenocrysts PRIMELT1 computes theprimary magma composition by adding olivine to a lavathat has experienced olivine fractionation in this way itinverts the effects of fractional crystallization The onlyrequirement is that the olivine phenocrysts are not acci-dental or exotic to the lava compositions Support for thisprocedure can be seen in the calculated olivine liquid-line-of-descent shown by the curved arrays with 5 olivineaddition increments (Fig 12c) Fractional crystallization ofolivine from model primary magmas having 85^95wt FeO and 153^175wt MgO will yield arrays of deriva-tive liquids and randomly sorted olivines that in most butnot all cases connect the picrites and high-MgO basalts Anewer version of the PRIMELT software (Herzberg ampAsimow 2008) can perform olivine addition to and sub-traction from lavas with highly variably olivine phenocrystcontents and yields results that converge with those shownin Fig 12
Fig 11 Relationship between abundance of olivine phenocrysts and whole-rock MgO contents for the Keogh Lake picrites (a) Tracings ofolivine grains (gray) in six high-MgO samples made from high resolution (2400 dpi) scans of petrographic thin-sections Scale bars represent10mm sample numbers are indicated at the upper left (b) Modal olivine vs whole-rock MgO Continuous line is the best-fit line for 10 high-MgO samplesThe areal proportion of olivine was calculated from raster images of olivine grains using ImageJ image analysis software whichprovides an acceptable estimation of the modal abundance (eg Chayes 1954)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
25
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
Gray lines = final melting pressure
L + Ol + Opx
07
08
09
Pressure (GPa)
10
3
4
5
6
1
SOLIDUS
01
04
0506
L + Ol
2
03
02
PeridotiteKR-4003
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
Thick black lines = initial melting pressure
Dashed lines = melt fraction
Melt Fraction
2
3 4 5 6
L+Ol+Opx
SolidusSpinel
SolidusGarnet Peridotite
7
L+Ol
0 10 20 30 4040
45
50
55
60
MgO (wt)
SiO2
(wt)
PeridotiteKR-4003
Melt Fraction
0 10 20 305
10
15
CaO(wt)
MgO (wt)
3
4 5 6 7
2
46
2
Peridotite Partial Melts
Thick black line = initial melting pressureGray lines = final melting pressure
Peridotite
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
02
Pressure (GPa)
0302
04
Pressure (GPa)
Picrite
High-MgO basalt
Tholeiitic basalt
(c)
(d)
Karmutsen flood basaltsOlivine addition modelOlivine addition (5 increment)Primary magma
PicriteHigh-MgO basaltTholeiitic basalt
(a)
(b)
800 850 900
910
920
930
940
950
Karmutsen flood basalts
Olivine addition modelOlivine addition (5 increment)Primary magma
Gray lines = Fo contents of olivine
Fig 12 Estimated primary magma compositions for three Keogh Lake picrites and high-MgO basalts (samples 4723A4 4723A135616A7) using the forward and inverse modeling technique of Herzberg et al (2007) Karmutsen compositions and modeling results areoverlain on diagrams provided by C Herzberg (a) Whole-rock FeO vs MgO for Karmutsen samples from this study Total iron estimatedto be FeO is 090 Gray lines show olivine compositions that would precipitate from liquid of a given MgO^FeO composition Black lineswith crosses show results of olivine addition (inverse model) using PRIMELT1 (Herzberg et al 2007) Gray field highlights olivinecompositions with Fo contents of 91^92 predicted to precipitate (b) FeO vs MgO showing Karmutsen lava compositions and results ofa forward model for accumulated fractional melting of fertile peridotite KR-4003 (Kettle River Peridotite Walter 1998) (c) SiO2 vsMgO for Karmtusen lavas and model results (d) CaO vs MgO showing Karmtusen lava compositions and model results To brieflysummarize the technique [see Herzberg et al (2007) for complete description] potential parental magma compositions for the high-MgOlava series were selected (highest MgO and appropriate CaO) and using PRIMELT1 software olivine was incrementally added tothe selected compositions to show an array of potential primary magma compositions (inverse model) Then using PRIMELT1 theresults from the inverse model were compared with a range of accumulated fractional melts for fertile peridotite derived fromparameterization of the experimental results for KR-4003 from Walter (1998) forward model Herzberg amp OrsquoHara (2002) A melt fractionwas sought that was unique to both the inverse and forward models (Herzberg et al 2007) A unique solution was found when there was a com-mon melt fraction for both models in FeO^MgO and CaO^MgO^Al2O3^SiO2 (CMAS) projection space This modeling assumes thatolivine was the only phase crystallizing and ignores chromite precipitation and possible augite fractionation in the mantle (Herzbergamp OrsquoHara 2002) Results are best for a residue of spinel lherzolite (not pyroxenite) The presence of accumulated olivine in samples ofthe Keogh Lake picrites used as starting compositions does not significantly affect the results because we are modeling addition of olivine(see text for further description) The tholeiitic basalts cannot be used for modeling because they are all saturated in plagthorn cpxthornolThick black line for initial melting pressure at 4 GPa is not shown in (c) for clarity Gray area in (a) indicates parental magmas thatwill crystallize olivine with Fo 91^92 at the surface Dashed lines in (c) indicate melt fraction Solidi for spinel and garnet peridotite arelabelled in (c)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
The PRIMELT1 results indicate that if the Karmutsenpicritic lavas were derived from accumulated fractionalmelting of fertile peridotite the primary magmas wouldhave contained 15^17wt MgO and 10wt CaO(Fig 12 Table 6) formed from 23^27 partial meltingand would have crystallized olivine with a Fo content of 91 (Fig 12 Table 6) Calculated mantle temperatures forthe Karmutsen picrites (14908C) indicate melting ofanomalously hot mantle (100^2008C above ambient) com-pared with ambient mantle that produces mid-ocean ridgebasalt (MORB 1280^14008C Herzberg et al 2007)which initiated between 3 and 4 GPa Several picritesappear to be near-primary melts and required very littleaddition of olivine (eg sample 93G171 with measured158wt MgO) These temperature and melting esti-mates of the Karmutsen picrites are consistent withdecompression of hot mantle peridotite in an actively con-vecting plume head (ie plume initiation model) Threesamples (picrite 5614A1 and high-MgO basalts 5614A3and 5614A5) are lower in iron than the other Karmutsenlavas with FeO in the 65^80wt range (Fig 12c) andhave MgO and FeO contents that are similar to primitiveMORB from the Sequeiros fracture zone of the EastPacific Rise (EPR Perfit et al 1996) These samples
however differ from EPR MORB in having higher SiO2
and lower Al2O3 and they are restricted geographicallyto one area (Sara Lake Fig 2)We therefore have evidence for primary magmas with
MgO contents that range from a possible low of 110wt to as high as 174wt with inferred mantle potential tem-peratures from 1340 to 15208C This is similar to the rangedisplayed by lavas from the Ontong Java Plateau deter-mined by Herzberg (2004) who found that primarymagma compositions assuming a peridotite sourcewould contain 17wt MgO and 10wt CaO(Table 6) and that these magmas would have formed by27 melting at a mantle potential temperature of15008C (first melting occurring at 36 GPa) Fitton ampGodard (2004) estimated 30 partial melting for theOntong Java lavas based on the Zr contents of primarymagmas of Kroenke-type basalt calculated by incrementaladdition of equilibrium olivine However if a considerableamount of eclogite was involved in the formation of theOntong Java plateau magmas the excess temperatures esti-mated by Herzberg et al (2007) may not be required(Korenaga 2005) The variability in mantle potential tem-perature for the Keogh Lake picrites is also similar to thewide range of temperatures that have been calculated for
Table 6 Estimated primary magma compositions for Karmutsen basalts and other oceanic plateaus or islands
Sample 4723A4 4723A13 5616A7 93G171 Average OJP Mauna Keay Gorgonaz
Ontong Java primary magma composition for accumulated fraction melting (AFM) from Herzberg (2004)yMauna Kea primary magma composition is average of four samples of Herzberg (2006 table 1)zGorgona primary magma composition for AFM for 1F fertile source from Herzberg amp OrsquoHara (2002 table 4)Total iron estimated to be FeO is 090
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
27
lavas from single volcanoes in the Galapagos Hawaii andelsewhere by Herzberg amp Asimow (2008) Those workersinterpreted the range to reflect the transport of magmasfrom both the cool periphery and the hotter axis of aplume A thermally heterogeneous plume will yield pri-mary magmas with different MgO contents and the rangeof primary magma compositions and their inferred mantlepotential temperatures for Karmutsen lavas may reflectmelting from different parts of the plume
Source of the Karmutsen Formation lavasThe Sr^Nd^Hf^Pb isotopic compositions of theKarmutsen volcanic rocks on Vancouver Island indicatethat they formed from plume-type mantle containing adepleted component that is compositionally similar to butdistinct from other oceanic plateaus that formed in thePacific Ocean (eg Ontong Java and Caribbean Fig 13)Trace element and isotopic constraints can be used to testwhether the depleted component of the KarmutsenFormation was intrinsic to the mantle plume and distinctfrom the source of MORB or the result of mixing or invol-vement of ambient depleted MORB mantle with plume-type mantle The Karmutsen tholeiitic basalts have initialeNd and
87Sr86Sr ratios that fall within the range of basaltsfrom the Caribbean Plateau (eg Kerr et al 1996 1997Hauff et al 2000 Kerr 2003) and a range of Hawaiian iso-tope data (eg Frey et al 2005) and lie within and abovethe range for the Ontong Java Plateau (Tejada et al 2004Fig 13a) Karmutsen tholeiitic basalts with the highest ini-tial eNd and lowest initial 87Sr86Sr lie within and justbelow the field for age-corrected northern EPR MORB at230 Ma (eg Niu et al1999 Regelous et al1999) Initial Hfand Nd isotopic compositions for the Karmutsen samplesare noticeably offset to the low side of the ocean islandbasalt (OIB) array (Vervoort et al 1999) and lie belowand slightly overlap the low eHf end of fields for Hawaiiand the Caribbean Plateau (Fig 13c) the Karmutsen sam-ples mostly fall between and slightly overlap a field forage-corrected Pacific MORB at 230 Ma and Ontong Java(Mahoney et al 1992 1994 Nowell et al 1998 Chauvel ampBlichert-Toft 2001) Karmutsen lava Pb isotope ratios over-lap the broad range for the Caribbean Plateau and thelinear trends for the Karmutsen samples intersect thefields for EPR MORB Hawaii and Ontong Java in208Pb^206Pb space but do not intersect these fields in207Pb^206Pb space (Fig 13d and e) The more radiogenic207Pb204Pb for a given 206Pb204Pb of the Karmutsenbasalts requires high UPb ratios at an ancient time when235U was abundant similar to the Caribbean Plateau andenriched in comparison with EPR MORB Hawaii andOntong JavaThe low eHf^low eNd component of the Karmutsen
basalts that places them below the OIB mantle array andpartly within the field for age-corrected Pacific MORBcan form from low ratios of LuHf and high SmNd over
time (eg Salters amp White 1998) MORB will develop aHf^Nd isotopic signature over time that falls below themantle array Low LuHf can also lead to low 176Hf177Hffrom remelting of residues produced by melting in theabsence of garnet (eg Salters amp Hart 1991 Salters ampWhite 1998) The Hf^Nd isotopic compositions of theKarmutsen Formation indicate the possible involvementof depleted MORB mantle however similar to Iceland(Fitton et al 2003) Nb^Zr^Y variations provide a way todistinguish between a depleted component within theplume and a MORB-like component The Karmutsenbasalts and picrites fall within the Iceland array and donot overlap the MORB field in a logarithmic plot of NbYvs ZrY (Fig 13b) using the fields established by Fittonet al (1997) Similar to the depleted component within theCaribbean Plateau (Thompson et al 2003) the trend ofHf^Nd isotopic compositions towards Pacific MORB-likecompositions is not followed by MORB values in Nb^Zr^Y systematics and this suggests that the depleted compo-nent in the source of the Karmutsen basalts is distinctfrom the source of MORB and probably an intrinsic partof the plume The relatively radiogenic Pb isotopic ratiosand REE patterns distinct from MORB also support thisinterpretationTrace-element characteristics suggest the possible invol-
vement of sub-arc lithospheric mantle in the generation ofthe Karmutsen picrites Lassiter et al (1995) suggested thatmixing of a plume-type source with eNdthorn 6 to thorn 7 witharc material with low NbTh could reproduce the varia-tions observed in the Karmutsen basalts however theyconcluded that the absence of low NbLa ratios in theirsample of tholeiitic basalts restricts the amount of arc litho-sphere involved The study of Lassiter et al (1995) wasbased on major and trace element and Sr Nd and Pb iso-topic data for a suite of 29 samples from Buttle Lake in theStrathcona Provincial Park on central Vancouver Island(Fig 1) Whereas there are no clear HFSE depletions inthe Karmutsen tholeiitic basalts the low NbLa (and TaLa) of the Karmutsen picritic lavas in this study indicatethe possible involvement of Paleozoic arc lithosphere onVancouver Island (Fig 8)
REE modeling dynamic melting andsource mineralogyThe combined isotopic and trace-element variations of thepicritic and tholeiitic Karmutsen lavas indicate that differ-ences in trace elements may have originated from differentmelting histories For example the LuHf ratios of theKarmutsen picritic lavas (mean 008) are considerablyhigher than those in the tholeiitic basalts (mean 002)however the small range of overlapping initial eHf valuesfor the two lava suites suggests that these differences devel-oped during melting within the plume (Fig 9b) Below wemodel the effects of source mineralogy and depth of
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
28
melting on differences in trace-element concentrations inthe tholeiitic basalts and picritesDynamic melting models simulate progressive decom-
pression melting where some of the melt fraction isretained in the residue and only when the degree of partialmelting is greater than the critical mass porosity and the
source becomes permeable is excess melt extracted fromthe residue (Zou1998) For the Karmutsen lavas evolutionof trace element concentrations was simulated using theincongruent dynamic melting model developed by Zou ampReid (2001) Melting of the mantle at pressures greater than1 GPa involves incongruent melting (eg Longhi 2002)
OIB array
PacificMORB(230 Ma)
(0 Ma)
EPR(230 Ma)
-4
-2
0
2
4
8
10
12
14
16
18
20
22
-4 -2 0 2 4 6 8 10 12 14
minus2
0
2
4
6
8
10
12
14
0702 0703 0704 0705 0706
IndianMORB
87Sr 86Sr (initial)
Hf (initial)
(a) (c)
Nd (initial)
Karmutsen tholeiitic basalt
HawaiiOntong JavaCaribbean Plateau
Karmutsen high-MgO lava
high-MgO lavasKarmutsen
1540
1545
1550
1555
1560
1565
175 180 185 190 195 200375
380
385
390
395
175 180 185 190 195 200
(d) (e)
EPR
EPR
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb
Karmutsen Formation (initial)
HawaiiOntong JavaCaribbean Plateau
East Pacific Rise
Karmutsen Formation (measured)
Juan de FucaGorda
ε Nd
(initial)
Karmutsen Formation (different age-correction for 3 picrites)
(0 Ma)
NbY
001
01
1
1 10
ZrY
OIB
NMORB
(b)
Iceland array
6
Fig 13 Comparison of age-corrected (230 Ma) Sr^Nd^Hf^Pb isotopic compositions for Karmutsen flood basalts onVancouver Island withage-corrected OIB and MORB and Nb^Zr^Y variation (a) Initial eNd vs 87Sr86Sr (b) NbYand ZrY variation (after Fitton et al 1997)(c) Initial eHf vs eNd (d) Measured and initial 207Pb204Pb vs 206Pb204Pb (e) Measured and initial 208Pb204Pb vs 206Pb204Pb References fordata sources are listed in Electronic Appendix 4 Most of the compiled data were extracted from the GEOROC database (httpgeorocmpch-mainzgwdgdegeoroc) OIB array line in (c) is fromVervoort et al (1999) EPR East Pacific Rise Several high-MgO samples affected bysecondary alteration were not plotted for clarity in Pb isotope plots Estimation of fields for age-corrected Pacific MORB and EPR at 230 Mawere made using parent^daughter concentrations (in ppm) from Salters amp Stracke (2004) of Lufrac14 0063 Hffrac14 0202 Smfrac14 027 Ndfrac14 071Rbfrac14 0086 and Srfrac14 836 In the NbYvs ZrYplot the normal (N)-MORB and OIB fields not including Icelandic OIB are from data com-pilations of Fitton et al (1997 2003) The OIB field is data from 780 basalts (MgO45wt ) from major ocean islands given by Fitton et al(2003) The N-MORB field is based on data from the Reykjanes Ridge EPR and Southwest Indian Ridge from Fitton et al (1997) The twoparallel gray lines in (b) (Iceland array) are the limits of data for Icelandic rocks
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
29
with olivine being produced during the reactioncpxthornopxthorn spfrac14meltthornol for spinel peridotites The mod-eling used coefficients for incongruent melting reactionsbased on experiments on lherzolite melting (eg Kinzleramp Grove 1992 Longhi 2002) and was separated into (1)melting of spinel lherzolite (2) two combined intervals ofmelting for a garnet lherzolite (gt lherz) and spinel lher-zolite (sp lherz) source (3) melting of garnet lherzoliteThe model solutions are non-unique and a range indegree of melting partition coefficients melt reactioncoefficients and proportion of mixed source componentsis possible to achieve acceptable solutions (see Fig 14 cap-tion for description of modeling parameters anduncertainties)The LREE-depleted high-MgO lavas require melting of
a depleted spinel lherzolite source (Fig 14a) The modelingresults indicate a high degree of melting (22^25) similarto the results from PRIMELT1 based on major-elementvariations (discussed above) of a LREE-depleted sourceThe high-MgO lavas lack a residual garnet signature Theenriched tholeiitic basalts involved melting of both garnetand spinel lherzolite and represent aggregate melts pro-duced from continuous melting throughout the depth ofthe melting column (Fig 14b) Trace-element concentra-tions in the tholeiitic and picritic lavas are not consistentwith low-degree melting of garnet lherzolite alone(Fig 14c) The modeling results indicate that the aggregatemelts that formed the tholeiitic basalts would haveinvolved lesser proportions of enriched small-degreemelts (1^6 melting) generated at depth and greateramounts of depleted high-degree melts (12^25) gener-ated at lower pressure (a ratio of garnet lherzolite tospinel lherzolite melt of 14 provides the best fits seeFig 14b and description in figure caption) The totaldegree of melting for the tholeiitic basalts was high(23^27) similar to the degree of partial melting in thespinel lherzolite facies for the primary melts that eruptedas the Keogh Lake picrites of a source that was also prob-ably depleted in LREE
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
(c)
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
Melting of garnet lherzolite
High-MgO lavas
Melting of spinel lherzolite
source (07DM+03PM)
source (07DM+03PM)
6 melting
30 melting
(b)
(a)
30 melting
Sam
ple
Cho
ndrit
eS
ampl
eC
hond
rite 20 melting
1 melting
Melting of garnet lherzoliteand spinel lherzolite
x gt lherz + 4x sp lherz
5 meltingx gt lherz + 4x sp lherz
Sam
ple
Cho
ndrit
e
Tholeiitic lavas
Tholeiitic lavas
High-MgO lavas
source (07DM+03PM)
Fig 14 Trace-element modeling results for incongruent dynamicmantle melting for picritic and tholeiitic lavas from the KarmutsenFormation Three steps of modeling are shown (a) melting of spinellherzolite compared with high-MgO lavas (b) melting of garnet lher-zolite and spinel lherzolite compared with tholeiitic lavas (c) meltingof garnet lherzolite compared with tholeiitic and picritic lavasShaded fields in each panel for tholeiitic basalts and picrites arelabeled Patterns with symbols in each panel are modeling results(patterns are 5 melting increments in (a) and (b) and 1 incre-ments in (c) PM primitive mantle DM depleted MORB mantle gtlherz garnet lherzolite sp lherz spinel lherzolite In (b) the ratios ofper cent melting for garnet and spinel lherzolite are indicated (x gtlherzthorn 4x sp lherz means that for 5 melting 1 melt in garnetfacies and 4 melt in spinel facies where x is per cent melting in theinterval of modeling) The ratio of the respective melt proportions inthe aggregate melt (x gt lherzthorn 4x sp lherz) was kept constant andthis ratio of melt contribution provided the best fit to the data Arange of proportion of source components (07DMthorn 03PM to
09DMthorn 01PM) can reproduce the variation in the high-MgOlavas Melting modeling uses the formulation of Zou amp Reid (2001)an example calculation is shown in their Appendix PM fromMcDonough amp Sun (1995) and DM from Salters amp Stracke (2004)Melt reaction coefficients of spinel lherzolite from Kinzler amp Grove(1992) and garnet lherzolite fromWalter (1998) Partition coefficientsfrom Salters amp Stracke (2004) and Shaw (2000) were kept constantSource mineralogy for spinel lherzolite (018cpx027opx052ol003sp) from Kinzler (1997) and for garnet lherzolite(034cpx008opx053ol005gt) from Salters amp Stracke (2004) Arange of source compositions for melting of spinel lherzolite(07DMthorn 03PM to 03DMthorn 07PM) reproduce REE patternssimilar to those of the high-MgO lavas with best fits for 22ndash25melting and 07DMthorn 03PM source composition Concentrationsof tholeiitic basalts cannot be reproduced from an entirelyprimitive or depleted source
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
30
The uncertainty in the composition of the source in thespinel lherzolite melting model for the high-MgO lavasprecludes determining whether the source was moredepleted in incompatible trace elements than the source ofthe tholeiitic lavas (see Fig 14 caption for description) It ispossible that early high-pressure low-degree melting left aresidue within parts of the plume (possibly the hotter inte-rior portion) that was depleted in trace elements (egElliott et al 1991) further decompression and high-degreemelting of such depleted regions could then have generatedthe depleted high-MgO Karmutsen lavas Shallow high-degree melting would preferentially sample depletedregions whereas deeper low-degree melting may notsample more refractory depleted regions (eg CaribbeanEscuder-Viruete et al 2007) The picrites lie at the highend of the range of Nd isotopic compositions and havelow NbZr similar to depleted Icelandic basalts (Fittonet al 2003) therefore the picrites may represent the partsof the mantle plume that were the most depleted prior tothe initiation of melting As Fitton et al (2003) suggestedthese depleted melts may only rarely be erupted withoutcombining with more voluminous less depleted melts andthey may be detected only as undifferentiated small-volume flows like the Keogh Lake picrites within theKarmutsen Formation on Vancouver IslandDecompression melting within the mantle plume initiatedwithin the garnet stability field (35^25 GPa) at highmantle potential temperatures (Tp414508C) and pro-ceeded beneath oceanic arc lithosphere within the spinelstability field (25^09 GPa) where more extensivedegrees of melting could occur
Magmatic evolution of Karmutsentholeiitic basaltsPrimary magmas in large igneous provinces leave exten-sive coarse-grained cumulate residues within or below thecrust these mafic and ultramafic plutonic sequences repre-sent a significant proportion of the original magma thatpartially crystallized at depth (eg Farnetani et al 1996)For example seismic and petrological studies of OntongJava (eg Farnetani et al 1996) combined with the use ofMELTS (Ghiorso amp Sack 1995) indicate that the volcanicsequence may represent only 40 of the total magmareaching the Moho Neal et al (1997) estimated that30^45 fractional crystallization took place for theOntong Java magmas and produced cumulate thicknessesof 7^14 km corresponding to 11^22 km of flood basaltsdepending on the presence of pre-existing oceanic crustand the total crustal thickness (25^35 km) Interpretationsof seismic velocity measurements suggest the presence ofpyroxene and gabbroic cumulates 9^16 km thick beneaththe Ontong Java Plateau (eg Hussong et al 1979)Wrangellia is characterized by high crustal velocities in
seismic refraction lines which is indicative of mafic
plutonic rocks extending to depth beneath VancouverIsland (Clowes et al 1995)Wrangellia crust is 25^30 kmthick beneath Vancouver Island and is underlain by astrongly reflective zone of high velocity and density thathas been interpreted as a major shear zone where lowerWrangellia lithosphere was detached (Clowes et al 1995)The vast majority of the Karmutsen flows are evolved tho-leiitic lavas (eg low MgO high FeOMgO) indicating animportant role for partial crystallization of melts at lowpressure and a significant portion of the crust beneathVancouver Island has seismic properties that are consistentwith crystalline residues from partially crystallizedKarmutsen magmas To test the proportion of the primarymagma that fractionated within the crust MELTS(Ghiorso amp Sack 1995) was used to simulate fractionalcrystallization using several estimated primary magmacompositions from the PRIMELT1 modeling results Themajor-element composition of the primary magmas couldnot be estimated for the tholeiitic basalts because of exten-sive plagthorn cpxthornol fractionation and thus the MELTSmodeling of the picrites is used as a proxy for the evolutionof major elements in the volcanic sequence A pressure of 1kbar was used with variable water contents (anhydrousand 02wt H2O) calculated at the quartz^fayalite^magnetite (QFM) oxygen buffer Experimental resultsfrom Ontong Java indicate that most crystallizationoccurs in magma chambers shallower than 6 km deep(52 kbar Sano amp Yamashita 2004) however some crys-tallization may take place at greater depths (3^5 kbarFarnetani et al 1996)A selection of the MELTS results are shown in Fig 15
Olivine and spinel crystallize at high temperature until15^20wt of the liquid mass has fractionated and theresidual magma contains 9^10wt MgO (Fig 15f) At1235^12258C plagioclase begins crystallizing and between1235 and 11908C olivine ceases crystallizing and clinopyr-oxene saturates (Fig 15f) As expected the addition of clin-opyroxene and plagioclase to the crystallization sequencecauses substantial changes in the composition of the resid-ual liquid FeO and TiO2 increase and Al2O3 and CaOdecrease while the decrease in MgO lessens considerably(Fig 15) The near-vertical trends for the Karmutsen tho-leiitic basalts especially for CaO do not match the calcu-lated liquid-line-of-descent very well which misses manyof the high-CaO high-Al2O3 low-FeO basalts A positivecorrelation between Al2O3CaO and SrNd indicates thataccumulation of plagioclase played a major role in the evo-lution of the tholeiitic basalts (Fig 15g) Increasing thecrystallization pressure slightly and the water content ofthe melts does not systematically improve the match ofthe MELTS models to the observed dataThe MELTS results indicate that a significant propor-
tion of the crystallization takes place before plagioclase(15^20 crystallization) and clinopyroxene (35^45
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
31
crystallization) join the crystallization sequence Theseresults suggest that445 crystallization of the primarymagmas occurred (estimated from the residual liquid percent in Fig 15f) and if the original flood basalt stratigra-phy on Vancouver Island was 6 km thick at least anequivalent amount of complementary plutonic rocksshould have crystallized The Karmutsen basalts and theirplutonic residues thus represent a significant addition ofcrust (perhaps 412 km thickness) previously thickenedduring the formation of the underlying Paleozoic arc
Growth of the Wrangellia oceanic plateauon Vancouver IslandThe extrusive part of the Wrangellia oceanic plateau onVancouver Island was constructed in a three-layersequence of mostly tholeiitic basalts with a restrictedrange of composition (Fig 16) The plateau overliesDevonian to Mississippian arc volcanic sequences andmarine sediments of Mississippian to Permian ages thatare intruded by gabbroic rocks related to the flood basaltsThe volcanic stratigraphy of theWrangellia plateau formed
00
05
10
15
20
25
30
35
40
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
16
0 2 4 6 8 10 12 14 16 18 20
10
11
12
13
14
15
16
17
18
19
20
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
0 2 4 6 8 10 12 14 16 18 20
MgO (wt)
0 10 20 30 40 50
percent of total mass
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
) Mineral proportions
Sp (e)
Ol
CpxPlag
PicriteHigh-MgO basaltTholeiitic basalt
TiO2 (wt)
MgO (wt)
Al2O3 (wt) CaO (wt)
FeO (wt)
MgO(a) (b)
(c) (d)
MgO (wt)MgO (wt)
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
)
Residual liquid ()
1220
1190
Ol + Sp
Plag+Ol+Sp Plag+Cpx+Sp
02 wt H2O
no H2O
Residual liquid and crystallizing phases
(f)
0102030405060708090100
plagioclaseaccumulation
plagioclaseaccumulation
plagioclase
accumulatio
n
10
12
14
16
0 10 20 30 40
Al2O3CaO
SrNd
tholeiitic basalts
(g)
Fig 15 Forward fractional crystallization modeling results for major elements from MELTS (Ghiorso amp Sack 1995) compared with picriticand tholeiitic lavas from the Karmutsen Formation (a^d) Major-element compositions of residual liquids compared with Karmutsen lavacompositions For clarity the results are shown only for one starting composition (open circle) using the composition of the estimated primarymagma for the high-MgO lava series from the modeling technique of Herzberg et al (2007) for sample 4723A13 as shown in Fig 12 The esti-mated primary magmas for samples 4723A3 and 93G171 yield similar results A pressure of 1 kbar was used with no H2O and 02wt H2Ocalculated at QFM black line is anhydrous and grey line is 02wt H2OThe end-point for the anhydrous result is at 78 fractionation and11308C and the end-point for the result with 02wt H2O is 75 fractionation and11108C (e) Mineral proportions of crystallized phases fromone MELTrun (sample 4723A13) with no H2O (f) Per cent residual liquid vs temperature labelled with intervals of crystallizing phases (g)Al2O3CaO vs SrNd for the tholeiitic basalts
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
32
in a single eruptive phase (c 231^225 Ma) as an emergentbasalt sequence during a deeper-water shallow-water andsubaerial stage and subsided below sea level relativelysoon after its formation (Fig 16) The deeper-water stagecontains pillowed and massive submarine flows (3000mthick) that built up the submarine volcanic edifice from theseafloor The lowest pillow basalts were emplaced onunconsolidated and lithified fine-grained siliciclastic andcarbonaceous sediments containing Daonella beds(c 235^232 Ma) that are intruded by abundantKarmutsen mafic sills (Fig 16a) The different pillowedand massive submarine flows resulted primarily from dif-ferent effusive rates and local topography Some of themassive submarine flows commonly recognized by radiat-ing columnar jointing represent master tubes for the deliv-ery of lava to distal parts of submarine flow fields Thepillowed flows which commonly have uneven topography
in modern sequences may have served as levees to themassive submarine flows rather than allowing unob-structed spreading and inflation typical of the massive sub-aerial flows that form on shallow slopes in continentalflood basalt provinces (eg Self et al 1997) The rapid suc-cession of flows is evident by the near absence of sedimentbetween submarine flows Minor volumes of picritic pillowlavas occur in the upper part of the pillowed flow sequencein areas of the plateau exposed on northern VancouverIsland (Fig 16 Nixon et al 2008)The growth of the plateau in shallow water led to
emplacement of an increasing proportion of volcaniclastics(pillow breccias and hyaloclastites) that accumulated up to1500m thick in some areas (Fig 16b Nixon et al 2008)Horizons of pillowed flows occur locally within the volca-niclastic stratigraphy Sedimentary structures (eg gradedbedding fluidization structures) and petrographic textures
Late Devonian-Mississippian arc
Daonella beds
Middle Triassic
deeper water
Late Devonian-Mississippian arc
pillowed flows massive submarine
flows
pillow brecciahyaloclastiteshallow water
~4 km
~6 km
Late Triassic limestoneinterflow limestonelenses
Karmutsen Formation
mafic intrusions sediment-sill complex
Karmutsen Formation
(a) (b)
(c) (d)
subaerial flows
Late Devonian-Mississippian arc Late Devonian-Mississippian arc
picritic pillow basalts
Miss to Permian marine sediments
sills
Fig 16 Growth of the Wrangellia oceanic plateau onVancouver Island (a) Deep-water eruptive phase of pillowed and massive submarineflows with almost no intervening sediments onto MiddleTriassic shale containing Daonella beds (c 235^232 Ma) The basement of the plateauis composed of Devonian to Mississippian arc sequences overlain by Mississippian to Permian carbonates and siliciclastic sediments(b) Shallow-water flow emplacement with increasing proportion of volcaniclastics resulted primarily from cooling^contraction granulationPicritic pillow basalts occur primarily near the transition between pillowed submarine flows and volcaniclastic units Pillowed and massiveflows occur locally within the volcaniclastic unit (c) Emergent stage of growth consists of overlapping subaerial flow fields similar to continen-tal flood basalts with pahoehoe flow structures Multi-tiered columnar jointing occurrs locally in response to ponding of flows (eg Bondre et al2004) Interflow carbonate deposits and plagioclase megacrystic flows developed locally during the waning stage of volcanism (d) Post-volcanicsubsidence and deposition of Late Carnian to Norian (c 212^223 Ma) platform carbonates on top of the plateau Units are labeled in the dia-grams Thicknesses of the underlying units are not to scale
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
33
within the volcaniclastic rocks indicate formation primar-ily via cooling^contraction granulation magma^water^steam interaction autobrecciation resedimentation andmass-wasting rather than pyroclastic eruption (Nixonet al 2008)A broad subaerial platform was constructed by emplace-
ment of subaerial lava flows during an emergent subaerialstage (Fig 16c) The subaerial unit predominantly com-prises well-layered sheet flows that formed in a similarway to continental flood basalts as hundred(s) of pahoehoeflow fields erupted on shallow slopes (Self et al 1997) how-ever volcaniclastic and pillowed flow units are locally pre-sent mostly in the upper parts of the stratigraphyLocalized ponding of flows led to the formation of multi-tiered columnar jointing in some subaerial flows (egBondre et al 2004) As volcanism waned local interflowcarbonate deposits developed and plagioclase megacrysticflows were erupted Subsidence of the plateau is recordedby the interflow lenses and deposits of platform carbonates(40^1500m thick) on top of the Karmutsen basalts thatcontain Late Carnian to Early Norian fossils (Fig 16d c216^227 Ma Nixon et al 2008)The evolution of the internal architecture of the
Wrangellia oceanic plateau was primarily related to theconditions of emplacement (ie water depth and eruptionenvironment) There does not appear to have been a sys-tematic coincident evolution or shift of lava compositionsonVancouver Island as has been recognized in most con-tinental flood basalts and parts of the Wrangellia oceanicplateau in Alaska and Yukon (Greene et al 2008 2009)In general with the exception of the picritic pillow lavasin the submarine stratigraphy and local variations betweenhigh-MgO and tholeiitic lavas there are no significantcompositional differences between lavas erupted duringthe different stages of growth of the plateau (ie submarineand subaerial flows) nor are substantial compositional dif-ferences generally found between massive submarine flowsin contact with pillowed flows tholeiitic basalt flows nearthe base of the Karmutsen Formation are compositionallysimilar to flows sampled near the top of the volcanicsequence Where it was possible to sample continuous sec-tions of volcanic stratigraphy greater than several hun-dreds of meters (eg Keogh^Maynard Lake area west ofthe Karmutsen Range Fig 2) there is no clear relationshipbetween stratigraphic position and compositionalvariation
CONCLUSIONSThe Karmutsen Formation covers extensive areas ofVancouver Island (20 000 km2) and forms part of a largeoceanic plateau erupted over a geologically short interval(c 225^231 Ma) The tripartite volcanic stratigraphyexposed on Vancouver Island is at least 6 km thick andcomprises a succession of submarine flows volcaniclastic
deposits and massive subaerial flowsThese volcanologicaldifferences are primarily related to the eruption environ-ment (deep-water shallow-water or subaerial) The rapidgrowth of the plateau prevented significant accumulationsof intervening sediment except in the uppermost stratigra-phy where isolated limestone lenses commonly associatedwith pillow basalt and hyaloclastite deposits preserve arecord of the final stages of subsidence of the plateauThe Wrangellia plateau on Vancouver Island was con-
structed dominantly of tholeiitic basalt with restrictedmajor- and trace-element and isotopic compositional var-iations Minor volumes of picritic pillow basalts eruptedlate in the submarine phase The Karmutsen basalts wereemplaced onto older Paleozoic arc volcanic sequences andlow NbLa for the picrites may indicate a contributionfrom the sub-arc lithospheric mantle Modeling of the pet-rogenesis of the Karmutsen picrites indicates melting ofanomalously hot mantle (15008C) and extensive degreesof partial melting (23^27) and is consistent with aplume initiation model Combined Sr^Nd^Hf^Pb isotopicsystematics indicate that the picritic and tholeiitic lavaswere derived from a depleted plume-type mantle sourcedistinct from the source of MORB that has compositionalsimilarities to the source of the lavas forming theCaribbean Plateau Picrites were generated by melting ofspinel lherzolite whereas the tholeiitic basalts formed asaggregate melts of both garnet and spinel lherzolite Thetholeiitic basalts underwent extensive low-pressure fractio-nation (52^3 kbar) and seismic studies indicate that someof theWrangellia crust beneathVancouver Island may cor-respond to the plutonic residues of the Karmutsen floodbasalts
ACKNOWLEDGEMENTSWe are grateful to Nick Arndt for helping us get this pro-ject started and Nick Massey for insights into VancouverIsland geology We greatly appreciate the assistance withPRIMELT modeling from Claude Herzberg We thankMikkel Schau for his insight and enthusiasm during field-work Jane Barling assisted with analyses by MC-ICP-MSThoughtful and thorough reviews of the manuscript byJohn Mahoney Fred Frey Wendy Bohrson DougalJerram and Teal Riley are greatly appreciated Fundingwas generously provided by the Rocks to Riches Programadministered by the BC amp Yukon Chamber of Mines in2004 by the BC Geological Survey and by NSERCDiscovery Grants to JSS and DW ARG was supportedby a University Graduate Fellowship at UBC
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
34
REFERENCESBarker F Brown A S Budahn J R amp Plafker G (1989) Back-arc
with frontal-arc component origin of Triassic Karmutsen basaltBritish Columbia Canada Chemical Geology 75 81^102
Blichert-Toft JWeis D Maerschalk C Agranier A amp Albare de F(2003) Hawaiian hot spot dynamics as inferred from the Hf and Pbisotope evolution of Mauna Kea volcano Geochemistry GeophysicsGeosystems 4(2) 1^27 doi1010292002GC000340
Bondre N R Duraiswami R A amp Dole G (2004) Morphologyand emplacement of flows from the Deccan Volcanic ProvinceIndia Bulletin of Volcanology 66 29^45
Carlisle D (1963) Pillow breccias and their aquagene tuffs QuadraIsland British Columbia Journal of Geology 71 48^71
Carlisle D (1972) Late Paleozoic to Mid-Triassic sedimentary-volcanic sequence on Northeastern Vancouver Island Geological
Society of Canada Paper Report of Activities 72-1 Part B 24^30Carlisle D amp SuzukiT (1974) Emergent basalt and submergent car-
bonate^clastic sequences including the Upper Triassic Dilleri andWelleri zones on Vancouver Island Canadian Journal of Earth
Sciences 11 254^279Chauvel C amp Blichert-Toft J (2001) A hafnium isotope and trace
element perspective on melting of the depleted mantle Earth and
Planetary Science Letters 190 137^151Chayes F (1954)The theory of thin-section analysis Journal of Geology
62 92^101Cho M Liou J G amp Maruyama S (1987) Transition from the zeo-
lite to prehnite^pumpellyite facies in the Karmutsen metabasitesVancouver Island British Columbia Journal of Petrology 27(2)467^494
Clowes R M Zelt C A Amor J R amp Ellis R M (1995)Lithospheric structure in the southern Canadian Cordillera froma network of seismic refraction lines Canadian Journal of Earth
Sciences 32(10) 1485^1513Coffin M F amp Eldholm O (1994) Large igneous provinces Crustal
structure dimensions and external consequences Reviews of
Geophysics 32(1) 1^36Cox K G (1980) A model for flood basalt volcanism Journal of
Petrology 21 629^650Eldholm O amp Coffin M F (2000) Large igneous provinces
and plate tectonics In Richards M A Gordon R G amp vander Hilst R D (eds) The History and Dynamics of Global
Plate Motions American Geophysical Union Geophysical Monograph 121309^326
Elliott T R Hawkesworth C J amp Grolaquo nvold K (1991) Dynamicmelting of the Iceland plume Nature 351 201^206
Escuder-Viruete J Pecurren rez-Estaucurren n A Contreras F Joubert MWeis D Ullrich T D amp Spadea P (2007) Plume mantle sourceheterogeneity through time Insights from the Duarte ComplexHispaniola northeastern Caribbean Journal of Geophysical Research112(B04203) doi1010292006JB004323
Farnetani C G Richards M A amp Ghiorso M S (1996)Petrological models of magma evolution and deep crustal structurebeneath hotspots and flood basalt provinces Earth and Planetary
Science Letters 143 81^94Fitton J G amp Godard M (2004) Origin and evolution of magmas
on the Ontong Java Plateau In Fitton J G Mahoney J JWallace P J amp Saunders A D (eds) Origin and Evolution of the
Ontong Java Plateau Geological Society London Special Publications 229151^178
Fitton J G Saunders A D Norry M J Hardarson B S ampTaylor R N (1997) Thermal and chemical structure of theIceland plume Earth and Planetary Science Letters 153 197^208
Fitton J G Saunders A D Kempton P D amp Hardarson B S(2003) Does depleted mantle form an intrinsic part of the Icelandplume Geochemistry Geophysics Geosystems 4(3) 1032 doi1010292002GC000424
Frey F A Huang S Blichert-Toft J Regelous M amp Boyet M(2005) Origin of depleted components in basalt related to theHawaiian hotspot Evidence from isotopic and incompatible ele-ment ratios Geochemistry Geophysics Geosystems 6(1) 23 doi1010292004GC000757
Galer S J G amp Abouchami W (1998) Practical application of leadtriple spiking for correction of instrumental mass discriminationMineralogical Magazine 62A 491^492
Ghiorso M S amp Sack R O (1995) Chemical mass transfer in mag-matic processes IV A revised and internally consistent thermody-namic model for the interpolation and extrapolation of liquid^solidequilibria in magmatic systems at elevated temperatures and pres-sures Contributions to Mineralogy and Petrology 119 197^212
Greene A R Scoates J S amp Weis D (2005)WrangelliaTerrane onVancouver Island Distribution of flood basalts with implicationsfor potential Ni^Cu^PGE mineralization In Grant B (ed)Geological Fieldwork 2004 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2005-1 209^220Greene A R Scoates J S Nixon G T amp Weis D (2006) Picritic
lavas and basal sills in the Karmutsen flood basalt provinceWrangellia northern Vancouver Island In Grant B (ed)Geological Fieldwork 2005 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2006-1 39^52Greene A R Scoates J S amp Weis D (2008) Wrangellia flood
basalts in Alaska A record of plume^lithosphere interaction in aLate Triassic accreted oceanic plateau Geochemistry Geophysics
Geosystems 9(Q12004) doi1010292008GC002092Greene A R Scoates J S Weis D amp Israel S (2009)
Geochemistry of triassic flood basalts from the Yukon (Canada)segment of the accreted Wrangellia oceanic plateau Lithosdoi101016jlithos200811010
Hauff F Hoernle K Tilton G Graham D W amp Kerr A C(2000) Large volume recycling of oceanic lithosphere over shorttime scales geochemical constraints from the Caribbean LargeIgneous Province Earth and Planetary Science Letters 174 247^263
Hauff F Hoernle K amp Schmidt A (2003) Sr^Nd^Pb compositionof Mesozoic Pacific oceanic crust (Site 1149 and 801 ODP Leg185)Implications for alteration of ocean crust and the input into theIzu^Bonin^Mariana subduction system Geochemistry Geophysics
Geosystems 4(8) doi1010292002GC000421Herzberg C (2004) Partial melting below the Ontong Java Plateau
In Fitton J G Mahoney J J Wallace P J amp Saunders A D(eds) Origin and Evolution of the Ontong Java Plateau Geological SocietyLondon Special Publications 229 179^183
Herzberg C (2006) Petrology and thermal structure of the Hawaiianplume from Mauna Kea volcano Nature 444(7119) 605
Herzberg C amp Asimow P D (2008) Petrology of some oceanicisland basalts PRIMELT2XLS software for primary magma cal-culation Geochemistry Geophysics Geosystems 9(Q09001) doi1010292008GC002057
Herzberg C amp OrsquoHara M J (2002) Plume-associated ultramaficmagmas of Phanerozoic age Journal of Petrology 43(10) 1857^1883
Herzberg C Asimow P D Arndt N Niu Y Lesher C MFitton J G Cheadle M J amp Saunders A D (2007)Temperatures in ambient mantle and plumes Constraints frombasalts picrites and komatiites Geochemistry Geophysics Geosystems8(Q02006) doi1010292006GC001390
Huang S amp Frey F A (2005) Trace element abundances of MaunaKea basalt from phase 2 of the Hawaii Scientific Drilling Project
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
35
Petrogenetic implications of correlations with major element con-tent and isotopic ratios Geochemistry Geophysics Geosystems 4(6)doi1010292002GC000322
Hussong D M Wipperman L K amp Kroenke L W (1979) Thecrustal structure of the Ontong Java and Manihiki OceanicPlateaus Journal of Geophysical Research 84(B11) 6003^6010
Jerram D A amp Widdowson M (2005) The anatomy of ContinentalFlood Basalt Provinces geological constraints on the processes andproducts of flood volcanism Lithos 79(3^4) 385^405
Jones D L Silberling N J amp Hillhouse J (1977)Wrangellia a dis-placed terrane in northwestern North America CanadianJournal ofEarth Sciences 14(11) 2565^2577
Kerr A C (2003) Oceanic Plateaus In Rudnick R (ed) The Crust
Treatise on GeochemistryVol 3 Oxford Elsevier Science pp 537^565Kerr A C amp Mahoney J J (2007) Oceanic plateaus Problematic
plumes potential paradigms Chemical Geology 241 332^353Kerr A CTarney J Marriner G F Klaver GT Saunders A D
amp Thirlwall M F (1996) The geochemistry and petrogenesis ofthe Late-Cretaceous picrites and basalts of Curacao NetherlandsAntilles a remnant of an oceanic plateau Contributions to
Mineralogy and Petrology 124(1) 29^43Kerr A C Marriner G F Tarney J Nivia A Saunders A D
Thirlwall M F amp Sinton A C (1997) Cretaceous basaltic ter-ranes in Western Colombia Elemental chronological and Sr-Ndisotopic constraints on petrogenesis Journal of Petrology 38(6)677^702
Kinzler R J (1997) Melting of mantle peridotite at pressuresapproaching the spinel to garnet transition Application to mid-ocean ridge basalt petrogenesis Journal of Geophysical Research
102(B1) 853^874Kinzler R J amp Grove T L (1992) Primary magmas of mid-ocean
ridge basalts 1 Experiments and methods Journal of Geophysical
Research 97(B5) 6885^6906Korenaga J (2005) Why did not the Ontong Java Plateau form sub-
aerially Earth and Planetary Science Letters 234 385^399Kuniyoshi S (1972) Petrology of the Karmutsen (Volcanic) Group
northeastern Vancouver Island British Columbia PhD disserta-tion University of California Los Angeles 242 pp
Lassiter J C DePaolo D J amp Mahoney J J (1995) Geochemistry ofthe Wrangellia flood basalt province Implications for the role ofcontinental and oceanic lithosphere in flood basalt genesis Journalof Petrology 36(4) 983^1009
LeBas M J (2000) IUGS reclassification of the high-Mg and picriticvolcanic rocks Journal of Petrology 41(10) 1467^1470
Longhi J (2002) Some phase equilibrium systematics of lherzolitemelting I Geochemistry Geophysics Geosystems 3(3) doi1010292001GC000204
MacDonald G A amp Katsura T (1964) Chemical composition ofHawaiian lavas Journal of Petrology 5 82^133
Mahoney J J (1987) An isotopic survey of Pacific oceanic plateausimplications for their nature and origin In Keating B HFryer P Batiza R amp Boehlert GW (eds) Seamounts Islands andAtolls American Geophysical Union Geophysical Monograph 43 207^220
Mahoney J LeRoex A P Peng Z Fisher R L amp Natland J H(1992) Southwestern limits of Indian Ocean Ridge mantle and theorigin of low 206Pb204Pb mid-ocean ridge basalt isotope systema-tics of the central Southwest Indian Ridge (178^508E) Journal ofGeophysical Research 97(B13) 19771^19790
Mahoney J J Sinton J M Kurz M D MacDougall J DSpencer K J amp Lugmair GW (1994) Isotope and trace elementcharacteristics of a super-fast spreading ridge East Pacific rise13^238S Earth and Planetary Science Letters 121 173^193
Massey N W D (1995a) Geology and mineral resources of theAlberni^Nanaimo Lakes sheet Vancouver Island 92F1W 92F2Eand part of 92F7E British Columbia Ministry of Energy Mines and
Petroleum Resources Paper 1992-2 132 ppMassey N W D (1995b) Geology and mineral resources of the
Cowichan Lake sheet Vancouver Island 92C16 British Columbia
Ministry of Energy Mines and Petroleum Resources Paper 1992-3 112 ppMassey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005a) Digital Geology Map of British Columbia Tile NM9 MidCoast BC British Columbia Ministry of Energy and Mines Geofile
2005-2Massey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005b) Digital Geology Map of British Columbia Tile NM10Southwest BC British Columbia Ministry of Energy and Mines Geofile
2005-3McDonough W F amp Sun S (1995) The composition of the Earth
Chemical Geology 120 223^253Monger JW H amp Journeay J M (1994) Basement geology and tec-
tonic evolution of theVancouver region In Monger JW H (ed)Geology and Geological Hazards of the Vancouver Region Southwestern
British Columbia Geological Survey of Canada Bulletin 481 3^25Muller J E (1977) Geology of Vancouver Island Geological Survey of
Canada Open File Map 463Muller J E (1980)The Paleozoic Sicker Group of Vancouver Island British
Columbia Geological Survey of Canada Paper 79-30 22 ppMuller J E Northcote K E amp Carlisle D (1974) Geology and
mineral deposits of Alert Bay^Cape Scott map area VancouverIsland British Columbia Geological Survey of Canada Paper 74-8 77
Neal C R Mahoney J J Kroenke L W Duncan R A ampPetterson M G (1997) The Ontong^Java Plateau InMahoney J J amp Coffin M F (eds) Large Igneous Provinces
Continental Oceanic and Planetary FloodVolcanism American Geophysical
Union Geophysical Monograph 100 183^216Niu Y Collerson K D Batiza R Wendt J I amp Regelous M
(1999) Origin of enriched-typed mid-ocean ridge basalt at ridgesfar from mantle plumes The East Pacific Rise at 11820rsquoN Journalof Geophysical Research 104 7067^7088
Nixon G T amp Orr A J (2007) Recent revisions to the EarlyMesozoic stratigraphy of Northern Vancouver Island (NTS 102I092L) and metallogenic implicatinos British Columbia InGrant B (ed) Geological Fieldwork 2006 British Columbia Ministry of
Energy Mines and Petroleum Resources Paper 2007-1 163^177Nixon GT Hammack J L KoyanagiV M Payie G J Haggart
JW Orchard M JTozerT Friedman R M Archibald D APalfy J amp Cordey F (2006a) Geology of the Quatsino^PortMcNeill area northernVancouver Island British Columbia Ministry
of Energy Mines and Petroleum Resources Geoscience Map 2006-2 scale150 000
Nixon G T Kelman M C Stevenson D Stokes L A ampJohnston K A (2006b) Preliminary geology of the Nimpkishmap area (NTS 092L07) northern Vancouver Island BritishColumbia In Grant B amp Newell J M (eds) Geological Fieldwork2005 British Columbia Ministry of Energy Mines and Petroleum Resources
Paper 2006-1 135^152Nixon G T Laroque J Pals A Styan J Greene A R amp
Scoates J S (2008) High-Mg lavas in the Karmutsen flood basaltsnorthernVancouver Island (NTS 092L) Stratigraphic setting andmetallogenic significance In Grant B (ed) Geological Fieldwork
2007 British Columbia Ministry of Energy Mines and Petroleum
Resources Paper 2008-1 175^190Nowell G M Kempton P D Noble S R Fitton J G
Saunders A Mahoney J J amp Taylor R N (1998) High precisionHf isotope measurements of MORB and OIB by thermal ionisation
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
36
mass spectrometry insights into the depleted mantle Chemical
Geology 149 211^233Parrish R R amp McNicoll V J (1992) U^Pb age determinations
from the southern Vancouver Island area British Columbia InRadiogenic Age and Isotopic Studies Report 5 Geological Survey of
Canada Paper 91-2 79^86Peate DW (1997) The Parana^Etendeka Province In Coffin M F
amp Mahoney J (eds) Large Igneous Provinces Continental Oceanic andPlanetary Flood Volcanism American Geophysical Union Geophysical
Monograph 100 217^245Perfit M R Fornari D J Ridley W I Kirk P D Casey J
Kastens K A Reynolds J R Edwards M Desonie DShuster R amp Paradis S (1996) Recent volcanism in the Siqueirostransform fault picritic basalts and implications for MORBmagma genesis Earth and Planetary Science Letters 141(1^4) 91^108
Pretorius W Weis D Williams G Hanano D Kieffer B ampScoates J S (2006) Complete trace elemental characterization ofgranitoid (USGSG-2GSP-2) reference materials by high resolutioninductively coupled plasma-mass spectrometry Geostandards and
Geoanalytical Research 30(1) 39^54Regelous M Niu Y Wendt J I Batiza R Greig A amp
Collerson K D (1999) Variations in the geochemistry of magma-tism on the East Pacific Rise at 10830rsquoN since 800 ka Earth and
Planetary Science Letters 168 45^63Recurren villon S Chauvel C Arndt N T Pik R Martineau F
Fourcade S amp Marty B (2002) Heterogeneity of the Caribbeanplateau mantle source Sr O and He isotopic compositions of oli-vine and clinopyroxene from Gorgona Island Earth and Planetary
Science Letters 205(1^2) 91Rhodes J M (1996) Geochemical stratigraphy of lava flows samples
by the Hawaii Scientific Drilling Project Journal of Geophysical
Research 101(B5) 11729^11746Rhodes J M amp Vollinger M J (2004) Composition of basaltic lavas
sampled by phase-2 of the Hawaii Scientific Drilling ProjectGeochemical stratigraphy and magma types Geochemistry
Geophysics Geosystems 5(3) doi1010292002GC000434Richards M A Jones D L Duncan R A amp DePaolo D J (1991)
A mantle plume initiation model for the Wrangellia flood basaltand other oceanic plateaus Science 254 263^267
Salters V J M amp Hart S R (1991) The mantle sources of oceanridges islands and arcs the Hf-isotope connection Earth and
Planetary Science Letters 104 364^380Salters V J M amp Stracke A (2004) Composition of the depleted
SaltersV J amp WhiteW (1998) Hf isotope constraints on mantle evo-lution Chemical Geology 145 447^460
SanoT amp Yamashita S (2004) Experimental petrology of basementlavas from Ocean Drilling Program Leg 192 implications for dif-ferentiation processes in Ontong Java Plateau magmas InFitton J G Mahoney J JWallace P J amp Saunders A D (eds)Origin and Evolution of the Ontong Java Plateau Geological Society
London Special Publications 229 185^218Saunders A D (2005) Large igneous provinces Origin and environ-
mental consequences Elements 1 259^263Self S ThordarsonT amp Keszthely L (1997) Emplacement of conti-
nental flood basalt lava flows In Mahoney J J amp Coffin M F(eds) Large Igneous Provinces Continental Oceanic and Planetary Flood
Volcanism American Geophysical Union Geophysical Monograph 100381^410
Shaw D M (2000) Continuous (dynamic) melting theory revisitedCanadian Mineralogist 38(5) 1041^1063
Sluggett C L (2003) Uranium^lead age and geochemical constraintson Paleozoic and Early Mesozoic magmatism in WrangelliaTerrane Saltspring Island British Columbia BSc thesisUniversity of British ColumbiaVancouver 84 pp
Storey M Mahoney J J amp Saunders A D (1997) Cretaceousbasalts in Madagascar and the transition between plume and con-tinental lithospheric mantle sources In Mahoney J J ampCoffin M F (eds) Large Igneous Provinces Continental Oceanic and
Planetary Flood Volcanism Aamerican Geophysical Union Geophysical
Monograph 100 95^122Surdam R C (1967) Low-grade metamorphism of the Karmutsen
Group PhD dissertation University of California Los Angeles288 pp
Sutherland-Brown A Yorath C J Anderson R G amp Dom K(1986) Geological maps of southern Vancouver IslandLITHOPROBE I 92C10 11 14 16 92F1 2 7 8 Geological Survey ofCanada Open File 1272
Tejada M L G Mahoney J J Castillo P R Ingle S P Sheth HC amp Weis D (2004) Pin-pricking the elephant evidence on theorigin of the Ontong Java Plateau from Pb^Sr^Hf^Nd isotopiccharacteristics of ODP Leg 192 basalts In Fitton J GMahoney J J Wallace P J amp Saunders A D (eds) Origin and
Evolution of the Ontong Java Plateau Geological Society London Special
Publications 229 133^150Thompson P M E Kempton P D amp Kerr A C (2008) Evaluation
of the effects of alteration and leaching on Sm^Nd and Lu^Hf sys-tematics in submarine mafic rocks Lithos 104(1^4) 164^176
Thompson P M E Kempton P D White R V Kerr A CTarney J Saunders A D Fitton J G amp McBirney A (2003)Hf-Nd isotope constraints on the origin of the CretaceousCaribbean plateau and its relationship to the Galapagos plumeEarth and Planetary Science Letters 217(1^2) 59^75
Vervoort J D Patchett P Blichert-Toft J amp Albare de F (1999)Relationships between Lu^Hf and Sm^Nd isotopic systems in theglobal sedimentary system Earth and Planetary Science Letters 16879^99
Walter M J (1998) Melting of garnet peridotite and the origin ofkomatiite and depleted lithosphere Journal of Petrology 39(1) 29^60
Weis D Kieffer B Maerschalk C Barling J de Jong JWilliams G A Hanano D Mattielli N Scoates J SGoolaerts A Friedman R A amp Mahoney J B (2006) High-pre-cision isotopic characterization of USGS reference materials byTIMS and MC-ICP-MS Geochemistry Geophysics Geosystems
7(Q08006) doi1010292006GC001283Weis D Kieffer B Hanano D Silva I N Barling J
Pretorius W Maerschalk C amp Mattielli N (2007) Hf isotopecompositions of US Geological Survey reference materialsGeochemistry Geophysics Geosystems 8(Q06006) doi1010292006GC001473
Wheeler J O amp McFeely P (1991) Tectonic assemblage map of theCanadian Cordillera and adjacent part of the United States ofAmerica Geological Survey of Canada Map 1712A
WhiteW M Albare de F amp Tecurren louk P (2000) High-precision analy-sis of Pb isotope ratios by multi-collector ICP-MS Chemical Geology167 257^270
Yorath C J Sutherland Brown A amp Massey N W D (1999)LITHOPROBE southern Vancouver Island British Columbia Geological
Survey of Canada Bulletin 498 145 ppZou H (1998) Trace element fractionation during modal and
non-model dynamic melting and open-system melting Amathematical treatment Geochimica et Cosmochimica Acta 62(11)1937^1945
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
37
Zou H amp Reid M R (2001) Quantitative modeling of trace elementfractionation during incongruent dynamic melting Geochimica et
Cosmochimica Acta 65(1) 153^162
APPENDIX SAMPLEPREPARAT ION ANDANALYT ICAL METHODSOnly the freshest rocks in the field were sampled and onlythe least altered samples were selected for chemical andisotopic analysis based on thorough petrographic inspec-tion Fifty-nine of the 129 collected samples were crushed(400 g) into small pieces 52mm in diameter in aRocklabs hydraulic piston crusher between WC plates tominimize contamination The coarse-crush was mixedand 100 g was powdered in a planetary mill using agatejars and balls cleaned with quartz sand between samples
ActLabs analytical methodsThe major- and trace-element compositions of the whole-rock powders were determined at Activation LaboratoriesLtd (Actlabs) in Ancaster Ontario Analytical techniquesand detection limits are also available from Actlabs (httpwwwactlabscommethsub_code4erehtm) The analyticalmethod for each of the elements analyzed is indicated inTable 2 For the major elements a 02 g sample was mixedwith a lithium metaborate^lithium tetraborate mixtureand fused in a graphite crucible The molten mixture waspoured into a 5 HNO3 solution and shaken until dis-solved (30min) The samples were analyzed for majoroxides and selected trace elements by inductively coupledplasma optical emission spectrometry (ICP-OES) on acombination simultaneous^sequential Thermo Jarrell-AshEnviro II system Internal calibration was achieved usinga variety of international reference materials (eg W-2BIR-1 DNC-1) and independent control samplesAdditional trace elements were analyzed by both instru-mental neutron activation analysis (INAA) and induc-tively couple plasma mass spectrometry (ICP-MS) ForINAA15^25 g of sample was weighed into small polyeth-ylene vials and irradiated with a control international ref-erence material CANMET WMS-1 and NiCr flux wires ata thermal neutron flux of 71012 neutronscm2 per s in theMcMaster Nuclear Reactor Following a 7 day waitingperiod the samples were measured on an Ortec high-purity Ge detector linked to a Canberra Series 95 multi-channel analyzer Activities for each element weredecay- and weight-corrected and compared with a detectorcalibration developed from multiple international certifiedreference materials For ICP-MS 025 g of sample wasdigested in HF followed by HNO3^HClO4 heated andtaken to dryness The samples were brought back into solu-tion with HCl Samples were analyzed using a PerkinElmer Optima 3000 ICP In-lab standards or certified ref-erence materials (eg W-2 BIR-1 DNC-1) were used for
quality control A total of 15 blind duplicates were ana-lyzed from the same sample powders to assessreproducibility
University of Massachusetts analyticalmethodsFourteen sample duplicate powders taken from the same100 g sample aliquot (high-MgO basalts and picrites)were also analyzed at the Ronald B Gilmore X-RayFluorescence (XRF) Laboratory at the University ofMassachusetts Major elements were measured on a fusedLa-bearing lithium borate glass disc using a SiemensMRS-400 spectrometer Trace element concentrations(Rb Sr Ba Ce Nb Zr Y Pb Zn Ga Ni Cr V) weremeasured on a separate powder pellet using a PhilipsPW2400 sequential spectrometer using a Rh tube LOIand ferrous iron measurements were made as describedby Rhodes amp Vollinger (2004) Precision and accuracy esti-mates for the data have been described by Rhodes (1996)and Rhodes amp Vollinger (2004) Results for each sampleare the average of two separate analyses and are shown inElectronic Appendix 2 for a comparison of analyses ofduplicate powders using XRFand ICP-MS
PCIGR analytical methodsA subset of 19 samples was selected for high-precisiontrace-element analysis and Sr Nd Pb and Hf isotopicanalysis at the Pacific Centre for Isotopic andGeochemical Research (PCIGR) at the University ofBritish Columbia (UBC) Samples were selected from the59 samples analyzed for whole-rock chemistry at ActLabsbased on major- and trace-element chemistry alteration(low LOI and petrographic alteration index) and samplelocation Samples were prepared for trace-element analysisat the PCIGR by the technique described by Pretorius et al(2006) on unleached rock powders Sample powders(100mg) were weighed in 7ml screw-top Savillex bea-kers and dissolved in 1ml 14N HNO3 and 5ml 48 HFon a hotplate for 48 h at 1308C with periodic ultrasonica-tion Samples were dried and redissolved in 6ml 6N HClon a hotplate for 24 h and then dried and redissolved in1ml concentrated HNO3 for 24 h before final dryingTrace element abundances were measured by high-resolution inductively coupled plasma-mass spectrometry(HR-ICP-MS) with aThermo Finnigan Element2 systemfollowing the procedures described by Pretorius et al(2006) within 24 h of redissolution HFSE and LILE weremeasured in medium-resolution mode at 2000 dilutionusing a PFATeflon spray chamber washed with aqua regiafor 3min between samples REE were measured in high-resolution mode and U and Pb in low-resolution modeat 2000 dilution using a glass spray chamber washedwith 2 HNO3 between samples Total procedural blanksand reference materials (BCR-2 BHVO-2) were analyzedwith the batch of samples Indium was used as an internal
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
38
standard in all samples and standard solutionsBackground and standard solutions were analyzed afterevery five samples to detect memory effects and massdrift Results for PCIGR trace-element analyses areshown in Electronic Appendix 3 for the full set of analyzedtrace elementsSample digestion for purification of Sr Nd Hf and Pb
for column chemistry began by weighing each samplepowder (400^500mg) prior to leaching All samples wereleached with 6N HCl and placed in an ultrasonic bath for15min Samples were rinsed twice with 18 M-cm H2Obetween each leaching step (15 total) until the supernatantliquid was clear [following the method of Mahoney (1987)]Samples were then dried on a hotplate for 24 h andweighed again Sample solutions were then prepared bydissolving 100^250mg of the leached powder dissolvedin 1ml 14N HNO3 and 10ml 48 HF on a hotplate for48 h at 1308C with periodic ultrasonication Samples weredried and redissolved in 6ml 6N HCl on a hotplate for24 h and then dried Pb was separated using anion exhangecolumns and the discard was used for Sr REE and Hfseparation Nd was separated from the other REE and Hfrequired two additional purification steps Detailed proce-dures for column chemistry for separating Sr Nd and Pbat the PCIGR have been described byWeis et al (2006) andHf purification has been described byWeis et al (2007) Srand Nd isotope ratios were measured by thermal ioniza-tion mass spectrometry (TIMS) on a Thermo FinniganTriton system in static mode with relay matrix rotation ona single Ta and double Re^Ta filament respectively Fourto five filaments per barrel of 21 were occupied by stan-dards (NBS 987 for Sr and LaJolla for Nd) for each barrelwhere samples were run Sample Sr and Nd isotopic com-positions were corrected for mass fractionation using86Sr88Srfrac14 01194 and 146Nd144Ndfrac14 07219 Each samplewas then normalized using the barrel average of the refer-ence material relative to the values of143Nd144Ndfrac14 0511858 and 87Sr86Srfrac14 0710248 (Weiset al 2006) During the course of the Vancouver Islandanalyses the La Jolla Nd standard gave an average valueof 05118566 (nfrac14 7) and NBS987 standard gave an aver-age of 07102408 (nfrac1411) (2 error is reported as times106) 147Sm144Nd errors are 15 or 0006 USGeological Survey (USGS) reference material BHVO-2was processed with the samples and yielded Sr and Nd iso-topic ratios of 07034607 and 05129786 respectivelyThese are in agreement with the published values of070347920 and 051298411 respectively (Weis et al2006)
Pb and Hf isotopic compositions were analyzed by mul-tiple collector-inductively coupled plasma-mass spectrome-try (MC-ICP-MS) on a static multi-collection on a NuPlasma (Nu Instruments) system The detailed analyticalprocedure for Pb isotopic analyses on the Nu Plasma atthe PCIGR has been described by Weis et al (2006) Theconfiguration for Pb analyses allows for collection of PbTl and Hg togetherTl and Hg are used to monitor instru-mental mass discrimination and isobaric overlap repec-tively All sample solutions were analyzed withapproximately the same PbTl ratio (4) as the referencematerial NIST SRM 981 To accomplish this a small ali-quot of each sample solution from the Pb columns was ana-lyzed on the Element2 to determine the precise amount ofPb available for analysis on the Nu Plasma The SRM 981standard was run after every two samples on the NuPlasma During the time samples were run analyses ofthe SRM 981 Pb reference material gave values of206Pb204Pbfrac1416940322 207Pb204Pbfrac1415495823and 208Pb204Pbfrac14 36713164 (nfrac14 61 2 error is reportedas times 104) these values are in excellent agreement withreported TIMS triple-spike values of Galer amp Abouchami(1998) Fractionation-corrected Pb isotopic ratios were fur-ther corrected by the sample-standard bracketing methodor the ln^ln correction method described by White et al(2000) and Blichert-Toft et al (2003) Leached powders ofUSGS reference material BHVO-2 yielded Pb isotopicratios of 206Pb204Pbfrac14186454 8 207Pb204Pbfrac14 1549105 and 208Pb204Pbfrac14 38222514 and BCR-2yielded 206Pb204Pbfrac141880466 207Pb204Pbfrac14 1562518 and 208Pb204Pbfrac14 3883496 (2 error isreported as times 104) These values are in agreement withthose for leached residues of BHVO-2 and BCR-2 fromWeis et al (2006)Hf isotopic compositions were analyzed following the
procedures detailed byWeis et al (2007) The configurationfor Hf analyses monitored Lu mass 175 andYb mass 172 toallow for interference correction to masses 174 and 176 Hfisotopic ratios were normalized internally for mass fractio-nation to a 179Hf177Hf ratio of 07325 using an exponentialcorrection Standards were run after every two samplesand sample results were normalized to the ratio of the in-run daily average and a 176Hf177Hf ratio for JMC-475 of0282160 During the course of analyses the Hf standardJMC-475 gave an average value 02821679 (nfrac14130)USGS reference materials BCR-2 and BHVO-2 were pro-cessed with the samples and yielded Hf isotopic ratios of02828675 and 02831005 respectively Publishedvalues for BCR-2 and BHVO-2 are 02828717 and0283104 8 respectively (Weis et al 2007)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
1Ni concentrations for these samples were determined by XRFTHOL tholeiitic basalt PIC picrite HI-MG high MgO basalt CG coarse-grained (sill or gabbro) MIN SIL mineralizedsill OUTLIER anomalous pillowed flow in plots MA Mount Arrowsmith SL Schoen Lake KR Karmutsen Range QIQuadra Island Sample locations are given using the Universal Transverse Mercator (UTM) coordinate system (NAD83zones 9 and 10) Analyses were performed at Activation Laboratory (ActLabs) Fe2O3
is total iron expressed as Fe2O3LOI loss on ignition All major elements Sr V and Y were measured by ICP quadrupole OES on solutions of fusedsamples Cu Ni Pb and Zn were measured by total dilution ICP Cs Ga Ge Hf Nb Rb Ta Th U Zr and REE weremeasured by magnetic-sector ICP on solutions of fused samples Co Cr and Sc were measured by INAA Blanks arebelow detection limit See Electronic Appendices 2 and 3 for complete XRF data and PCIGR trace element datarespectively Major elements for sample 93G17 are normalized anhydrous Sample 5615A7 was analyzed by XRFSamples from Quadra Island are not shown in figures
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
17
picritic and high-MgO basalt pillow lavas extend to20wt MgO (Fig 6 and references listed in thecaption)The tholeiitic basalts from the Karmutsen Range dis-
play a tight range of parallel light rare earth element(LREE)-enriched REE patterns (LaYbCNfrac14 21^24mean 2202) whereas the picrites and high-MgObasalts are characterized by sub-parallel LREE-depleted
patterns (LaYbCNfrac14 03^07 mean 06 02) with lowerREE abundances than the tholeiitic basalts (Fig 7)Tholeiitic basalts from the Schoen Lake and MountArrowsmith areas have similar REE patterns tosamples from the Karmutsen Range (Schoen Lake LaYbCNfrac14 21^26 mean 2303 Mount Arrowsmith LaYbCNfrac1419^25 mean 2204) and the coarse-grainedmafic rocks have similar REE patterns to the tholeiitic
Depleted MORB average(Salters amp Stracke 2004)
Schoen LakeSchoen Lake
Mount Arrowsmith
Karmutsen RangeS
ampl
eC
hond
rite
Sam
ple
Cho
ndrit
e
Sam
ple
Prim
itive
Man
tleS
ampl
eP
rimiti
ve M
antle
Sam
ple
Prim
itive
Man
tle
Mount Arrowsmith
Karmutsen Range
(a) (b)
(c) (d)
(e) (f )
Sam
ple
Cho
ndrit
e
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained mafic rock
Pillowed flowMineralized sill
1
10
100
1
10
100
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
01
1
10
100
1
10
100
1
10
100
Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
Fig 7 Whole-rock REE and other incompatible-element concentrations for the Karmutsen Formation (a) (c) and (e) are chondrite-normal-ized REE patterns for samples from each of the three main field areas onVancouver Island (b) (d) and (f) are primitive mantle-normalizedtrace-element patterns for samples from each of the three main field areas All normalization values are from McDonough amp Sun (1995) Theclear distinction between LREE-enriched tholeiitic basalts and LREE-depleted picrites the effects of alteration on the LILE (especially loss ofK and Rb) and the different range for the vertical scale in (b) (extends down to 01) should be noted
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
18
basalts (LaYbCNfrac14 21^29 mean 2508) Two LREE-depleted coarse-grained mafic rocks from the SchoenLake area have similar REE patterns (LaYbCNfrac14 07) tothe picrites from the Keogh Lake area indicating that thisdistinctive suite of rocks is exposed as far south asSchoen Lake that is 75 km away (Fig 7) The anomalouspillowed flow from the Karmutsen Range has lower LaYbCN (13^18) than the main group of tholeiitic basaltsfrom the Karmutsen Range (Fig 7) The mineralized sillfrom the Schoen Lake area is distinctly LREE-enriched(LaYbCNfrac14 33) with the highest REE abundances(LaCNfrac14 940) of all Karmutsen samples this may reflectcontamination by adjacent sediment also evident in thin-section where sulfide blebs are cored by quartz (Greeneet al 2006)The primitive mantle-normalized trace-element pat-
terns (Fig 7) and trace-element variations and ratios(Fig 8) highlight the differences in trace-element
concentrations between the four main groups ofKarmutsen flood basalts onVancouver Island The tholeii-tic basalts from all three areas have relatively smooth par-allel trace-element patterns some samples have smallpositive Sr anomalies relative to Nd and Sm and manysamples have prominent negative K anomalies relative toU and Nb reflecting K loss during alteration (Fig 7) Thelarge ion lithophile elements (LILE) in the tholeiiticbasalts (mainly Rb and K not Cs) are depleted relativeto the high field strength elements (HFSE) and LREELILE segments especially for the Schoen Lake area(Fig 7d) are remarkably parallel The picrites andhigh-MgO basalts form a tight band of parallel trace-element patterns and are depleted in HFSE and LREEwith positive Sr anomalies and relatively low Rb Thecoarse-grained mafic rocks from the Karmutsen Rangehave identical trace-element patterns to the tholeiiticbasalts Samples from the Schoen Lake area have similar
000
025
050
075
100
125
0 1 2 3 4 5
00
04
08
12
16
0 5 10 15 20 25
MgO (wt)
0
5
10
15
0 50 100 150
Hf (ppm)
Th (ppm)
NbLaNb (ppm)
Zr (ppm)
(a)
(c)
(b)
(d)
Th (ppm)
00
02
04
06
08
10
12
00 01 02 03 04 05
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
NbLa=1
4723A4 5615A12
U (ppm)
Fig 8 Whole-rock trace-element concentrations and ratios for the Karmutsen Formation [except (b) which is vs wt MgO] (a) Nb vs Zr(b) NbLa vs MgO (c) Th vs Hf (d) Th vs U There is a clear distinction between the tholeiitic basalts and picrites in Nb and Zr both inconcentration and the slope of each trend
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
19
trace-element patterns to the Karmutsen Range except forthe two LREE-depleted coarse-grained mafic rocks andthe mineralized sill (Fig 7d)
Sr^Nd^Hf^Pb isotopic compositionsA total of 19 samples from the four main groups ofthe Karmutsen Formation were selected for radiogenicisotopic geochemistry on the basis of major- and trace-element variation stratigraphic position and geographicallocation The samples have overlapping age-correctedHf and Nd isotope ratios and distinct ranges of Sr isotopiccompositions (Fig 9) The tholeiitic basalts picriteshigh-MgO basalt and coarse-grained mafic rocks have
initial eHffrac14thorn87 to thorn126 and eNdfrac14thorn66 to thorn88 cor-rected for in situ radioactive decay since 230 Ma (Fig 9Tables 3 and 4) All the Karmutsen samples form a well-defined linear array in a Lu^Hf isochron diagram corre-sponding to an age of 24115 Ma (Fig 9) This is withinerror of the accepted age of the Karmutsen Formationand indicates that the Hf isotope systematics behaved as aclosed system since c 230 Ma The anomalous pillowedflow has similar initial eHf (thorn98) and slightly lower initialeNd (thorn62) compared with the other Karmutsen samplesThe picrites and high-MgO basalt have higher initial87Sr86Sr (070398^070518) than the tholeiitic basalts(initial 87Sr86Srfrac14 070306^070381) and the coarse-grained mafic rocks (initial 87Sr86Srfrac14 070265^070428)
05128
05129
05130
05131
05132
015 017 019 021 023 025
02829
02830
02831
02832
02833
02834
000 002 004 006 008 010
4
6
8
10
0702 0703 0704 0705 07066
8
10
12
14
5 6 7 8 9 10
12
0
1
2
3
4
5
6
7
0702 0703 0704 0705 0706
(d)
LOI (wt )
87Sr 86Sr
4723A2
Nd
143Nd144Nd
147Sm144Nd
87Sr 86Sr (230 Ma)
(230 Ma)
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
(a)
(c)
176Hf177Hf
Age=241 plusmn 15 Ma=+990 plusmn 064
176Lu177Hf
Hf (230 Ma)
Nd (230 Ma)
(e)
Hf (i)
(b)
Fig 9 Whole-rock Sr Nd and Hf isotopic compositions for the Karmutsen Formation (a) 143Nd144Nd vs 147Sm144Nd (b) 176Hf177Hf vs176Lu177Hf The slope of the best-fit line for all samples corresponds to an age of 24115 Ma (c) Initial eNd vs 87Sr86Sr Age-correction to230 Ma (d) LOI vs 87Sr86Sr (e) Initial eHf vs eNd Average 2 error bars are shown in a corner of each panel Complete chemical duplicatesshown inTables 3 and 4 (samples 4720A7 and 4722A4) are circled in each plot
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
overlap the ranges of the picrites and tholeiitic basalts(Fig 9 Table 3)The measured Pb isotopic compositions of the tholeiitic
basalts are more radiogenic than those of the picrites andthe most magnesian Keogh Lake picrites have the leastradiogenic Pb isotopic compositions The range ofinitial Pb isotope ratios for the picrites is206Pb204Pbfrac1417868^18580 207Pb204Pbfrac1415547^15562and 208Pb204Pbfrac14 37873^38257 and the range for thetholeiitic basalts is 206Pb204Pbfrac1418782^19098207Pb204Pbfrac1415570^15584 and 208Pb204Pbfrac14 37312^38587 (Fig 10 Table 5) The coarse-grained mafic rocksoverlap the range of initial Pb isotopic compositions forthe tholeiitic basalts with 206Pb204Pbfrac1418652^19155207Pb204Pbfrac1415568^15588 and 208Pb204Pbfrac14 38153^38735 One high-MgO basalt has the highest initial206Pb204Pb (19242) and 207Pb204Pb (15590) and thelowest initial 208Pb204Pb (37676) The Pb isotopic ratiosfor Karmutsen samples define broadly linearrelationships in Pb isotope plots The age-corrected Pb
isotopic compositions of several picrites have been affectedby U and Pb (andor Th) mobility during alteration(Fig 10)
ALTERAT IONThe Karmutsen basalts have retained most of their origi-nal igneous structures and textures however secondaryalteration and low-grade metamorphism have producedzeolitic- and prehnite^pumpellyite-bearing mineral assem-blages (Table 1 Cho et al 1987) Alteration and metamor-phism have primarily affected the distribution of the LILEand the Sr isotopic systematics The Keogh Lake picriteshave been affected more by alteration than the tholeiiticbasalts and have higher LOI (up to 55wt ) variableLILE and higher measured Sr Nd and Hf and lowermeasured Pb isotope ratios than the tholeiitic basalts Srand Pb isotope compositions for picrites and tholeiiticbasalts show a relationship with LOI (Figs 9 and 10)whereas Nd and Hf isotopic compositions do not correlate
Table 3 Sr and Nd isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Rb Sr 87Sr86Sr 2m87Rb86Sr 87Sr86Srt Sm Nd 143Nd144Nd 2m eNd
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses carried out at the PCIGR thecomplete trace element analyses are given in Electronic Appendix 3
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
21
with LOI (not shown) The relatively high apparent initialSr isotopic compositions for the high-MgO lavas (up to07052) probably resulted from post-magmatic loss of Rbor an increase in 87Sr86Sr through addition of seawater Sr(eg Hauff et al 2003) High-MgO lavas from theCaribbean plateau have similarly high 87Sr86Sr comparedwith basalts (Recurren villon et al 2002)The correction for in situdecay on initial Pb isotopic ratios has been affected bymobilization of U and Th (and Pb) in the whole-rockssince their formation A thorough acid leaching duringsample preparation was used that has been shown to effec-tively remove alteration phases (Weis et al 2006 NobreSilva et al in preparation) Whole-rock SmNd ratiosand age-correction of Nd isotopes may be affected bythe leaching procedure for submarine rocks older than50 Ma (Thompson et al 2008 Fig 9) there ishowever very little variation in Nd isotopic compositionsof the picrites or tholeiitic basalts and this system does notappear to be disturbed by alteration The HFSE abun-dances for tholeiitic and picritic basalts exhibit clear
linear correlations in binary diagrams (Fig 8) whereasplots of LILE vs HFSE are highly scattered (not shown)because of the mobility of some of the alkali and alkalineearth LILE (Cs Rb Ba K) and Sr for most samplesduring alterationThe degree of alteration in the Karmutsen samples does
not appear to be related to eruption environment or depthof burial Pillow rims are compositionally different frompillow cores (Surdam1967 Kuniyoshi1972) and aquagenetuffs are chemically different from isolated pillows withinpillow breccias however submarine and subaerial basaltsexhibit a similar degree of alterationThere is no clear cor-relation between the submarine and subaerial basalts andsome commonly used chemical alteration indices (eg BaRb vs K2OP2O5 Huang amp Frey 2005) There is also nodefinitive correlation between the petrographic alterationindex (Table 1) and chemical alteration indices although10 of 13 tholeiitic basalts with the highest K and LILEabundances have the highest petrographic alterationindex of three (seeTable 1)
Table 4 Hf isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Lu Hf 176Hf177Hf 2m eHf176Lu177Hf 176Hf177Hft eHf(t)
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
OLIV INE ACCUMULAT ION INH IGH-MgO AND PICR IT IC LAVASGeochemical trends and petrographic characteristics indi-cate that accumulation of olivine played an important rolein the formation of the high-MgO basalts and picrites fromthe Keogh Lake area on northern Vancouver Island TheKeogh Lake samples show a strong linear correlation inplots of Al2O3 TiO2 ScYb and Ni vs MgO and many ofthe picrites have abundant clusters of olivine pseudomorphs(Table1 Fig11)There is a strong linear correlationbetween
modal per cent olivine and whole-rock magnesium contentmagnesium contents range between 10 and 19wt MgOand the proportion of olivine phenocrysts in these samplesvaries from 0 to 42 vol (Fig 11)With a few exceptionsboth the size range and median size of the olivine pheno-crysts in most samples are comparable The clear correla-tion between the proportion of olivine phenocrysts andwhole-rock magnesium contents indicates that accumula-tion of olivine was directly responsible for the high MgOcontents (410wt ) in most of the picritic lavas
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
207Pb204Pb
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
208Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
(a) (b)
(c) (e)
4723A24723A2
4723A2
4723A2
4722A4
4723A4
4722A4
4723A4
4723A134723A3
4723A3
4723A13
1556
1558
1560
1562
1564
1566
1568
186 190 194 198 202 206 2101550
1552
1554
1556
1558
1560
178 180 182 184 186 188 190 192 194
380
384
388
392
396
186 190 194 198 202 206 210374
378
382
386
390
178 180 182 184 186 188 190 192 194
206Pb204Pb(d)
LOI (wt )
0
1
2
3
4
5
6
7
186 190 194 198 202 206 210
4723A2
Fig 10 Pb isotopic compositions of leached whole-rock samples measured by MC-ICP-MS for the Karmutsen Formation Error bars are smal-ler than symbols (a) Measured 207Pb204Pb vs 206Pb204Pb (b) Initial 207Pb204Pb vs 206Pb204Pb Age-correction to 230 Ma (c) Measured208Pb204Pb vs 206Pb204Pb (d) LOI vs 206Pb204Pb (e) Initial 208Pb204Pb vs 206Pb204Pb Complete chemical duplicates shown in Table 5(samples 4720A7 and 4722A4) are circled in each plot The dashed lines in (b) and (e) show the differences in age-corrections for two picrites(4723A4 4722A4) using the measured UTh and Pb concentrations for each sample and the age-corrections when concentrations are used fromthe two picrites (4723A3 4723A13) that appear to be least affected by alteration
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
23
DISCUSS IONThe compositional and spatial development of the eruptedlava sequences of the Wrangellia oceanic plateau onVancouver Island provide information about the evolutionof a mantle plume upwelling beneath oceanic lithospherein an analogous way to the interpretation of continentalflood basalts The Caribbean and Ontong Java plateausare the two primary examples of accreted parts of oceanicplateaus that have been studied to date and comparisonscan be made with the Wrangellia oceanic plateauHowever the Wrangellia oceanic plateau is a distinctiveoccurrence of an oceanic plateau because it formed on thecrust of a Devonian to Mississippian intra-oceanic islandarc overlain by Mississippian to Permian limestones andpelagic sediments The nature and thickness of oceaniclithospheric mantle are considered to play a fundamentalrole in the decompression and evolution of a mantleplume and thus the compositions of the erupted lavasequences (Greene et al 2008) The geochemical and
stratigraphic relationships observed in the KarmutsenFormation onVancouver Island and the focus of the subse-quent discussion sections provide constraints on (1) thetemperature and degree of melting required to generatethe primary picritic magmas (2) the composition of themantle source (3) the depth of melting and residual min-eralogy in the source region (4) the low-pressure evolutionof the Karmutsen tholeiitic basalts (5) the growth of theWrangellia oceanic plateau
Melting conditions and major-elementcomposition of the primary magmasThe primary magmas to most flood basalt provinces arebelieved to be picritic (eg Cox 1980) and they have beenfound in many continental and oceanic flood basaltsequences worldwide (eg Siberia Karoo Paranacurren ^Etendeka Caribbean Deccan Saunders 2005) Near-pri-mary picritic lavas with low total alkali abundances
Table 5 Pb isotopic compositions of Karmutsen basaltsVancouver Island BC
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
require high-degree partial melting of the mantle source(eg Herzberg amp OrsquoHara 2002) The Keogh Lake picritesfrom northernVancouver Island are the best candidates foridentifying the least modified partial melts of the mantleplume source despite having accumulated olivine pheno-crysts and can be used to estimate conditions of meltingand the composition of the primary magmas for theKarmutsen FormationPrimary magma compositions mantle potential tem-
peratures and source melt fractions for the picritic lavas ofthe Karmutsen Formation were calculated from primitivewhole-rock compositions using PRIMELT1XLS software(Herzberg et al 2007) and are presented in Fig 12 andTable 6 A detailed discussion of the computationalmethod has been given by Herzberg amp OrsquoHara (2002)and Herzberg et al (2007) key features of the approachare outlined belowPrimary magma composition is calculated for a primi-
tive lava by incremental addition of olivine and compari-son with primary melts of mantle peridotite Lavas thathad experienced plagioclase andor clinopyroxene fractio-nation are excluded from this analysis All calculated pri-mary magma compositions are assumed to be derived byaccumulated fractional melting of fertile peridotite KR-4003 Uncertainties in fertile peridotite composition donot propagate to significant variations in melt fractionand mantle potential temperature melting of depletedperidotite propagates to calculated melt fractions that aretoo high but with a negligible error in mantle potential
temperature PRIMELT1XLS calculates the olivine liqui-dus temperature at 1atm which should be similar to theactual eruption temperature of the primary magmaand is accurate to 308C at the 2 level of confidenceMantle potential temperature is model dependent witha precision that is similar to the accuracy of the olivineliquidus temperature (C Herzberg personal communica-tion 2008)With regard to the evidence for accumulated olivine in
the Keogh Lake picrites it should not matter whether themodeled lava composition is aphyric or laden with accu-mulated olivine phenocrysts PRIMELT1 computes theprimary magma composition by adding olivine to a lavathat has experienced olivine fractionation in this way itinverts the effects of fractional crystallization The onlyrequirement is that the olivine phenocrysts are not acci-dental or exotic to the lava compositions Support for thisprocedure can be seen in the calculated olivine liquid-line-of-descent shown by the curved arrays with 5 olivineaddition increments (Fig 12c) Fractional crystallization ofolivine from model primary magmas having 85^95wt FeO and 153^175wt MgO will yield arrays of deriva-tive liquids and randomly sorted olivines that in most butnot all cases connect the picrites and high-MgO basalts Anewer version of the PRIMELT software (Herzberg ampAsimow 2008) can perform olivine addition to and sub-traction from lavas with highly variably olivine phenocrystcontents and yields results that converge with those shownin Fig 12
Fig 11 Relationship between abundance of olivine phenocrysts and whole-rock MgO contents for the Keogh Lake picrites (a) Tracings ofolivine grains (gray) in six high-MgO samples made from high resolution (2400 dpi) scans of petrographic thin-sections Scale bars represent10mm sample numbers are indicated at the upper left (b) Modal olivine vs whole-rock MgO Continuous line is the best-fit line for 10 high-MgO samplesThe areal proportion of olivine was calculated from raster images of olivine grains using ImageJ image analysis software whichprovides an acceptable estimation of the modal abundance (eg Chayes 1954)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
25
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
Gray lines = final melting pressure
L + Ol + Opx
07
08
09
Pressure (GPa)
10
3
4
5
6
1
SOLIDUS
01
04
0506
L + Ol
2
03
02
PeridotiteKR-4003
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
Thick black lines = initial melting pressure
Dashed lines = melt fraction
Melt Fraction
2
3 4 5 6
L+Ol+Opx
SolidusSpinel
SolidusGarnet Peridotite
7
L+Ol
0 10 20 30 4040
45
50
55
60
MgO (wt)
SiO2
(wt)
PeridotiteKR-4003
Melt Fraction
0 10 20 305
10
15
CaO(wt)
MgO (wt)
3
4 5 6 7
2
46
2
Peridotite Partial Melts
Thick black line = initial melting pressureGray lines = final melting pressure
Peridotite
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
02
Pressure (GPa)
0302
04
Pressure (GPa)
Picrite
High-MgO basalt
Tholeiitic basalt
(c)
(d)
Karmutsen flood basaltsOlivine addition modelOlivine addition (5 increment)Primary magma
PicriteHigh-MgO basaltTholeiitic basalt
(a)
(b)
800 850 900
910
920
930
940
950
Karmutsen flood basalts
Olivine addition modelOlivine addition (5 increment)Primary magma
Gray lines = Fo contents of olivine
Fig 12 Estimated primary magma compositions for three Keogh Lake picrites and high-MgO basalts (samples 4723A4 4723A135616A7) using the forward and inverse modeling technique of Herzberg et al (2007) Karmutsen compositions and modeling results areoverlain on diagrams provided by C Herzberg (a) Whole-rock FeO vs MgO for Karmutsen samples from this study Total iron estimatedto be FeO is 090 Gray lines show olivine compositions that would precipitate from liquid of a given MgO^FeO composition Black lineswith crosses show results of olivine addition (inverse model) using PRIMELT1 (Herzberg et al 2007) Gray field highlights olivinecompositions with Fo contents of 91^92 predicted to precipitate (b) FeO vs MgO showing Karmutsen lava compositions and results ofa forward model for accumulated fractional melting of fertile peridotite KR-4003 (Kettle River Peridotite Walter 1998) (c) SiO2 vsMgO for Karmtusen lavas and model results (d) CaO vs MgO showing Karmtusen lava compositions and model results To brieflysummarize the technique [see Herzberg et al (2007) for complete description] potential parental magma compositions for the high-MgOlava series were selected (highest MgO and appropriate CaO) and using PRIMELT1 software olivine was incrementally added tothe selected compositions to show an array of potential primary magma compositions (inverse model) Then using PRIMELT1 theresults from the inverse model were compared with a range of accumulated fractional melts for fertile peridotite derived fromparameterization of the experimental results for KR-4003 from Walter (1998) forward model Herzberg amp OrsquoHara (2002) A melt fractionwas sought that was unique to both the inverse and forward models (Herzberg et al 2007) A unique solution was found when there was a com-mon melt fraction for both models in FeO^MgO and CaO^MgO^Al2O3^SiO2 (CMAS) projection space This modeling assumes thatolivine was the only phase crystallizing and ignores chromite precipitation and possible augite fractionation in the mantle (Herzbergamp OrsquoHara 2002) Results are best for a residue of spinel lherzolite (not pyroxenite) The presence of accumulated olivine in samples ofthe Keogh Lake picrites used as starting compositions does not significantly affect the results because we are modeling addition of olivine(see text for further description) The tholeiitic basalts cannot be used for modeling because they are all saturated in plagthorn cpxthornolThick black line for initial melting pressure at 4 GPa is not shown in (c) for clarity Gray area in (a) indicates parental magmas thatwill crystallize olivine with Fo 91^92 at the surface Dashed lines in (c) indicate melt fraction Solidi for spinel and garnet peridotite arelabelled in (c)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
The PRIMELT1 results indicate that if the Karmutsenpicritic lavas were derived from accumulated fractionalmelting of fertile peridotite the primary magmas wouldhave contained 15^17wt MgO and 10wt CaO(Fig 12 Table 6) formed from 23^27 partial meltingand would have crystallized olivine with a Fo content of 91 (Fig 12 Table 6) Calculated mantle temperatures forthe Karmutsen picrites (14908C) indicate melting ofanomalously hot mantle (100^2008C above ambient) com-pared with ambient mantle that produces mid-ocean ridgebasalt (MORB 1280^14008C Herzberg et al 2007)which initiated between 3 and 4 GPa Several picritesappear to be near-primary melts and required very littleaddition of olivine (eg sample 93G171 with measured158wt MgO) These temperature and melting esti-mates of the Karmutsen picrites are consistent withdecompression of hot mantle peridotite in an actively con-vecting plume head (ie plume initiation model) Threesamples (picrite 5614A1 and high-MgO basalts 5614A3and 5614A5) are lower in iron than the other Karmutsenlavas with FeO in the 65^80wt range (Fig 12c) andhave MgO and FeO contents that are similar to primitiveMORB from the Sequeiros fracture zone of the EastPacific Rise (EPR Perfit et al 1996) These samples
however differ from EPR MORB in having higher SiO2
and lower Al2O3 and they are restricted geographicallyto one area (Sara Lake Fig 2)We therefore have evidence for primary magmas with
MgO contents that range from a possible low of 110wt to as high as 174wt with inferred mantle potential tem-peratures from 1340 to 15208C This is similar to the rangedisplayed by lavas from the Ontong Java Plateau deter-mined by Herzberg (2004) who found that primarymagma compositions assuming a peridotite sourcewould contain 17wt MgO and 10wt CaO(Table 6) and that these magmas would have formed by27 melting at a mantle potential temperature of15008C (first melting occurring at 36 GPa) Fitton ampGodard (2004) estimated 30 partial melting for theOntong Java lavas based on the Zr contents of primarymagmas of Kroenke-type basalt calculated by incrementaladdition of equilibrium olivine However if a considerableamount of eclogite was involved in the formation of theOntong Java plateau magmas the excess temperatures esti-mated by Herzberg et al (2007) may not be required(Korenaga 2005) The variability in mantle potential tem-perature for the Keogh Lake picrites is also similar to thewide range of temperatures that have been calculated for
Table 6 Estimated primary magma compositions for Karmutsen basalts and other oceanic plateaus or islands
Sample 4723A4 4723A13 5616A7 93G171 Average OJP Mauna Keay Gorgonaz
Ontong Java primary magma composition for accumulated fraction melting (AFM) from Herzberg (2004)yMauna Kea primary magma composition is average of four samples of Herzberg (2006 table 1)zGorgona primary magma composition for AFM for 1F fertile source from Herzberg amp OrsquoHara (2002 table 4)Total iron estimated to be FeO is 090
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
27
lavas from single volcanoes in the Galapagos Hawaii andelsewhere by Herzberg amp Asimow (2008) Those workersinterpreted the range to reflect the transport of magmasfrom both the cool periphery and the hotter axis of aplume A thermally heterogeneous plume will yield pri-mary magmas with different MgO contents and the rangeof primary magma compositions and their inferred mantlepotential temperatures for Karmutsen lavas may reflectmelting from different parts of the plume
Source of the Karmutsen Formation lavasThe Sr^Nd^Hf^Pb isotopic compositions of theKarmutsen volcanic rocks on Vancouver Island indicatethat they formed from plume-type mantle containing adepleted component that is compositionally similar to butdistinct from other oceanic plateaus that formed in thePacific Ocean (eg Ontong Java and Caribbean Fig 13)Trace element and isotopic constraints can be used to testwhether the depleted component of the KarmutsenFormation was intrinsic to the mantle plume and distinctfrom the source of MORB or the result of mixing or invol-vement of ambient depleted MORB mantle with plume-type mantle The Karmutsen tholeiitic basalts have initialeNd and
87Sr86Sr ratios that fall within the range of basaltsfrom the Caribbean Plateau (eg Kerr et al 1996 1997Hauff et al 2000 Kerr 2003) and a range of Hawaiian iso-tope data (eg Frey et al 2005) and lie within and abovethe range for the Ontong Java Plateau (Tejada et al 2004Fig 13a) Karmutsen tholeiitic basalts with the highest ini-tial eNd and lowest initial 87Sr86Sr lie within and justbelow the field for age-corrected northern EPR MORB at230 Ma (eg Niu et al1999 Regelous et al1999) Initial Hfand Nd isotopic compositions for the Karmutsen samplesare noticeably offset to the low side of the ocean islandbasalt (OIB) array (Vervoort et al 1999) and lie belowand slightly overlap the low eHf end of fields for Hawaiiand the Caribbean Plateau (Fig 13c) the Karmutsen sam-ples mostly fall between and slightly overlap a field forage-corrected Pacific MORB at 230 Ma and Ontong Java(Mahoney et al 1992 1994 Nowell et al 1998 Chauvel ampBlichert-Toft 2001) Karmutsen lava Pb isotope ratios over-lap the broad range for the Caribbean Plateau and thelinear trends for the Karmutsen samples intersect thefields for EPR MORB Hawaii and Ontong Java in208Pb^206Pb space but do not intersect these fields in207Pb^206Pb space (Fig 13d and e) The more radiogenic207Pb204Pb for a given 206Pb204Pb of the Karmutsenbasalts requires high UPb ratios at an ancient time when235U was abundant similar to the Caribbean Plateau andenriched in comparison with EPR MORB Hawaii andOntong JavaThe low eHf^low eNd component of the Karmutsen
basalts that places them below the OIB mantle array andpartly within the field for age-corrected Pacific MORBcan form from low ratios of LuHf and high SmNd over
time (eg Salters amp White 1998) MORB will develop aHf^Nd isotopic signature over time that falls below themantle array Low LuHf can also lead to low 176Hf177Hffrom remelting of residues produced by melting in theabsence of garnet (eg Salters amp Hart 1991 Salters ampWhite 1998) The Hf^Nd isotopic compositions of theKarmutsen Formation indicate the possible involvementof depleted MORB mantle however similar to Iceland(Fitton et al 2003) Nb^Zr^Y variations provide a way todistinguish between a depleted component within theplume and a MORB-like component The Karmutsenbasalts and picrites fall within the Iceland array and donot overlap the MORB field in a logarithmic plot of NbYvs ZrY (Fig 13b) using the fields established by Fittonet al (1997) Similar to the depleted component within theCaribbean Plateau (Thompson et al 2003) the trend ofHf^Nd isotopic compositions towards Pacific MORB-likecompositions is not followed by MORB values in Nb^Zr^Y systematics and this suggests that the depleted compo-nent in the source of the Karmutsen basalts is distinctfrom the source of MORB and probably an intrinsic partof the plume The relatively radiogenic Pb isotopic ratiosand REE patterns distinct from MORB also support thisinterpretationTrace-element characteristics suggest the possible invol-
vement of sub-arc lithospheric mantle in the generation ofthe Karmutsen picrites Lassiter et al (1995) suggested thatmixing of a plume-type source with eNdthorn 6 to thorn 7 witharc material with low NbTh could reproduce the varia-tions observed in the Karmutsen basalts however theyconcluded that the absence of low NbLa ratios in theirsample of tholeiitic basalts restricts the amount of arc litho-sphere involved The study of Lassiter et al (1995) wasbased on major and trace element and Sr Nd and Pb iso-topic data for a suite of 29 samples from Buttle Lake in theStrathcona Provincial Park on central Vancouver Island(Fig 1) Whereas there are no clear HFSE depletions inthe Karmutsen tholeiitic basalts the low NbLa (and TaLa) of the Karmutsen picritic lavas in this study indicatethe possible involvement of Paleozoic arc lithosphere onVancouver Island (Fig 8)
REE modeling dynamic melting andsource mineralogyThe combined isotopic and trace-element variations of thepicritic and tholeiitic Karmutsen lavas indicate that differ-ences in trace elements may have originated from differentmelting histories For example the LuHf ratios of theKarmutsen picritic lavas (mean 008) are considerablyhigher than those in the tholeiitic basalts (mean 002)however the small range of overlapping initial eHf valuesfor the two lava suites suggests that these differences devel-oped during melting within the plume (Fig 9b) Below wemodel the effects of source mineralogy and depth of
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
28
melting on differences in trace-element concentrations inthe tholeiitic basalts and picritesDynamic melting models simulate progressive decom-
pression melting where some of the melt fraction isretained in the residue and only when the degree of partialmelting is greater than the critical mass porosity and the
source becomes permeable is excess melt extracted fromthe residue (Zou1998) For the Karmutsen lavas evolutionof trace element concentrations was simulated using theincongruent dynamic melting model developed by Zou ampReid (2001) Melting of the mantle at pressures greater than1 GPa involves incongruent melting (eg Longhi 2002)
OIB array
PacificMORB(230 Ma)
(0 Ma)
EPR(230 Ma)
-4
-2
0
2
4
8
10
12
14
16
18
20
22
-4 -2 0 2 4 6 8 10 12 14
minus2
0
2
4
6
8
10
12
14
0702 0703 0704 0705 0706
IndianMORB
87Sr 86Sr (initial)
Hf (initial)
(a) (c)
Nd (initial)
Karmutsen tholeiitic basalt
HawaiiOntong JavaCaribbean Plateau
Karmutsen high-MgO lava
high-MgO lavasKarmutsen
1540
1545
1550
1555
1560
1565
175 180 185 190 195 200375
380
385
390
395
175 180 185 190 195 200
(d) (e)
EPR
EPR
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb
Karmutsen Formation (initial)
HawaiiOntong JavaCaribbean Plateau
East Pacific Rise
Karmutsen Formation (measured)
Juan de FucaGorda
ε Nd
(initial)
Karmutsen Formation (different age-correction for 3 picrites)
(0 Ma)
NbY
001
01
1
1 10
ZrY
OIB
NMORB
(b)
Iceland array
6
Fig 13 Comparison of age-corrected (230 Ma) Sr^Nd^Hf^Pb isotopic compositions for Karmutsen flood basalts onVancouver Island withage-corrected OIB and MORB and Nb^Zr^Y variation (a) Initial eNd vs 87Sr86Sr (b) NbYand ZrY variation (after Fitton et al 1997)(c) Initial eHf vs eNd (d) Measured and initial 207Pb204Pb vs 206Pb204Pb (e) Measured and initial 208Pb204Pb vs 206Pb204Pb References fordata sources are listed in Electronic Appendix 4 Most of the compiled data were extracted from the GEOROC database (httpgeorocmpch-mainzgwdgdegeoroc) OIB array line in (c) is fromVervoort et al (1999) EPR East Pacific Rise Several high-MgO samples affected bysecondary alteration were not plotted for clarity in Pb isotope plots Estimation of fields for age-corrected Pacific MORB and EPR at 230 Mawere made using parent^daughter concentrations (in ppm) from Salters amp Stracke (2004) of Lufrac14 0063 Hffrac14 0202 Smfrac14 027 Ndfrac14 071Rbfrac14 0086 and Srfrac14 836 In the NbYvs ZrYplot the normal (N)-MORB and OIB fields not including Icelandic OIB are from data com-pilations of Fitton et al (1997 2003) The OIB field is data from 780 basalts (MgO45wt ) from major ocean islands given by Fitton et al(2003) The N-MORB field is based on data from the Reykjanes Ridge EPR and Southwest Indian Ridge from Fitton et al (1997) The twoparallel gray lines in (b) (Iceland array) are the limits of data for Icelandic rocks
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
29
with olivine being produced during the reactioncpxthornopxthorn spfrac14meltthornol for spinel peridotites The mod-eling used coefficients for incongruent melting reactionsbased on experiments on lherzolite melting (eg Kinzleramp Grove 1992 Longhi 2002) and was separated into (1)melting of spinel lherzolite (2) two combined intervals ofmelting for a garnet lherzolite (gt lherz) and spinel lher-zolite (sp lherz) source (3) melting of garnet lherzoliteThe model solutions are non-unique and a range indegree of melting partition coefficients melt reactioncoefficients and proportion of mixed source componentsis possible to achieve acceptable solutions (see Fig 14 cap-tion for description of modeling parameters anduncertainties)The LREE-depleted high-MgO lavas require melting of
a depleted spinel lherzolite source (Fig 14a) The modelingresults indicate a high degree of melting (22^25) similarto the results from PRIMELT1 based on major-elementvariations (discussed above) of a LREE-depleted sourceThe high-MgO lavas lack a residual garnet signature Theenriched tholeiitic basalts involved melting of both garnetand spinel lherzolite and represent aggregate melts pro-duced from continuous melting throughout the depth ofthe melting column (Fig 14b) Trace-element concentra-tions in the tholeiitic and picritic lavas are not consistentwith low-degree melting of garnet lherzolite alone(Fig 14c) The modeling results indicate that the aggregatemelts that formed the tholeiitic basalts would haveinvolved lesser proportions of enriched small-degreemelts (1^6 melting) generated at depth and greateramounts of depleted high-degree melts (12^25) gener-ated at lower pressure (a ratio of garnet lherzolite tospinel lherzolite melt of 14 provides the best fits seeFig 14b and description in figure caption) The totaldegree of melting for the tholeiitic basalts was high(23^27) similar to the degree of partial melting in thespinel lherzolite facies for the primary melts that eruptedas the Keogh Lake picrites of a source that was also prob-ably depleted in LREE
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
(c)
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
Melting of garnet lherzolite
High-MgO lavas
Melting of spinel lherzolite
source (07DM+03PM)
source (07DM+03PM)
6 melting
30 melting
(b)
(a)
30 melting
Sam
ple
Cho
ndrit
eS
ampl
eC
hond
rite 20 melting
1 melting
Melting of garnet lherzoliteand spinel lherzolite
x gt lherz + 4x sp lherz
5 meltingx gt lherz + 4x sp lherz
Sam
ple
Cho
ndrit
e
Tholeiitic lavas
Tholeiitic lavas
High-MgO lavas
source (07DM+03PM)
Fig 14 Trace-element modeling results for incongruent dynamicmantle melting for picritic and tholeiitic lavas from the KarmutsenFormation Three steps of modeling are shown (a) melting of spinellherzolite compared with high-MgO lavas (b) melting of garnet lher-zolite and spinel lherzolite compared with tholeiitic lavas (c) meltingof garnet lherzolite compared with tholeiitic and picritic lavasShaded fields in each panel for tholeiitic basalts and picrites arelabeled Patterns with symbols in each panel are modeling results(patterns are 5 melting increments in (a) and (b) and 1 incre-ments in (c) PM primitive mantle DM depleted MORB mantle gtlherz garnet lherzolite sp lherz spinel lherzolite In (b) the ratios ofper cent melting for garnet and spinel lherzolite are indicated (x gtlherzthorn 4x sp lherz means that for 5 melting 1 melt in garnetfacies and 4 melt in spinel facies where x is per cent melting in theinterval of modeling) The ratio of the respective melt proportions inthe aggregate melt (x gt lherzthorn 4x sp lherz) was kept constant andthis ratio of melt contribution provided the best fit to the data Arange of proportion of source components (07DMthorn 03PM to
09DMthorn 01PM) can reproduce the variation in the high-MgOlavas Melting modeling uses the formulation of Zou amp Reid (2001)an example calculation is shown in their Appendix PM fromMcDonough amp Sun (1995) and DM from Salters amp Stracke (2004)Melt reaction coefficients of spinel lherzolite from Kinzler amp Grove(1992) and garnet lherzolite fromWalter (1998) Partition coefficientsfrom Salters amp Stracke (2004) and Shaw (2000) were kept constantSource mineralogy for spinel lherzolite (018cpx027opx052ol003sp) from Kinzler (1997) and for garnet lherzolite(034cpx008opx053ol005gt) from Salters amp Stracke (2004) Arange of source compositions for melting of spinel lherzolite(07DMthorn 03PM to 03DMthorn 07PM) reproduce REE patternssimilar to those of the high-MgO lavas with best fits for 22ndash25melting and 07DMthorn 03PM source composition Concentrationsof tholeiitic basalts cannot be reproduced from an entirelyprimitive or depleted source
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
30
The uncertainty in the composition of the source in thespinel lherzolite melting model for the high-MgO lavasprecludes determining whether the source was moredepleted in incompatible trace elements than the source ofthe tholeiitic lavas (see Fig 14 caption for description) It ispossible that early high-pressure low-degree melting left aresidue within parts of the plume (possibly the hotter inte-rior portion) that was depleted in trace elements (egElliott et al 1991) further decompression and high-degreemelting of such depleted regions could then have generatedthe depleted high-MgO Karmutsen lavas Shallow high-degree melting would preferentially sample depletedregions whereas deeper low-degree melting may notsample more refractory depleted regions (eg CaribbeanEscuder-Viruete et al 2007) The picrites lie at the highend of the range of Nd isotopic compositions and havelow NbZr similar to depleted Icelandic basalts (Fittonet al 2003) therefore the picrites may represent the partsof the mantle plume that were the most depleted prior tothe initiation of melting As Fitton et al (2003) suggestedthese depleted melts may only rarely be erupted withoutcombining with more voluminous less depleted melts andthey may be detected only as undifferentiated small-volume flows like the Keogh Lake picrites within theKarmutsen Formation on Vancouver IslandDecompression melting within the mantle plume initiatedwithin the garnet stability field (35^25 GPa) at highmantle potential temperatures (Tp414508C) and pro-ceeded beneath oceanic arc lithosphere within the spinelstability field (25^09 GPa) where more extensivedegrees of melting could occur
Magmatic evolution of Karmutsentholeiitic basaltsPrimary magmas in large igneous provinces leave exten-sive coarse-grained cumulate residues within or below thecrust these mafic and ultramafic plutonic sequences repre-sent a significant proportion of the original magma thatpartially crystallized at depth (eg Farnetani et al 1996)For example seismic and petrological studies of OntongJava (eg Farnetani et al 1996) combined with the use ofMELTS (Ghiorso amp Sack 1995) indicate that the volcanicsequence may represent only 40 of the total magmareaching the Moho Neal et al (1997) estimated that30^45 fractional crystallization took place for theOntong Java magmas and produced cumulate thicknessesof 7^14 km corresponding to 11^22 km of flood basaltsdepending on the presence of pre-existing oceanic crustand the total crustal thickness (25^35 km) Interpretationsof seismic velocity measurements suggest the presence ofpyroxene and gabbroic cumulates 9^16 km thick beneaththe Ontong Java Plateau (eg Hussong et al 1979)Wrangellia is characterized by high crustal velocities in
seismic refraction lines which is indicative of mafic
plutonic rocks extending to depth beneath VancouverIsland (Clowes et al 1995)Wrangellia crust is 25^30 kmthick beneath Vancouver Island and is underlain by astrongly reflective zone of high velocity and density thathas been interpreted as a major shear zone where lowerWrangellia lithosphere was detached (Clowes et al 1995)The vast majority of the Karmutsen flows are evolved tho-leiitic lavas (eg low MgO high FeOMgO) indicating animportant role for partial crystallization of melts at lowpressure and a significant portion of the crust beneathVancouver Island has seismic properties that are consistentwith crystalline residues from partially crystallizedKarmutsen magmas To test the proportion of the primarymagma that fractionated within the crust MELTS(Ghiorso amp Sack 1995) was used to simulate fractionalcrystallization using several estimated primary magmacompositions from the PRIMELT1 modeling results Themajor-element composition of the primary magmas couldnot be estimated for the tholeiitic basalts because of exten-sive plagthorn cpxthornol fractionation and thus the MELTSmodeling of the picrites is used as a proxy for the evolutionof major elements in the volcanic sequence A pressure of 1kbar was used with variable water contents (anhydrousand 02wt H2O) calculated at the quartz^fayalite^magnetite (QFM) oxygen buffer Experimental resultsfrom Ontong Java indicate that most crystallizationoccurs in magma chambers shallower than 6 km deep(52 kbar Sano amp Yamashita 2004) however some crys-tallization may take place at greater depths (3^5 kbarFarnetani et al 1996)A selection of the MELTS results are shown in Fig 15
Olivine and spinel crystallize at high temperature until15^20wt of the liquid mass has fractionated and theresidual magma contains 9^10wt MgO (Fig 15f) At1235^12258C plagioclase begins crystallizing and between1235 and 11908C olivine ceases crystallizing and clinopyr-oxene saturates (Fig 15f) As expected the addition of clin-opyroxene and plagioclase to the crystallization sequencecauses substantial changes in the composition of the resid-ual liquid FeO and TiO2 increase and Al2O3 and CaOdecrease while the decrease in MgO lessens considerably(Fig 15) The near-vertical trends for the Karmutsen tho-leiitic basalts especially for CaO do not match the calcu-lated liquid-line-of-descent very well which misses manyof the high-CaO high-Al2O3 low-FeO basalts A positivecorrelation between Al2O3CaO and SrNd indicates thataccumulation of plagioclase played a major role in the evo-lution of the tholeiitic basalts (Fig 15g) Increasing thecrystallization pressure slightly and the water content ofthe melts does not systematically improve the match ofthe MELTS models to the observed dataThe MELTS results indicate that a significant propor-
tion of the crystallization takes place before plagioclase(15^20 crystallization) and clinopyroxene (35^45
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
31
crystallization) join the crystallization sequence Theseresults suggest that445 crystallization of the primarymagmas occurred (estimated from the residual liquid percent in Fig 15f) and if the original flood basalt stratigra-phy on Vancouver Island was 6 km thick at least anequivalent amount of complementary plutonic rocksshould have crystallized The Karmutsen basalts and theirplutonic residues thus represent a significant addition ofcrust (perhaps 412 km thickness) previously thickenedduring the formation of the underlying Paleozoic arc
Growth of the Wrangellia oceanic plateauon Vancouver IslandThe extrusive part of the Wrangellia oceanic plateau onVancouver Island was constructed in a three-layersequence of mostly tholeiitic basalts with a restrictedrange of composition (Fig 16) The plateau overliesDevonian to Mississippian arc volcanic sequences andmarine sediments of Mississippian to Permian ages thatare intruded by gabbroic rocks related to the flood basaltsThe volcanic stratigraphy of theWrangellia plateau formed
00
05
10
15
20
25
30
35
40
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
16
0 2 4 6 8 10 12 14 16 18 20
10
11
12
13
14
15
16
17
18
19
20
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
0 2 4 6 8 10 12 14 16 18 20
MgO (wt)
0 10 20 30 40 50
percent of total mass
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
) Mineral proportions
Sp (e)
Ol
CpxPlag
PicriteHigh-MgO basaltTholeiitic basalt
TiO2 (wt)
MgO (wt)
Al2O3 (wt) CaO (wt)
FeO (wt)
MgO(a) (b)
(c) (d)
MgO (wt)MgO (wt)
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
)
Residual liquid ()
1220
1190
Ol + Sp
Plag+Ol+Sp Plag+Cpx+Sp
02 wt H2O
no H2O
Residual liquid and crystallizing phases
(f)
0102030405060708090100
plagioclaseaccumulation
plagioclaseaccumulation
plagioclase
accumulatio
n
10
12
14
16
0 10 20 30 40
Al2O3CaO
SrNd
tholeiitic basalts
(g)
Fig 15 Forward fractional crystallization modeling results for major elements from MELTS (Ghiorso amp Sack 1995) compared with picriticand tholeiitic lavas from the Karmutsen Formation (a^d) Major-element compositions of residual liquids compared with Karmutsen lavacompositions For clarity the results are shown only for one starting composition (open circle) using the composition of the estimated primarymagma for the high-MgO lava series from the modeling technique of Herzberg et al (2007) for sample 4723A13 as shown in Fig 12 The esti-mated primary magmas for samples 4723A3 and 93G171 yield similar results A pressure of 1 kbar was used with no H2O and 02wt H2Ocalculated at QFM black line is anhydrous and grey line is 02wt H2OThe end-point for the anhydrous result is at 78 fractionation and11308C and the end-point for the result with 02wt H2O is 75 fractionation and11108C (e) Mineral proportions of crystallized phases fromone MELTrun (sample 4723A13) with no H2O (f) Per cent residual liquid vs temperature labelled with intervals of crystallizing phases (g)Al2O3CaO vs SrNd for the tholeiitic basalts
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
32
in a single eruptive phase (c 231^225 Ma) as an emergentbasalt sequence during a deeper-water shallow-water andsubaerial stage and subsided below sea level relativelysoon after its formation (Fig 16) The deeper-water stagecontains pillowed and massive submarine flows (3000mthick) that built up the submarine volcanic edifice from theseafloor The lowest pillow basalts were emplaced onunconsolidated and lithified fine-grained siliciclastic andcarbonaceous sediments containing Daonella beds(c 235^232 Ma) that are intruded by abundantKarmutsen mafic sills (Fig 16a) The different pillowedand massive submarine flows resulted primarily from dif-ferent effusive rates and local topography Some of themassive submarine flows commonly recognized by radiat-ing columnar jointing represent master tubes for the deliv-ery of lava to distal parts of submarine flow fields Thepillowed flows which commonly have uneven topography
in modern sequences may have served as levees to themassive submarine flows rather than allowing unob-structed spreading and inflation typical of the massive sub-aerial flows that form on shallow slopes in continentalflood basalt provinces (eg Self et al 1997) The rapid suc-cession of flows is evident by the near absence of sedimentbetween submarine flows Minor volumes of picritic pillowlavas occur in the upper part of the pillowed flow sequencein areas of the plateau exposed on northern VancouverIsland (Fig 16 Nixon et al 2008)The growth of the plateau in shallow water led to
emplacement of an increasing proportion of volcaniclastics(pillow breccias and hyaloclastites) that accumulated up to1500m thick in some areas (Fig 16b Nixon et al 2008)Horizons of pillowed flows occur locally within the volca-niclastic stratigraphy Sedimentary structures (eg gradedbedding fluidization structures) and petrographic textures
Late Devonian-Mississippian arc
Daonella beds
Middle Triassic
deeper water
Late Devonian-Mississippian arc
pillowed flows massive submarine
flows
pillow brecciahyaloclastiteshallow water
~4 km
~6 km
Late Triassic limestoneinterflow limestonelenses
Karmutsen Formation
mafic intrusions sediment-sill complex
Karmutsen Formation
(a) (b)
(c) (d)
subaerial flows
Late Devonian-Mississippian arc Late Devonian-Mississippian arc
picritic pillow basalts
Miss to Permian marine sediments
sills
Fig 16 Growth of the Wrangellia oceanic plateau onVancouver Island (a) Deep-water eruptive phase of pillowed and massive submarineflows with almost no intervening sediments onto MiddleTriassic shale containing Daonella beds (c 235^232 Ma) The basement of the plateauis composed of Devonian to Mississippian arc sequences overlain by Mississippian to Permian carbonates and siliciclastic sediments(b) Shallow-water flow emplacement with increasing proportion of volcaniclastics resulted primarily from cooling^contraction granulationPicritic pillow basalts occur primarily near the transition between pillowed submarine flows and volcaniclastic units Pillowed and massiveflows occur locally within the volcaniclastic unit (c) Emergent stage of growth consists of overlapping subaerial flow fields similar to continen-tal flood basalts with pahoehoe flow structures Multi-tiered columnar jointing occurrs locally in response to ponding of flows (eg Bondre et al2004) Interflow carbonate deposits and plagioclase megacrystic flows developed locally during the waning stage of volcanism (d) Post-volcanicsubsidence and deposition of Late Carnian to Norian (c 212^223 Ma) platform carbonates on top of the plateau Units are labeled in the dia-grams Thicknesses of the underlying units are not to scale
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
33
within the volcaniclastic rocks indicate formation primar-ily via cooling^contraction granulation magma^water^steam interaction autobrecciation resedimentation andmass-wasting rather than pyroclastic eruption (Nixonet al 2008)A broad subaerial platform was constructed by emplace-
ment of subaerial lava flows during an emergent subaerialstage (Fig 16c) The subaerial unit predominantly com-prises well-layered sheet flows that formed in a similarway to continental flood basalts as hundred(s) of pahoehoeflow fields erupted on shallow slopes (Self et al 1997) how-ever volcaniclastic and pillowed flow units are locally pre-sent mostly in the upper parts of the stratigraphyLocalized ponding of flows led to the formation of multi-tiered columnar jointing in some subaerial flows (egBondre et al 2004) As volcanism waned local interflowcarbonate deposits developed and plagioclase megacrysticflows were erupted Subsidence of the plateau is recordedby the interflow lenses and deposits of platform carbonates(40^1500m thick) on top of the Karmutsen basalts thatcontain Late Carnian to Early Norian fossils (Fig 16d c216^227 Ma Nixon et al 2008)The evolution of the internal architecture of the
Wrangellia oceanic plateau was primarily related to theconditions of emplacement (ie water depth and eruptionenvironment) There does not appear to have been a sys-tematic coincident evolution or shift of lava compositionsonVancouver Island as has been recognized in most con-tinental flood basalts and parts of the Wrangellia oceanicplateau in Alaska and Yukon (Greene et al 2008 2009)In general with the exception of the picritic pillow lavasin the submarine stratigraphy and local variations betweenhigh-MgO and tholeiitic lavas there are no significantcompositional differences between lavas erupted duringthe different stages of growth of the plateau (ie submarineand subaerial flows) nor are substantial compositional dif-ferences generally found between massive submarine flowsin contact with pillowed flows tholeiitic basalt flows nearthe base of the Karmutsen Formation are compositionallysimilar to flows sampled near the top of the volcanicsequence Where it was possible to sample continuous sec-tions of volcanic stratigraphy greater than several hun-dreds of meters (eg Keogh^Maynard Lake area west ofthe Karmutsen Range Fig 2) there is no clear relationshipbetween stratigraphic position and compositionalvariation
CONCLUSIONSThe Karmutsen Formation covers extensive areas ofVancouver Island (20 000 km2) and forms part of a largeoceanic plateau erupted over a geologically short interval(c 225^231 Ma) The tripartite volcanic stratigraphyexposed on Vancouver Island is at least 6 km thick andcomprises a succession of submarine flows volcaniclastic
deposits and massive subaerial flowsThese volcanologicaldifferences are primarily related to the eruption environ-ment (deep-water shallow-water or subaerial) The rapidgrowth of the plateau prevented significant accumulationsof intervening sediment except in the uppermost stratigra-phy where isolated limestone lenses commonly associatedwith pillow basalt and hyaloclastite deposits preserve arecord of the final stages of subsidence of the plateauThe Wrangellia plateau on Vancouver Island was con-
structed dominantly of tholeiitic basalt with restrictedmajor- and trace-element and isotopic compositional var-iations Minor volumes of picritic pillow basalts eruptedlate in the submarine phase The Karmutsen basalts wereemplaced onto older Paleozoic arc volcanic sequences andlow NbLa for the picrites may indicate a contributionfrom the sub-arc lithospheric mantle Modeling of the pet-rogenesis of the Karmutsen picrites indicates melting ofanomalously hot mantle (15008C) and extensive degreesof partial melting (23^27) and is consistent with aplume initiation model Combined Sr^Nd^Hf^Pb isotopicsystematics indicate that the picritic and tholeiitic lavaswere derived from a depleted plume-type mantle sourcedistinct from the source of MORB that has compositionalsimilarities to the source of the lavas forming theCaribbean Plateau Picrites were generated by melting ofspinel lherzolite whereas the tholeiitic basalts formed asaggregate melts of both garnet and spinel lherzolite Thetholeiitic basalts underwent extensive low-pressure fractio-nation (52^3 kbar) and seismic studies indicate that someof theWrangellia crust beneathVancouver Island may cor-respond to the plutonic residues of the Karmutsen floodbasalts
ACKNOWLEDGEMENTSWe are grateful to Nick Arndt for helping us get this pro-ject started and Nick Massey for insights into VancouverIsland geology We greatly appreciate the assistance withPRIMELT modeling from Claude Herzberg We thankMikkel Schau for his insight and enthusiasm during field-work Jane Barling assisted with analyses by MC-ICP-MSThoughtful and thorough reviews of the manuscript byJohn Mahoney Fred Frey Wendy Bohrson DougalJerram and Teal Riley are greatly appreciated Fundingwas generously provided by the Rocks to Riches Programadministered by the BC amp Yukon Chamber of Mines in2004 by the BC Geological Survey and by NSERCDiscovery Grants to JSS and DW ARG was supportedby a University Graduate Fellowship at UBC
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
34
REFERENCESBarker F Brown A S Budahn J R amp Plafker G (1989) Back-arc
with frontal-arc component origin of Triassic Karmutsen basaltBritish Columbia Canada Chemical Geology 75 81^102
Blichert-Toft JWeis D Maerschalk C Agranier A amp Albare de F(2003) Hawaiian hot spot dynamics as inferred from the Hf and Pbisotope evolution of Mauna Kea volcano Geochemistry GeophysicsGeosystems 4(2) 1^27 doi1010292002GC000340
Bondre N R Duraiswami R A amp Dole G (2004) Morphologyand emplacement of flows from the Deccan Volcanic ProvinceIndia Bulletin of Volcanology 66 29^45
Carlisle D (1963) Pillow breccias and their aquagene tuffs QuadraIsland British Columbia Journal of Geology 71 48^71
Carlisle D (1972) Late Paleozoic to Mid-Triassic sedimentary-volcanic sequence on Northeastern Vancouver Island Geological
Society of Canada Paper Report of Activities 72-1 Part B 24^30Carlisle D amp SuzukiT (1974) Emergent basalt and submergent car-
bonate^clastic sequences including the Upper Triassic Dilleri andWelleri zones on Vancouver Island Canadian Journal of Earth
Sciences 11 254^279Chauvel C amp Blichert-Toft J (2001) A hafnium isotope and trace
element perspective on melting of the depleted mantle Earth and
Planetary Science Letters 190 137^151Chayes F (1954)The theory of thin-section analysis Journal of Geology
62 92^101Cho M Liou J G amp Maruyama S (1987) Transition from the zeo-
lite to prehnite^pumpellyite facies in the Karmutsen metabasitesVancouver Island British Columbia Journal of Petrology 27(2)467^494
Clowes R M Zelt C A Amor J R amp Ellis R M (1995)Lithospheric structure in the southern Canadian Cordillera froma network of seismic refraction lines Canadian Journal of Earth
Sciences 32(10) 1485^1513Coffin M F amp Eldholm O (1994) Large igneous provinces Crustal
structure dimensions and external consequences Reviews of
Geophysics 32(1) 1^36Cox K G (1980) A model for flood basalt volcanism Journal of
Petrology 21 629^650Eldholm O amp Coffin M F (2000) Large igneous provinces
and plate tectonics In Richards M A Gordon R G amp vander Hilst R D (eds) The History and Dynamics of Global
Plate Motions American Geophysical Union Geophysical Monograph 121309^326
Elliott T R Hawkesworth C J amp Grolaquo nvold K (1991) Dynamicmelting of the Iceland plume Nature 351 201^206
Escuder-Viruete J Pecurren rez-Estaucurren n A Contreras F Joubert MWeis D Ullrich T D amp Spadea P (2007) Plume mantle sourceheterogeneity through time Insights from the Duarte ComplexHispaniola northeastern Caribbean Journal of Geophysical Research112(B04203) doi1010292006JB004323
Farnetani C G Richards M A amp Ghiorso M S (1996)Petrological models of magma evolution and deep crustal structurebeneath hotspots and flood basalt provinces Earth and Planetary
Science Letters 143 81^94Fitton J G amp Godard M (2004) Origin and evolution of magmas
on the Ontong Java Plateau In Fitton J G Mahoney J JWallace P J amp Saunders A D (eds) Origin and Evolution of the
Ontong Java Plateau Geological Society London Special Publications 229151^178
Fitton J G Saunders A D Norry M J Hardarson B S ampTaylor R N (1997) Thermal and chemical structure of theIceland plume Earth and Planetary Science Letters 153 197^208
Fitton J G Saunders A D Kempton P D amp Hardarson B S(2003) Does depleted mantle form an intrinsic part of the Icelandplume Geochemistry Geophysics Geosystems 4(3) 1032 doi1010292002GC000424
Frey F A Huang S Blichert-Toft J Regelous M amp Boyet M(2005) Origin of depleted components in basalt related to theHawaiian hotspot Evidence from isotopic and incompatible ele-ment ratios Geochemistry Geophysics Geosystems 6(1) 23 doi1010292004GC000757
Galer S J G amp Abouchami W (1998) Practical application of leadtriple spiking for correction of instrumental mass discriminationMineralogical Magazine 62A 491^492
Ghiorso M S amp Sack R O (1995) Chemical mass transfer in mag-matic processes IV A revised and internally consistent thermody-namic model for the interpolation and extrapolation of liquid^solidequilibria in magmatic systems at elevated temperatures and pres-sures Contributions to Mineralogy and Petrology 119 197^212
Greene A R Scoates J S amp Weis D (2005)WrangelliaTerrane onVancouver Island Distribution of flood basalts with implicationsfor potential Ni^Cu^PGE mineralization In Grant B (ed)Geological Fieldwork 2004 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2005-1 209^220Greene A R Scoates J S Nixon G T amp Weis D (2006) Picritic
lavas and basal sills in the Karmutsen flood basalt provinceWrangellia northern Vancouver Island In Grant B (ed)Geological Fieldwork 2005 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2006-1 39^52Greene A R Scoates J S amp Weis D (2008) Wrangellia flood
basalts in Alaska A record of plume^lithosphere interaction in aLate Triassic accreted oceanic plateau Geochemistry Geophysics
Geosystems 9(Q12004) doi1010292008GC002092Greene A R Scoates J S Weis D amp Israel S (2009)
Geochemistry of triassic flood basalts from the Yukon (Canada)segment of the accreted Wrangellia oceanic plateau Lithosdoi101016jlithos200811010
Hauff F Hoernle K Tilton G Graham D W amp Kerr A C(2000) Large volume recycling of oceanic lithosphere over shorttime scales geochemical constraints from the Caribbean LargeIgneous Province Earth and Planetary Science Letters 174 247^263
Hauff F Hoernle K amp Schmidt A (2003) Sr^Nd^Pb compositionof Mesozoic Pacific oceanic crust (Site 1149 and 801 ODP Leg185)Implications for alteration of ocean crust and the input into theIzu^Bonin^Mariana subduction system Geochemistry Geophysics
Geosystems 4(8) doi1010292002GC000421Herzberg C (2004) Partial melting below the Ontong Java Plateau
In Fitton J G Mahoney J J Wallace P J amp Saunders A D(eds) Origin and Evolution of the Ontong Java Plateau Geological SocietyLondon Special Publications 229 179^183
Herzberg C (2006) Petrology and thermal structure of the Hawaiianplume from Mauna Kea volcano Nature 444(7119) 605
Herzberg C amp Asimow P D (2008) Petrology of some oceanicisland basalts PRIMELT2XLS software for primary magma cal-culation Geochemistry Geophysics Geosystems 9(Q09001) doi1010292008GC002057
Herzberg C amp OrsquoHara M J (2002) Plume-associated ultramaficmagmas of Phanerozoic age Journal of Petrology 43(10) 1857^1883
Herzberg C Asimow P D Arndt N Niu Y Lesher C MFitton J G Cheadle M J amp Saunders A D (2007)Temperatures in ambient mantle and plumes Constraints frombasalts picrites and komatiites Geochemistry Geophysics Geosystems8(Q02006) doi1010292006GC001390
Huang S amp Frey F A (2005) Trace element abundances of MaunaKea basalt from phase 2 of the Hawaii Scientific Drilling Project
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
35
Petrogenetic implications of correlations with major element con-tent and isotopic ratios Geochemistry Geophysics Geosystems 4(6)doi1010292002GC000322
Hussong D M Wipperman L K amp Kroenke L W (1979) Thecrustal structure of the Ontong Java and Manihiki OceanicPlateaus Journal of Geophysical Research 84(B11) 6003^6010
Jerram D A amp Widdowson M (2005) The anatomy of ContinentalFlood Basalt Provinces geological constraints on the processes andproducts of flood volcanism Lithos 79(3^4) 385^405
Jones D L Silberling N J amp Hillhouse J (1977)Wrangellia a dis-placed terrane in northwestern North America CanadianJournal ofEarth Sciences 14(11) 2565^2577
Kerr A C (2003) Oceanic Plateaus In Rudnick R (ed) The Crust
Treatise on GeochemistryVol 3 Oxford Elsevier Science pp 537^565Kerr A C amp Mahoney J J (2007) Oceanic plateaus Problematic
plumes potential paradigms Chemical Geology 241 332^353Kerr A CTarney J Marriner G F Klaver GT Saunders A D
amp Thirlwall M F (1996) The geochemistry and petrogenesis ofthe Late-Cretaceous picrites and basalts of Curacao NetherlandsAntilles a remnant of an oceanic plateau Contributions to
Mineralogy and Petrology 124(1) 29^43Kerr A C Marriner G F Tarney J Nivia A Saunders A D
Thirlwall M F amp Sinton A C (1997) Cretaceous basaltic ter-ranes in Western Colombia Elemental chronological and Sr-Ndisotopic constraints on petrogenesis Journal of Petrology 38(6)677^702
Kinzler R J (1997) Melting of mantle peridotite at pressuresapproaching the spinel to garnet transition Application to mid-ocean ridge basalt petrogenesis Journal of Geophysical Research
102(B1) 853^874Kinzler R J amp Grove T L (1992) Primary magmas of mid-ocean
ridge basalts 1 Experiments and methods Journal of Geophysical
Research 97(B5) 6885^6906Korenaga J (2005) Why did not the Ontong Java Plateau form sub-
aerially Earth and Planetary Science Letters 234 385^399Kuniyoshi S (1972) Petrology of the Karmutsen (Volcanic) Group
northeastern Vancouver Island British Columbia PhD disserta-tion University of California Los Angeles 242 pp
Lassiter J C DePaolo D J amp Mahoney J J (1995) Geochemistry ofthe Wrangellia flood basalt province Implications for the role ofcontinental and oceanic lithosphere in flood basalt genesis Journalof Petrology 36(4) 983^1009
LeBas M J (2000) IUGS reclassification of the high-Mg and picriticvolcanic rocks Journal of Petrology 41(10) 1467^1470
Longhi J (2002) Some phase equilibrium systematics of lherzolitemelting I Geochemistry Geophysics Geosystems 3(3) doi1010292001GC000204
MacDonald G A amp Katsura T (1964) Chemical composition ofHawaiian lavas Journal of Petrology 5 82^133
Mahoney J J (1987) An isotopic survey of Pacific oceanic plateausimplications for their nature and origin In Keating B HFryer P Batiza R amp Boehlert GW (eds) Seamounts Islands andAtolls American Geophysical Union Geophysical Monograph 43 207^220
Mahoney J LeRoex A P Peng Z Fisher R L amp Natland J H(1992) Southwestern limits of Indian Ocean Ridge mantle and theorigin of low 206Pb204Pb mid-ocean ridge basalt isotope systema-tics of the central Southwest Indian Ridge (178^508E) Journal ofGeophysical Research 97(B13) 19771^19790
Mahoney J J Sinton J M Kurz M D MacDougall J DSpencer K J amp Lugmair GW (1994) Isotope and trace elementcharacteristics of a super-fast spreading ridge East Pacific rise13^238S Earth and Planetary Science Letters 121 173^193
Massey N W D (1995a) Geology and mineral resources of theAlberni^Nanaimo Lakes sheet Vancouver Island 92F1W 92F2Eand part of 92F7E British Columbia Ministry of Energy Mines and
Petroleum Resources Paper 1992-2 132 ppMassey N W D (1995b) Geology and mineral resources of the
Cowichan Lake sheet Vancouver Island 92C16 British Columbia
Ministry of Energy Mines and Petroleum Resources Paper 1992-3 112 ppMassey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005a) Digital Geology Map of British Columbia Tile NM9 MidCoast BC British Columbia Ministry of Energy and Mines Geofile
2005-2Massey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005b) Digital Geology Map of British Columbia Tile NM10Southwest BC British Columbia Ministry of Energy and Mines Geofile
2005-3McDonough W F amp Sun S (1995) The composition of the Earth
Chemical Geology 120 223^253Monger JW H amp Journeay J M (1994) Basement geology and tec-
tonic evolution of theVancouver region In Monger JW H (ed)Geology and Geological Hazards of the Vancouver Region Southwestern
British Columbia Geological Survey of Canada Bulletin 481 3^25Muller J E (1977) Geology of Vancouver Island Geological Survey of
Canada Open File Map 463Muller J E (1980)The Paleozoic Sicker Group of Vancouver Island British
Columbia Geological Survey of Canada Paper 79-30 22 ppMuller J E Northcote K E amp Carlisle D (1974) Geology and
mineral deposits of Alert Bay^Cape Scott map area VancouverIsland British Columbia Geological Survey of Canada Paper 74-8 77
Neal C R Mahoney J J Kroenke L W Duncan R A ampPetterson M G (1997) The Ontong^Java Plateau InMahoney J J amp Coffin M F (eds) Large Igneous Provinces
Continental Oceanic and Planetary FloodVolcanism American Geophysical
Union Geophysical Monograph 100 183^216Niu Y Collerson K D Batiza R Wendt J I amp Regelous M
(1999) Origin of enriched-typed mid-ocean ridge basalt at ridgesfar from mantle plumes The East Pacific Rise at 11820rsquoN Journalof Geophysical Research 104 7067^7088
Nixon G T amp Orr A J (2007) Recent revisions to the EarlyMesozoic stratigraphy of Northern Vancouver Island (NTS 102I092L) and metallogenic implicatinos British Columbia InGrant B (ed) Geological Fieldwork 2006 British Columbia Ministry of
Energy Mines and Petroleum Resources Paper 2007-1 163^177Nixon GT Hammack J L KoyanagiV M Payie G J Haggart
JW Orchard M JTozerT Friedman R M Archibald D APalfy J amp Cordey F (2006a) Geology of the Quatsino^PortMcNeill area northernVancouver Island British Columbia Ministry
of Energy Mines and Petroleum Resources Geoscience Map 2006-2 scale150 000
Nixon G T Kelman M C Stevenson D Stokes L A ampJohnston K A (2006b) Preliminary geology of the Nimpkishmap area (NTS 092L07) northern Vancouver Island BritishColumbia In Grant B amp Newell J M (eds) Geological Fieldwork2005 British Columbia Ministry of Energy Mines and Petroleum Resources
Paper 2006-1 135^152Nixon G T Laroque J Pals A Styan J Greene A R amp
Scoates J S (2008) High-Mg lavas in the Karmutsen flood basaltsnorthernVancouver Island (NTS 092L) Stratigraphic setting andmetallogenic significance In Grant B (ed) Geological Fieldwork
2007 British Columbia Ministry of Energy Mines and Petroleum
Resources Paper 2008-1 175^190Nowell G M Kempton P D Noble S R Fitton J G
Saunders A Mahoney J J amp Taylor R N (1998) High precisionHf isotope measurements of MORB and OIB by thermal ionisation
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
36
mass spectrometry insights into the depleted mantle Chemical
Geology 149 211^233Parrish R R amp McNicoll V J (1992) U^Pb age determinations
from the southern Vancouver Island area British Columbia InRadiogenic Age and Isotopic Studies Report 5 Geological Survey of
Canada Paper 91-2 79^86Peate DW (1997) The Parana^Etendeka Province In Coffin M F
amp Mahoney J (eds) Large Igneous Provinces Continental Oceanic andPlanetary Flood Volcanism American Geophysical Union Geophysical
Monograph 100 217^245Perfit M R Fornari D J Ridley W I Kirk P D Casey J
Kastens K A Reynolds J R Edwards M Desonie DShuster R amp Paradis S (1996) Recent volcanism in the Siqueirostransform fault picritic basalts and implications for MORBmagma genesis Earth and Planetary Science Letters 141(1^4) 91^108
Pretorius W Weis D Williams G Hanano D Kieffer B ampScoates J S (2006) Complete trace elemental characterization ofgranitoid (USGSG-2GSP-2) reference materials by high resolutioninductively coupled plasma-mass spectrometry Geostandards and
Geoanalytical Research 30(1) 39^54Regelous M Niu Y Wendt J I Batiza R Greig A amp
Collerson K D (1999) Variations in the geochemistry of magma-tism on the East Pacific Rise at 10830rsquoN since 800 ka Earth and
Planetary Science Letters 168 45^63Recurren villon S Chauvel C Arndt N T Pik R Martineau F
Fourcade S amp Marty B (2002) Heterogeneity of the Caribbeanplateau mantle source Sr O and He isotopic compositions of oli-vine and clinopyroxene from Gorgona Island Earth and Planetary
Science Letters 205(1^2) 91Rhodes J M (1996) Geochemical stratigraphy of lava flows samples
by the Hawaii Scientific Drilling Project Journal of Geophysical
Research 101(B5) 11729^11746Rhodes J M amp Vollinger M J (2004) Composition of basaltic lavas
sampled by phase-2 of the Hawaii Scientific Drilling ProjectGeochemical stratigraphy and magma types Geochemistry
Geophysics Geosystems 5(3) doi1010292002GC000434Richards M A Jones D L Duncan R A amp DePaolo D J (1991)
A mantle plume initiation model for the Wrangellia flood basaltand other oceanic plateaus Science 254 263^267
Salters V J M amp Hart S R (1991) The mantle sources of oceanridges islands and arcs the Hf-isotope connection Earth and
Planetary Science Letters 104 364^380Salters V J M amp Stracke A (2004) Composition of the depleted
SaltersV J amp WhiteW (1998) Hf isotope constraints on mantle evo-lution Chemical Geology 145 447^460
SanoT amp Yamashita S (2004) Experimental petrology of basementlavas from Ocean Drilling Program Leg 192 implications for dif-ferentiation processes in Ontong Java Plateau magmas InFitton J G Mahoney J JWallace P J amp Saunders A D (eds)Origin and Evolution of the Ontong Java Plateau Geological Society
London Special Publications 229 185^218Saunders A D (2005) Large igneous provinces Origin and environ-
mental consequences Elements 1 259^263Self S ThordarsonT amp Keszthely L (1997) Emplacement of conti-
nental flood basalt lava flows In Mahoney J J amp Coffin M F(eds) Large Igneous Provinces Continental Oceanic and Planetary Flood
Volcanism American Geophysical Union Geophysical Monograph 100381^410
Shaw D M (2000) Continuous (dynamic) melting theory revisitedCanadian Mineralogist 38(5) 1041^1063
Sluggett C L (2003) Uranium^lead age and geochemical constraintson Paleozoic and Early Mesozoic magmatism in WrangelliaTerrane Saltspring Island British Columbia BSc thesisUniversity of British ColumbiaVancouver 84 pp
Storey M Mahoney J J amp Saunders A D (1997) Cretaceousbasalts in Madagascar and the transition between plume and con-tinental lithospheric mantle sources In Mahoney J J ampCoffin M F (eds) Large Igneous Provinces Continental Oceanic and
Planetary Flood Volcanism Aamerican Geophysical Union Geophysical
Monograph 100 95^122Surdam R C (1967) Low-grade metamorphism of the Karmutsen
Group PhD dissertation University of California Los Angeles288 pp
Sutherland-Brown A Yorath C J Anderson R G amp Dom K(1986) Geological maps of southern Vancouver IslandLITHOPROBE I 92C10 11 14 16 92F1 2 7 8 Geological Survey ofCanada Open File 1272
Tejada M L G Mahoney J J Castillo P R Ingle S P Sheth HC amp Weis D (2004) Pin-pricking the elephant evidence on theorigin of the Ontong Java Plateau from Pb^Sr^Hf^Nd isotopiccharacteristics of ODP Leg 192 basalts In Fitton J GMahoney J J Wallace P J amp Saunders A D (eds) Origin and
Evolution of the Ontong Java Plateau Geological Society London Special
Publications 229 133^150Thompson P M E Kempton P D amp Kerr A C (2008) Evaluation
of the effects of alteration and leaching on Sm^Nd and Lu^Hf sys-tematics in submarine mafic rocks Lithos 104(1^4) 164^176
Thompson P M E Kempton P D White R V Kerr A CTarney J Saunders A D Fitton J G amp McBirney A (2003)Hf-Nd isotope constraints on the origin of the CretaceousCaribbean plateau and its relationship to the Galapagos plumeEarth and Planetary Science Letters 217(1^2) 59^75
Vervoort J D Patchett P Blichert-Toft J amp Albare de F (1999)Relationships between Lu^Hf and Sm^Nd isotopic systems in theglobal sedimentary system Earth and Planetary Science Letters 16879^99
Walter M J (1998) Melting of garnet peridotite and the origin ofkomatiite and depleted lithosphere Journal of Petrology 39(1) 29^60
Weis D Kieffer B Maerschalk C Barling J de Jong JWilliams G A Hanano D Mattielli N Scoates J SGoolaerts A Friedman R A amp Mahoney J B (2006) High-pre-cision isotopic characterization of USGS reference materials byTIMS and MC-ICP-MS Geochemistry Geophysics Geosystems
7(Q08006) doi1010292006GC001283Weis D Kieffer B Hanano D Silva I N Barling J
Pretorius W Maerschalk C amp Mattielli N (2007) Hf isotopecompositions of US Geological Survey reference materialsGeochemistry Geophysics Geosystems 8(Q06006) doi1010292006GC001473
Wheeler J O amp McFeely P (1991) Tectonic assemblage map of theCanadian Cordillera and adjacent part of the United States ofAmerica Geological Survey of Canada Map 1712A
WhiteW M Albare de F amp Tecurren louk P (2000) High-precision analy-sis of Pb isotope ratios by multi-collector ICP-MS Chemical Geology167 257^270
Yorath C J Sutherland Brown A amp Massey N W D (1999)LITHOPROBE southern Vancouver Island British Columbia Geological
Survey of Canada Bulletin 498 145 ppZou H (1998) Trace element fractionation during modal and
non-model dynamic melting and open-system melting Amathematical treatment Geochimica et Cosmochimica Acta 62(11)1937^1945
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
37
Zou H amp Reid M R (2001) Quantitative modeling of trace elementfractionation during incongruent dynamic melting Geochimica et
Cosmochimica Acta 65(1) 153^162
APPENDIX SAMPLEPREPARAT ION ANDANALYT ICAL METHODSOnly the freshest rocks in the field were sampled and onlythe least altered samples were selected for chemical andisotopic analysis based on thorough petrographic inspec-tion Fifty-nine of the 129 collected samples were crushed(400 g) into small pieces 52mm in diameter in aRocklabs hydraulic piston crusher between WC plates tominimize contamination The coarse-crush was mixedand 100 g was powdered in a planetary mill using agatejars and balls cleaned with quartz sand between samples
ActLabs analytical methodsThe major- and trace-element compositions of the whole-rock powders were determined at Activation LaboratoriesLtd (Actlabs) in Ancaster Ontario Analytical techniquesand detection limits are also available from Actlabs (httpwwwactlabscommethsub_code4erehtm) The analyticalmethod for each of the elements analyzed is indicated inTable 2 For the major elements a 02 g sample was mixedwith a lithium metaborate^lithium tetraborate mixtureand fused in a graphite crucible The molten mixture waspoured into a 5 HNO3 solution and shaken until dis-solved (30min) The samples were analyzed for majoroxides and selected trace elements by inductively coupledplasma optical emission spectrometry (ICP-OES) on acombination simultaneous^sequential Thermo Jarrell-AshEnviro II system Internal calibration was achieved usinga variety of international reference materials (eg W-2BIR-1 DNC-1) and independent control samplesAdditional trace elements were analyzed by both instru-mental neutron activation analysis (INAA) and induc-tively couple plasma mass spectrometry (ICP-MS) ForINAA15^25 g of sample was weighed into small polyeth-ylene vials and irradiated with a control international ref-erence material CANMET WMS-1 and NiCr flux wires ata thermal neutron flux of 71012 neutronscm2 per s in theMcMaster Nuclear Reactor Following a 7 day waitingperiod the samples were measured on an Ortec high-purity Ge detector linked to a Canberra Series 95 multi-channel analyzer Activities for each element weredecay- and weight-corrected and compared with a detectorcalibration developed from multiple international certifiedreference materials For ICP-MS 025 g of sample wasdigested in HF followed by HNO3^HClO4 heated andtaken to dryness The samples were brought back into solu-tion with HCl Samples were analyzed using a PerkinElmer Optima 3000 ICP In-lab standards or certified ref-erence materials (eg W-2 BIR-1 DNC-1) were used for
quality control A total of 15 blind duplicates were ana-lyzed from the same sample powders to assessreproducibility
University of Massachusetts analyticalmethodsFourteen sample duplicate powders taken from the same100 g sample aliquot (high-MgO basalts and picrites)were also analyzed at the Ronald B Gilmore X-RayFluorescence (XRF) Laboratory at the University ofMassachusetts Major elements were measured on a fusedLa-bearing lithium borate glass disc using a SiemensMRS-400 spectrometer Trace element concentrations(Rb Sr Ba Ce Nb Zr Y Pb Zn Ga Ni Cr V) weremeasured on a separate powder pellet using a PhilipsPW2400 sequential spectrometer using a Rh tube LOIand ferrous iron measurements were made as describedby Rhodes amp Vollinger (2004) Precision and accuracy esti-mates for the data have been described by Rhodes (1996)and Rhodes amp Vollinger (2004) Results for each sampleare the average of two separate analyses and are shown inElectronic Appendix 2 for a comparison of analyses ofduplicate powders using XRFand ICP-MS
PCIGR analytical methodsA subset of 19 samples was selected for high-precisiontrace-element analysis and Sr Nd Pb and Hf isotopicanalysis at the Pacific Centre for Isotopic andGeochemical Research (PCIGR) at the University ofBritish Columbia (UBC) Samples were selected from the59 samples analyzed for whole-rock chemistry at ActLabsbased on major- and trace-element chemistry alteration(low LOI and petrographic alteration index) and samplelocation Samples were prepared for trace-element analysisat the PCIGR by the technique described by Pretorius et al(2006) on unleached rock powders Sample powders(100mg) were weighed in 7ml screw-top Savillex bea-kers and dissolved in 1ml 14N HNO3 and 5ml 48 HFon a hotplate for 48 h at 1308C with periodic ultrasonica-tion Samples were dried and redissolved in 6ml 6N HClon a hotplate for 24 h and then dried and redissolved in1ml concentrated HNO3 for 24 h before final dryingTrace element abundances were measured by high-resolution inductively coupled plasma-mass spectrometry(HR-ICP-MS) with aThermo Finnigan Element2 systemfollowing the procedures described by Pretorius et al(2006) within 24 h of redissolution HFSE and LILE weremeasured in medium-resolution mode at 2000 dilutionusing a PFATeflon spray chamber washed with aqua regiafor 3min between samples REE were measured in high-resolution mode and U and Pb in low-resolution modeat 2000 dilution using a glass spray chamber washedwith 2 HNO3 between samples Total procedural blanksand reference materials (BCR-2 BHVO-2) were analyzedwith the batch of samples Indium was used as an internal
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
38
standard in all samples and standard solutionsBackground and standard solutions were analyzed afterevery five samples to detect memory effects and massdrift Results for PCIGR trace-element analyses areshown in Electronic Appendix 3 for the full set of analyzedtrace elementsSample digestion for purification of Sr Nd Hf and Pb
for column chemistry began by weighing each samplepowder (400^500mg) prior to leaching All samples wereleached with 6N HCl and placed in an ultrasonic bath for15min Samples were rinsed twice with 18 M-cm H2Obetween each leaching step (15 total) until the supernatantliquid was clear [following the method of Mahoney (1987)]Samples were then dried on a hotplate for 24 h andweighed again Sample solutions were then prepared bydissolving 100^250mg of the leached powder dissolvedin 1ml 14N HNO3 and 10ml 48 HF on a hotplate for48 h at 1308C with periodic ultrasonication Samples weredried and redissolved in 6ml 6N HCl on a hotplate for24 h and then dried Pb was separated using anion exhangecolumns and the discard was used for Sr REE and Hfseparation Nd was separated from the other REE and Hfrequired two additional purification steps Detailed proce-dures for column chemistry for separating Sr Nd and Pbat the PCIGR have been described byWeis et al (2006) andHf purification has been described byWeis et al (2007) Srand Nd isotope ratios were measured by thermal ioniza-tion mass spectrometry (TIMS) on a Thermo FinniganTriton system in static mode with relay matrix rotation ona single Ta and double Re^Ta filament respectively Fourto five filaments per barrel of 21 were occupied by stan-dards (NBS 987 for Sr and LaJolla for Nd) for each barrelwhere samples were run Sample Sr and Nd isotopic com-positions were corrected for mass fractionation using86Sr88Srfrac14 01194 and 146Nd144Ndfrac14 07219 Each samplewas then normalized using the barrel average of the refer-ence material relative to the values of143Nd144Ndfrac14 0511858 and 87Sr86Srfrac14 0710248 (Weiset al 2006) During the course of the Vancouver Islandanalyses the La Jolla Nd standard gave an average valueof 05118566 (nfrac14 7) and NBS987 standard gave an aver-age of 07102408 (nfrac1411) (2 error is reported as times106) 147Sm144Nd errors are 15 or 0006 USGeological Survey (USGS) reference material BHVO-2was processed with the samples and yielded Sr and Nd iso-topic ratios of 07034607 and 05129786 respectivelyThese are in agreement with the published values of070347920 and 051298411 respectively (Weis et al2006)
Pb and Hf isotopic compositions were analyzed by mul-tiple collector-inductively coupled plasma-mass spectrome-try (MC-ICP-MS) on a static multi-collection on a NuPlasma (Nu Instruments) system The detailed analyticalprocedure for Pb isotopic analyses on the Nu Plasma atthe PCIGR has been described by Weis et al (2006) Theconfiguration for Pb analyses allows for collection of PbTl and Hg togetherTl and Hg are used to monitor instru-mental mass discrimination and isobaric overlap repec-tively All sample solutions were analyzed withapproximately the same PbTl ratio (4) as the referencematerial NIST SRM 981 To accomplish this a small ali-quot of each sample solution from the Pb columns was ana-lyzed on the Element2 to determine the precise amount ofPb available for analysis on the Nu Plasma The SRM 981standard was run after every two samples on the NuPlasma During the time samples were run analyses ofthe SRM 981 Pb reference material gave values of206Pb204Pbfrac1416940322 207Pb204Pbfrac1415495823and 208Pb204Pbfrac14 36713164 (nfrac14 61 2 error is reportedas times 104) these values are in excellent agreement withreported TIMS triple-spike values of Galer amp Abouchami(1998) Fractionation-corrected Pb isotopic ratios were fur-ther corrected by the sample-standard bracketing methodor the ln^ln correction method described by White et al(2000) and Blichert-Toft et al (2003) Leached powders ofUSGS reference material BHVO-2 yielded Pb isotopicratios of 206Pb204Pbfrac14186454 8 207Pb204Pbfrac14 1549105 and 208Pb204Pbfrac14 38222514 and BCR-2yielded 206Pb204Pbfrac141880466 207Pb204Pbfrac14 1562518 and 208Pb204Pbfrac14 3883496 (2 error isreported as times 104) These values are in agreement withthose for leached residues of BHVO-2 and BCR-2 fromWeis et al (2006)Hf isotopic compositions were analyzed following the
procedures detailed byWeis et al (2007) The configurationfor Hf analyses monitored Lu mass 175 andYb mass 172 toallow for interference correction to masses 174 and 176 Hfisotopic ratios were normalized internally for mass fractio-nation to a 179Hf177Hf ratio of 07325 using an exponentialcorrection Standards were run after every two samplesand sample results were normalized to the ratio of the in-run daily average and a 176Hf177Hf ratio for JMC-475 of0282160 During the course of analyses the Hf standardJMC-475 gave an average value 02821679 (nfrac14130)USGS reference materials BCR-2 and BHVO-2 were pro-cessed with the samples and yielded Hf isotopic ratios of02828675 and 02831005 respectively Publishedvalues for BCR-2 and BHVO-2 are 02828717 and0283104 8 respectively (Weis et al 2007)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
1Ni concentrations for these samples were determined by XRFTHOL tholeiitic basalt PIC picrite HI-MG high MgO basalt CG coarse-grained (sill or gabbro) MIN SIL mineralizedsill OUTLIER anomalous pillowed flow in plots MA Mount Arrowsmith SL Schoen Lake KR Karmutsen Range QIQuadra Island Sample locations are given using the Universal Transverse Mercator (UTM) coordinate system (NAD83zones 9 and 10) Analyses were performed at Activation Laboratory (ActLabs) Fe2O3
is total iron expressed as Fe2O3LOI loss on ignition All major elements Sr V and Y were measured by ICP quadrupole OES on solutions of fusedsamples Cu Ni Pb and Zn were measured by total dilution ICP Cs Ga Ge Hf Nb Rb Ta Th U Zr and REE weremeasured by magnetic-sector ICP on solutions of fused samples Co Cr and Sc were measured by INAA Blanks arebelow detection limit See Electronic Appendices 2 and 3 for complete XRF data and PCIGR trace element datarespectively Major elements for sample 93G17 are normalized anhydrous Sample 5615A7 was analyzed by XRFSamples from Quadra Island are not shown in figures
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
17
picritic and high-MgO basalt pillow lavas extend to20wt MgO (Fig 6 and references listed in thecaption)The tholeiitic basalts from the Karmutsen Range dis-
play a tight range of parallel light rare earth element(LREE)-enriched REE patterns (LaYbCNfrac14 21^24mean 2202) whereas the picrites and high-MgObasalts are characterized by sub-parallel LREE-depleted
patterns (LaYbCNfrac14 03^07 mean 06 02) with lowerREE abundances than the tholeiitic basalts (Fig 7)Tholeiitic basalts from the Schoen Lake and MountArrowsmith areas have similar REE patterns tosamples from the Karmutsen Range (Schoen Lake LaYbCNfrac14 21^26 mean 2303 Mount Arrowsmith LaYbCNfrac1419^25 mean 2204) and the coarse-grainedmafic rocks have similar REE patterns to the tholeiitic
Depleted MORB average(Salters amp Stracke 2004)
Schoen LakeSchoen Lake
Mount Arrowsmith
Karmutsen RangeS
ampl
eC
hond
rite
Sam
ple
Cho
ndrit
e
Sam
ple
Prim
itive
Man
tleS
ampl
eP
rimiti
ve M
antle
Sam
ple
Prim
itive
Man
tle
Mount Arrowsmith
Karmutsen Range
(a) (b)
(c) (d)
(e) (f )
Sam
ple
Cho
ndrit
e
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained mafic rock
Pillowed flowMineralized sill
1
10
100
1
10
100
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
01
1
10
100
1
10
100
1
10
100
Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
Fig 7 Whole-rock REE and other incompatible-element concentrations for the Karmutsen Formation (a) (c) and (e) are chondrite-normal-ized REE patterns for samples from each of the three main field areas onVancouver Island (b) (d) and (f) are primitive mantle-normalizedtrace-element patterns for samples from each of the three main field areas All normalization values are from McDonough amp Sun (1995) Theclear distinction between LREE-enriched tholeiitic basalts and LREE-depleted picrites the effects of alteration on the LILE (especially loss ofK and Rb) and the different range for the vertical scale in (b) (extends down to 01) should be noted
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
18
basalts (LaYbCNfrac14 21^29 mean 2508) Two LREE-depleted coarse-grained mafic rocks from the SchoenLake area have similar REE patterns (LaYbCNfrac14 07) tothe picrites from the Keogh Lake area indicating that thisdistinctive suite of rocks is exposed as far south asSchoen Lake that is 75 km away (Fig 7) The anomalouspillowed flow from the Karmutsen Range has lower LaYbCN (13^18) than the main group of tholeiitic basaltsfrom the Karmutsen Range (Fig 7) The mineralized sillfrom the Schoen Lake area is distinctly LREE-enriched(LaYbCNfrac14 33) with the highest REE abundances(LaCNfrac14 940) of all Karmutsen samples this may reflectcontamination by adjacent sediment also evident in thin-section where sulfide blebs are cored by quartz (Greeneet al 2006)The primitive mantle-normalized trace-element pat-
terns (Fig 7) and trace-element variations and ratios(Fig 8) highlight the differences in trace-element
concentrations between the four main groups ofKarmutsen flood basalts onVancouver Island The tholeii-tic basalts from all three areas have relatively smooth par-allel trace-element patterns some samples have smallpositive Sr anomalies relative to Nd and Sm and manysamples have prominent negative K anomalies relative toU and Nb reflecting K loss during alteration (Fig 7) Thelarge ion lithophile elements (LILE) in the tholeiiticbasalts (mainly Rb and K not Cs) are depleted relativeto the high field strength elements (HFSE) and LREELILE segments especially for the Schoen Lake area(Fig 7d) are remarkably parallel The picrites andhigh-MgO basalts form a tight band of parallel trace-element patterns and are depleted in HFSE and LREEwith positive Sr anomalies and relatively low Rb Thecoarse-grained mafic rocks from the Karmutsen Rangehave identical trace-element patterns to the tholeiiticbasalts Samples from the Schoen Lake area have similar
000
025
050
075
100
125
0 1 2 3 4 5
00
04
08
12
16
0 5 10 15 20 25
MgO (wt)
0
5
10
15
0 50 100 150
Hf (ppm)
Th (ppm)
NbLaNb (ppm)
Zr (ppm)
(a)
(c)
(b)
(d)
Th (ppm)
00
02
04
06
08
10
12
00 01 02 03 04 05
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
NbLa=1
4723A4 5615A12
U (ppm)
Fig 8 Whole-rock trace-element concentrations and ratios for the Karmutsen Formation [except (b) which is vs wt MgO] (a) Nb vs Zr(b) NbLa vs MgO (c) Th vs Hf (d) Th vs U There is a clear distinction between the tholeiitic basalts and picrites in Nb and Zr both inconcentration and the slope of each trend
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
19
trace-element patterns to the Karmutsen Range except forthe two LREE-depleted coarse-grained mafic rocks andthe mineralized sill (Fig 7d)
Sr^Nd^Hf^Pb isotopic compositionsA total of 19 samples from the four main groups ofthe Karmutsen Formation were selected for radiogenicisotopic geochemistry on the basis of major- and trace-element variation stratigraphic position and geographicallocation The samples have overlapping age-correctedHf and Nd isotope ratios and distinct ranges of Sr isotopiccompositions (Fig 9) The tholeiitic basalts picriteshigh-MgO basalt and coarse-grained mafic rocks have
initial eHffrac14thorn87 to thorn126 and eNdfrac14thorn66 to thorn88 cor-rected for in situ radioactive decay since 230 Ma (Fig 9Tables 3 and 4) All the Karmutsen samples form a well-defined linear array in a Lu^Hf isochron diagram corre-sponding to an age of 24115 Ma (Fig 9) This is withinerror of the accepted age of the Karmutsen Formationand indicates that the Hf isotope systematics behaved as aclosed system since c 230 Ma The anomalous pillowedflow has similar initial eHf (thorn98) and slightly lower initialeNd (thorn62) compared with the other Karmutsen samplesThe picrites and high-MgO basalt have higher initial87Sr86Sr (070398^070518) than the tholeiitic basalts(initial 87Sr86Srfrac14 070306^070381) and the coarse-grained mafic rocks (initial 87Sr86Srfrac14 070265^070428)
05128
05129
05130
05131
05132
015 017 019 021 023 025
02829
02830
02831
02832
02833
02834
000 002 004 006 008 010
4
6
8
10
0702 0703 0704 0705 07066
8
10
12
14
5 6 7 8 9 10
12
0
1
2
3
4
5
6
7
0702 0703 0704 0705 0706
(d)
LOI (wt )
87Sr 86Sr
4723A2
Nd
143Nd144Nd
147Sm144Nd
87Sr 86Sr (230 Ma)
(230 Ma)
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
(a)
(c)
176Hf177Hf
Age=241 plusmn 15 Ma=+990 plusmn 064
176Lu177Hf
Hf (230 Ma)
Nd (230 Ma)
(e)
Hf (i)
(b)
Fig 9 Whole-rock Sr Nd and Hf isotopic compositions for the Karmutsen Formation (a) 143Nd144Nd vs 147Sm144Nd (b) 176Hf177Hf vs176Lu177Hf The slope of the best-fit line for all samples corresponds to an age of 24115 Ma (c) Initial eNd vs 87Sr86Sr Age-correction to230 Ma (d) LOI vs 87Sr86Sr (e) Initial eHf vs eNd Average 2 error bars are shown in a corner of each panel Complete chemical duplicatesshown inTables 3 and 4 (samples 4720A7 and 4722A4) are circled in each plot
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
overlap the ranges of the picrites and tholeiitic basalts(Fig 9 Table 3)The measured Pb isotopic compositions of the tholeiitic
basalts are more radiogenic than those of the picrites andthe most magnesian Keogh Lake picrites have the leastradiogenic Pb isotopic compositions The range ofinitial Pb isotope ratios for the picrites is206Pb204Pbfrac1417868^18580 207Pb204Pbfrac1415547^15562and 208Pb204Pbfrac14 37873^38257 and the range for thetholeiitic basalts is 206Pb204Pbfrac1418782^19098207Pb204Pbfrac1415570^15584 and 208Pb204Pbfrac14 37312^38587 (Fig 10 Table 5) The coarse-grained mafic rocksoverlap the range of initial Pb isotopic compositions forthe tholeiitic basalts with 206Pb204Pbfrac1418652^19155207Pb204Pbfrac1415568^15588 and 208Pb204Pbfrac14 38153^38735 One high-MgO basalt has the highest initial206Pb204Pb (19242) and 207Pb204Pb (15590) and thelowest initial 208Pb204Pb (37676) The Pb isotopic ratiosfor Karmutsen samples define broadly linearrelationships in Pb isotope plots The age-corrected Pb
isotopic compositions of several picrites have been affectedby U and Pb (andor Th) mobility during alteration(Fig 10)
ALTERAT IONThe Karmutsen basalts have retained most of their origi-nal igneous structures and textures however secondaryalteration and low-grade metamorphism have producedzeolitic- and prehnite^pumpellyite-bearing mineral assem-blages (Table 1 Cho et al 1987) Alteration and metamor-phism have primarily affected the distribution of the LILEand the Sr isotopic systematics The Keogh Lake picriteshave been affected more by alteration than the tholeiiticbasalts and have higher LOI (up to 55wt ) variableLILE and higher measured Sr Nd and Hf and lowermeasured Pb isotope ratios than the tholeiitic basalts Srand Pb isotope compositions for picrites and tholeiiticbasalts show a relationship with LOI (Figs 9 and 10)whereas Nd and Hf isotopic compositions do not correlate
Table 3 Sr and Nd isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Rb Sr 87Sr86Sr 2m87Rb86Sr 87Sr86Srt Sm Nd 143Nd144Nd 2m eNd
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses carried out at the PCIGR thecomplete trace element analyses are given in Electronic Appendix 3
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
21
with LOI (not shown) The relatively high apparent initialSr isotopic compositions for the high-MgO lavas (up to07052) probably resulted from post-magmatic loss of Rbor an increase in 87Sr86Sr through addition of seawater Sr(eg Hauff et al 2003) High-MgO lavas from theCaribbean plateau have similarly high 87Sr86Sr comparedwith basalts (Recurren villon et al 2002)The correction for in situdecay on initial Pb isotopic ratios has been affected bymobilization of U and Th (and Pb) in the whole-rockssince their formation A thorough acid leaching duringsample preparation was used that has been shown to effec-tively remove alteration phases (Weis et al 2006 NobreSilva et al in preparation) Whole-rock SmNd ratiosand age-correction of Nd isotopes may be affected bythe leaching procedure for submarine rocks older than50 Ma (Thompson et al 2008 Fig 9) there ishowever very little variation in Nd isotopic compositionsof the picrites or tholeiitic basalts and this system does notappear to be disturbed by alteration The HFSE abun-dances for tholeiitic and picritic basalts exhibit clear
linear correlations in binary diagrams (Fig 8) whereasplots of LILE vs HFSE are highly scattered (not shown)because of the mobility of some of the alkali and alkalineearth LILE (Cs Rb Ba K) and Sr for most samplesduring alterationThe degree of alteration in the Karmutsen samples does
not appear to be related to eruption environment or depthof burial Pillow rims are compositionally different frompillow cores (Surdam1967 Kuniyoshi1972) and aquagenetuffs are chemically different from isolated pillows withinpillow breccias however submarine and subaerial basaltsexhibit a similar degree of alterationThere is no clear cor-relation between the submarine and subaerial basalts andsome commonly used chemical alteration indices (eg BaRb vs K2OP2O5 Huang amp Frey 2005) There is also nodefinitive correlation between the petrographic alterationindex (Table 1) and chemical alteration indices although10 of 13 tholeiitic basalts with the highest K and LILEabundances have the highest petrographic alterationindex of three (seeTable 1)
Table 4 Hf isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Lu Hf 176Hf177Hf 2m eHf176Lu177Hf 176Hf177Hft eHf(t)
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
OLIV INE ACCUMULAT ION INH IGH-MgO AND PICR IT IC LAVASGeochemical trends and petrographic characteristics indi-cate that accumulation of olivine played an important rolein the formation of the high-MgO basalts and picrites fromthe Keogh Lake area on northern Vancouver Island TheKeogh Lake samples show a strong linear correlation inplots of Al2O3 TiO2 ScYb and Ni vs MgO and many ofthe picrites have abundant clusters of olivine pseudomorphs(Table1 Fig11)There is a strong linear correlationbetween
modal per cent olivine and whole-rock magnesium contentmagnesium contents range between 10 and 19wt MgOand the proportion of olivine phenocrysts in these samplesvaries from 0 to 42 vol (Fig 11)With a few exceptionsboth the size range and median size of the olivine pheno-crysts in most samples are comparable The clear correla-tion between the proportion of olivine phenocrysts andwhole-rock magnesium contents indicates that accumula-tion of olivine was directly responsible for the high MgOcontents (410wt ) in most of the picritic lavas
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
207Pb204Pb
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
208Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
(a) (b)
(c) (e)
4723A24723A2
4723A2
4723A2
4722A4
4723A4
4722A4
4723A4
4723A134723A3
4723A3
4723A13
1556
1558
1560
1562
1564
1566
1568
186 190 194 198 202 206 2101550
1552
1554
1556
1558
1560
178 180 182 184 186 188 190 192 194
380
384
388
392
396
186 190 194 198 202 206 210374
378
382
386
390
178 180 182 184 186 188 190 192 194
206Pb204Pb(d)
LOI (wt )
0
1
2
3
4
5
6
7
186 190 194 198 202 206 210
4723A2
Fig 10 Pb isotopic compositions of leached whole-rock samples measured by MC-ICP-MS for the Karmutsen Formation Error bars are smal-ler than symbols (a) Measured 207Pb204Pb vs 206Pb204Pb (b) Initial 207Pb204Pb vs 206Pb204Pb Age-correction to 230 Ma (c) Measured208Pb204Pb vs 206Pb204Pb (d) LOI vs 206Pb204Pb (e) Initial 208Pb204Pb vs 206Pb204Pb Complete chemical duplicates shown in Table 5(samples 4720A7 and 4722A4) are circled in each plot The dashed lines in (b) and (e) show the differences in age-corrections for two picrites(4723A4 4722A4) using the measured UTh and Pb concentrations for each sample and the age-corrections when concentrations are used fromthe two picrites (4723A3 4723A13) that appear to be least affected by alteration
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
23
DISCUSS IONThe compositional and spatial development of the eruptedlava sequences of the Wrangellia oceanic plateau onVancouver Island provide information about the evolutionof a mantle plume upwelling beneath oceanic lithospherein an analogous way to the interpretation of continentalflood basalts The Caribbean and Ontong Java plateausare the two primary examples of accreted parts of oceanicplateaus that have been studied to date and comparisonscan be made with the Wrangellia oceanic plateauHowever the Wrangellia oceanic plateau is a distinctiveoccurrence of an oceanic plateau because it formed on thecrust of a Devonian to Mississippian intra-oceanic islandarc overlain by Mississippian to Permian limestones andpelagic sediments The nature and thickness of oceaniclithospheric mantle are considered to play a fundamentalrole in the decompression and evolution of a mantleplume and thus the compositions of the erupted lavasequences (Greene et al 2008) The geochemical and
stratigraphic relationships observed in the KarmutsenFormation onVancouver Island and the focus of the subse-quent discussion sections provide constraints on (1) thetemperature and degree of melting required to generatethe primary picritic magmas (2) the composition of themantle source (3) the depth of melting and residual min-eralogy in the source region (4) the low-pressure evolutionof the Karmutsen tholeiitic basalts (5) the growth of theWrangellia oceanic plateau
Melting conditions and major-elementcomposition of the primary magmasThe primary magmas to most flood basalt provinces arebelieved to be picritic (eg Cox 1980) and they have beenfound in many continental and oceanic flood basaltsequences worldwide (eg Siberia Karoo Paranacurren ^Etendeka Caribbean Deccan Saunders 2005) Near-pri-mary picritic lavas with low total alkali abundances
Table 5 Pb isotopic compositions of Karmutsen basaltsVancouver Island BC
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
require high-degree partial melting of the mantle source(eg Herzberg amp OrsquoHara 2002) The Keogh Lake picritesfrom northernVancouver Island are the best candidates foridentifying the least modified partial melts of the mantleplume source despite having accumulated olivine pheno-crysts and can be used to estimate conditions of meltingand the composition of the primary magmas for theKarmutsen FormationPrimary magma compositions mantle potential tem-
peratures and source melt fractions for the picritic lavas ofthe Karmutsen Formation were calculated from primitivewhole-rock compositions using PRIMELT1XLS software(Herzberg et al 2007) and are presented in Fig 12 andTable 6 A detailed discussion of the computationalmethod has been given by Herzberg amp OrsquoHara (2002)and Herzberg et al (2007) key features of the approachare outlined belowPrimary magma composition is calculated for a primi-
tive lava by incremental addition of olivine and compari-son with primary melts of mantle peridotite Lavas thathad experienced plagioclase andor clinopyroxene fractio-nation are excluded from this analysis All calculated pri-mary magma compositions are assumed to be derived byaccumulated fractional melting of fertile peridotite KR-4003 Uncertainties in fertile peridotite composition donot propagate to significant variations in melt fractionand mantle potential temperature melting of depletedperidotite propagates to calculated melt fractions that aretoo high but with a negligible error in mantle potential
temperature PRIMELT1XLS calculates the olivine liqui-dus temperature at 1atm which should be similar to theactual eruption temperature of the primary magmaand is accurate to 308C at the 2 level of confidenceMantle potential temperature is model dependent witha precision that is similar to the accuracy of the olivineliquidus temperature (C Herzberg personal communica-tion 2008)With regard to the evidence for accumulated olivine in
the Keogh Lake picrites it should not matter whether themodeled lava composition is aphyric or laden with accu-mulated olivine phenocrysts PRIMELT1 computes theprimary magma composition by adding olivine to a lavathat has experienced olivine fractionation in this way itinverts the effects of fractional crystallization The onlyrequirement is that the olivine phenocrysts are not acci-dental or exotic to the lava compositions Support for thisprocedure can be seen in the calculated olivine liquid-line-of-descent shown by the curved arrays with 5 olivineaddition increments (Fig 12c) Fractional crystallization ofolivine from model primary magmas having 85^95wt FeO and 153^175wt MgO will yield arrays of deriva-tive liquids and randomly sorted olivines that in most butnot all cases connect the picrites and high-MgO basalts Anewer version of the PRIMELT software (Herzberg ampAsimow 2008) can perform olivine addition to and sub-traction from lavas with highly variably olivine phenocrystcontents and yields results that converge with those shownin Fig 12
Fig 11 Relationship between abundance of olivine phenocrysts and whole-rock MgO contents for the Keogh Lake picrites (a) Tracings ofolivine grains (gray) in six high-MgO samples made from high resolution (2400 dpi) scans of petrographic thin-sections Scale bars represent10mm sample numbers are indicated at the upper left (b) Modal olivine vs whole-rock MgO Continuous line is the best-fit line for 10 high-MgO samplesThe areal proportion of olivine was calculated from raster images of olivine grains using ImageJ image analysis software whichprovides an acceptable estimation of the modal abundance (eg Chayes 1954)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
25
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
Gray lines = final melting pressure
L + Ol + Opx
07
08
09
Pressure (GPa)
10
3
4
5
6
1
SOLIDUS
01
04
0506
L + Ol
2
03
02
PeridotiteKR-4003
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
Thick black lines = initial melting pressure
Dashed lines = melt fraction
Melt Fraction
2
3 4 5 6
L+Ol+Opx
SolidusSpinel
SolidusGarnet Peridotite
7
L+Ol
0 10 20 30 4040
45
50
55
60
MgO (wt)
SiO2
(wt)
PeridotiteKR-4003
Melt Fraction
0 10 20 305
10
15
CaO(wt)
MgO (wt)
3
4 5 6 7
2
46
2
Peridotite Partial Melts
Thick black line = initial melting pressureGray lines = final melting pressure
Peridotite
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
02
Pressure (GPa)
0302
04
Pressure (GPa)
Picrite
High-MgO basalt
Tholeiitic basalt
(c)
(d)
Karmutsen flood basaltsOlivine addition modelOlivine addition (5 increment)Primary magma
PicriteHigh-MgO basaltTholeiitic basalt
(a)
(b)
800 850 900
910
920
930
940
950
Karmutsen flood basalts
Olivine addition modelOlivine addition (5 increment)Primary magma
Gray lines = Fo contents of olivine
Fig 12 Estimated primary magma compositions for three Keogh Lake picrites and high-MgO basalts (samples 4723A4 4723A135616A7) using the forward and inverse modeling technique of Herzberg et al (2007) Karmutsen compositions and modeling results areoverlain on diagrams provided by C Herzberg (a) Whole-rock FeO vs MgO for Karmutsen samples from this study Total iron estimatedto be FeO is 090 Gray lines show olivine compositions that would precipitate from liquid of a given MgO^FeO composition Black lineswith crosses show results of olivine addition (inverse model) using PRIMELT1 (Herzberg et al 2007) Gray field highlights olivinecompositions with Fo contents of 91^92 predicted to precipitate (b) FeO vs MgO showing Karmutsen lava compositions and results ofa forward model for accumulated fractional melting of fertile peridotite KR-4003 (Kettle River Peridotite Walter 1998) (c) SiO2 vsMgO for Karmtusen lavas and model results (d) CaO vs MgO showing Karmtusen lava compositions and model results To brieflysummarize the technique [see Herzberg et al (2007) for complete description] potential parental magma compositions for the high-MgOlava series were selected (highest MgO and appropriate CaO) and using PRIMELT1 software olivine was incrementally added tothe selected compositions to show an array of potential primary magma compositions (inverse model) Then using PRIMELT1 theresults from the inverse model were compared with a range of accumulated fractional melts for fertile peridotite derived fromparameterization of the experimental results for KR-4003 from Walter (1998) forward model Herzberg amp OrsquoHara (2002) A melt fractionwas sought that was unique to both the inverse and forward models (Herzberg et al 2007) A unique solution was found when there was a com-mon melt fraction for both models in FeO^MgO and CaO^MgO^Al2O3^SiO2 (CMAS) projection space This modeling assumes thatolivine was the only phase crystallizing and ignores chromite precipitation and possible augite fractionation in the mantle (Herzbergamp OrsquoHara 2002) Results are best for a residue of spinel lherzolite (not pyroxenite) The presence of accumulated olivine in samples ofthe Keogh Lake picrites used as starting compositions does not significantly affect the results because we are modeling addition of olivine(see text for further description) The tholeiitic basalts cannot be used for modeling because they are all saturated in plagthorn cpxthornolThick black line for initial melting pressure at 4 GPa is not shown in (c) for clarity Gray area in (a) indicates parental magmas thatwill crystallize olivine with Fo 91^92 at the surface Dashed lines in (c) indicate melt fraction Solidi for spinel and garnet peridotite arelabelled in (c)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
The PRIMELT1 results indicate that if the Karmutsenpicritic lavas were derived from accumulated fractionalmelting of fertile peridotite the primary magmas wouldhave contained 15^17wt MgO and 10wt CaO(Fig 12 Table 6) formed from 23^27 partial meltingand would have crystallized olivine with a Fo content of 91 (Fig 12 Table 6) Calculated mantle temperatures forthe Karmutsen picrites (14908C) indicate melting ofanomalously hot mantle (100^2008C above ambient) com-pared with ambient mantle that produces mid-ocean ridgebasalt (MORB 1280^14008C Herzberg et al 2007)which initiated between 3 and 4 GPa Several picritesappear to be near-primary melts and required very littleaddition of olivine (eg sample 93G171 with measured158wt MgO) These temperature and melting esti-mates of the Karmutsen picrites are consistent withdecompression of hot mantle peridotite in an actively con-vecting plume head (ie plume initiation model) Threesamples (picrite 5614A1 and high-MgO basalts 5614A3and 5614A5) are lower in iron than the other Karmutsenlavas with FeO in the 65^80wt range (Fig 12c) andhave MgO and FeO contents that are similar to primitiveMORB from the Sequeiros fracture zone of the EastPacific Rise (EPR Perfit et al 1996) These samples
however differ from EPR MORB in having higher SiO2
and lower Al2O3 and they are restricted geographicallyto one area (Sara Lake Fig 2)We therefore have evidence for primary magmas with
MgO contents that range from a possible low of 110wt to as high as 174wt with inferred mantle potential tem-peratures from 1340 to 15208C This is similar to the rangedisplayed by lavas from the Ontong Java Plateau deter-mined by Herzberg (2004) who found that primarymagma compositions assuming a peridotite sourcewould contain 17wt MgO and 10wt CaO(Table 6) and that these magmas would have formed by27 melting at a mantle potential temperature of15008C (first melting occurring at 36 GPa) Fitton ampGodard (2004) estimated 30 partial melting for theOntong Java lavas based on the Zr contents of primarymagmas of Kroenke-type basalt calculated by incrementaladdition of equilibrium olivine However if a considerableamount of eclogite was involved in the formation of theOntong Java plateau magmas the excess temperatures esti-mated by Herzberg et al (2007) may not be required(Korenaga 2005) The variability in mantle potential tem-perature for the Keogh Lake picrites is also similar to thewide range of temperatures that have been calculated for
Table 6 Estimated primary magma compositions for Karmutsen basalts and other oceanic plateaus or islands
Sample 4723A4 4723A13 5616A7 93G171 Average OJP Mauna Keay Gorgonaz
Ontong Java primary magma composition for accumulated fraction melting (AFM) from Herzberg (2004)yMauna Kea primary magma composition is average of four samples of Herzberg (2006 table 1)zGorgona primary magma composition for AFM for 1F fertile source from Herzberg amp OrsquoHara (2002 table 4)Total iron estimated to be FeO is 090
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
27
lavas from single volcanoes in the Galapagos Hawaii andelsewhere by Herzberg amp Asimow (2008) Those workersinterpreted the range to reflect the transport of magmasfrom both the cool periphery and the hotter axis of aplume A thermally heterogeneous plume will yield pri-mary magmas with different MgO contents and the rangeof primary magma compositions and their inferred mantlepotential temperatures for Karmutsen lavas may reflectmelting from different parts of the plume
Source of the Karmutsen Formation lavasThe Sr^Nd^Hf^Pb isotopic compositions of theKarmutsen volcanic rocks on Vancouver Island indicatethat they formed from plume-type mantle containing adepleted component that is compositionally similar to butdistinct from other oceanic plateaus that formed in thePacific Ocean (eg Ontong Java and Caribbean Fig 13)Trace element and isotopic constraints can be used to testwhether the depleted component of the KarmutsenFormation was intrinsic to the mantle plume and distinctfrom the source of MORB or the result of mixing or invol-vement of ambient depleted MORB mantle with plume-type mantle The Karmutsen tholeiitic basalts have initialeNd and
87Sr86Sr ratios that fall within the range of basaltsfrom the Caribbean Plateau (eg Kerr et al 1996 1997Hauff et al 2000 Kerr 2003) and a range of Hawaiian iso-tope data (eg Frey et al 2005) and lie within and abovethe range for the Ontong Java Plateau (Tejada et al 2004Fig 13a) Karmutsen tholeiitic basalts with the highest ini-tial eNd and lowest initial 87Sr86Sr lie within and justbelow the field for age-corrected northern EPR MORB at230 Ma (eg Niu et al1999 Regelous et al1999) Initial Hfand Nd isotopic compositions for the Karmutsen samplesare noticeably offset to the low side of the ocean islandbasalt (OIB) array (Vervoort et al 1999) and lie belowand slightly overlap the low eHf end of fields for Hawaiiand the Caribbean Plateau (Fig 13c) the Karmutsen sam-ples mostly fall between and slightly overlap a field forage-corrected Pacific MORB at 230 Ma and Ontong Java(Mahoney et al 1992 1994 Nowell et al 1998 Chauvel ampBlichert-Toft 2001) Karmutsen lava Pb isotope ratios over-lap the broad range for the Caribbean Plateau and thelinear trends for the Karmutsen samples intersect thefields for EPR MORB Hawaii and Ontong Java in208Pb^206Pb space but do not intersect these fields in207Pb^206Pb space (Fig 13d and e) The more radiogenic207Pb204Pb for a given 206Pb204Pb of the Karmutsenbasalts requires high UPb ratios at an ancient time when235U was abundant similar to the Caribbean Plateau andenriched in comparison with EPR MORB Hawaii andOntong JavaThe low eHf^low eNd component of the Karmutsen
basalts that places them below the OIB mantle array andpartly within the field for age-corrected Pacific MORBcan form from low ratios of LuHf and high SmNd over
time (eg Salters amp White 1998) MORB will develop aHf^Nd isotopic signature over time that falls below themantle array Low LuHf can also lead to low 176Hf177Hffrom remelting of residues produced by melting in theabsence of garnet (eg Salters amp Hart 1991 Salters ampWhite 1998) The Hf^Nd isotopic compositions of theKarmutsen Formation indicate the possible involvementof depleted MORB mantle however similar to Iceland(Fitton et al 2003) Nb^Zr^Y variations provide a way todistinguish between a depleted component within theplume and a MORB-like component The Karmutsenbasalts and picrites fall within the Iceland array and donot overlap the MORB field in a logarithmic plot of NbYvs ZrY (Fig 13b) using the fields established by Fittonet al (1997) Similar to the depleted component within theCaribbean Plateau (Thompson et al 2003) the trend ofHf^Nd isotopic compositions towards Pacific MORB-likecompositions is not followed by MORB values in Nb^Zr^Y systematics and this suggests that the depleted compo-nent in the source of the Karmutsen basalts is distinctfrom the source of MORB and probably an intrinsic partof the plume The relatively radiogenic Pb isotopic ratiosand REE patterns distinct from MORB also support thisinterpretationTrace-element characteristics suggest the possible invol-
vement of sub-arc lithospheric mantle in the generation ofthe Karmutsen picrites Lassiter et al (1995) suggested thatmixing of a plume-type source with eNdthorn 6 to thorn 7 witharc material with low NbTh could reproduce the varia-tions observed in the Karmutsen basalts however theyconcluded that the absence of low NbLa ratios in theirsample of tholeiitic basalts restricts the amount of arc litho-sphere involved The study of Lassiter et al (1995) wasbased on major and trace element and Sr Nd and Pb iso-topic data for a suite of 29 samples from Buttle Lake in theStrathcona Provincial Park on central Vancouver Island(Fig 1) Whereas there are no clear HFSE depletions inthe Karmutsen tholeiitic basalts the low NbLa (and TaLa) of the Karmutsen picritic lavas in this study indicatethe possible involvement of Paleozoic arc lithosphere onVancouver Island (Fig 8)
REE modeling dynamic melting andsource mineralogyThe combined isotopic and trace-element variations of thepicritic and tholeiitic Karmutsen lavas indicate that differ-ences in trace elements may have originated from differentmelting histories For example the LuHf ratios of theKarmutsen picritic lavas (mean 008) are considerablyhigher than those in the tholeiitic basalts (mean 002)however the small range of overlapping initial eHf valuesfor the two lava suites suggests that these differences devel-oped during melting within the plume (Fig 9b) Below wemodel the effects of source mineralogy and depth of
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
28
melting on differences in trace-element concentrations inthe tholeiitic basalts and picritesDynamic melting models simulate progressive decom-
pression melting where some of the melt fraction isretained in the residue and only when the degree of partialmelting is greater than the critical mass porosity and the
source becomes permeable is excess melt extracted fromthe residue (Zou1998) For the Karmutsen lavas evolutionof trace element concentrations was simulated using theincongruent dynamic melting model developed by Zou ampReid (2001) Melting of the mantle at pressures greater than1 GPa involves incongruent melting (eg Longhi 2002)
OIB array
PacificMORB(230 Ma)
(0 Ma)
EPR(230 Ma)
-4
-2
0
2
4
8
10
12
14
16
18
20
22
-4 -2 0 2 4 6 8 10 12 14
minus2
0
2
4
6
8
10
12
14
0702 0703 0704 0705 0706
IndianMORB
87Sr 86Sr (initial)
Hf (initial)
(a) (c)
Nd (initial)
Karmutsen tholeiitic basalt
HawaiiOntong JavaCaribbean Plateau
Karmutsen high-MgO lava
high-MgO lavasKarmutsen
1540
1545
1550
1555
1560
1565
175 180 185 190 195 200375
380
385
390
395
175 180 185 190 195 200
(d) (e)
EPR
EPR
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb
Karmutsen Formation (initial)
HawaiiOntong JavaCaribbean Plateau
East Pacific Rise
Karmutsen Formation (measured)
Juan de FucaGorda
ε Nd
(initial)
Karmutsen Formation (different age-correction for 3 picrites)
(0 Ma)
NbY
001
01
1
1 10
ZrY
OIB
NMORB
(b)
Iceland array
6
Fig 13 Comparison of age-corrected (230 Ma) Sr^Nd^Hf^Pb isotopic compositions for Karmutsen flood basalts onVancouver Island withage-corrected OIB and MORB and Nb^Zr^Y variation (a) Initial eNd vs 87Sr86Sr (b) NbYand ZrY variation (after Fitton et al 1997)(c) Initial eHf vs eNd (d) Measured and initial 207Pb204Pb vs 206Pb204Pb (e) Measured and initial 208Pb204Pb vs 206Pb204Pb References fordata sources are listed in Electronic Appendix 4 Most of the compiled data were extracted from the GEOROC database (httpgeorocmpch-mainzgwdgdegeoroc) OIB array line in (c) is fromVervoort et al (1999) EPR East Pacific Rise Several high-MgO samples affected bysecondary alteration were not plotted for clarity in Pb isotope plots Estimation of fields for age-corrected Pacific MORB and EPR at 230 Mawere made using parent^daughter concentrations (in ppm) from Salters amp Stracke (2004) of Lufrac14 0063 Hffrac14 0202 Smfrac14 027 Ndfrac14 071Rbfrac14 0086 and Srfrac14 836 In the NbYvs ZrYplot the normal (N)-MORB and OIB fields not including Icelandic OIB are from data com-pilations of Fitton et al (1997 2003) The OIB field is data from 780 basalts (MgO45wt ) from major ocean islands given by Fitton et al(2003) The N-MORB field is based on data from the Reykjanes Ridge EPR and Southwest Indian Ridge from Fitton et al (1997) The twoparallel gray lines in (b) (Iceland array) are the limits of data for Icelandic rocks
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
29
with olivine being produced during the reactioncpxthornopxthorn spfrac14meltthornol for spinel peridotites The mod-eling used coefficients for incongruent melting reactionsbased on experiments on lherzolite melting (eg Kinzleramp Grove 1992 Longhi 2002) and was separated into (1)melting of spinel lherzolite (2) two combined intervals ofmelting for a garnet lherzolite (gt lherz) and spinel lher-zolite (sp lherz) source (3) melting of garnet lherzoliteThe model solutions are non-unique and a range indegree of melting partition coefficients melt reactioncoefficients and proportion of mixed source componentsis possible to achieve acceptable solutions (see Fig 14 cap-tion for description of modeling parameters anduncertainties)The LREE-depleted high-MgO lavas require melting of
a depleted spinel lherzolite source (Fig 14a) The modelingresults indicate a high degree of melting (22^25) similarto the results from PRIMELT1 based on major-elementvariations (discussed above) of a LREE-depleted sourceThe high-MgO lavas lack a residual garnet signature Theenriched tholeiitic basalts involved melting of both garnetand spinel lherzolite and represent aggregate melts pro-duced from continuous melting throughout the depth ofthe melting column (Fig 14b) Trace-element concentra-tions in the tholeiitic and picritic lavas are not consistentwith low-degree melting of garnet lherzolite alone(Fig 14c) The modeling results indicate that the aggregatemelts that formed the tholeiitic basalts would haveinvolved lesser proportions of enriched small-degreemelts (1^6 melting) generated at depth and greateramounts of depleted high-degree melts (12^25) gener-ated at lower pressure (a ratio of garnet lherzolite tospinel lherzolite melt of 14 provides the best fits seeFig 14b and description in figure caption) The totaldegree of melting for the tholeiitic basalts was high(23^27) similar to the degree of partial melting in thespinel lherzolite facies for the primary melts that eruptedas the Keogh Lake picrites of a source that was also prob-ably depleted in LREE
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
(c)
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
Melting of garnet lherzolite
High-MgO lavas
Melting of spinel lherzolite
source (07DM+03PM)
source (07DM+03PM)
6 melting
30 melting
(b)
(a)
30 melting
Sam
ple
Cho
ndrit
eS
ampl
eC
hond
rite 20 melting
1 melting
Melting of garnet lherzoliteand spinel lherzolite
x gt lherz + 4x sp lherz
5 meltingx gt lherz + 4x sp lherz
Sam
ple
Cho
ndrit
e
Tholeiitic lavas
Tholeiitic lavas
High-MgO lavas
source (07DM+03PM)
Fig 14 Trace-element modeling results for incongruent dynamicmantle melting for picritic and tholeiitic lavas from the KarmutsenFormation Three steps of modeling are shown (a) melting of spinellherzolite compared with high-MgO lavas (b) melting of garnet lher-zolite and spinel lherzolite compared with tholeiitic lavas (c) meltingof garnet lherzolite compared with tholeiitic and picritic lavasShaded fields in each panel for tholeiitic basalts and picrites arelabeled Patterns with symbols in each panel are modeling results(patterns are 5 melting increments in (a) and (b) and 1 incre-ments in (c) PM primitive mantle DM depleted MORB mantle gtlherz garnet lherzolite sp lherz spinel lherzolite In (b) the ratios ofper cent melting for garnet and spinel lherzolite are indicated (x gtlherzthorn 4x sp lherz means that for 5 melting 1 melt in garnetfacies and 4 melt in spinel facies where x is per cent melting in theinterval of modeling) The ratio of the respective melt proportions inthe aggregate melt (x gt lherzthorn 4x sp lherz) was kept constant andthis ratio of melt contribution provided the best fit to the data Arange of proportion of source components (07DMthorn 03PM to
09DMthorn 01PM) can reproduce the variation in the high-MgOlavas Melting modeling uses the formulation of Zou amp Reid (2001)an example calculation is shown in their Appendix PM fromMcDonough amp Sun (1995) and DM from Salters amp Stracke (2004)Melt reaction coefficients of spinel lherzolite from Kinzler amp Grove(1992) and garnet lherzolite fromWalter (1998) Partition coefficientsfrom Salters amp Stracke (2004) and Shaw (2000) were kept constantSource mineralogy for spinel lherzolite (018cpx027opx052ol003sp) from Kinzler (1997) and for garnet lherzolite(034cpx008opx053ol005gt) from Salters amp Stracke (2004) Arange of source compositions for melting of spinel lherzolite(07DMthorn 03PM to 03DMthorn 07PM) reproduce REE patternssimilar to those of the high-MgO lavas with best fits for 22ndash25melting and 07DMthorn 03PM source composition Concentrationsof tholeiitic basalts cannot be reproduced from an entirelyprimitive or depleted source
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
30
The uncertainty in the composition of the source in thespinel lherzolite melting model for the high-MgO lavasprecludes determining whether the source was moredepleted in incompatible trace elements than the source ofthe tholeiitic lavas (see Fig 14 caption for description) It ispossible that early high-pressure low-degree melting left aresidue within parts of the plume (possibly the hotter inte-rior portion) that was depleted in trace elements (egElliott et al 1991) further decompression and high-degreemelting of such depleted regions could then have generatedthe depleted high-MgO Karmutsen lavas Shallow high-degree melting would preferentially sample depletedregions whereas deeper low-degree melting may notsample more refractory depleted regions (eg CaribbeanEscuder-Viruete et al 2007) The picrites lie at the highend of the range of Nd isotopic compositions and havelow NbZr similar to depleted Icelandic basalts (Fittonet al 2003) therefore the picrites may represent the partsof the mantle plume that were the most depleted prior tothe initiation of melting As Fitton et al (2003) suggestedthese depleted melts may only rarely be erupted withoutcombining with more voluminous less depleted melts andthey may be detected only as undifferentiated small-volume flows like the Keogh Lake picrites within theKarmutsen Formation on Vancouver IslandDecompression melting within the mantle plume initiatedwithin the garnet stability field (35^25 GPa) at highmantle potential temperatures (Tp414508C) and pro-ceeded beneath oceanic arc lithosphere within the spinelstability field (25^09 GPa) where more extensivedegrees of melting could occur
Magmatic evolution of Karmutsentholeiitic basaltsPrimary magmas in large igneous provinces leave exten-sive coarse-grained cumulate residues within or below thecrust these mafic and ultramafic plutonic sequences repre-sent a significant proportion of the original magma thatpartially crystallized at depth (eg Farnetani et al 1996)For example seismic and petrological studies of OntongJava (eg Farnetani et al 1996) combined with the use ofMELTS (Ghiorso amp Sack 1995) indicate that the volcanicsequence may represent only 40 of the total magmareaching the Moho Neal et al (1997) estimated that30^45 fractional crystallization took place for theOntong Java magmas and produced cumulate thicknessesof 7^14 km corresponding to 11^22 km of flood basaltsdepending on the presence of pre-existing oceanic crustand the total crustal thickness (25^35 km) Interpretationsof seismic velocity measurements suggest the presence ofpyroxene and gabbroic cumulates 9^16 km thick beneaththe Ontong Java Plateau (eg Hussong et al 1979)Wrangellia is characterized by high crustal velocities in
seismic refraction lines which is indicative of mafic
plutonic rocks extending to depth beneath VancouverIsland (Clowes et al 1995)Wrangellia crust is 25^30 kmthick beneath Vancouver Island and is underlain by astrongly reflective zone of high velocity and density thathas been interpreted as a major shear zone where lowerWrangellia lithosphere was detached (Clowes et al 1995)The vast majority of the Karmutsen flows are evolved tho-leiitic lavas (eg low MgO high FeOMgO) indicating animportant role for partial crystallization of melts at lowpressure and a significant portion of the crust beneathVancouver Island has seismic properties that are consistentwith crystalline residues from partially crystallizedKarmutsen magmas To test the proportion of the primarymagma that fractionated within the crust MELTS(Ghiorso amp Sack 1995) was used to simulate fractionalcrystallization using several estimated primary magmacompositions from the PRIMELT1 modeling results Themajor-element composition of the primary magmas couldnot be estimated for the tholeiitic basalts because of exten-sive plagthorn cpxthornol fractionation and thus the MELTSmodeling of the picrites is used as a proxy for the evolutionof major elements in the volcanic sequence A pressure of 1kbar was used with variable water contents (anhydrousand 02wt H2O) calculated at the quartz^fayalite^magnetite (QFM) oxygen buffer Experimental resultsfrom Ontong Java indicate that most crystallizationoccurs in magma chambers shallower than 6 km deep(52 kbar Sano amp Yamashita 2004) however some crys-tallization may take place at greater depths (3^5 kbarFarnetani et al 1996)A selection of the MELTS results are shown in Fig 15
Olivine and spinel crystallize at high temperature until15^20wt of the liquid mass has fractionated and theresidual magma contains 9^10wt MgO (Fig 15f) At1235^12258C plagioclase begins crystallizing and between1235 and 11908C olivine ceases crystallizing and clinopyr-oxene saturates (Fig 15f) As expected the addition of clin-opyroxene and plagioclase to the crystallization sequencecauses substantial changes in the composition of the resid-ual liquid FeO and TiO2 increase and Al2O3 and CaOdecrease while the decrease in MgO lessens considerably(Fig 15) The near-vertical trends for the Karmutsen tho-leiitic basalts especially for CaO do not match the calcu-lated liquid-line-of-descent very well which misses manyof the high-CaO high-Al2O3 low-FeO basalts A positivecorrelation between Al2O3CaO and SrNd indicates thataccumulation of plagioclase played a major role in the evo-lution of the tholeiitic basalts (Fig 15g) Increasing thecrystallization pressure slightly and the water content ofthe melts does not systematically improve the match ofthe MELTS models to the observed dataThe MELTS results indicate that a significant propor-
tion of the crystallization takes place before plagioclase(15^20 crystallization) and clinopyroxene (35^45
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
31
crystallization) join the crystallization sequence Theseresults suggest that445 crystallization of the primarymagmas occurred (estimated from the residual liquid percent in Fig 15f) and if the original flood basalt stratigra-phy on Vancouver Island was 6 km thick at least anequivalent amount of complementary plutonic rocksshould have crystallized The Karmutsen basalts and theirplutonic residues thus represent a significant addition ofcrust (perhaps 412 km thickness) previously thickenedduring the formation of the underlying Paleozoic arc
Growth of the Wrangellia oceanic plateauon Vancouver IslandThe extrusive part of the Wrangellia oceanic plateau onVancouver Island was constructed in a three-layersequence of mostly tholeiitic basalts with a restrictedrange of composition (Fig 16) The plateau overliesDevonian to Mississippian arc volcanic sequences andmarine sediments of Mississippian to Permian ages thatare intruded by gabbroic rocks related to the flood basaltsThe volcanic stratigraphy of theWrangellia plateau formed
00
05
10
15
20
25
30
35
40
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
16
0 2 4 6 8 10 12 14 16 18 20
10
11
12
13
14
15
16
17
18
19
20
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
0 2 4 6 8 10 12 14 16 18 20
MgO (wt)
0 10 20 30 40 50
percent of total mass
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
) Mineral proportions
Sp (e)
Ol
CpxPlag
PicriteHigh-MgO basaltTholeiitic basalt
TiO2 (wt)
MgO (wt)
Al2O3 (wt) CaO (wt)
FeO (wt)
MgO(a) (b)
(c) (d)
MgO (wt)MgO (wt)
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
)
Residual liquid ()
1220
1190
Ol + Sp
Plag+Ol+Sp Plag+Cpx+Sp
02 wt H2O
no H2O
Residual liquid and crystallizing phases
(f)
0102030405060708090100
plagioclaseaccumulation
plagioclaseaccumulation
plagioclase
accumulatio
n
10
12
14
16
0 10 20 30 40
Al2O3CaO
SrNd
tholeiitic basalts
(g)
Fig 15 Forward fractional crystallization modeling results for major elements from MELTS (Ghiorso amp Sack 1995) compared with picriticand tholeiitic lavas from the Karmutsen Formation (a^d) Major-element compositions of residual liquids compared with Karmutsen lavacompositions For clarity the results are shown only for one starting composition (open circle) using the composition of the estimated primarymagma for the high-MgO lava series from the modeling technique of Herzberg et al (2007) for sample 4723A13 as shown in Fig 12 The esti-mated primary magmas for samples 4723A3 and 93G171 yield similar results A pressure of 1 kbar was used with no H2O and 02wt H2Ocalculated at QFM black line is anhydrous and grey line is 02wt H2OThe end-point for the anhydrous result is at 78 fractionation and11308C and the end-point for the result with 02wt H2O is 75 fractionation and11108C (e) Mineral proportions of crystallized phases fromone MELTrun (sample 4723A13) with no H2O (f) Per cent residual liquid vs temperature labelled with intervals of crystallizing phases (g)Al2O3CaO vs SrNd for the tholeiitic basalts
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
32
in a single eruptive phase (c 231^225 Ma) as an emergentbasalt sequence during a deeper-water shallow-water andsubaerial stage and subsided below sea level relativelysoon after its formation (Fig 16) The deeper-water stagecontains pillowed and massive submarine flows (3000mthick) that built up the submarine volcanic edifice from theseafloor The lowest pillow basalts were emplaced onunconsolidated and lithified fine-grained siliciclastic andcarbonaceous sediments containing Daonella beds(c 235^232 Ma) that are intruded by abundantKarmutsen mafic sills (Fig 16a) The different pillowedand massive submarine flows resulted primarily from dif-ferent effusive rates and local topography Some of themassive submarine flows commonly recognized by radiat-ing columnar jointing represent master tubes for the deliv-ery of lava to distal parts of submarine flow fields Thepillowed flows which commonly have uneven topography
in modern sequences may have served as levees to themassive submarine flows rather than allowing unob-structed spreading and inflation typical of the massive sub-aerial flows that form on shallow slopes in continentalflood basalt provinces (eg Self et al 1997) The rapid suc-cession of flows is evident by the near absence of sedimentbetween submarine flows Minor volumes of picritic pillowlavas occur in the upper part of the pillowed flow sequencein areas of the plateau exposed on northern VancouverIsland (Fig 16 Nixon et al 2008)The growth of the plateau in shallow water led to
emplacement of an increasing proportion of volcaniclastics(pillow breccias and hyaloclastites) that accumulated up to1500m thick in some areas (Fig 16b Nixon et al 2008)Horizons of pillowed flows occur locally within the volca-niclastic stratigraphy Sedimentary structures (eg gradedbedding fluidization structures) and petrographic textures
Late Devonian-Mississippian arc
Daonella beds
Middle Triassic
deeper water
Late Devonian-Mississippian arc
pillowed flows massive submarine
flows
pillow brecciahyaloclastiteshallow water
~4 km
~6 km
Late Triassic limestoneinterflow limestonelenses
Karmutsen Formation
mafic intrusions sediment-sill complex
Karmutsen Formation
(a) (b)
(c) (d)
subaerial flows
Late Devonian-Mississippian arc Late Devonian-Mississippian arc
picritic pillow basalts
Miss to Permian marine sediments
sills
Fig 16 Growth of the Wrangellia oceanic plateau onVancouver Island (a) Deep-water eruptive phase of pillowed and massive submarineflows with almost no intervening sediments onto MiddleTriassic shale containing Daonella beds (c 235^232 Ma) The basement of the plateauis composed of Devonian to Mississippian arc sequences overlain by Mississippian to Permian carbonates and siliciclastic sediments(b) Shallow-water flow emplacement with increasing proportion of volcaniclastics resulted primarily from cooling^contraction granulationPicritic pillow basalts occur primarily near the transition between pillowed submarine flows and volcaniclastic units Pillowed and massiveflows occur locally within the volcaniclastic unit (c) Emergent stage of growth consists of overlapping subaerial flow fields similar to continen-tal flood basalts with pahoehoe flow structures Multi-tiered columnar jointing occurrs locally in response to ponding of flows (eg Bondre et al2004) Interflow carbonate deposits and plagioclase megacrystic flows developed locally during the waning stage of volcanism (d) Post-volcanicsubsidence and deposition of Late Carnian to Norian (c 212^223 Ma) platform carbonates on top of the plateau Units are labeled in the dia-grams Thicknesses of the underlying units are not to scale
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
33
within the volcaniclastic rocks indicate formation primar-ily via cooling^contraction granulation magma^water^steam interaction autobrecciation resedimentation andmass-wasting rather than pyroclastic eruption (Nixonet al 2008)A broad subaerial platform was constructed by emplace-
ment of subaerial lava flows during an emergent subaerialstage (Fig 16c) The subaerial unit predominantly com-prises well-layered sheet flows that formed in a similarway to continental flood basalts as hundred(s) of pahoehoeflow fields erupted on shallow slopes (Self et al 1997) how-ever volcaniclastic and pillowed flow units are locally pre-sent mostly in the upper parts of the stratigraphyLocalized ponding of flows led to the formation of multi-tiered columnar jointing in some subaerial flows (egBondre et al 2004) As volcanism waned local interflowcarbonate deposits developed and plagioclase megacrysticflows were erupted Subsidence of the plateau is recordedby the interflow lenses and deposits of platform carbonates(40^1500m thick) on top of the Karmutsen basalts thatcontain Late Carnian to Early Norian fossils (Fig 16d c216^227 Ma Nixon et al 2008)The evolution of the internal architecture of the
Wrangellia oceanic plateau was primarily related to theconditions of emplacement (ie water depth and eruptionenvironment) There does not appear to have been a sys-tematic coincident evolution or shift of lava compositionsonVancouver Island as has been recognized in most con-tinental flood basalts and parts of the Wrangellia oceanicplateau in Alaska and Yukon (Greene et al 2008 2009)In general with the exception of the picritic pillow lavasin the submarine stratigraphy and local variations betweenhigh-MgO and tholeiitic lavas there are no significantcompositional differences between lavas erupted duringthe different stages of growth of the plateau (ie submarineand subaerial flows) nor are substantial compositional dif-ferences generally found between massive submarine flowsin contact with pillowed flows tholeiitic basalt flows nearthe base of the Karmutsen Formation are compositionallysimilar to flows sampled near the top of the volcanicsequence Where it was possible to sample continuous sec-tions of volcanic stratigraphy greater than several hun-dreds of meters (eg Keogh^Maynard Lake area west ofthe Karmutsen Range Fig 2) there is no clear relationshipbetween stratigraphic position and compositionalvariation
CONCLUSIONSThe Karmutsen Formation covers extensive areas ofVancouver Island (20 000 km2) and forms part of a largeoceanic plateau erupted over a geologically short interval(c 225^231 Ma) The tripartite volcanic stratigraphyexposed on Vancouver Island is at least 6 km thick andcomprises a succession of submarine flows volcaniclastic
deposits and massive subaerial flowsThese volcanologicaldifferences are primarily related to the eruption environ-ment (deep-water shallow-water or subaerial) The rapidgrowth of the plateau prevented significant accumulationsof intervening sediment except in the uppermost stratigra-phy where isolated limestone lenses commonly associatedwith pillow basalt and hyaloclastite deposits preserve arecord of the final stages of subsidence of the plateauThe Wrangellia plateau on Vancouver Island was con-
structed dominantly of tholeiitic basalt with restrictedmajor- and trace-element and isotopic compositional var-iations Minor volumes of picritic pillow basalts eruptedlate in the submarine phase The Karmutsen basalts wereemplaced onto older Paleozoic arc volcanic sequences andlow NbLa for the picrites may indicate a contributionfrom the sub-arc lithospheric mantle Modeling of the pet-rogenesis of the Karmutsen picrites indicates melting ofanomalously hot mantle (15008C) and extensive degreesof partial melting (23^27) and is consistent with aplume initiation model Combined Sr^Nd^Hf^Pb isotopicsystematics indicate that the picritic and tholeiitic lavaswere derived from a depleted plume-type mantle sourcedistinct from the source of MORB that has compositionalsimilarities to the source of the lavas forming theCaribbean Plateau Picrites were generated by melting ofspinel lherzolite whereas the tholeiitic basalts formed asaggregate melts of both garnet and spinel lherzolite Thetholeiitic basalts underwent extensive low-pressure fractio-nation (52^3 kbar) and seismic studies indicate that someof theWrangellia crust beneathVancouver Island may cor-respond to the plutonic residues of the Karmutsen floodbasalts
ACKNOWLEDGEMENTSWe are grateful to Nick Arndt for helping us get this pro-ject started and Nick Massey for insights into VancouverIsland geology We greatly appreciate the assistance withPRIMELT modeling from Claude Herzberg We thankMikkel Schau for his insight and enthusiasm during field-work Jane Barling assisted with analyses by MC-ICP-MSThoughtful and thorough reviews of the manuscript byJohn Mahoney Fred Frey Wendy Bohrson DougalJerram and Teal Riley are greatly appreciated Fundingwas generously provided by the Rocks to Riches Programadministered by the BC amp Yukon Chamber of Mines in2004 by the BC Geological Survey and by NSERCDiscovery Grants to JSS and DW ARG was supportedby a University Graduate Fellowship at UBC
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
34
REFERENCESBarker F Brown A S Budahn J R amp Plafker G (1989) Back-arc
with frontal-arc component origin of Triassic Karmutsen basaltBritish Columbia Canada Chemical Geology 75 81^102
Blichert-Toft JWeis D Maerschalk C Agranier A amp Albare de F(2003) Hawaiian hot spot dynamics as inferred from the Hf and Pbisotope evolution of Mauna Kea volcano Geochemistry GeophysicsGeosystems 4(2) 1^27 doi1010292002GC000340
Bondre N R Duraiswami R A amp Dole G (2004) Morphologyand emplacement of flows from the Deccan Volcanic ProvinceIndia Bulletin of Volcanology 66 29^45
Carlisle D (1963) Pillow breccias and their aquagene tuffs QuadraIsland British Columbia Journal of Geology 71 48^71
Carlisle D (1972) Late Paleozoic to Mid-Triassic sedimentary-volcanic sequence on Northeastern Vancouver Island Geological
Society of Canada Paper Report of Activities 72-1 Part B 24^30Carlisle D amp SuzukiT (1974) Emergent basalt and submergent car-
bonate^clastic sequences including the Upper Triassic Dilleri andWelleri zones on Vancouver Island Canadian Journal of Earth
Sciences 11 254^279Chauvel C amp Blichert-Toft J (2001) A hafnium isotope and trace
element perspective on melting of the depleted mantle Earth and
Planetary Science Letters 190 137^151Chayes F (1954)The theory of thin-section analysis Journal of Geology
62 92^101Cho M Liou J G amp Maruyama S (1987) Transition from the zeo-
lite to prehnite^pumpellyite facies in the Karmutsen metabasitesVancouver Island British Columbia Journal of Petrology 27(2)467^494
Clowes R M Zelt C A Amor J R amp Ellis R M (1995)Lithospheric structure in the southern Canadian Cordillera froma network of seismic refraction lines Canadian Journal of Earth
Sciences 32(10) 1485^1513Coffin M F amp Eldholm O (1994) Large igneous provinces Crustal
structure dimensions and external consequences Reviews of
Geophysics 32(1) 1^36Cox K G (1980) A model for flood basalt volcanism Journal of
Petrology 21 629^650Eldholm O amp Coffin M F (2000) Large igneous provinces
and plate tectonics In Richards M A Gordon R G amp vander Hilst R D (eds) The History and Dynamics of Global
Plate Motions American Geophysical Union Geophysical Monograph 121309^326
Elliott T R Hawkesworth C J amp Grolaquo nvold K (1991) Dynamicmelting of the Iceland plume Nature 351 201^206
Escuder-Viruete J Pecurren rez-Estaucurren n A Contreras F Joubert MWeis D Ullrich T D amp Spadea P (2007) Plume mantle sourceheterogeneity through time Insights from the Duarte ComplexHispaniola northeastern Caribbean Journal of Geophysical Research112(B04203) doi1010292006JB004323
Farnetani C G Richards M A amp Ghiorso M S (1996)Petrological models of magma evolution and deep crustal structurebeneath hotspots and flood basalt provinces Earth and Planetary
Science Letters 143 81^94Fitton J G amp Godard M (2004) Origin and evolution of magmas
on the Ontong Java Plateau In Fitton J G Mahoney J JWallace P J amp Saunders A D (eds) Origin and Evolution of the
Ontong Java Plateau Geological Society London Special Publications 229151^178
Fitton J G Saunders A D Norry M J Hardarson B S ampTaylor R N (1997) Thermal and chemical structure of theIceland plume Earth and Planetary Science Letters 153 197^208
Fitton J G Saunders A D Kempton P D amp Hardarson B S(2003) Does depleted mantle form an intrinsic part of the Icelandplume Geochemistry Geophysics Geosystems 4(3) 1032 doi1010292002GC000424
Frey F A Huang S Blichert-Toft J Regelous M amp Boyet M(2005) Origin of depleted components in basalt related to theHawaiian hotspot Evidence from isotopic and incompatible ele-ment ratios Geochemistry Geophysics Geosystems 6(1) 23 doi1010292004GC000757
Galer S J G amp Abouchami W (1998) Practical application of leadtriple spiking for correction of instrumental mass discriminationMineralogical Magazine 62A 491^492
Ghiorso M S amp Sack R O (1995) Chemical mass transfer in mag-matic processes IV A revised and internally consistent thermody-namic model for the interpolation and extrapolation of liquid^solidequilibria in magmatic systems at elevated temperatures and pres-sures Contributions to Mineralogy and Petrology 119 197^212
Greene A R Scoates J S amp Weis D (2005)WrangelliaTerrane onVancouver Island Distribution of flood basalts with implicationsfor potential Ni^Cu^PGE mineralization In Grant B (ed)Geological Fieldwork 2004 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2005-1 209^220Greene A R Scoates J S Nixon G T amp Weis D (2006) Picritic
lavas and basal sills in the Karmutsen flood basalt provinceWrangellia northern Vancouver Island In Grant B (ed)Geological Fieldwork 2005 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2006-1 39^52Greene A R Scoates J S amp Weis D (2008) Wrangellia flood
basalts in Alaska A record of plume^lithosphere interaction in aLate Triassic accreted oceanic plateau Geochemistry Geophysics
Geosystems 9(Q12004) doi1010292008GC002092Greene A R Scoates J S Weis D amp Israel S (2009)
Geochemistry of triassic flood basalts from the Yukon (Canada)segment of the accreted Wrangellia oceanic plateau Lithosdoi101016jlithos200811010
Hauff F Hoernle K Tilton G Graham D W amp Kerr A C(2000) Large volume recycling of oceanic lithosphere over shorttime scales geochemical constraints from the Caribbean LargeIgneous Province Earth and Planetary Science Letters 174 247^263
Hauff F Hoernle K amp Schmidt A (2003) Sr^Nd^Pb compositionof Mesozoic Pacific oceanic crust (Site 1149 and 801 ODP Leg185)Implications for alteration of ocean crust and the input into theIzu^Bonin^Mariana subduction system Geochemistry Geophysics
Geosystems 4(8) doi1010292002GC000421Herzberg C (2004) Partial melting below the Ontong Java Plateau
In Fitton J G Mahoney J J Wallace P J amp Saunders A D(eds) Origin and Evolution of the Ontong Java Plateau Geological SocietyLondon Special Publications 229 179^183
Herzberg C (2006) Petrology and thermal structure of the Hawaiianplume from Mauna Kea volcano Nature 444(7119) 605
Herzberg C amp Asimow P D (2008) Petrology of some oceanicisland basalts PRIMELT2XLS software for primary magma cal-culation Geochemistry Geophysics Geosystems 9(Q09001) doi1010292008GC002057
Herzberg C amp OrsquoHara M J (2002) Plume-associated ultramaficmagmas of Phanerozoic age Journal of Petrology 43(10) 1857^1883
Herzberg C Asimow P D Arndt N Niu Y Lesher C MFitton J G Cheadle M J amp Saunders A D (2007)Temperatures in ambient mantle and plumes Constraints frombasalts picrites and komatiites Geochemistry Geophysics Geosystems8(Q02006) doi1010292006GC001390
Huang S amp Frey F A (2005) Trace element abundances of MaunaKea basalt from phase 2 of the Hawaii Scientific Drilling Project
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
35
Petrogenetic implications of correlations with major element con-tent and isotopic ratios Geochemistry Geophysics Geosystems 4(6)doi1010292002GC000322
Hussong D M Wipperman L K amp Kroenke L W (1979) Thecrustal structure of the Ontong Java and Manihiki OceanicPlateaus Journal of Geophysical Research 84(B11) 6003^6010
Jerram D A amp Widdowson M (2005) The anatomy of ContinentalFlood Basalt Provinces geological constraints on the processes andproducts of flood volcanism Lithos 79(3^4) 385^405
Jones D L Silberling N J amp Hillhouse J (1977)Wrangellia a dis-placed terrane in northwestern North America CanadianJournal ofEarth Sciences 14(11) 2565^2577
Kerr A C (2003) Oceanic Plateaus In Rudnick R (ed) The Crust
Treatise on GeochemistryVol 3 Oxford Elsevier Science pp 537^565Kerr A C amp Mahoney J J (2007) Oceanic plateaus Problematic
plumes potential paradigms Chemical Geology 241 332^353Kerr A CTarney J Marriner G F Klaver GT Saunders A D
amp Thirlwall M F (1996) The geochemistry and petrogenesis ofthe Late-Cretaceous picrites and basalts of Curacao NetherlandsAntilles a remnant of an oceanic plateau Contributions to
Mineralogy and Petrology 124(1) 29^43Kerr A C Marriner G F Tarney J Nivia A Saunders A D
Thirlwall M F amp Sinton A C (1997) Cretaceous basaltic ter-ranes in Western Colombia Elemental chronological and Sr-Ndisotopic constraints on petrogenesis Journal of Petrology 38(6)677^702
Kinzler R J (1997) Melting of mantle peridotite at pressuresapproaching the spinel to garnet transition Application to mid-ocean ridge basalt petrogenesis Journal of Geophysical Research
102(B1) 853^874Kinzler R J amp Grove T L (1992) Primary magmas of mid-ocean
ridge basalts 1 Experiments and methods Journal of Geophysical
Research 97(B5) 6885^6906Korenaga J (2005) Why did not the Ontong Java Plateau form sub-
aerially Earth and Planetary Science Letters 234 385^399Kuniyoshi S (1972) Petrology of the Karmutsen (Volcanic) Group
northeastern Vancouver Island British Columbia PhD disserta-tion University of California Los Angeles 242 pp
Lassiter J C DePaolo D J amp Mahoney J J (1995) Geochemistry ofthe Wrangellia flood basalt province Implications for the role ofcontinental and oceanic lithosphere in flood basalt genesis Journalof Petrology 36(4) 983^1009
LeBas M J (2000) IUGS reclassification of the high-Mg and picriticvolcanic rocks Journal of Petrology 41(10) 1467^1470
Longhi J (2002) Some phase equilibrium systematics of lherzolitemelting I Geochemistry Geophysics Geosystems 3(3) doi1010292001GC000204
MacDonald G A amp Katsura T (1964) Chemical composition ofHawaiian lavas Journal of Petrology 5 82^133
Mahoney J J (1987) An isotopic survey of Pacific oceanic plateausimplications for their nature and origin In Keating B HFryer P Batiza R amp Boehlert GW (eds) Seamounts Islands andAtolls American Geophysical Union Geophysical Monograph 43 207^220
Mahoney J LeRoex A P Peng Z Fisher R L amp Natland J H(1992) Southwestern limits of Indian Ocean Ridge mantle and theorigin of low 206Pb204Pb mid-ocean ridge basalt isotope systema-tics of the central Southwest Indian Ridge (178^508E) Journal ofGeophysical Research 97(B13) 19771^19790
Mahoney J J Sinton J M Kurz M D MacDougall J DSpencer K J amp Lugmair GW (1994) Isotope and trace elementcharacteristics of a super-fast spreading ridge East Pacific rise13^238S Earth and Planetary Science Letters 121 173^193
Massey N W D (1995a) Geology and mineral resources of theAlberni^Nanaimo Lakes sheet Vancouver Island 92F1W 92F2Eand part of 92F7E British Columbia Ministry of Energy Mines and
Petroleum Resources Paper 1992-2 132 ppMassey N W D (1995b) Geology and mineral resources of the
Cowichan Lake sheet Vancouver Island 92C16 British Columbia
Ministry of Energy Mines and Petroleum Resources Paper 1992-3 112 ppMassey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005a) Digital Geology Map of British Columbia Tile NM9 MidCoast BC British Columbia Ministry of Energy and Mines Geofile
2005-2Massey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005b) Digital Geology Map of British Columbia Tile NM10Southwest BC British Columbia Ministry of Energy and Mines Geofile
2005-3McDonough W F amp Sun S (1995) The composition of the Earth
Chemical Geology 120 223^253Monger JW H amp Journeay J M (1994) Basement geology and tec-
tonic evolution of theVancouver region In Monger JW H (ed)Geology and Geological Hazards of the Vancouver Region Southwestern
British Columbia Geological Survey of Canada Bulletin 481 3^25Muller J E (1977) Geology of Vancouver Island Geological Survey of
Canada Open File Map 463Muller J E (1980)The Paleozoic Sicker Group of Vancouver Island British
Columbia Geological Survey of Canada Paper 79-30 22 ppMuller J E Northcote K E amp Carlisle D (1974) Geology and
mineral deposits of Alert Bay^Cape Scott map area VancouverIsland British Columbia Geological Survey of Canada Paper 74-8 77
Neal C R Mahoney J J Kroenke L W Duncan R A ampPetterson M G (1997) The Ontong^Java Plateau InMahoney J J amp Coffin M F (eds) Large Igneous Provinces
Continental Oceanic and Planetary FloodVolcanism American Geophysical
Union Geophysical Monograph 100 183^216Niu Y Collerson K D Batiza R Wendt J I amp Regelous M
(1999) Origin of enriched-typed mid-ocean ridge basalt at ridgesfar from mantle plumes The East Pacific Rise at 11820rsquoN Journalof Geophysical Research 104 7067^7088
Nixon G T amp Orr A J (2007) Recent revisions to the EarlyMesozoic stratigraphy of Northern Vancouver Island (NTS 102I092L) and metallogenic implicatinos British Columbia InGrant B (ed) Geological Fieldwork 2006 British Columbia Ministry of
Energy Mines and Petroleum Resources Paper 2007-1 163^177Nixon GT Hammack J L KoyanagiV M Payie G J Haggart
JW Orchard M JTozerT Friedman R M Archibald D APalfy J amp Cordey F (2006a) Geology of the Quatsino^PortMcNeill area northernVancouver Island British Columbia Ministry
of Energy Mines and Petroleum Resources Geoscience Map 2006-2 scale150 000
Nixon G T Kelman M C Stevenson D Stokes L A ampJohnston K A (2006b) Preliminary geology of the Nimpkishmap area (NTS 092L07) northern Vancouver Island BritishColumbia In Grant B amp Newell J M (eds) Geological Fieldwork2005 British Columbia Ministry of Energy Mines and Petroleum Resources
Paper 2006-1 135^152Nixon G T Laroque J Pals A Styan J Greene A R amp
Scoates J S (2008) High-Mg lavas in the Karmutsen flood basaltsnorthernVancouver Island (NTS 092L) Stratigraphic setting andmetallogenic significance In Grant B (ed) Geological Fieldwork
2007 British Columbia Ministry of Energy Mines and Petroleum
Resources Paper 2008-1 175^190Nowell G M Kempton P D Noble S R Fitton J G
Saunders A Mahoney J J amp Taylor R N (1998) High precisionHf isotope measurements of MORB and OIB by thermal ionisation
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
36
mass spectrometry insights into the depleted mantle Chemical
Geology 149 211^233Parrish R R amp McNicoll V J (1992) U^Pb age determinations
from the southern Vancouver Island area British Columbia InRadiogenic Age and Isotopic Studies Report 5 Geological Survey of
Canada Paper 91-2 79^86Peate DW (1997) The Parana^Etendeka Province In Coffin M F
amp Mahoney J (eds) Large Igneous Provinces Continental Oceanic andPlanetary Flood Volcanism American Geophysical Union Geophysical
Monograph 100 217^245Perfit M R Fornari D J Ridley W I Kirk P D Casey J
Kastens K A Reynolds J R Edwards M Desonie DShuster R amp Paradis S (1996) Recent volcanism in the Siqueirostransform fault picritic basalts and implications for MORBmagma genesis Earth and Planetary Science Letters 141(1^4) 91^108
Pretorius W Weis D Williams G Hanano D Kieffer B ampScoates J S (2006) Complete trace elemental characterization ofgranitoid (USGSG-2GSP-2) reference materials by high resolutioninductively coupled plasma-mass spectrometry Geostandards and
Geoanalytical Research 30(1) 39^54Regelous M Niu Y Wendt J I Batiza R Greig A amp
Collerson K D (1999) Variations in the geochemistry of magma-tism on the East Pacific Rise at 10830rsquoN since 800 ka Earth and
Planetary Science Letters 168 45^63Recurren villon S Chauvel C Arndt N T Pik R Martineau F
Fourcade S amp Marty B (2002) Heterogeneity of the Caribbeanplateau mantle source Sr O and He isotopic compositions of oli-vine and clinopyroxene from Gorgona Island Earth and Planetary
Science Letters 205(1^2) 91Rhodes J M (1996) Geochemical stratigraphy of lava flows samples
by the Hawaii Scientific Drilling Project Journal of Geophysical
Research 101(B5) 11729^11746Rhodes J M amp Vollinger M J (2004) Composition of basaltic lavas
sampled by phase-2 of the Hawaii Scientific Drilling ProjectGeochemical stratigraphy and magma types Geochemistry
Geophysics Geosystems 5(3) doi1010292002GC000434Richards M A Jones D L Duncan R A amp DePaolo D J (1991)
A mantle plume initiation model for the Wrangellia flood basaltand other oceanic plateaus Science 254 263^267
Salters V J M amp Hart S R (1991) The mantle sources of oceanridges islands and arcs the Hf-isotope connection Earth and
Planetary Science Letters 104 364^380Salters V J M amp Stracke A (2004) Composition of the depleted
SaltersV J amp WhiteW (1998) Hf isotope constraints on mantle evo-lution Chemical Geology 145 447^460
SanoT amp Yamashita S (2004) Experimental petrology of basementlavas from Ocean Drilling Program Leg 192 implications for dif-ferentiation processes in Ontong Java Plateau magmas InFitton J G Mahoney J JWallace P J amp Saunders A D (eds)Origin and Evolution of the Ontong Java Plateau Geological Society
London Special Publications 229 185^218Saunders A D (2005) Large igneous provinces Origin and environ-
mental consequences Elements 1 259^263Self S ThordarsonT amp Keszthely L (1997) Emplacement of conti-
nental flood basalt lava flows In Mahoney J J amp Coffin M F(eds) Large Igneous Provinces Continental Oceanic and Planetary Flood
Volcanism American Geophysical Union Geophysical Monograph 100381^410
Shaw D M (2000) Continuous (dynamic) melting theory revisitedCanadian Mineralogist 38(5) 1041^1063
Sluggett C L (2003) Uranium^lead age and geochemical constraintson Paleozoic and Early Mesozoic magmatism in WrangelliaTerrane Saltspring Island British Columbia BSc thesisUniversity of British ColumbiaVancouver 84 pp
Storey M Mahoney J J amp Saunders A D (1997) Cretaceousbasalts in Madagascar and the transition between plume and con-tinental lithospheric mantle sources In Mahoney J J ampCoffin M F (eds) Large Igneous Provinces Continental Oceanic and
Planetary Flood Volcanism Aamerican Geophysical Union Geophysical
Monograph 100 95^122Surdam R C (1967) Low-grade metamorphism of the Karmutsen
Group PhD dissertation University of California Los Angeles288 pp
Sutherland-Brown A Yorath C J Anderson R G amp Dom K(1986) Geological maps of southern Vancouver IslandLITHOPROBE I 92C10 11 14 16 92F1 2 7 8 Geological Survey ofCanada Open File 1272
Tejada M L G Mahoney J J Castillo P R Ingle S P Sheth HC amp Weis D (2004) Pin-pricking the elephant evidence on theorigin of the Ontong Java Plateau from Pb^Sr^Hf^Nd isotopiccharacteristics of ODP Leg 192 basalts In Fitton J GMahoney J J Wallace P J amp Saunders A D (eds) Origin and
Evolution of the Ontong Java Plateau Geological Society London Special
Publications 229 133^150Thompson P M E Kempton P D amp Kerr A C (2008) Evaluation
of the effects of alteration and leaching on Sm^Nd and Lu^Hf sys-tematics in submarine mafic rocks Lithos 104(1^4) 164^176
Thompson P M E Kempton P D White R V Kerr A CTarney J Saunders A D Fitton J G amp McBirney A (2003)Hf-Nd isotope constraints on the origin of the CretaceousCaribbean plateau and its relationship to the Galapagos plumeEarth and Planetary Science Letters 217(1^2) 59^75
Vervoort J D Patchett P Blichert-Toft J amp Albare de F (1999)Relationships between Lu^Hf and Sm^Nd isotopic systems in theglobal sedimentary system Earth and Planetary Science Letters 16879^99
Walter M J (1998) Melting of garnet peridotite and the origin ofkomatiite and depleted lithosphere Journal of Petrology 39(1) 29^60
Weis D Kieffer B Maerschalk C Barling J de Jong JWilliams G A Hanano D Mattielli N Scoates J SGoolaerts A Friedman R A amp Mahoney J B (2006) High-pre-cision isotopic characterization of USGS reference materials byTIMS and MC-ICP-MS Geochemistry Geophysics Geosystems
7(Q08006) doi1010292006GC001283Weis D Kieffer B Hanano D Silva I N Barling J
Pretorius W Maerschalk C amp Mattielli N (2007) Hf isotopecompositions of US Geological Survey reference materialsGeochemistry Geophysics Geosystems 8(Q06006) doi1010292006GC001473
Wheeler J O amp McFeely P (1991) Tectonic assemblage map of theCanadian Cordillera and adjacent part of the United States ofAmerica Geological Survey of Canada Map 1712A
WhiteW M Albare de F amp Tecurren louk P (2000) High-precision analy-sis of Pb isotope ratios by multi-collector ICP-MS Chemical Geology167 257^270
Yorath C J Sutherland Brown A amp Massey N W D (1999)LITHOPROBE southern Vancouver Island British Columbia Geological
Survey of Canada Bulletin 498 145 ppZou H (1998) Trace element fractionation during modal and
non-model dynamic melting and open-system melting Amathematical treatment Geochimica et Cosmochimica Acta 62(11)1937^1945
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
37
Zou H amp Reid M R (2001) Quantitative modeling of trace elementfractionation during incongruent dynamic melting Geochimica et
Cosmochimica Acta 65(1) 153^162
APPENDIX SAMPLEPREPARAT ION ANDANALYT ICAL METHODSOnly the freshest rocks in the field were sampled and onlythe least altered samples were selected for chemical andisotopic analysis based on thorough petrographic inspec-tion Fifty-nine of the 129 collected samples were crushed(400 g) into small pieces 52mm in diameter in aRocklabs hydraulic piston crusher between WC plates tominimize contamination The coarse-crush was mixedand 100 g was powdered in a planetary mill using agatejars and balls cleaned with quartz sand between samples
ActLabs analytical methodsThe major- and trace-element compositions of the whole-rock powders were determined at Activation LaboratoriesLtd (Actlabs) in Ancaster Ontario Analytical techniquesand detection limits are also available from Actlabs (httpwwwactlabscommethsub_code4erehtm) The analyticalmethod for each of the elements analyzed is indicated inTable 2 For the major elements a 02 g sample was mixedwith a lithium metaborate^lithium tetraborate mixtureand fused in a graphite crucible The molten mixture waspoured into a 5 HNO3 solution and shaken until dis-solved (30min) The samples were analyzed for majoroxides and selected trace elements by inductively coupledplasma optical emission spectrometry (ICP-OES) on acombination simultaneous^sequential Thermo Jarrell-AshEnviro II system Internal calibration was achieved usinga variety of international reference materials (eg W-2BIR-1 DNC-1) and independent control samplesAdditional trace elements were analyzed by both instru-mental neutron activation analysis (INAA) and induc-tively couple plasma mass spectrometry (ICP-MS) ForINAA15^25 g of sample was weighed into small polyeth-ylene vials and irradiated with a control international ref-erence material CANMET WMS-1 and NiCr flux wires ata thermal neutron flux of 71012 neutronscm2 per s in theMcMaster Nuclear Reactor Following a 7 day waitingperiod the samples were measured on an Ortec high-purity Ge detector linked to a Canberra Series 95 multi-channel analyzer Activities for each element weredecay- and weight-corrected and compared with a detectorcalibration developed from multiple international certifiedreference materials For ICP-MS 025 g of sample wasdigested in HF followed by HNO3^HClO4 heated andtaken to dryness The samples were brought back into solu-tion with HCl Samples were analyzed using a PerkinElmer Optima 3000 ICP In-lab standards or certified ref-erence materials (eg W-2 BIR-1 DNC-1) were used for
quality control A total of 15 blind duplicates were ana-lyzed from the same sample powders to assessreproducibility
University of Massachusetts analyticalmethodsFourteen sample duplicate powders taken from the same100 g sample aliquot (high-MgO basalts and picrites)were also analyzed at the Ronald B Gilmore X-RayFluorescence (XRF) Laboratory at the University ofMassachusetts Major elements were measured on a fusedLa-bearing lithium borate glass disc using a SiemensMRS-400 spectrometer Trace element concentrations(Rb Sr Ba Ce Nb Zr Y Pb Zn Ga Ni Cr V) weremeasured on a separate powder pellet using a PhilipsPW2400 sequential spectrometer using a Rh tube LOIand ferrous iron measurements were made as describedby Rhodes amp Vollinger (2004) Precision and accuracy esti-mates for the data have been described by Rhodes (1996)and Rhodes amp Vollinger (2004) Results for each sampleare the average of two separate analyses and are shown inElectronic Appendix 2 for a comparison of analyses ofduplicate powders using XRFand ICP-MS
PCIGR analytical methodsA subset of 19 samples was selected for high-precisiontrace-element analysis and Sr Nd Pb and Hf isotopicanalysis at the Pacific Centre for Isotopic andGeochemical Research (PCIGR) at the University ofBritish Columbia (UBC) Samples were selected from the59 samples analyzed for whole-rock chemistry at ActLabsbased on major- and trace-element chemistry alteration(low LOI and petrographic alteration index) and samplelocation Samples were prepared for trace-element analysisat the PCIGR by the technique described by Pretorius et al(2006) on unleached rock powders Sample powders(100mg) were weighed in 7ml screw-top Savillex bea-kers and dissolved in 1ml 14N HNO3 and 5ml 48 HFon a hotplate for 48 h at 1308C with periodic ultrasonica-tion Samples were dried and redissolved in 6ml 6N HClon a hotplate for 24 h and then dried and redissolved in1ml concentrated HNO3 for 24 h before final dryingTrace element abundances were measured by high-resolution inductively coupled plasma-mass spectrometry(HR-ICP-MS) with aThermo Finnigan Element2 systemfollowing the procedures described by Pretorius et al(2006) within 24 h of redissolution HFSE and LILE weremeasured in medium-resolution mode at 2000 dilutionusing a PFATeflon spray chamber washed with aqua regiafor 3min between samples REE were measured in high-resolution mode and U and Pb in low-resolution modeat 2000 dilution using a glass spray chamber washedwith 2 HNO3 between samples Total procedural blanksand reference materials (BCR-2 BHVO-2) were analyzedwith the batch of samples Indium was used as an internal
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
38
standard in all samples and standard solutionsBackground and standard solutions were analyzed afterevery five samples to detect memory effects and massdrift Results for PCIGR trace-element analyses areshown in Electronic Appendix 3 for the full set of analyzedtrace elementsSample digestion for purification of Sr Nd Hf and Pb
for column chemistry began by weighing each samplepowder (400^500mg) prior to leaching All samples wereleached with 6N HCl and placed in an ultrasonic bath for15min Samples were rinsed twice with 18 M-cm H2Obetween each leaching step (15 total) until the supernatantliquid was clear [following the method of Mahoney (1987)]Samples were then dried on a hotplate for 24 h andweighed again Sample solutions were then prepared bydissolving 100^250mg of the leached powder dissolvedin 1ml 14N HNO3 and 10ml 48 HF on a hotplate for48 h at 1308C with periodic ultrasonication Samples weredried and redissolved in 6ml 6N HCl on a hotplate for24 h and then dried Pb was separated using anion exhangecolumns and the discard was used for Sr REE and Hfseparation Nd was separated from the other REE and Hfrequired two additional purification steps Detailed proce-dures for column chemistry for separating Sr Nd and Pbat the PCIGR have been described byWeis et al (2006) andHf purification has been described byWeis et al (2007) Srand Nd isotope ratios were measured by thermal ioniza-tion mass spectrometry (TIMS) on a Thermo FinniganTriton system in static mode with relay matrix rotation ona single Ta and double Re^Ta filament respectively Fourto five filaments per barrel of 21 were occupied by stan-dards (NBS 987 for Sr and LaJolla for Nd) for each barrelwhere samples were run Sample Sr and Nd isotopic com-positions were corrected for mass fractionation using86Sr88Srfrac14 01194 and 146Nd144Ndfrac14 07219 Each samplewas then normalized using the barrel average of the refer-ence material relative to the values of143Nd144Ndfrac14 0511858 and 87Sr86Srfrac14 0710248 (Weiset al 2006) During the course of the Vancouver Islandanalyses the La Jolla Nd standard gave an average valueof 05118566 (nfrac14 7) and NBS987 standard gave an aver-age of 07102408 (nfrac1411) (2 error is reported as times106) 147Sm144Nd errors are 15 or 0006 USGeological Survey (USGS) reference material BHVO-2was processed with the samples and yielded Sr and Nd iso-topic ratios of 07034607 and 05129786 respectivelyThese are in agreement with the published values of070347920 and 051298411 respectively (Weis et al2006)
Pb and Hf isotopic compositions were analyzed by mul-tiple collector-inductively coupled plasma-mass spectrome-try (MC-ICP-MS) on a static multi-collection on a NuPlasma (Nu Instruments) system The detailed analyticalprocedure for Pb isotopic analyses on the Nu Plasma atthe PCIGR has been described by Weis et al (2006) Theconfiguration for Pb analyses allows for collection of PbTl and Hg togetherTl and Hg are used to monitor instru-mental mass discrimination and isobaric overlap repec-tively All sample solutions were analyzed withapproximately the same PbTl ratio (4) as the referencematerial NIST SRM 981 To accomplish this a small ali-quot of each sample solution from the Pb columns was ana-lyzed on the Element2 to determine the precise amount ofPb available for analysis on the Nu Plasma The SRM 981standard was run after every two samples on the NuPlasma During the time samples were run analyses ofthe SRM 981 Pb reference material gave values of206Pb204Pbfrac1416940322 207Pb204Pbfrac1415495823and 208Pb204Pbfrac14 36713164 (nfrac14 61 2 error is reportedas times 104) these values are in excellent agreement withreported TIMS triple-spike values of Galer amp Abouchami(1998) Fractionation-corrected Pb isotopic ratios were fur-ther corrected by the sample-standard bracketing methodor the ln^ln correction method described by White et al(2000) and Blichert-Toft et al (2003) Leached powders ofUSGS reference material BHVO-2 yielded Pb isotopicratios of 206Pb204Pbfrac14186454 8 207Pb204Pbfrac14 1549105 and 208Pb204Pbfrac14 38222514 and BCR-2yielded 206Pb204Pbfrac141880466 207Pb204Pbfrac14 1562518 and 208Pb204Pbfrac14 3883496 (2 error isreported as times 104) These values are in agreement withthose for leached residues of BHVO-2 and BCR-2 fromWeis et al (2006)Hf isotopic compositions were analyzed following the
procedures detailed byWeis et al (2007) The configurationfor Hf analyses monitored Lu mass 175 andYb mass 172 toallow for interference correction to masses 174 and 176 Hfisotopic ratios were normalized internally for mass fractio-nation to a 179Hf177Hf ratio of 07325 using an exponentialcorrection Standards were run after every two samplesand sample results were normalized to the ratio of the in-run daily average and a 176Hf177Hf ratio for JMC-475 of0282160 During the course of analyses the Hf standardJMC-475 gave an average value 02821679 (nfrac14130)USGS reference materials BCR-2 and BHVO-2 were pro-cessed with the samples and yielded Hf isotopic ratios of02828675 and 02831005 respectively Publishedvalues for BCR-2 and BHVO-2 are 02828717 and0283104 8 respectively (Weis et al 2007)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
1Ni concentrations for these samples were determined by XRFTHOL tholeiitic basalt PIC picrite HI-MG high MgO basalt CG coarse-grained (sill or gabbro) MIN SIL mineralizedsill OUTLIER anomalous pillowed flow in plots MA Mount Arrowsmith SL Schoen Lake KR Karmutsen Range QIQuadra Island Sample locations are given using the Universal Transverse Mercator (UTM) coordinate system (NAD83zones 9 and 10) Analyses were performed at Activation Laboratory (ActLabs) Fe2O3
is total iron expressed as Fe2O3LOI loss on ignition All major elements Sr V and Y were measured by ICP quadrupole OES on solutions of fusedsamples Cu Ni Pb and Zn were measured by total dilution ICP Cs Ga Ge Hf Nb Rb Ta Th U Zr and REE weremeasured by magnetic-sector ICP on solutions of fused samples Co Cr and Sc were measured by INAA Blanks arebelow detection limit See Electronic Appendices 2 and 3 for complete XRF data and PCIGR trace element datarespectively Major elements for sample 93G17 are normalized anhydrous Sample 5615A7 was analyzed by XRFSamples from Quadra Island are not shown in figures
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
17
picritic and high-MgO basalt pillow lavas extend to20wt MgO (Fig 6 and references listed in thecaption)The tholeiitic basalts from the Karmutsen Range dis-
play a tight range of parallel light rare earth element(LREE)-enriched REE patterns (LaYbCNfrac14 21^24mean 2202) whereas the picrites and high-MgObasalts are characterized by sub-parallel LREE-depleted
patterns (LaYbCNfrac14 03^07 mean 06 02) with lowerREE abundances than the tholeiitic basalts (Fig 7)Tholeiitic basalts from the Schoen Lake and MountArrowsmith areas have similar REE patterns tosamples from the Karmutsen Range (Schoen Lake LaYbCNfrac14 21^26 mean 2303 Mount Arrowsmith LaYbCNfrac1419^25 mean 2204) and the coarse-grainedmafic rocks have similar REE patterns to the tholeiitic
Depleted MORB average(Salters amp Stracke 2004)
Schoen LakeSchoen Lake
Mount Arrowsmith
Karmutsen RangeS
ampl
eC
hond
rite
Sam
ple
Cho
ndrit
e
Sam
ple
Prim
itive
Man
tleS
ampl
eP
rimiti
ve M
antle
Sam
ple
Prim
itive
Man
tle
Mount Arrowsmith
Karmutsen Range
(a) (b)
(c) (d)
(e) (f )
Sam
ple
Cho
ndrit
e
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained mafic rock
Pillowed flowMineralized sill
1
10
100
1
10
100
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
01
1
10
100
1
10
100
1
10
100
Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
Fig 7 Whole-rock REE and other incompatible-element concentrations for the Karmutsen Formation (a) (c) and (e) are chondrite-normal-ized REE patterns for samples from each of the three main field areas onVancouver Island (b) (d) and (f) are primitive mantle-normalizedtrace-element patterns for samples from each of the three main field areas All normalization values are from McDonough amp Sun (1995) Theclear distinction between LREE-enriched tholeiitic basalts and LREE-depleted picrites the effects of alteration on the LILE (especially loss ofK and Rb) and the different range for the vertical scale in (b) (extends down to 01) should be noted
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
18
basalts (LaYbCNfrac14 21^29 mean 2508) Two LREE-depleted coarse-grained mafic rocks from the SchoenLake area have similar REE patterns (LaYbCNfrac14 07) tothe picrites from the Keogh Lake area indicating that thisdistinctive suite of rocks is exposed as far south asSchoen Lake that is 75 km away (Fig 7) The anomalouspillowed flow from the Karmutsen Range has lower LaYbCN (13^18) than the main group of tholeiitic basaltsfrom the Karmutsen Range (Fig 7) The mineralized sillfrom the Schoen Lake area is distinctly LREE-enriched(LaYbCNfrac14 33) with the highest REE abundances(LaCNfrac14 940) of all Karmutsen samples this may reflectcontamination by adjacent sediment also evident in thin-section where sulfide blebs are cored by quartz (Greeneet al 2006)The primitive mantle-normalized trace-element pat-
terns (Fig 7) and trace-element variations and ratios(Fig 8) highlight the differences in trace-element
concentrations between the four main groups ofKarmutsen flood basalts onVancouver Island The tholeii-tic basalts from all three areas have relatively smooth par-allel trace-element patterns some samples have smallpositive Sr anomalies relative to Nd and Sm and manysamples have prominent negative K anomalies relative toU and Nb reflecting K loss during alteration (Fig 7) Thelarge ion lithophile elements (LILE) in the tholeiiticbasalts (mainly Rb and K not Cs) are depleted relativeto the high field strength elements (HFSE) and LREELILE segments especially for the Schoen Lake area(Fig 7d) are remarkably parallel The picrites andhigh-MgO basalts form a tight band of parallel trace-element patterns and are depleted in HFSE and LREEwith positive Sr anomalies and relatively low Rb Thecoarse-grained mafic rocks from the Karmutsen Rangehave identical trace-element patterns to the tholeiiticbasalts Samples from the Schoen Lake area have similar
000
025
050
075
100
125
0 1 2 3 4 5
00
04
08
12
16
0 5 10 15 20 25
MgO (wt)
0
5
10
15
0 50 100 150
Hf (ppm)
Th (ppm)
NbLaNb (ppm)
Zr (ppm)
(a)
(c)
(b)
(d)
Th (ppm)
00
02
04
06
08
10
12
00 01 02 03 04 05
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
NbLa=1
4723A4 5615A12
U (ppm)
Fig 8 Whole-rock trace-element concentrations and ratios for the Karmutsen Formation [except (b) which is vs wt MgO] (a) Nb vs Zr(b) NbLa vs MgO (c) Th vs Hf (d) Th vs U There is a clear distinction between the tholeiitic basalts and picrites in Nb and Zr both inconcentration and the slope of each trend
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
19
trace-element patterns to the Karmutsen Range except forthe two LREE-depleted coarse-grained mafic rocks andthe mineralized sill (Fig 7d)
Sr^Nd^Hf^Pb isotopic compositionsA total of 19 samples from the four main groups ofthe Karmutsen Formation were selected for radiogenicisotopic geochemistry on the basis of major- and trace-element variation stratigraphic position and geographicallocation The samples have overlapping age-correctedHf and Nd isotope ratios and distinct ranges of Sr isotopiccompositions (Fig 9) The tholeiitic basalts picriteshigh-MgO basalt and coarse-grained mafic rocks have
initial eHffrac14thorn87 to thorn126 and eNdfrac14thorn66 to thorn88 cor-rected for in situ radioactive decay since 230 Ma (Fig 9Tables 3 and 4) All the Karmutsen samples form a well-defined linear array in a Lu^Hf isochron diagram corre-sponding to an age of 24115 Ma (Fig 9) This is withinerror of the accepted age of the Karmutsen Formationand indicates that the Hf isotope systematics behaved as aclosed system since c 230 Ma The anomalous pillowedflow has similar initial eHf (thorn98) and slightly lower initialeNd (thorn62) compared with the other Karmutsen samplesThe picrites and high-MgO basalt have higher initial87Sr86Sr (070398^070518) than the tholeiitic basalts(initial 87Sr86Srfrac14 070306^070381) and the coarse-grained mafic rocks (initial 87Sr86Srfrac14 070265^070428)
05128
05129
05130
05131
05132
015 017 019 021 023 025
02829
02830
02831
02832
02833
02834
000 002 004 006 008 010
4
6
8
10
0702 0703 0704 0705 07066
8
10
12
14
5 6 7 8 9 10
12
0
1
2
3
4
5
6
7
0702 0703 0704 0705 0706
(d)
LOI (wt )
87Sr 86Sr
4723A2
Nd
143Nd144Nd
147Sm144Nd
87Sr 86Sr (230 Ma)
(230 Ma)
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
(a)
(c)
176Hf177Hf
Age=241 plusmn 15 Ma=+990 plusmn 064
176Lu177Hf
Hf (230 Ma)
Nd (230 Ma)
(e)
Hf (i)
(b)
Fig 9 Whole-rock Sr Nd and Hf isotopic compositions for the Karmutsen Formation (a) 143Nd144Nd vs 147Sm144Nd (b) 176Hf177Hf vs176Lu177Hf The slope of the best-fit line for all samples corresponds to an age of 24115 Ma (c) Initial eNd vs 87Sr86Sr Age-correction to230 Ma (d) LOI vs 87Sr86Sr (e) Initial eHf vs eNd Average 2 error bars are shown in a corner of each panel Complete chemical duplicatesshown inTables 3 and 4 (samples 4720A7 and 4722A4) are circled in each plot
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
overlap the ranges of the picrites and tholeiitic basalts(Fig 9 Table 3)The measured Pb isotopic compositions of the tholeiitic
basalts are more radiogenic than those of the picrites andthe most magnesian Keogh Lake picrites have the leastradiogenic Pb isotopic compositions The range ofinitial Pb isotope ratios for the picrites is206Pb204Pbfrac1417868^18580 207Pb204Pbfrac1415547^15562and 208Pb204Pbfrac14 37873^38257 and the range for thetholeiitic basalts is 206Pb204Pbfrac1418782^19098207Pb204Pbfrac1415570^15584 and 208Pb204Pbfrac14 37312^38587 (Fig 10 Table 5) The coarse-grained mafic rocksoverlap the range of initial Pb isotopic compositions forthe tholeiitic basalts with 206Pb204Pbfrac1418652^19155207Pb204Pbfrac1415568^15588 and 208Pb204Pbfrac14 38153^38735 One high-MgO basalt has the highest initial206Pb204Pb (19242) and 207Pb204Pb (15590) and thelowest initial 208Pb204Pb (37676) The Pb isotopic ratiosfor Karmutsen samples define broadly linearrelationships in Pb isotope plots The age-corrected Pb
isotopic compositions of several picrites have been affectedby U and Pb (andor Th) mobility during alteration(Fig 10)
ALTERAT IONThe Karmutsen basalts have retained most of their origi-nal igneous structures and textures however secondaryalteration and low-grade metamorphism have producedzeolitic- and prehnite^pumpellyite-bearing mineral assem-blages (Table 1 Cho et al 1987) Alteration and metamor-phism have primarily affected the distribution of the LILEand the Sr isotopic systematics The Keogh Lake picriteshave been affected more by alteration than the tholeiiticbasalts and have higher LOI (up to 55wt ) variableLILE and higher measured Sr Nd and Hf and lowermeasured Pb isotope ratios than the tholeiitic basalts Srand Pb isotope compositions for picrites and tholeiiticbasalts show a relationship with LOI (Figs 9 and 10)whereas Nd and Hf isotopic compositions do not correlate
Table 3 Sr and Nd isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Rb Sr 87Sr86Sr 2m87Rb86Sr 87Sr86Srt Sm Nd 143Nd144Nd 2m eNd
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses carried out at the PCIGR thecomplete trace element analyses are given in Electronic Appendix 3
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
21
with LOI (not shown) The relatively high apparent initialSr isotopic compositions for the high-MgO lavas (up to07052) probably resulted from post-magmatic loss of Rbor an increase in 87Sr86Sr through addition of seawater Sr(eg Hauff et al 2003) High-MgO lavas from theCaribbean plateau have similarly high 87Sr86Sr comparedwith basalts (Recurren villon et al 2002)The correction for in situdecay on initial Pb isotopic ratios has been affected bymobilization of U and Th (and Pb) in the whole-rockssince their formation A thorough acid leaching duringsample preparation was used that has been shown to effec-tively remove alteration phases (Weis et al 2006 NobreSilva et al in preparation) Whole-rock SmNd ratiosand age-correction of Nd isotopes may be affected bythe leaching procedure for submarine rocks older than50 Ma (Thompson et al 2008 Fig 9) there ishowever very little variation in Nd isotopic compositionsof the picrites or tholeiitic basalts and this system does notappear to be disturbed by alteration The HFSE abun-dances for tholeiitic and picritic basalts exhibit clear
linear correlations in binary diagrams (Fig 8) whereasplots of LILE vs HFSE are highly scattered (not shown)because of the mobility of some of the alkali and alkalineearth LILE (Cs Rb Ba K) and Sr for most samplesduring alterationThe degree of alteration in the Karmutsen samples does
not appear to be related to eruption environment or depthof burial Pillow rims are compositionally different frompillow cores (Surdam1967 Kuniyoshi1972) and aquagenetuffs are chemically different from isolated pillows withinpillow breccias however submarine and subaerial basaltsexhibit a similar degree of alterationThere is no clear cor-relation between the submarine and subaerial basalts andsome commonly used chemical alteration indices (eg BaRb vs K2OP2O5 Huang amp Frey 2005) There is also nodefinitive correlation between the petrographic alterationindex (Table 1) and chemical alteration indices although10 of 13 tholeiitic basalts with the highest K and LILEabundances have the highest petrographic alterationindex of three (seeTable 1)
Table 4 Hf isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Lu Hf 176Hf177Hf 2m eHf176Lu177Hf 176Hf177Hft eHf(t)
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
OLIV INE ACCUMULAT ION INH IGH-MgO AND PICR IT IC LAVASGeochemical trends and petrographic characteristics indi-cate that accumulation of olivine played an important rolein the formation of the high-MgO basalts and picrites fromthe Keogh Lake area on northern Vancouver Island TheKeogh Lake samples show a strong linear correlation inplots of Al2O3 TiO2 ScYb and Ni vs MgO and many ofthe picrites have abundant clusters of olivine pseudomorphs(Table1 Fig11)There is a strong linear correlationbetween
modal per cent olivine and whole-rock magnesium contentmagnesium contents range between 10 and 19wt MgOand the proportion of olivine phenocrysts in these samplesvaries from 0 to 42 vol (Fig 11)With a few exceptionsboth the size range and median size of the olivine pheno-crysts in most samples are comparable The clear correla-tion between the proportion of olivine phenocrysts andwhole-rock magnesium contents indicates that accumula-tion of olivine was directly responsible for the high MgOcontents (410wt ) in most of the picritic lavas
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
207Pb204Pb
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
208Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
(a) (b)
(c) (e)
4723A24723A2
4723A2
4723A2
4722A4
4723A4
4722A4
4723A4
4723A134723A3
4723A3
4723A13
1556
1558
1560
1562
1564
1566
1568
186 190 194 198 202 206 2101550
1552
1554
1556
1558
1560
178 180 182 184 186 188 190 192 194
380
384
388
392
396
186 190 194 198 202 206 210374
378
382
386
390
178 180 182 184 186 188 190 192 194
206Pb204Pb(d)
LOI (wt )
0
1
2
3
4
5
6
7
186 190 194 198 202 206 210
4723A2
Fig 10 Pb isotopic compositions of leached whole-rock samples measured by MC-ICP-MS for the Karmutsen Formation Error bars are smal-ler than symbols (a) Measured 207Pb204Pb vs 206Pb204Pb (b) Initial 207Pb204Pb vs 206Pb204Pb Age-correction to 230 Ma (c) Measured208Pb204Pb vs 206Pb204Pb (d) LOI vs 206Pb204Pb (e) Initial 208Pb204Pb vs 206Pb204Pb Complete chemical duplicates shown in Table 5(samples 4720A7 and 4722A4) are circled in each plot The dashed lines in (b) and (e) show the differences in age-corrections for two picrites(4723A4 4722A4) using the measured UTh and Pb concentrations for each sample and the age-corrections when concentrations are used fromthe two picrites (4723A3 4723A13) that appear to be least affected by alteration
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
23
DISCUSS IONThe compositional and spatial development of the eruptedlava sequences of the Wrangellia oceanic plateau onVancouver Island provide information about the evolutionof a mantle plume upwelling beneath oceanic lithospherein an analogous way to the interpretation of continentalflood basalts The Caribbean and Ontong Java plateausare the two primary examples of accreted parts of oceanicplateaus that have been studied to date and comparisonscan be made with the Wrangellia oceanic plateauHowever the Wrangellia oceanic plateau is a distinctiveoccurrence of an oceanic plateau because it formed on thecrust of a Devonian to Mississippian intra-oceanic islandarc overlain by Mississippian to Permian limestones andpelagic sediments The nature and thickness of oceaniclithospheric mantle are considered to play a fundamentalrole in the decompression and evolution of a mantleplume and thus the compositions of the erupted lavasequences (Greene et al 2008) The geochemical and
stratigraphic relationships observed in the KarmutsenFormation onVancouver Island and the focus of the subse-quent discussion sections provide constraints on (1) thetemperature and degree of melting required to generatethe primary picritic magmas (2) the composition of themantle source (3) the depth of melting and residual min-eralogy in the source region (4) the low-pressure evolutionof the Karmutsen tholeiitic basalts (5) the growth of theWrangellia oceanic plateau
Melting conditions and major-elementcomposition of the primary magmasThe primary magmas to most flood basalt provinces arebelieved to be picritic (eg Cox 1980) and they have beenfound in many continental and oceanic flood basaltsequences worldwide (eg Siberia Karoo Paranacurren ^Etendeka Caribbean Deccan Saunders 2005) Near-pri-mary picritic lavas with low total alkali abundances
Table 5 Pb isotopic compositions of Karmutsen basaltsVancouver Island BC
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
require high-degree partial melting of the mantle source(eg Herzberg amp OrsquoHara 2002) The Keogh Lake picritesfrom northernVancouver Island are the best candidates foridentifying the least modified partial melts of the mantleplume source despite having accumulated olivine pheno-crysts and can be used to estimate conditions of meltingand the composition of the primary magmas for theKarmutsen FormationPrimary magma compositions mantle potential tem-
peratures and source melt fractions for the picritic lavas ofthe Karmutsen Formation were calculated from primitivewhole-rock compositions using PRIMELT1XLS software(Herzberg et al 2007) and are presented in Fig 12 andTable 6 A detailed discussion of the computationalmethod has been given by Herzberg amp OrsquoHara (2002)and Herzberg et al (2007) key features of the approachare outlined belowPrimary magma composition is calculated for a primi-
tive lava by incremental addition of olivine and compari-son with primary melts of mantle peridotite Lavas thathad experienced plagioclase andor clinopyroxene fractio-nation are excluded from this analysis All calculated pri-mary magma compositions are assumed to be derived byaccumulated fractional melting of fertile peridotite KR-4003 Uncertainties in fertile peridotite composition donot propagate to significant variations in melt fractionand mantle potential temperature melting of depletedperidotite propagates to calculated melt fractions that aretoo high but with a negligible error in mantle potential
temperature PRIMELT1XLS calculates the olivine liqui-dus temperature at 1atm which should be similar to theactual eruption temperature of the primary magmaand is accurate to 308C at the 2 level of confidenceMantle potential temperature is model dependent witha precision that is similar to the accuracy of the olivineliquidus temperature (C Herzberg personal communica-tion 2008)With regard to the evidence for accumulated olivine in
the Keogh Lake picrites it should not matter whether themodeled lava composition is aphyric or laden with accu-mulated olivine phenocrysts PRIMELT1 computes theprimary magma composition by adding olivine to a lavathat has experienced olivine fractionation in this way itinverts the effects of fractional crystallization The onlyrequirement is that the olivine phenocrysts are not acci-dental or exotic to the lava compositions Support for thisprocedure can be seen in the calculated olivine liquid-line-of-descent shown by the curved arrays with 5 olivineaddition increments (Fig 12c) Fractional crystallization ofolivine from model primary magmas having 85^95wt FeO and 153^175wt MgO will yield arrays of deriva-tive liquids and randomly sorted olivines that in most butnot all cases connect the picrites and high-MgO basalts Anewer version of the PRIMELT software (Herzberg ampAsimow 2008) can perform olivine addition to and sub-traction from lavas with highly variably olivine phenocrystcontents and yields results that converge with those shownin Fig 12
Fig 11 Relationship between abundance of olivine phenocrysts and whole-rock MgO contents for the Keogh Lake picrites (a) Tracings ofolivine grains (gray) in six high-MgO samples made from high resolution (2400 dpi) scans of petrographic thin-sections Scale bars represent10mm sample numbers are indicated at the upper left (b) Modal olivine vs whole-rock MgO Continuous line is the best-fit line for 10 high-MgO samplesThe areal proportion of olivine was calculated from raster images of olivine grains using ImageJ image analysis software whichprovides an acceptable estimation of the modal abundance (eg Chayes 1954)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
25
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
Gray lines = final melting pressure
L + Ol + Opx
07
08
09
Pressure (GPa)
10
3
4
5
6
1
SOLIDUS
01
04
0506
L + Ol
2
03
02
PeridotiteKR-4003
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
Thick black lines = initial melting pressure
Dashed lines = melt fraction
Melt Fraction
2
3 4 5 6
L+Ol+Opx
SolidusSpinel
SolidusGarnet Peridotite
7
L+Ol
0 10 20 30 4040
45
50
55
60
MgO (wt)
SiO2
(wt)
PeridotiteKR-4003
Melt Fraction
0 10 20 305
10
15
CaO(wt)
MgO (wt)
3
4 5 6 7
2
46
2
Peridotite Partial Melts
Thick black line = initial melting pressureGray lines = final melting pressure
Peridotite
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
02
Pressure (GPa)
0302
04
Pressure (GPa)
Picrite
High-MgO basalt
Tholeiitic basalt
(c)
(d)
Karmutsen flood basaltsOlivine addition modelOlivine addition (5 increment)Primary magma
PicriteHigh-MgO basaltTholeiitic basalt
(a)
(b)
800 850 900
910
920
930
940
950
Karmutsen flood basalts
Olivine addition modelOlivine addition (5 increment)Primary magma
Gray lines = Fo contents of olivine
Fig 12 Estimated primary magma compositions for three Keogh Lake picrites and high-MgO basalts (samples 4723A4 4723A135616A7) using the forward and inverse modeling technique of Herzberg et al (2007) Karmutsen compositions and modeling results areoverlain on diagrams provided by C Herzberg (a) Whole-rock FeO vs MgO for Karmutsen samples from this study Total iron estimatedto be FeO is 090 Gray lines show olivine compositions that would precipitate from liquid of a given MgO^FeO composition Black lineswith crosses show results of olivine addition (inverse model) using PRIMELT1 (Herzberg et al 2007) Gray field highlights olivinecompositions with Fo contents of 91^92 predicted to precipitate (b) FeO vs MgO showing Karmutsen lava compositions and results ofa forward model for accumulated fractional melting of fertile peridotite KR-4003 (Kettle River Peridotite Walter 1998) (c) SiO2 vsMgO for Karmtusen lavas and model results (d) CaO vs MgO showing Karmtusen lava compositions and model results To brieflysummarize the technique [see Herzberg et al (2007) for complete description] potential parental magma compositions for the high-MgOlava series were selected (highest MgO and appropriate CaO) and using PRIMELT1 software olivine was incrementally added tothe selected compositions to show an array of potential primary magma compositions (inverse model) Then using PRIMELT1 theresults from the inverse model were compared with a range of accumulated fractional melts for fertile peridotite derived fromparameterization of the experimental results for KR-4003 from Walter (1998) forward model Herzberg amp OrsquoHara (2002) A melt fractionwas sought that was unique to both the inverse and forward models (Herzberg et al 2007) A unique solution was found when there was a com-mon melt fraction for both models in FeO^MgO and CaO^MgO^Al2O3^SiO2 (CMAS) projection space This modeling assumes thatolivine was the only phase crystallizing and ignores chromite precipitation and possible augite fractionation in the mantle (Herzbergamp OrsquoHara 2002) Results are best for a residue of spinel lherzolite (not pyroxenite) The presence of accumulated olivine in samples ofthe Keogh Lake picrites used as starting compositions does not significantly affect the results because we are modeling addition of olivine(see text for further description) The tholeiitic basalts cannot be used for modeling because they are all saturated in plagthorn cpxthornolThick black line for initial melting pressure at 4 GPa is not shown in (c) for clarity Gray area in (a) indicates parental magmas thatwill crystallize olivine with Fo 91^92 at the surface Dashed lines in (c) indicate melt fraction Solidi for spinel and garnet peridotite arelabelled in (c)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
The PRIMELT1 results indicate that if the Karmutsenpicritic lavas were derived from accumulated fractionalmelting of fertile peridotite the primary magmas wouldhave contained 15^17wt MgO and 10wt CaO(Fig 12 Table 6) formed from 23^27 partial meltingand would have crystallized olivine with a Fo content of 91 (Fig 12 Table 6) Calculated mantle temperatures forthe Karmutsen picrites (14908C) indicate melting ofanomalously hot mantle (100^2008C above ambient) com-pared with ambient mantle that produces mid-ocean ridgebasalt (MORB 1280^14008C Herzberg et al 2007)which initiated between 3 and 4 GPa Several picritesappear to be near-primary melts and required very littleaddition of olivine (eg sample 93G171 with measured158wt MgO) These temperature and melting esti-mates of the Karmutsen picrites are consistent withdecompression of hot mantle peridotite in an actively con-vecting plume head (ie plume initiation model) Threesamples (picrite 5614A1 and high-MgO basalts 5614A3and 5614A5) are lower in iron than the other Karmutsenlavas with FeO in the 65^80wt range (Fig 12c) andhave MgO and FeO contents that are similar to primitiveMORB from the Sequeiros fracture zone of the EastPacific Rise (EPR Perfit et al 1996) These samples
however differ from EPR MORB in having higher SiO2
and lower Al2O3 and they are restricted geographicallyto one area (Sara Lake Fig 2)We therefore have evidence for primary magmas with
MgO contents that range from a possible low of 110wt to as high as 174wt with inferred mantle potential tem-peratures from 1340 to 15208C This is similar to the rangedisplayed by lavas from the Ontong Java Plateau deter-mined by Herzberg (2004) who found that primarymagma compositions assuming a peridotite sourcewould contain 17wt MgO and 10wt CaO(Table 6) and that these magmas would have formed by27 melting at a mantle potential temperature of15008C (first melting occurring at 36 GPa) Fitton ampGodard (2004) estimated 30 partial melting for theOntong Java lavas based on the Zr contents of primarymagmas of Kroenke-type basalt calculated by incrementaladdition of equilibrium olivine However if a considerableamount of eclogite was involved in the formation of theOntong Java plateau magmas the excess temperatures esti-mated by Herzberg et al (2007) may not be required(Korenaga 2005) The variability in mantle potential tem-perature for the Keogh Lake picrites is also similar to thewide range of temperatures that have been calculated for
Table 6 Estimated primary magma compositions for Karmutsen basalts and other oceanic plateaus or islands
Sample 4723A4 4723A13 5616A7 93G171 Average OJP Mauna Keay Gorgonaz
Ontong Java primary magma composition for accumulated fraction melting (AFM) from Herzberg (2004)yMauna Kea primary magma composition is average of four samples of Herzberg (2006 table 1)zGorgona primary magma composition for AFM for 1F fertile source from Herzberg amp OrsquoHara (2002 table 4)Total iron estimated to be FeO is 090
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
27
lavas from single volcanoes in the Galapagos Hawaii andelsewhere by Herzberg amp Asimow (2008) Those workersinterpreted the range to reflect the transport of magmasfrom both the cool periphery and the hotter axis of aplume A thermally heterogeneous plume will yield pri-mary magmas with different MgO contents and the rangeof primary magma compositions and their inferred mantlepotential temperatures for Karmutsen lavas may reflectmelting from different parts of the plume
Source of the Karmutsen Formation lavasThe Sr^Nd^Hf^Pb isotopic compositions of theKarmutsen volcanic rocks on Vancouver Island indicatethat they formed from plume-type mantle containing adepleted component that is compositionally similar to butdistinct from other oceanic plateaus that formed in thePacific Ocean (eg Ontong Java and Caribbean Fig 13)Trace element and isotopic constraints can be used to testwhether the depleted component of the KarmutsenFormation was intrinsic to the mantle plume and distinctfrom the source of MORB or the result of mixing or invol-vement of ambient depleted MORB mantle with plume-type mantle The Karmutsen tholeiitic basalts have initialeNd and
87Sr86Sr ratios that fall within the range of basaltsfrom the Caribbean Plateau (eg Kerr et al 1996 1997Hauff et al 2000 Kerr 2003) and a range of Hawaiian iso-tope data (eg Frey et al 2005) and lie within and abovethe range for the Ontong Java Plateau (Tejada et al 2004Fig 13a) Karmutsen tholeiitic basalts with the highest ini-tial eNd and lowest initial 87Sr86Sr lie within and justbelow the field for age-corrected northern EPR MORB at230 Ma (eg Niu et al1999 Regelous et al1999) Initial Hfand Nd isotopic compositions for the Karmutsen samplesare noticeably offset to the low side of the ocean islandbasalt (OIB) array (Vervoort et al 1999) and lie belowand slightly overlap the low eHf end of fields for Hawaiiand the Caribbean Plateau (Fig 13c) the Karmutsen sam-ples mostly fall between and slightly overlap a field forage-corrected Pacific MORB at 230 Ma and Ontong Java(Mahoney et al 1992 1994 Nowell et al 1998 Chauvel ampBlichert-Toft 2001) Karmutsen lava Pb isotope ratios over-lap the broad range for the Caribbean Plateau and thelinear trends for the Karmutsen samples intersect thefields for EPR MORB Hawaii and Ontong Java in208Pb^206Pb space but do not intersect these fields in207Pb^206Pb space (Fig 13d and e) The more radiogenic207Pb204Pb for a given 206Pb204Pb of the Karmutsenbasalts requires high UPb ratios at an ancient time when235U was abundant similar to the Caribbean Plateau andenriched in comparison with EPR MORB Hawaii andOntong JavaThe low eHf^low eNd component of the Karmutsen
basalts that places them below the OIB mantle array andpartly within the field for age-corrected Pacific MORBcan form from low ratios of LuHf and high SmNd over
time (eg Salters amp White 1998) MORB will develop aHf^Nd isotopic signature over time that falls below themantle array Low LuHf can also lead to low 176Hf177Hffrom remelting of residues produced by melting in theabsence of garnet (eg Salters amp Hart 1991 Salters ampWhite 1998) The Hf^Nd isotopic compositions of theKarmutsen Formation indicate the possible involvementof depleted MORB mantle however similar to Iceland(Fitton et al 2003) Nb^Zr^Y variations provide a way todistinguish between a depleted component within theplume and a MORB-like component The Karmutsenbasalts and picrites fall within the Iceland array and donot overlap the MORB field in a logarithmic plot of NbYvs ZrY (Fig 13b) using the fields established by Fittonet al (1997) Similar to the depleted component within theCaribbean Plateau (Thompson et al 2003) the trend ofHf^Nd isotopic compositions towards Pacific MORB-likecompositions is not followed by MORB values in Nb^Zr^Y systematics and this suggests that the depleted compo-nent in the source of the Karmutsen basalts is distinctfrom the source of MORB and probably an intrinsic partof the plume The relatively radiogenic Pb isotopic ratiosand REE patterns distinct from MORB also support thisinterpretationTrace-element characteristics suggest the possible invol-
vement of sub-arc lithospheric mantle in the generation ofthe Karmutsen picrites Lassiter et al (1995) suggested thatmixing of a plume-type source with eNdthorn 6 to thorn 7 witharc material with low NbTh could reproduce the varia-tions observed in the Karmutsen basalts however theyconcluded that the absence of low NbLa ratios in theirsample of tholeiitic basalts restricts the amount of arc litho-sphere involved The study of Lassiter et al (1995) wasbased on major and trace element and Sr Nd and Pb iso-topic data for a suite of 29 samples from Buttle Lake in theStrathcona Provincial Park on central Vancouver Island(Fig 1) Whereas there are no clear HFSE depletions inthe Karmutsen tholeiitic basalts the low NbLa (and TaLa) of the Karmutsen picritic lavas in this study indicatethe possible involvement of Paleozoic arc lithosphere onVancouver Island (Fig 8)
REE modeling dynamic melting andsource mineralogyThe combined isotopic and trace-element variations of thepicritic and tholeiitic Karmutsen lavas indicate that differ-ences in trace elements may have originated from differentmelting histories For example the LuHf ratios of theKarmutsen picritic lavas (mean 008) are considerablyhigher than those in the tholeiitic basalts (mean 002)however the small range of overlapping initial eHf valuesfor the two lava suites suggests that these differences devel-oped during melting within the plume (Fig 9b) Below wemodel the effects of source mineralogy and depth of
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
28
melting on differences in trace-element concentrations inthe tholeiitic basalts and picritesDynamic melting models simulate progressive decom-
pression melting where some of the melt fraction isretained in the residue and only when the degree of partialmelting is greater than the critical mass porosity and the
source becomes permeable is excess melt extracted fromthe residue (Zou1998) For the Karmutsen lavas evolutionof trace element concentrations was simulated using theincongruent dynamic melting model developed by Zou ampReid (2001) Melting of the mantle at pressures greater than1 GPa involves incongruent melting (eg Longhi 2002)
OIB array
PacificMORB(230 Ma)
(0 Ma)
EPR(230 Ma)
-4
-2
0
2
4
8
10
12
14
16
18
20
22
-4 -2 0 2 4 6 8 10 12 14
minus2
0
2
4
6
8
10
12
14
0702 0703 0704 0705 0706
IndianMORB
87Sr 86Sr (initial)
Hf (initial)
(a) (c)
Nd (initial)
Karmutsen tholeiitic basalt
HawaiiOntong JavaCaribbean Plateau
Karmutsen high-MgO lava
high-MgO lavasKarmutsen
1540
1545
1550
1555
1560
1565
175 180 185 190 195 200375
380
385
390
395
175 180 185 190 195 200
(d) (e)
EPR
EPR
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb
Karmutsen Formation (initial)
HawaiiOntong JavaCaribbean Plateau
East Pacific Rise
Karmutsen Formation (measured)
Juan de FucaGorda
ε Nd
(initial)
Karmutsen Formation (different age-correction for 3 picrites)
(0 Ma)
NbY
001
01
1
1 10
ZrY
OIB
NMORB
(b)
Iceland array
6
Fig 13 Comparison of age-corrected (230 Ma) Sr^Nd^Hf^Pb isotopic compositions for Karmutsen flood basalts onVancouver Island withage-corrected OIB and MORB and Nb^Zr^Y variation (a) Initial eNd vs 87Sr86Sr (b) NbYand ZrY variation (after Fitton et al 1997)(c) Initial eHf vs eNd (d) Measured and initial 207Pb204Pb vs 206Pb204Pb (e) Measured and initial 208Pb204Pb vs 206Pb204Pb References fordata sources are listed in Electronic Appendix 4 Most of the compiled data were extracted from the GEOROC database (httpgeorocmpch-mainzgwdgdegeoroc) OIB array line in (c) is fromVervoort et al (1999) EPR East Pacific Rise Several high-MgO samples affected bysecondary alteration were not plotted for clarity in Pb isotope plots Estimation of fields for age-corrected Pacific MORB and EPR at 230 Mawere made using parent^daughter concentrations (in ppm) from Salters amp Stracke (2004) of Lufrac14 0063 Hffrac14 0202 Smfrac14 027 Ndfrac14 071Rbfrac14 0086 and Srfrac14 836 In the NbYvs ZrYplot the normal (N)-MORB and OIB fields not including Icelandic OIB are from data com-pilations of Fitton et al (1997 2003) The OIB field is data from 780 basalts (MgO45wt ) from major ocean islands given by Fitton et al(2003) The N-MORB field is based on data from the Reykjanes Ridge EPR and Southwest Indian Ridge from Fitton et al (1997) The twoparallel gray lines in (b) (Iceland array) are the limits of data for Icelandic rocks
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
29
with olivine being produced during the reactioncpxthornopxthorn spfrac14meltthornol for spinel peridotites The mod-eling used coefficients for incongruent melting reactionsbased on experiments on lherzolite melting (eg Kinzleramp Grove 1992 Longhi 2002) and was separated into (1)melting of spinel lherzolite (2) two combined intervals ofmelting for a garnet lherzolite (gt lherz) and spinel lher-zolite (sp lherz) source (3) melting of garnet lherzoliteThe model solutions are non-unique and a range indegree of melting partition coefficients melt reactioncoefficients and proportion of mixed source componentsis possible to achieve acceptable solutions (see Fig 14 cap-tion for description of modeling parameters anduncertainties)The LREE-depleted high-MgO lavas require melting of
a depleted spinel lherzolite source (Fig 14a) The modelingresults indicate a high degree of melting (22^25) similarto the results from PRIMELT1 based on major-elementvariations (discussed above) of a LREE-depleted sourceThe high-MgO lavas lack a residual garnet signature Theenriched tholeiitic basalts involved melting of both garnetand spinel lherzolite and represent aggregate melts pro-duced from continuous melting throughout the depth ofthe melting column (Fig 14b) Trace-element concentra-tions in the tholeiitic and picritic lavas are not consistentwith low-degree melting of garnet lherzolite alone(Fig 14c) The modeling results indicate that the aggregatemelts that formed the tholeiitic basalts would haveinvolved lesser proportions of enriched small-degreemelts (1^6 melting) generated at depth and greateramounts of depleted high-degree melts (12^25) gener-ated at lower pressure (a ratio of garnet lherzolite tospinel lherzolite melt of 14 provides the best fits seeFig 14b and description in figure caption) The totaldegree of melting for the tholeiitic basalts was high(23^27) similar to the degree of partial melting in thespinel lherzolite facies for the primary melts that eruptedas the Keogh Lake picrites of a source that was also prob-ably depleted in LREE
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
(c)
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
Melting of garnet lherzolite
High-MgO lavas
Melting of spinel lherzolite
source (07DM+03PM)
source (07DM+03PM)
6 melting
30 melting
(b)
(a)
30 melting
Sam
ple
Cho
ndrit
eS
ampl
eC
hond
rite 20 melting
1 melting
Melting of garnet lherzoliteand spinel lherzolite
x gt lherz + 4x sp lherz
5 meltingx gt lherz + 4x sp lherz
Sam
ple
Cho
ndrit
e
Tholeiitic lavas
Tholeiitic lavas
High-MgO lavas
source (07DM+03PM)
Fig 14 Trace-element modeling results for incongruent dynamicmantle melting for picritic and tholeiitic lavas from the KarmutsenFormation Three steps of modeling are shown (a) melting of spinellherzolite compared with high-MgO lavas (b) melting of garnet lher-zolite and spinel lherzolite compared with tholeiitic lavas (c) meltingof garnet lherzolite compared with tholeiitic and picritic lavasShaded fields in each panel for tholeiitic basalts and picrites arelabeled Patterns with symbols in each panel are modeling results(patterns are 5 melting increments in (a) and (b) and 1 incre-ments in (c) PM primitive mantle DM depleted MORB mantle gtlherz garnet lherzolite sp lherz spinel lherzolite In (b) the ratios ofper cent melting for garnet and spinel lherzolite are indicated (x gtlherzthorn 4x sp lherz means that for 5 melting 1 melt in garnetfacies and 4 melt in spinel facies where x is per cent melting in theinterval of modeling) The ratio of the respective melt proportions inthe aggregate melt (x gt lherzthorn 4x sp lherz) was kept constant andthis ratio of melt contribution provided the best fit to the data Arange of proportion of source components (07DMthorn 03PM to
09DMthorn 01PM) can reproduce the variation in the high-MgOlavas Melting modeling uses the formulation of Zou amp Reid (2001)an example calculation is shown in their Appendix PM fromMcDonough amp Sun (1995) and DM from Salters amp Stracke (2004)Melt reaction coefficients of spinel lherzolite from Kinzler amp Grove(1992) and garnet lherzolite fromWalter (1998) Partition coefficientsfrom Salters amp Stracke (2004) and Shaw (2000) were kept constantSource mineralogy for spinel lherzolite (018cpx027opx052ol003sp) from Kinzler (1997) and for garnet lherzolite(034cpx008opx053ol005gt) from Salters amp Stracke (2004) Arange of source compositions for melting of spinel lherzolite(07DMthorn 03PM to 03DMthorn 07PM) reproduce REE patternssimilar to those of the high-MgO lavas with best fits for 22ndash25melting and 07DMthorn 03PM source composition Concentrationsof tholeiitic basalts cannot be reproduced from an entirelyprimitive or depleted source
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
30
The uncertainty in the composition of the source in thespinel lherzolite melting model for the high-MgO lavasprecludes determining whether the source was moredepleted in incompatible trace elements than the source ofthe tholeiitic lavas (see Fig 14 caption for description) It ispossible that early high-pressure low-degree melting left aresidue within parts of the plume (possibly the hotter inte-rior portion) that was depleted in trace elements (egElliott et al 1991) further decompression and high-degreemelting of such depleted regions could then have generatedthe depleted high-MgO Karmutsen lavas Shallow high-degree melting would preferentially sample depletedregions whereas deeper low-degree melting may notsample more refractory depleted regions (eg CaribbeanEscuder-Viruete et al 2007) The picrites lie at the highend of the range of Nd isotopic compositions and havelow NbZr similar to depleted Icelandic basalts (Fittonet al 2003) therefore the picrites may represent the partsof the mantle plume that were the most depleted prior tothe initiation of melting As Fitton et al (2003) suggestedthese depleted melts may only rarely be erupted withoutcombining with more voluminous less depleted melts andthey may be detected only as undifferentiated small-volume flows like the Keogh Lake picrites within theKarmutsen Formation on Vancouver IslandDecompression melting within the mantle plume initiatedwithin the garnet stability field (35^25 GPa) at highmantle potential temperatures (Tp414508C) and pro-ceeded beneath oceanic arc lithosphere within the spinelstability field (25^09 GPa) where more extensivedegrees of melting could occur
Magmatic evolution of Karmutsentholeiitic basaltsPrimary magmas in large igneous provinces leave exten-sive coarse-grained cumulate residues within or below thecrust these mafic and ultramafic plutonic sequences repre-sent a significant proportion of the original magma thatpartially crystallized at depth (eg Farnetani et al 1996)For example seismic and petrological studies of OntongJava (eg Farnetani et al 1996) combined with the use ofMELTS (Ghiorso amp Sack 1995) indicate that the volcanicsequence may represent only 40 of the total magmareaching the Moho Neal et al (1997) estimated that30^45 fractional crystallization took place for theOntong Java magmas and produced cumulate thicknessesof 7^14 km corresponding to 11^22 km of flood basaltsdepending on the presence of pre-existing oceanic crustand the total crustal thickness (25^35 km) Interpretationsof seismic velocity measurements suggest the presence ofpyroxene and gabbroic cumulates 9^16 km thick beneaththe Ontong Java Plateau (eg Hussong et al 1979)Wrangellia is characterized by high crustal velocities in
seismic refraction lines which is indicative of mafic
plutonic rocks extending to depth beneath VancouverIsland (Clowes et al 1995)Wrangellia crust is 25^30 kmthick beneath Vancouver Island and is underlain by astrongly reflective zone of high velocity and density thathas been interpreted as a major shear zone where lowerWrangellia lithosphere was detached (Clowes et al 1995)The vast majority of the Karmutsen flows are evolved tho-leiitic lavas (eg low MgO high FeOMgO) indicating animportant role for partial crystallization of melts at lowpressure and a significant portion of the crust beneathVancouver Island has seismic properties that are consistentwith crystalline residues from partially crystallizedKarmutsen magmas To test the proportion of the primarymagma that fractionated within the crust MELTS(Ghiorso amp Sack 1995) was used to simulate fractionalcrystallization using several estimated primary magmacompositions from the PRIMELT1 modeling results Themajor-element composition of the primary magmas couldnot be estimated for the tholeiitic basalts because of exten-sive plagthorn cpxthornol fractionation and thus the MELTSmodeling of the picrites is used as a proxy for the evolutionof major elements in the volcanic sequence A pressure of 1kbar was used with variable water contents (anhydrousand 02wt H2O) calculated at the quartz^fayalite^magnetite (QFM) oxygen buffer Experimental resultsfrom Ontong Java indicate that most crystallizationoccurs in magma chambers shallower than 6 km deep(52 kbar Sano amp Yamashita 2004) however some crys-tallization may take place at greater depths (3^5 kbarFarnetani et al 1996)A selection of the MELTS results are shown in Fig 15
Olivine and spinel crystallize at high temperature until15^20wt of the liquid mass has fractionated and theresidual magma contains 9^10wt MgO (Fig 15f) At1235^12258C plagioclase begins crystallizing and between1235 and 11908C olivine ceases crystallizing and clinopyr-oxene saturates (Fig 15f) As expected the addition of clin-opyroxene and plagioclase to the crystallization sequencecauses substantial changes in the composition of the resid-ual liquid FeO and TiO2 increase and Al2O3 and CaOdecrease while the decrease in MgO lessens considerably(Fig 15) The near-vertical trends for the Karmutsen tho-leiitic basalts especially for CaO do not match the calcu-lated liquid-line-of-descent very well which misses manyof the high-CaO high-Al2O3 low-FeO basalts A positivecorrelation between Al2O3CaO and SrNd indicates thataccumulation of plagioclase played a major role in the evo-lution of the tholeiitic basalts (Fig 15g) Increasing thecrystallization pressure slightly and the water content ofthe melts does not systematically improve the match ofthe MELTS models to the observed dataThe MELTS results indicate that a significant propor-
tion of the crystallization takes place before plagioclase(15^20 crystallization) and clinopyroxene (35^45
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
31
crystallization) join the crystallization sequence Theseresults suggest that445 crystallization of the primarymagmas occurred (estimated from the residual liquid percent in Fig 15f) and if the original flood basalt stratigra-phy on Vancouver Island was 6 km thick at least anequivalent amount of complementary plutonic rocksshould have crystallized The Karmutsen basalts and theirplutonic residues thus represent a significant addition ofcrust (perhaps 412 km thickness) previously thickenedduring the formation of the underlying Paleozoic arc
Growth of the Wrangellia oceanic plateauon Vancouver IslandThe extrusive part of the Wrangellia oceanic plateau onVancouver Island was constructed in a three-layersequence of mostly tholeiitic basalts with a restrictedrange of composition (Fig 16) The plateau overliesDevonian to Mississippian arc volcanic sequences andmarine sediments of Mississippian to Permian ages thatare intruded by gabbroic rocks related to the flood basaltsThe volcanic stratigraphy of theWrangellia plateau formed
00
05
10
15
20
25
30
35
40
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
16
0 2 4 6 8 10 12 14 16 18 20
10
11
12
13
14
15
16
17
18
19
20
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
0 2 4 6 8 10 12 14 16 18 20
MgO (wt)
0 10 20 30 40 50
percent of total mass
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
) Mineral proportions
Sp (e)
Ol
CpxPlag
PicriteHigh-MgO basaltTholeiitic basalt
TiO2 (wt)
MgO (wt)
Al2O3 (wt) CaO (wt)
FeO (wt)
MgO(a) (b)
(c) (d)
MgO (wt)MgO (wt)
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
)
Residual liquid ()
1220
1190
Ol + Sp
Plag+Ol+Sp Plag+Cpx+Sp
02 wt H2O
no H2O
Residual liquid and crystallizing phases
(f)
0102030405060708090100
plagioclaseaccumulation
plagioclaseaccumulation
plagioclase
accumulatio
n
10
12
14
16
0 10 20 30 40
Al2O3CaO
SrNd
tholeiitic basalts
(g)
Fig 15 Forward fractional crystallization modeling results for major elements from MELTS (Ghiorso amp Sack 1995) compared with picriticand tholeiitic lavas from the Karmutsen Formation (a^d) Major-element compositions of residual liquids compared with Karmutsen lavacompositions For clarity the results are shown only for one starting composition (open circle) using the composition of the estimated primarymagma for the high-MgO lava series from the modeling technique of Herzberg et al (2007) for sample 4723A13 as shown in Fig 12 The esti-mated primary magmas for samples 4723A3 and 93G171 yield similar results A pressure of 1 kbar was used with no H2O and 02wt H2Ocalculated at QFM black line is anhydrous and grey line is 02wt H2OThe end-point for the anhydrous result is at 78 fractionation and11308C and the end-point for the result with 02wt H2O is 75 fractionation and11108C (e) Mineral proportions of crystallized phases fromone MELTrun (sample 4723A13) with no H2O (f) Per cent residual liquid vs temperature labelled with intervals of crystallizing phases (g)Al2O3CaO vs SrNd for the tholeiitic basalts
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
32
in a single eruptive phase (c 231^225 Ma) as an emergentbasalt sequence during a deeper-water shallow-water andsubaerial stage and subsided below sea level relativelysoon after its formation (Fig 16) The deeper-water stagecontains pillowed and massive submarine flows (3000mthick) that built up the submarine volcanic edifice from theseafloor The lowest pillow basalts were emplaced onunconsolidated and lithified fine-grained siliciclastic andcarbonaceous sediments containing Daonella beds(c 235^232 Ma) that are intruded by abundantKarmutsen mafic sills (Fig 16a) The different pillowedand massive submarine flows resulted primarily from dif-ferent effusive rates and local topography Some of themassive submarine flows commonly recognized by radiat-ing columnar jointing represent master tubes for the deliv-ery of lava to distal parts of submarine flow fields Thepillowed flows which commonly have uneven topography
in modern sequences may have served as levees to themassive submarine flows rather than allowing unob-structed spreading and inflation typical of the massive sub-aerial flows that form on shallow slopes in continentalflood basalt provinces (eg Self et al 1997) The rapid suc-cession of flows is evident by the near absence of sedimentbetween submarine flows Minor volumes of picritic pillowlavas occur in the upper part of the pillowed flow sequencein areas of the plateau exposed on northern VancouverIsland (Fig 16 Nixon et al 2008)The growth of the plateau in shallow water led to
emplacement of an increasing proportion of volcaniclastics(pillow breccias and hyaloclastites) that accumulated up to1500m thick in some areas (Fig 16b Nixon et al 2008)Horizons of pillowed flows occur locally within the volca-niclastic stratigraphy Sedimentary structures (eg gradedbedding fluidization structures) and petrographic textures
Late Devonian-Mississippian arc
Daonella beds
Middle Triassic
deeper water
Late Devonian-Mississippian arc
pillowed flows massive submarine
flows
pillow brecciahyaloclastiteshallow water
~4 km
~6 km
Late Triassic limestoneinterflow limestonelenses
Karmutsen Formation
mafic intrusions sediment-sill complex
Karmutsen Formation
(a) (b)
(c) (d)
subaerial flows
Late Devonian-Mississippian arc Late Devonian-Mississippian arc
picritic pillow basalts
Miss to Permian marine sediments
sills
Fig 16 Growth of the Wrangellia oceanic plateau onVancouver Island (a) Deep-water eruptive phase of pillowed and massive submarineflows with almost no intervening sediments onto MiddleTriassic shale containing Daonella beds (c 235^232 Ma) The basement of the plateauis composed of Devonian to Mississippian arc sequences overlain by Mississippian to Permian carbonates and siliciclastic sediments(b) Shallow-water flow emplacement with increasing proportion of volcaniclastics resulted primarily from cooling^contraction granulationPicritic pillow basalts occur primarily near the transition between pillowed submarine flows and volcaniclastic units Pillowed and massiveflows occur locally within the volcaniclastic unit (c) Emergent stage of growth consists of overlapping subaerial flow fields similar to continen-tal flood basalts with pahoehoe flow structures Multi-tiered columnar jointing occurrs locally in response to ponding of flows (eg Bondre et al2004) Interflow carbonate deposits and plagioclase megacrystic flows developed locally during the waning stage of volcanism (d) Post-volcanicsubsidence and deposition of Late Carnian to Norian (c 212^223 Ma) platform carbonates on top of the plateau Units are labeled in the dia-grams Thicknesses of the underlying units are not to scale
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
33
within the volcaniclastic rocks indicate formation primar-ily via cooling^contraction granulation magma^water^steam interaction autobrecciation resedimentation andmass-wasting rather than pyroclastic eruption (Nixonet al 2008)A broad subaerial platform was constructed by emplace-
ment of subaerial lava flows during an emergent subaerialstage (Fig 16c) The subaerial unit predominantly com-prises well-layered sheet flows that formed in a similarway to continental flood basalts as hundred(s) of pahoehoeflow fields erupted on shallow slopes (Self et al 1997) how-ever volcaniclastic and pillowed flow units are locally pre-sent mostly in the upper parts of the stratigraphyLocalized ponding of flows led to the formation of multi-tiered columnar jointing in some subaerial flows (egBondre et al 2004) As volcanism waned local interflowcarbonate deposits developed and plagioclase megacrysticflows were erupted Subsidence of the plateau is recordedby the interflow lenses and deposits of platform carbonates(40^1500m thick) on top of the Karmutsen basalts thatcontain Late Carnian to Early Norian fossils (Fig 16d c216^227 Ma Nixon et al 2008)The evolution of the internal architecture of the
Wrangellia oceanic plateau was primarily related to theconditions of emplacement (ie water depth and eruptionenvironment) There does not appear to have been a sys-tematic coincident evolution or shift of lava compositionsonVancouver Island as has been recognized in most con-tinental flood basalts and parts of the Wrangellia oceanicplateau in Alaska and Yukon (Greene et al 2008 2009)In general with the exception of the picritic pillow lavasin the submarine stratigraphy and local variations betweenhigh-MgO and tholeiitic lavas there are no significantcompositional differences between lavas erupted duringthe different stages of growth of the plateau (ie submarineand subaerial flows) nor are substantial compositional dif-ferences generally found between massive submarine flowsin contact with pillowed flows tholeiitic basalt flows nearthe base of the Karmutsen Formation are compositionallysimilar to flows sampled near the top of the volcanicsequence Where it was possible to sample continuous sec-tions of volcanic stratigraphy greater than several hun-dreds of meters (eg Keogh^Maynard Lake area west ofthe Karmutsen Range Fig 2) there is no clear relationshipbetween stratigraphic position and compositionalvariation
CONCLUSIONSThe Karmutsen Formation covers extensive areas ofVancouver Island (20 000 km2) and forms part of a largeoceanic plateau erupted over a geologically short interval(c 225^231 Ma) The tripartite volcanic stratigraphyexposed on Vancouver Island is at least 6 km thick andcomprises a succession of submarine flows volcaniclastic
deposits and massive subaerial flowsThese volcanologicaldifferences are primarily related to the eruption environ-ment (deep-water shallow-water or subaerial) The rapidgrowth of the plateau prevented significant accumulationsof intervening sediment except in the uppermost stratigra-phy where isolated limestone lenses commonly associatedwith pillow basalt and hyaloclastite deposits preserve arecord of the final stages of subsidence of the plateauThe Wrangellia plateau on Vancouver Island was con-
structed dominantly of tholeiitic basalt with restrictedmajor- and trace-element and isotopic compositional var-iations Minor volumes of picritic pillow basalts eruptedlate in the submarine phase The Karmutsen basalts wereemplaced onto older Paleozoic arc volcanic sequences andlow NbLa for the picrites may indicate a contributionfrom the sub-arc lithospheric mantle Modeling of the pet-rogenesis of the Karmutsen picrites indicates melting ofanomalously hot mantle (15008C) and extensive degreesof partial melting (23^27) and is consistent with aplume initiation model Combined Sr^Nd^Hf^Pb isotopicsystematics indicate that the picritic and tholeiitic lavaswere derived from a depleted plume-type mantle sourcedistinct from the source of MORB that has compositionalsimilarities to the source of the lavas forming theCaribbean Plateau Picrites were generated by melting ofspinel lherzolite whereas the tholeiitic basalts formed asaggregate melts of both garnet and spinel lherzolite Thetholeiitic basalts underwent extensive low-pressure fractio-nation (52^3 kbar) and seismic studies indicate that someof theWrangellia crust beneathVancouver Island may cor-respond to the plutonic residues of the Karmutsen floodbasalts
ACKNOWLEDGEMENTSWe are grateful to Nick Arndt for helping us get this pro-ject started and Nick Massey for insights into VancouverIsland geology We greatly appreciate the assistance withPRIMELT modeling from Claude Herzberg We thankMikkel Schau for his insight and enthusiasm during field-work Jane Barling assisted with analyses by MC-ICP-MSThoughtful and thorough reviews of the manuscript byJohn Mahoney Fred Frey Wendy Bohrson DougalJerram and Teal Riley are greatly appreciated Fundingwas generously provided by the Rocks to Riches Programadministered by the BC amp Yukon Chamber of Mines in2004 by the BC Geological Survey and by NSERCDiscovery Grants to JSS and DW ARG was supportedby a University Graduate Fellowship at UBC
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
34
REFERENCESBarker F Brown A S Budahn J R amp Plafker G (1989) Back-arc
with frontal-arc component origin of Triassic Karmutsen basaltBritish Columbia Canada Chemical Geology 75 81^102
Blichert-Toft JWeis D Maerschalk C Agranier A amp Albare de F(2003) Hawaiian hot spot dynamics as inferred from the Hf and Pbisotope evolution of Mauna Kea volcano Geochemistry GeophysicsGeosystems 4(2) 1^27 doi1010292002GC000340
Bondre N R Duraiswami R A amp Dole G (2004) Morphologyand emplacement of flows from the Deccan Volcanic ProvinceIndia Bulletin of Volcanology 66 29^45
Carlisle D (1963) Pillow breccias and their aquagene tuffs QuadraIsland British Columbia Journal of Geology 71 48^71
Carlisle D (1972) Late Paleozoic to Mid-Triassic sedimentary-volcanic sequence on Northeastern Vancouver Island Geological
Society of Canada Paper Report of Activities 72-1 Part B 24^30Carlisle D amp SuzukiT (1974) Emergent basalt and submergent car-
bonate^clastic sequences including the Upper Triassic Dilleri andWelleri zones on Vancouver Island Canadian Journal of Earth
Sciences 11 254^279Chauvel C amp Blichert-Toft J (2001) A hafnium isotope and trace
element perspective on melting of the depleted mantle Earth and
Planetary Science Letters 190 137^151Chayes F (1954)The theory of thin-section analysis Journal of Geology
62 92^101Cho M Liou J G amp Maruyama S (1987) Transition from the zeo-
lite to prehnite^pumpellyite facies in the Karmutsen metabasitesVancouver Island British Columbia Journal of Petrology 27(2)467^494
Clowes R M Zelt C A Amor J R amp Ellis R M (1995)Lithospheric structure in the southern Canadian Cordillera froma network of seismic refraction lines Canadian Journal of Earth
Sciences 32(10) 1485^1513Coffin M F amp Eldholm O (1994) Large igneous provinces Crustal
structure dimensions and external consequences Reviews of
Geophysics 32(1) 1^36Cox K G (1980) A model for flood basalt volcanism Journal of
Petrology 21 629^650Eldholm O amp Coffin M F (2000) Large igneous provinces
and plate tectonics In Richards M A Gordon R G amp vander Hilst R D (eds) The History and Dynamics of Global
Plate Motions American Geophysical Union Geophysical Monograph 121309^326
Elliott T R Hawkesworth C J amp Grolaquo nvold K (1991) Dynamicmelting of the Iceland plume Nature 351 201^206
Escuder-Viruete J Pecurren rez-Estaucurren n A Contreras F Joubert MWeis D Ullrich T D amp Spadea P (2007) Plume mantle sourceheterogeneity through time Insights from the Duarte ComplexHispaniola northeastern Caribbean Journal of Geophysical Research112(B04203) doi1010292006JB004323
Farnetani C G Richards M A amp Ghiorso M S (1996)Petrological models of magma evolution and deep crustal structurebeneath hotspots and flood basalt provinces Earth and Planetary
Science Letters 143 81^94Fitton J G amp Godard M (2004) Origin and evolution of magmas
on the Ontong Java Plateau In Fitton J G Mahoney J JWallace P J amp Saunders A D (eds) Origin and Evolution of the
Ontong Java Plateau Geological Society London Special Publications 229151^178
Fitton J G Saunders A D Norry M J Hardarson B S ampTaylor R N (1997) Thermal and chemical structure of theIceland plume Earth and Planetary Science Letters 153 197^208
Fitton J G Saunders A D Kempton P D amp Hardarson B S(2003) Does depleted mantle form an intrinsic part of the Icelandplume Geochemistry Geophysics Geosystems 4(3) 1032 doi1010292002GC000424
Frey F A Huang S Blichert-Toft J Regelous M amp Boyet M(2005) Origin of depleted components in basalt related to theHawaiian hotspot Evidence from isotopic and incompatible ele-ment ratios Geochemistry Geophysics Geosystems 6(1) 23 doi1010292004GC000757
Galer S J G amp Abouchami W (1998) Practical application of leadtriple spiking for correction of instrumental mass discriminationMineralogical Magazine 62A 491^492
Ghiorso M S amp Sack R O (1995) Chemical mass transfer in mag-matic processes IV A revised and internally consistent thermody-namic model for the interpolation and extrapolation of liquid^solidequilibria in magmatic systems at elevated temperatures and pres-sures Contributions to Mineralogy and Petrology 119 197^212
Greene A R Scoates J S amp Weis D (2005)WrangelliaTerrane onVancouver Island Distribution of flood basalts with implicationsfor potential Ni^Cu^PGE mineralization In Grant B (ed)Geological Fieldwork 2004 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2005-1 209^220Greene A R Scoates J S Nixon G T amp Weis D (2006) Picritic
lavas and basal sills in the Karmutsen flood basalt provinceWrangellia northern Vancouver Island In Grant B (ed)Geological Fieldwork 2005 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2006-1 39^52Greene A R Scoates J S amp Weis D (2008) Wrangellia flood
basalts in Alaska A record of plume^lithosphere interaction in aLate Triassic accreted oceanic plateau Geochemistry Geophysics
Geosystems 9(Q12004) doi1010292008GC002092Greene A R Scoates J S Weis D amp Israel S (2009)
Geochemistry of triassic flood basalts from the Yukon (Canada)segment of the accreted Wrangellia oceanic plateau Lithosdoi101016jlithos200811010
Hauff F Hoernle K Tilton G Graham D W amp Kerr A C(2000) Large volume recycling of oceanic lithosphere over shorttime scales geochemical constraints from the Caribbean LargeIgneous Province Earth and Planetary Science Letters 174 247^263
Hauff F Hoernle K amp Schmidt A (2003) Sr^Nd^Pb compositionof Mesozoic Pacific oceanic crust (Site 1149 and 801 ODP Leg185)Implications for alteration of ocean crust and the input into theIzu^Bonin^Mariana subduction system Geochemistry Geophysics
Geosystems 4(8) doi1010292002GC000421Herzberg C (2004) Partial melting below the Ontong Java Plateau
In Fitton J G Mahoney J J Wallace P J amp Saunders A D(eds) Origin and Evolution of the Ontong Java Plateau Geological SocietyLondon Special Publications 229 179^183
Herzberg C (2006) Petrology and thermal structure of the Hawaiianplume from Mauna Kea volcano Nature 444(7119) 605
Herzberg C amp Asimow P D (2008) Petrology of some oceanicisland basalts PRIMELT2XLS software for primary magma cal-culation Geochemistry Geophysics Geosystems 9(Q09001) doi1010292008GC002057
Herzberg C amp OrsquoHara M J (2002) Plume-associated ultramaficmagmas of Phanerozoic age Journal of Petrology 43(10) 1857^1883
Herzberg C Asimow P D Arndt N Niu Y Lesher C MFitton J G Cheadle M J amp Saunders A D (2007)Temperatures in ambient mantle and plumes Constraints frombasalts picrites and komatiites Geochemistry Geophysics Geosystems8(Q02006) doi1010292006GC001390
Huang S amp Frey F A (2005) Trace element abundances of MaunaKea basalt from phase 2 of the Hawaii Scientific Drilling Project
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
35
Petrogenetic implications of correlations with major element con-tent and isotopic ratios Geochemistry Geophysics Geosystems 4(6)doi1010292002GC000322
Hussong D M Wipperman L K amp Kroenke L W (1979) Thecrustal structure of the Ontong Java and Manihiki OceanicPlateaus Journal of Geophysical Research 84(B11) 6003^6010
Jerram D A amp Widdowson M (2005) The anatomy of ContinentalFlood Basalt Provinces geological constraints on the processes andproducts of flood volcanism Lithos 79(3^4) 385^405
Jones D L Silberling N J amp Hillhouse J (1977)Wrangellia a dis-placed terrane in northwestern North America CanadianJournal ofEarth Sciences 14(11) 2565^2577
Kerr A C (2003) Oceanic Plateaus In Rudnick R (ed) The Crust
Treatise on GeochemistryVol 3 Oxford Elsevier Science pp 537^565Kerr A C amp Mahoney J J (2007) Oceanic plateaus Problematic
plumes potential paradigms Chemical Geology 241 332^353Kerr A CTarney J Marriner G F Klaver GT Saunders A D
amp Thirlwall M F (1996) The geochemistry and petrogenesis ofthe Late-Cretaceous picrites and basalts of Curacao NetherlandsAntilles a remnant of an oceanic plateau Contributions to
Mineralogy and Petrology 124(1) 29^43Kerr A C Marriner G F Tarney J Nivia A Saunders A D
Thirlwall M F amp Sinton A C (1997) Cretaceous basaltic ter-ranes in Western Colombia Elemental chronological and Sr-Ndisotopic constraints on petrogenesis Journal of Petrology 38(6)677^702
Kinzler R J (1997) Melting of mantle peridotite at pressuresapproaching the spinel to garnet transition Application to mid-ocean ridge basalt petrogenesis Journal of Geophysical Research
102(B1) 853^874Kinzler R J amp Grove T L (1992) Primary magmas of mid-ocean
ridge basalts 1 Experiments and methods Journal of Geophysical
Research 97(B5) 6885^6906Korenaga J (2005) Why did not the Ontong Java Plateau form sub-
aerially Earth and Planetary Science Letters 234 385^399Kuniyoshi S (1972) Petrology of the Karmutsen (Volcanic) Group
northeastern Vancouver Island British Columbia PhD disserta-tion University of California Los Angeles 242 pp
Lassiter J C DePaolo D J amp Mahoney J J (1995) Geochemistry ofthe Wrangellia flood basalt province Implications for the role ofcontinental and oceanic lithosphere in flood basalt genesis Journalof Petrology 36(4) 983^1009
LeBas M J (2000) IUGS reclassification of the high-Mg and picriticvolcanic rocks Journal of Petrology 41(10) 1467^1470
Longhi J (2002) Some phase equilibrium systematics of lherzolitemelting I Geochemistry Geophysics Geosystems 3(3) doi1010292001GC000204
MacDonald G A amp Katsura T (1964) Chemical composition ofHawaiian lavas Journal of Petrology 5 82^133
Mahoney J J (1987) An isotopic survey of Pacific oceanic plateausimplications for their nature and origin In Keating B HFryer P Batiza R amp Boehlert GW (eds) Seamounts Islands andAtolls American Geophysical Union Geophysical Monograph 43 207^220
Mahoney J LeRoex A P Peng Z Fisher R L amp Natland J H(1992) Southwestern limits of Indian Ocean Ridge mantle and theorigin of low 206Pb204Pb mid-ocean ridge basalt isotope systema-tics of the central Southwest Indian Ridge (178^508E) Journal ofGeophysical Research 97(B13) 19771^19790
Mahoney J J Sinton J M Kurz M D MacDougall J DSpencer K J amp Lugmair GW (1994) Isotope and trace elementcharacteristics of a super-fast spreading ridge East Pacific rise13^238S Earth and Planetary Science Letters 121 173^193
Massey N W D (1995a) Geology and mineral resources of theAlberni^Nanaimo Lakes sheet Vancouver Island 92F1W 92F2Eand part of 92F7E British Columbia Ministry of Energy Mines and
Petroleum Resources Paper 1992-2 132 ppMassey N W D (1995b) Geology and mineral resources of the
Cowichan Lake sheet Vancouver Island 92C16 British Columbia
Ministry of Energy Mines and Petroleum Resources Paper 1992-3 112 ppMassey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005a) Digital Geology Map of British Columbia Tile NM9 MidCoast BC British Columbia Ministry of Energy and Mines Geofile
2005-2Massey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005b) Digital Geology Map of British Columbia Tile NM10Southwest BC British Columbia Ministry of Energy and Mines Geofile
2005-3McDonough W F amp Sun S (1995) The composition of the Earth
Chemical Geology 120 223^253Monger JW H amp Journeay J M (1994) Basement geology and tec-
tonic evolution of theVancouver region In Monger JW H (ed)Geology and Geological Hazards of the Vancouver Region Southwestern
British Columbia Geological Survey of Canada Bulletin 481 3^25Muller J E (1977) Geology of Vancouver Island Geological Survey of
Canada Open File Map 463Muller J E (1980)The Paleozoic Sicker Group of Vancouver Island British
Columbia Geological Survey of Canada Paper 79-30 22 ppMuller J E Northcote K E amp Carlisle D (1974) Geology and
mineral deposits of Alert Bay^Cape Scott map area VancouverIsland British Columbia Geological Survey of Canada Paper 74-8 77
Neal C R Mahoney J J Kroenke L W Duncan R A ampPetterson M G (1997) The Ontong^Java Plateau InMahoney J J amp Coffin M F (eds) Large Igneous Provinces
Continental Oceanic and Planetary FloodVolcanism American Geophysical
Union Geophysical Monograph 100 183^216Niu Y Collerson K D Batiza R Wendt J I amp Regelous M
(1999) Origin of enriched-typed mid-ocean ridge basalt at ridgesfar from mantle plumes The East Pacific Rise at 11820rsquoN Journalof Geophysical Research 104 7067^7088
Nixon G T amp Orr A J (2007) Recent revisions to the EarlyMesozoic stratigraphy of Northern Vancouver Island (NTS 102I092L) and metallogenic implicatinos British Columbia InGrant B (ed) Geological Fieldwork 2006 British Columbia Ministry of
Energy Mines and Petroleum Resources Paper 2007-1 163^177Nixon GT Hammack J L KoyanagiV M Payie G J Haggart
JW Orchard M JTozerT Friedman R M Archibald D APalfy J amp Cordey F (2006a) Geology of the Quatsino^PortMcNeill area northernVancouver Island British Columbia Ministry
of Energy Mines and Petroleum Resources Geoscience Map 2006-2 scale150 000
Nixon G T Kelman M C Stevenson D Stokes L A ampJohnston K A (2006b) Preliminary geology of the Nimpkishmap area (NTS 092L07) northern Vancouver Island BritishColumbia In Grant B amp Newell J M (eds) Geological Fieldwork2005 British Columbia Ministry of Energy Mines and Petroleum Resources
Paper 2006-1 135^152Nixon G T Laroque J Pals A Styan J Greene A R amp
Scoates J S (2008) High-Mg lavas in the Karmutsen flood basaltsnorthernVancouver Island (NTS 092L) Stratigraphic setting andmetallogenic significance In Grant B (ed) Geological Fieldwork
2007 British Columbia Ministry of Energy Mines and Petroleum
Resources Paper 2008-1 175^190Nowell G M Kempton P D Noble S R Fitton J G
Saunders A Mahoney J J amp Taylor R N (1998) High precisionHf isotope measurements of MORB and OIB by thermal ionisation
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
36
mass spectrometry insights into the depleted mantle Chemical
Geology 149 211^233Parrish R R amp McNicoll V J (1992) U^Pb age determinations
from the southern Vancouver Island area British Columbia InRadiogenic Age and Isotopic Studies Report 5 Geological Survey of
Canada Paper 91-2 79^86Peate DW (1997) The Parana^Etendeka Province In Coffin M F
amp Mahoney J (eds) Large Igneous Provinces Continental Oceanic andPlanetary Flood Volcanism American Geophysical Union Geophysical
Monograph 100 217^245Perfit M R Fornari D J Ridley W I Kirk P D Casey J
Kastens K A Reynolds J R Edwards M Desonie DShuster R amp Paradis S (1996) Recent volcanism in the Siqueirostransform fault picritic basalts and implications for MORBmagma genesis Earth and Planetary Science Letters 141(1^4) 91^108
Pretorius W Weis D Williams G Hanano D Kieffer B ampScoates J S (2006) Complete trace elemental characterization ofgranitoid (USGSG-2GSP-2) reference materials by high resolutioninductively coupled plasma-mass spectrometry Geostandards and
Geoanalytical Research 30(1) 39^54Regelous M Niu Y Wendt J I Batiza R Greig A amp
Collerson K D (1999) Variations in the geochemistry of magma-tism on the East Pacific Rise at 10830rsquoN since 800 ka Earth and
Planetary Science Letters 168 45^63Recurren villon S Chauvel C Arndt N T Pik R Martineau F
Fourcade S amp Marty B (2002) Heterogeneity of the Caribbeanplateau mantle source Sr O and He isotopic compositions of oli-vine and clinopyroxene from Gorgona Island Earth and Planetary
Science Letters 205(1^2) 91Rhodes J M (1996) Geochemical stratigraphy of lava flows samples
by the Hawaii Scientific Drilling Project Journal of Geophysical
Research 101(B5) 11729^11746Rhodes J M amp Vollinger M J (2004) Composition of basaltic lavas
sampled by phase-2 of the Hawaii Scientific Drilling ProjectGeochemical stratigraphy and magma types Geochemistry
Geophysics Geosystems 5(3) doi1010292002GC000434Richards M A Jones D L Duncan R A amp DePaolo D J (1991)
A mantle plume initiation model for the Wrangellia flood basaltand other oceanic plateaus Science 254 263^267
Salters V J M amp Hart S R (1991) The mantle sources of oceanridges islands and arcs the Hf-isotope connection Earth and
Planetary Science Letters 104 364^380Salters V J M amp Stracke A (2004) Composition of the depleted
SaltersV J amp WhiteW (1998) Hf isotope constraints on mantle evo-lution Chemical Geology 145 447^460
SanoT amp Yamashita S (2004) Experimental petrology of basementlavas from Ocean Drilling Program Leg 192 implications for dif-ferentiation processes in Ontong Java Plateau magmas InFitton J G Mahoney J JWallace P J amp Saunders A D (eds)Origin and Evolution of the Ontong Java Plateau Geological Society
London Special Publications 229 185^218Saunders A D (2005) Large igneous provinces Origin and environ-
mental consequences Elements 1 259^263Self S ThordarsonT amp Keszthely L (1997) Emplacement of conti-
nental flood basalt lava flows In Mahoney J J amp Coffin M F(eds) Large Igneous Provinces Continental Oceanic and Planetary Flood
Volcanism American Geophysical Union Geophysical Monograph 100381^410
Shaw D M (2000) Continuous (dynamic) melting theory revisitedCanadian Mineralogist 38(5) 1041^1063
Sluggett C L (2003) Uranium^lead age and geochemical constraintson Paleozoic and Early Mesozoic magmatism in WrangelliaTerrane Saltspring Island British Columbia BSc thesisUniversity of British ColumbiaVancouver 84 pp
Storey M Mahoney J J amp Saunders A D (1997) Cretaceousbasalts in Madagascar and the transition between plume and con-tinental lithospheric mantle sources In Mahoney J J ampCoffin M F (eds) Large Igneous Provinces Continental Oceanic and
Planetary Flood Volcanism Aamerican Geophysical Union Geophysical
Monograph 100 95^122Surdam R C (1967) Low-grade metamorphism of the Karmutsen
Group PhD dissertation University of California Los Angeles288 pp
Sutherland-Brown A Yorath C J Anderson R G amp Dom K(1986) Geological maps of southern Vancouver IslandLITHOPROBE I 92C10 11 14 16 92F1 2 7 8 Geological Survey ofCanada Open File 1272
Tejada M L G Mahoney J J Castillo P R Ingle S P Sheth HC amp Weis D (2004) Pin-pricking the elephant evidence on theorigin of the Ontong Java Plateau from Pb^Sr^Hf^Nd isotopiccharacteristics of ODP Leg 192 basalts In Fitton J GMahoney J J Wallace P J amp Saunders A D (eds) Origin and
Evolution of the Ontong Java Plateau Geological Society London Special
Publications 229 133^150Thompson P M E Kempton P D amp Kerr A C (2008) Evaluation
of the effects of alteration and leaching on Sm^Nd and Lu^Hf sys-tematics in submarine mafic rocks Lithos 104(1^4) 164^176
Thompson P M E Kempton P D White R V Kerr A CTarney J Saunders A D Fitton J G amp McBirney A (2003)Hf-Nd isotope constraints on the origin of the CretaceousCaribbean plateau and its relationship to the Galapagos plumeEarth and Planetary Science Letters 217(1^2) 59^75
Vervoort J D Patchett P Blichert-Toft J amp Albare de F (1999)Relationships between Lu^Hf and Sm^Nd isotopic systems in theglobal sedimentary system Earth and Planetary Science Letters 16879^99
Walter M J (1998) Melting of garnet peridotite and the origin ofkomatiite and depleted lithosphere Journal of Petrology 39(1) 29^60
Weis D Kieffer B Maerschalk C Barling J de Jong JWilliams G A Hanano D Mattielli N Scoates J SGoolaerts A Friedman R A amp Mahoney J B (2006) High-pre-cision isotopic characterization of USGS reference materials byTIMS and MC-ICP-MS Geochemistry Geophysics Geosystems
7(Q08006) doi1010292006GC001283Weis D Kieffer B Hanano D Silva I N Barling J
Pretorius W Maerschalk C amp Mattielli N (2007) Hf isotopecompositions of US Geological Survey reference materialsGeochemistry Geophysics Geosystems 8(Q06006) doi1010292006GC001473
Wheeler J O amp McFeely P (1991) Tectonic assemblage map of theCanadian Cordillera and adjacent part of the United States ofAmerica Geological Survey of Canada Map 1712A
WhiteW M Albare de F amp Tecurren louk P (2000) High-precision analy-sis of Pb isotope ratios by multi-collector ICP-MS Chemical Geology167 257^270
Yorath C J Sutherland Brown A amp Massey N W D (1999)LITHOPROBE southern Vancouver Island British Columbia Geological
Survey of Canada Bulletin 498 145 ppZou H (1998) Trace element fractionation during modal and
non-model dynamic melting and open-system melting Amathematical treatment Geochimica et Cosmochimica Acta 62(11)1937^1945
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
37
Zou H amp Reid M R (2001) Quantitative modeling of trace elementfractionation during incongruent dynamic melting Geochimica et
Cosmochimica Acta 65(1) 153^162
APPENDIX SAMPLEPREPARAT ION ANDANALYT ICAL METHODSOnly the freshest rocks in the field were sampled and onlythe least altered samples were selected for chemical andisotopic analysis based on thorough petrographic inspec-tion Fifty-nine of the 129 collected samples were crushed(400 g) into small pieces 52mm in diameter in aRocklabs hydraulic piston crusher between WC plates tominimize contamination The coarse-crush was mixedand 100 g was powdered in a planetary mill using agatejars and balls cleaned with quartz sand between samples
ActLabs analytical methodsThe major- and trace-element compositions of the whole-rock powders were determined at Activation LaboratoriesLtd (Actlabs) in Ancaster Ontario Analytical techniquesand detection limits are also available from Actlabs (httpwwwactlabscommethsub_code4erehtm) The analyticalmethod for each of the elements analyzed is indicated inTable 2 For the major elements a 02 g sample was mixedwith a lithium metaborate^lithium tetraborate mixtureand fused in a graphite crucible The molten mixture waspoured into a 5 HNO3 solution and shaken until dis-solved (30min) The samples were analyzed for majoroxides and selected trace elements by inductively coupledplasma optical emission spectrometry (ICP-OES) on acombination simultaneous^sequential Thermo Jarrell-AshEnviro II system Internal calibration was achieved usinga variety of international reference materials (eg W-2BIR-1 DNC-1) and independent control samplesAdditional trace elements were analyzed by both instru-mental neutron activation analysis (INAA) and induc-tively couple plasma mass spectrometry (ICP-MS) ForINAA15^25 g of sample was weighed into small polyeth-ylene vials and irradiated with a control international ref-erence material CANMET WMS-1 and NiCr flux wires ata thermal neutron flux of 71012 neutronscm2 per s in theMcMaster Nuclear Reactor Following a 7 day waitingperiod the samples were measured on an Ortec high-purity Ge detector linked to a Canberra Series 95 multi-channel analyzer Activities for each element weredecay- and weight-corrected and compared with a detectorcalibration developed from multiple international certifiedreference materials For ICP-MS 025 g of sample wasdigested in HF followed by HNO3^HClO4 heated andtaken to dryness The samples were brought back into solu-tion with HCl Samples were analyzed using a PerkinElmer Optima 3000 ICP In-lab standards or certified ref-erence materials (eg W-2 BIR-1 DNC-1) were used for
quality control A total of 15 blind duplicates were ana-lyzed from the same sample powders to assessreproducibility
University of Massachusetts analyticalmethodsFourteen sample duplicate powders taken from the same100 g sample aliquot (high-MgO basalts and picrites)were also analyzed at the Ronald B Gilmore X-RayFluorescence (XRF) Laboratory at the University ofMassachusetts Major elements were measured on a fusedLa-bearing lithium borate glass disc using a SiemensMRS-400 spectrometer Trace element concentrations(Rb Sr Ba Ce Nb Zr Y Pb Zn Ga Ni Cr V) weremeasured on a separate powder pellet using a PhilipsPW2400 sequential spectrometer using a Rh tube LOIand ferrous iron measurements were made as describedby Rhodes amp Vollinger (2004) Precision and accuracy esti-mates for the data have been described by Rhodes (1996)and Rhodes amp Vollinger (2004) Results for each sampleare the average of two separate analyses and are shown inElectronic Appendix 2 for a comparison of analyses ofduplicate powders using XRFand ICP-MS
PCIGR analytical methodsA subset of 19 samples was selected for high-precisiontrace-element analysis and Sr Nd Pb and Hf isotopicanalysis at the Pacific Centre for Isotopic andGeochemical Research (PCIGR) at the University ofBritish Columbia (UBC) Samples were selected from the59 samples analyzed for whole-rock chemistry at ActLabsbased on major- and trace-element chemistry alteration(low LOI and petrographic alteration index) and samplelocation Samples were prepared for trace-element analysisat the PCIGR by the technique described by Pretorius et al(2006) on unleached rock powders Sample powders(100mg) were weighed in 7ml screw-top Savillex bea-kers and dissolved in 1ml 14N HNO3 and 5ml 48 HFon a hotplate for 48 h at 1308C with periodic ultrasonica-tion Samples were dried and redissolved in 6ml 6N HClon a hotplate for 24 h and then dried and redissolved in1ml concentrated HNO3 for 24 h before final dryingTrace element abundances were measured by high-resolution inductively coupled plasma-mass spectrometry(HR-ICP-MS) with aThermo Finnigan Element2 systemfollowing the procedures described by Pretorius et al(2006) within 24 h of redissolution HFSE and LILE weremeasured in medium-resolution mode at 2000 dilutionusing a PFATeflon spray chamber washed with aqua regiafor 3min between samples REE were measured in high-resolution mode and U and Pb in low-resolution modeat 2000 dilution using a glass spray chamber washedwith 2 HNO3 between samples Total procedural blanksand reference materials (BCR-2 BHVO-2) were analyzedwith the batch of samples Indium was used as an internal
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
38
standard in all samples and standard solutionsBackground and standard solutions were analyzed afterevery five samples to detect memory effects and massdrift Results for PCIGR trace-element analyses areshown in Electronic Appendix 3 for the full set of analyzedtrace elementsSample digestion for purification of Sr Nd Hf and Pb
for column chemistry began by weighing each samplepowder (400^500mg) prior to leaching All samples wereleached with 6N HCl and placed in an ultrasonic bath for15min Samples were rinsed twice with 18 M-cm H2Obetween each leaching step (15 total) until the supernatantliquid was clear [following the method of Mahoney (1987)]Samples were then dried on a hotplate for 24 h andweighed again Sample solutions were then prepared bydissolving 100^250mg of the leached powder dissolvedin 1ml 14N HNO3 and 10ml 48 HF on a hotplate for48 h at 1308C with periodic ultrasonication Samples weredried and redissolved in 6ml 6N HCl on a hotplate for24 h and then dried Pb was separated using anion exhangecolumns and the discard was used for Sr REE and Hfseparation Nd was separated from the other REE and Hfrequired two additional purification steps Detailed proce-dures for column chemistry for separating Sr Nd and Pbat the PCIGR have been described byWeis et al (2006) andHf purification has been described byWeis et al (2007) Srand Nd isotope ratios were measured by thermal ioniza-tion mass spectrometry (TIMS) on a Thermo FinniganTriton system in static mode with relay matrix rotation ona single Ta and double Re^Ta filament respectively Fourto five filaments per barrel of 21 were occupied by stan-dards (NBS 987 for Sr and LaJolla for Nd) for each barrelwhere samples were run Sample Sr and Nd isotopic com-positions were corrected for mass fractionation using86Sr88Srfrac14 01194 and 146Nd144Ndfrac14 07219 Each samplewas then normalized using the barrel average of the refer-ence material relative to the values of143Nd144Ndfrac14 0511858 and 87Sr86Srfrac14 0710248 (Weiset al 2006) During the course of the Vancouver Islandanalyses the La Jolla Nd standard gave an average valueof 05118566 (nfrac14 7) and NBS987 standard gave an aver-age of 07102408 (nfrac1411) (2 error is reported as times106) 147Sm144Nd errors are 15 or 0006 USGeological Survey (USGS) reference material BHVO-2was processed with the samples and yielded Sr and Nd iso-topic ratios of 07034607 and 05129786 respectivelyThese are in agreement with the published values of070347920 and 051298411 respectively (Weis et al2006)
Pb and Hf isotopic compositions were analyzed by mul-tiple collector-inductively coupled plasma-mass spectrome-try (MC-ICP-MS) on a static multi-collection on a NuPlasma (Nu Instruments) system The detailed analyticalprocedure for Pb isotopic analyses on the Nu Plasma atthe PCIGR has been described by Weis et al (2006) Theconfiguration for Pb analyses allows for collection of PbTl and Hg togetherTl and Hg are used to monitor instru-mental mass discrimination and isobaric overlap repec-tively All sample solutions were analyzed withapproximately the same PbTl ratio (4) as the referencematerial NIST SRM 981 To accomplish this a small ali-quot of each sample solution from the Pb columns was ana-lyzed on the Element2 to determine the precise amount ofPb available for analysis on the Nu Plasma The SRM 981standard was run after every two samples on the NuPlasma During the time samples were run analyses ofthe SRM 981 Pb reference material gave values of206Pb204Pbfrac1416940322 207Pb204Pbfrac1415495823and 208Pb204Pbfrac14 36713164 (nfrac14 61 2 error is reportedas times 104) these values are in excellent agreement withreported TIMS triple-spike values of Galer amp Abouchami(1998) Fractionation-corrected Pb isotopic ratios were fur-ther corrected by the sample-standard bracketing methodor the ln^ln correction method described by White et al(2000) and Blichert-Toft et al (2003) Leached powders ofUSGS reference material BHVO-2 yielded Pb isotopicratios of 206Pb204Pbfrac14186454 8 207Pb204Pbfrac14 1549105 and 208Pb204Pbfrac14 38222514 and BCR-2yielded 206Pb204Pbfrac141880466 207Pb204Pbfrac14 1562518 and 208Pb204Pbfrac14 3883496 (2 error isreported as times 104) These values are in agreement withthose for leached residues of BHVO-2 and BCR-2 fromWeis et al (2006)Hf isotopic compositions were analyzed following the
procedures detailed byWeis et al (2007) The configurationfor Hf analyses monitored Lu mass 175 andYb mass 172 toallow for interference correction to masses 174 and 176 Hfisotopic ratios were normalized internally for mass fractio-nation to a 179Hf177Hf ratio of 07325 using an exponentialcorrection Standards were run after every two samplesand sample results were normalized to the ratio of the in-run daily average and a 176Hf177Hf ratio for JMC-475 of0282160 During the course of analyses the Hf standardJMC-475 gave an average value 02821679 (nfrac14130)USGS reference materials BCR-2 and BHVO-2 were pro-cessed with the samples and yielded Hf isotopic ratios of02828675 and 02831005 respectively Publishedvalues for BCR-2 and BHVO-2 are 02828717 and0283104 8 respectively (Weis et al 2007)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
1Ni concentrations for these samples were determined by XRFTHOL tholeiitic basalt PIC picrite HI-MG high MgO basalt CG coarse-grained (sill or gabbro) MIN SIL mineralizedsill OUTLIER anomalous pillowed flow in plots MA Mount Arrowsmith SL Schoen Lake KR Karmutsen Range QIQuadra Island Sample locations are given using the Universal Transverse Mercator (UTM) coordinate system (NAD83zones 9 and 10) Analyses were performed at Activation Laboratory (ActLabs) Fe2O3
is total iron expressed as Fe2O3LOI loss on ignition All major elements Sr V and Y were measured by ICP quadrupole OES on solutions of fusedsamples Cu Ni Pb and Zn were measured by total dilution ICP Cs Ga Ge Hf Nb Rb Ta Th U Zr and REE weremeasured by magnetic-sector ICP on solutions of fused samples Co Cr and Sc were measured by INAA Blanks arebelow detection limit See Electronic Appendices 2 and 3 for complete XRF data and PCIGR trace element datarespectively Major elements for sample 93G17 are normalized anhydrous Sample 5615A7 was analyzed by XRFSamples from Quadra Island are not shown in figures
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
17
picritic and high-MgO basalt pillow lavas extend to20wt MgO (Fig 6 and references listed in thecaption)The tholeiitic basalts from the Karmutsen Range dis-
play a tight range of parallel light rare earth element(LREE)-enriched REE patterns (LaYbCNfrac14 21^24mean 2202) whereas the picrites and high-MgObasalts are characterized by sub-parallel LREE-depleted
patterns (LaYbCNfrac14 03^07 mean 06 02) with lowerREE abundances than the tholeiitic basalts (Fig 7)Tholeiitic basalts from the Schoen Lake and MountArrowsmith areas have similar REE patterns tosamples from the Karmutsen Range (Schoen Lake LaYbCNfrac14 21^26 mean 2303 Mount Arrowsmith LaYbCNfrac1419^25 mean 2204) and the coarse-grainedmafic rocks have similar REE patterns to the tholeiitic
Depleted MORB average(Salters amp Stracke 2004)
Schoen LakeSchoen Lake
Mount Arrowsmith
Karmutsen RangeS
ampl
eC
hond
rite
Sam
ple
Cho
ndrit
e
Sam
ple
Prim
itive
Man
tleS
ampl
eP
rimiti
ve M
antle
Sam
ple
Prim
itive
Man
tle
Mount Arrowsmith
Karmutsen Range
(a) (b)
(c) (d)
(e) (f )
Sam
ple
Cho
ndrit
e
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained mafic rock
Pillowed flowMineralized sill
1
10
100
1
10
100
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
01
1
10
100
1
10
100
1
10
100
Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
Fig 7 Whole-rock REE and other incompatible-element concentrations for the Karmutsen Formation (a) (c) and (e) are chondrite-normal-ized REE patterns for samples from each of the three main field areas onVancouver Island (b) (d) and (f) are primitive mantle-normalizedtrace-element patterns for samples from each of the three main field areas All normalization values are from McDonough amp Sun (1995) Theclear distinction between LREE-enriched tholeiitic basalts and LREE-depleted picrites the effects of alteration on the LILE (especially loss ofK and Rb) and the different range for the vertical scale in (b) (extends down to 01) should be noted
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
18
basalts (LaYbCNfrac14 21^29 mean 2508) Two LREE-depleted coarse-grained mafic rocks from the SchoenLake area have similar REE patterns (LaYbCNfrac14 07) tothe picrites from the Keogh Lake area indicating that thisdistinctive suite of rocks is exposed as far south asSchoen Lake that is 75 km away (Fig 7) The anomalouspillowed flow from the Karmutsen Range has lower LaYbCN (13^18) than the main group of tholeiitic basaltsfrom the Karmutsen Range (Fig 7) The mineralized sillfrom the Schoen Lake area is distinctly LREE-enriched(LaYbCNfrac14 33) with the highest REE abundances(LaCNfrac14 940) of all Karmutsen samples this may reflectcontamination by adjacent sediment also evident in thin-section where sulfide blebs are cored by quartz (Greeneet al 2006)The primitive mantle-normalized trace-element pat-
terns (Fig 7) and trace-element variations and ratios(Fig 8) highlight the differences in trace-element
concentrations between the four main groups ofKarmutsen flood basalts onVancouver Island The tholeii-tic basalts from all three areas have relatively smooth par-allel trace-element patterns some samples have smallpositive Sr anomalies relative to Nd and Sm and manysamples have prominent negative K anomalies relative toU and Nb reflecting K loss during alteration (Fig 7) Thelarge ion lithophile elements (LILE) in the tholeiiticbasalts (mainly Rb and K not Cs) are depleted relativeto the high field strength elements (HFSE) and LREELILE segments especially for the Schoen Lake area(Fig 7d) are remarkably parallel The picrites andhigh-MgO basalts form a tight band of parallel trace-element patterns and are depleted in HFSE and LREEwith positive Sr anomalies and relatively low Rb Thecoarse-grained mafic rocks from the Karmutsen Rangehave identical trace-element patterns to the tholeiiticbasalts Samples from the Schoen Lake area have similar
000
025
050
075
100
125
0 1 2 3 4 5
00
04
08
12
16
0 5 10 15 20 25
MgO (wt)
0
5
10
15
0 50 100 150
Hf (ppm)
Th (ppm)
NbLaNb (ppm)
Zr (ppm)
(a)
(c)
(b)
(d)
Th (ppm)
00
02
04
06
08
10
12
00 01 02 03 04 05
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
NbLa=1
4723A4 5615A12
U (ppm)
Fig 8 Whole-rock trace-element concentrations and ratios for the Karmutsen Formation [except (b) which is vs wt MgO] (a) Nb vs Zr(b) NbLa vs MgO (c) Th vs Hf (d) Th vs U There is a clear distinction between the tholeiitic basalts and picrites in Nb and Zr both inconcentration and the slope of each trend
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
19
trace-element patterns to the Karmutsen Range except forthe two LREE-depleted coarse-grained mafic rocks andthe mineralized sill (Fig 7d)
Sr^Nd^Hf^Pb isotopic compositionsA total of 19 samples from the four main groups ofthe Karmutsen Formation were selected for radiogenicisotopic geochemistry on the basis of major- and trace-element variation stratigraphic position and geographicallocation The samples have overlapping age-correctedHf and Nd isotope ratios and distinct ranges of Sr isotopiccompositions (Fig 9) The tholeiitic basalts picriteshigh-MgO basalt and coarse-grained mafic rocks have
initial eHffrac14thorn87 to thorn126 and eNdfrac14thorn66 to thorn88 cor-rected for in situ radioactive decay since 230 Ma (Fig 9Tables 3 and 4) All the Karmutsen samples form a well-defined linear array in a Lu^Hf isochron diagram corre-sponding to an age of 24115 Ma (Fig 9) This is withinerror of the accepted age of the Karmutsen Formationand indicates that the Hf isotope systematics behaved as aclosed system since c 230 Ma The anomalous pillowedflow has similar initial eHf (thorn98) and slightly lower initialeNd (thorn62) compared with the other Karmutsen samplesThe picrites and high-MgO basalt have higher initial87Sr86Sr (070398^070518) than the tholeiitic basalts(initial 87Sr86Srfrac14 070306^070381) and the coarse-grained mafic rocks (initial 87Sr86Srfrac14 070265^070428)
05128
05129
05130
05131
05132
015 017 019 021 023 025
02829
02830
02831
02832
02833
02834
000 002 004 006 008 010
4
6
8
10
0702 0703 0704 0705 07066
8
10
12
14
5 6 7 8 9 10
12
0
1
2
3
4
5
6
7
0702 0703 0704 0705 0706
(d)
LOI (wt )
87Sr 86Sr
4723A2
Nd
143Nd144Nd
147Sm144Nd
87Sr 86Sr (230 Ma)
(230 Ma)
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
(a)
(c)
176Hf177Hf
Age=241 plusmn 15 Ma=+990 plusmn 064
176Lu177Hf
Hf (230 Ma)
Nd (230 Ma)
(e)
Hf (i)
(b)
Fig 9 Whole-rock Sr Nd and Hf isotopic compositions for the Karmutsen Formation (a) 143Nd144Nd vs 147Sm144Nd (b) 176Hf177Hf vs176Lu177Hf The slope of the best-fit line for all samples corresponds to an age of 24115 Ma (c) Initial eNd vs 87Sr86Sr Age-correction to230 Ma (d) LOI vs 87Sr86Sr (e) Initial eHf vs eNd Average 2 error bars are shown in a corner of each panel Complete chemical duplicatesshown inTables 3 and 4 (samples 4720A7 and 4722A4) are circled in each plot
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
overlap the ranges of the picrites and tholeiitic basalts(Fig 9 Table 3)The measured Pb isotopic compositions of the tholeiitic
basalts are more radiogenic than those of the picrites andthe most magnesian Keogh Lake picrites have the leastradiogenic Pb isotopic compositions The range ofinitial Pb isotope ratios for the picrites is206Pb204Pbfrac1417868^18580 207Pb204Pbfrac1415547^15562and 208Pb204Pbfrac14 37873^38257 and the range for thetholeiitic basalts is 206Pb204Pbfrac1418782^19098207Pb204Pbfrac1415570^15584 and 208Pb204Pbfrac14 37312^38587 (Fig 10 Table 5) The coarse-grained mafic rocksoverlap the range of initial Pb isotopic compositions forthe tholeiitic basalts with 206Pb204Pbfrac1418652^19155207Pb204Pbfrac1415568^15588 and 208Pb204Pbfrac14 38153^38735 One high-MgO basalt has the highest initial206Pb204Pb (19242) and 207Pb204Pb (15590) and thelowest initial 208Pb204Pb (37676) The Pb isotopic ratiosfor Karmutsen samples define broadly linearrelationships in Pb isotope plots The age-corrected Pb
isotopic compositions of several picrites have been affectedby U and Pb (andor Th) mobility during alteration(Fig 10)
ALTERAT IONThe Karmutsen basalts have retained most of their origi-nal igneous structures and textures however secondaryalteration and low-grade metamorphism have producedzeolitic- and prehnite^pumpellyite-bearing mineral assem-blages (Table 1 Cho et al 1987) Alteration and metamor-phism have primarily affected the distribution of the LILEand the Sr isotopic systematics The Keogh Lake picriteshave been affected more by alteration than the tholeiiticbasalts and have higher LOI (up to 55wt ) variableLILE and higher measured Sr Nd and Hf and lowermeasured Pb isotope ratios than the tholeiitic basalts Srand Pb isotope compositions for picrites and tholeiiticbasalts show a relationship with LOI (Figs 9 and 10)whereas Nd and Hf isotopic compositions do not correlate
Table 3 Sr and Nd isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Rb Sr 87Sr86Sr 2m87Rb86Sr 87Sr86Srt Sm Nd 143Nd144Nd 2m eNd
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses carried out at the PCIGR thecomplete trace element analyses are given in Electronic Appendix 3
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
21
with LOI (not shown) The relatively high apparent initialSr isotopic compositions for the high-MgO lavas (up to07052) probably resulted from post-magmatic loss of Rbor an increase in 87Sr86Sr through addition of seawater Sr(eg Hauff et al 2003) High-MgO lavas from theCaribbean plateau have similarly high 87Sr86Sr comparedwith basalts (Recurren villon et al 2002)The correction for in situdecay on initial Pb isotopic ratios has been affected bymobilization of U and Th (and Pb) in the whole-rockssince their formation A thorough acid leaching duringsample preparation was used that has been shown to effec-tively remove alteration phases (Weis et al 2006 NobreSilva et al in preparation) Whole-rock SmNd ratiosand age-correction of Nd isotopes may be affected bythe leaching procedure for submarine rocks older than50 Ma (Thompson et al 2008 Fig 9) there ishowever very little variation in Nd isotopic compositionsof the picrites or tholeiitic basalts and this system does notappear to be disturbed by alteration The HFSE abun-dances for tholeiitic and picritic basalts exhibit clear
linear correlations in binary diagrams (Fig 8) whereasplots of LILE vs HFSE are highly scattered (not shown)because of the mobility of some of the alkali and alkalineearth LILE (Cs Rb Ba K) and Sr for most samplesduring alterationThe degree of alteration in the Karmutsen samples does
not appear to be related to eruption environment or depthof burial Pillow rims are compositionally different frompillow cores (Surdam1967 Kuniyoshi1972) and aquagenetuffs are chemically different from isolated pillows withinpillow breccias however submarine and subaerial basaltsexhibit a similar degree of alterationThere is no clear cor-relation between the submarine and subaerial basalts andsome commonly used chemical alteration indices (eg BaRb vs K2OP2O5 Huang amp Frey 2005) There is also nodefinitive correlation between the petrographic alterationindex (Table 1) and chemical alteration indices although10 of 13 tholeiitic basalts with the highest K and LILEabundances have the highest petrographic alterationindex of three (seeTable 1)
Table 4 Hf isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Lu Hf 176Hf177Hf 2m eHf176Lu177Hf 176Hf177Hft eHf(t)
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
OLIV INE ACCUMULAT ION INH IGH-MgO AND PICR IT IC LAVASGeochemical trends and petrographic characteristics indi-cate that accumulation of olivine played an important rolein the formation of the high-MgO basalts and picrites fromthe Keogh Lake area on northern Vancouver Island TheKeogh Lake samples show a strong linear correlation inplots of Al2O3 TiO2 ScYb and Ni vs MgO and many ofthe picrites have abundant clusters of olivine pseudomorphs(Table1 Fig11)There is a strong linear correlationbetween
modal per cent olivine and whole-rock magnesium contentmagnesium contents range between 10 and 19wt MgOand the proportion of olivine phenocrysts in these samplesvaries from 0 to 42 vol (Fig 11)With a few exceptionsboth the size range and median size of the olivine pheno-crysts in most samples are comparable The clear correla-tion between the proportion of olivine phenocrysts andwhole-rock magnesium contents indicates that accumula-tion of olivine was directly responsible for the high MgOcontents (410wt ) in most of the picritic lavas
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
207Pb204Pb
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
208Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
(a) (b)
(c) (e)
4723A24723A2
4723A2
4723A2
4722A4
4723A4
4722A4
4723A4
4723A134723A3
4723A3
4723A13
1556
1558
1560
1562
1564
1566
1568
186 190 194 198 202 206 2101550
1552
1554
1556
1558
1560
178 180 182 184 186 188 190 192 194
380
384
388
392
396
186 190 194 198 202 206 210374
378
382
386
390
178 180 182 184 186 188 190 192 194
206Pb204Pb(d)
LOI (wt )
0
1
2
3
4
5
6
7
186 190 194 198 202 206 210
4723A2
Fig 10 Pb isotopic compositions of leached whole-rock samples measured by MC-ICP-MS for the Karmutsen Formation Error bars are smal-ler than symbols (a) Measured 207Pb204Pb vs 206Pb204Pb (b) Initial 207Pb204Pb vs 206Pb204Pb Age-correction to 230 Ma (c) Measured208Pb204Pb vs 206Pb204Pb (d) LOI vs 206Pb204Pb (e) Initial 208Pb204Pb vs 206Pb204Pb Complete chemical duplicates shown in Table 5(samples 4720A7 and 4722A4) are circled in each plot The dashed lines in (b) and (e) show the differences in age-corrections for two picrites(4723A4 4722A4) using the measured UTh and Pb concentrations for each sample and the age-corrections when concentrations are used fromthe two picrites (4723A3 4723A13) that appear to be least affected by alteration
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
23
DISCUSS IONThe compositional and spatial development of the eruptedlava sequences of the Wrangellia oceanic plateau onVancouver Island provide information about the evolutionof a mantle plume upwelling beneath oceanic lithospherein an analogous way to the interpretation of continentalflood basalts The Caribbean and Ontong Java plateausare the two primary examples of accreted parts of oceanicplateaus that have been studied to date and comparisonscan be made with the Wrangellia oceanic plateauHowever the Wrangellia oceanic plateau is a distinctiveoccurrence of an oceanic plateau because it formed on thecrust of a Devonian to Mississippian intra-oceanic islandarc overlain by Mississippian to Permian limestones andpelagic sediments The nature and thickness of oceaniclithospheric mantle are considered to play a fundamentalrole in the decompression and evolution of a mantleplume and thus the compositions of the erupted lavasequences (Greene et al 2008) The geochemical and
stratigraphic relationships observed in the KarmutsenFormation onVancouver Island and the focus of the subse-quent discussion sections provide constraints on (1) thetemperature and degree of melting required to generatethe primary picritic magmas (2) the composition of themantle source (3) the depth of melting and residual min-eralogy in the source region (4) the low-pressure evolutionof the Karmutsen tholeiitic basalts (5) the growth of theWrangellia oceanic plateau
Melting conditions and major-elementcomposition of the primary magmasThe primary magmas to most flood basalt provinces arebelieved to be picritic (eg Cox 1980) and they have beenfound in many continental and oceanic flood basaltsequences worldwide (eg Siberia Karoo Paranacurren ^Etendeka Caribbean Deccan Saunders 2005) Near-pri-mary picritic lavas with low total alkali abundances
Table 5 Pb isotopic compositions of Karmutsen basaltsVancouver Island BC
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
require high-degree partial melting of the mantle source(eg Herzberg amp OrsquoHara 2002) The Keogh Lake picritesfrom northernVancouver Island are the best candidates foridentifying the least modified partial melts of the mantleplume source despite having accumulated olivine pheno-crysts and can be used to estimate conditions of meltingand the composition of the primary magmas for theKarmutsen FormationPrimary magma compositions mantle potential tem-
peratures and source melt fractions for the picritic lavas ofthe Karmutsen Formation were calculated from primitivewhole-rock compositions using PRIMELT1XLS software(Herzberg et al 2007) and are presented in Fig 12 andTable 6 A detailed discussion of the computationalmethod has been given by Herzberg amp OrsquoHara (2002)and Herzberg et al (2007) key features of the approachare outlined belowPrimary magma composition is calculated for a primi-
tive lava by incremental addition of olivine and compari-son with primary melts of mantle peridotite Lavas thathad experienced plagioclase andor clinopyroxene fractio-nation are excluded from this analysis All calculated pri-mary magma compositions are assumed to be derived byaccumulated fractional melting of fertile peridotite KR-4003 Uncertainties in fertile peridotite composition donot propagate to significant variations in melt fractionand mantle potential temperature melting of depletedperidotite propagates to calculated melt fractions that aretoo high but with a negligible error in mantle potential
temperature PRIMELT1XLS calculates the olivine liqui-dus temperature at 1atm which should be similar to theactual eruption temperature of the primary magmaand is accurate to 308C at the 2 level of confidenceMantle potential temperature is model dependent witha precision that is similar to the accuracy of the olivineliquidus temperature (C Herzberg personal communica-tion 2008)With regard to the evidence for accumulated olivine in
the Keogh Lake picrites it should not matter whether themodeled lava composition is aphyric or laden with accu-mulated olivine phenocrysts PRIMELT1 computes theprimary magma composition by adding olivine to a lavathat has experienced olivine fractionation in this way itinverts the effects of fractional crystallization The onlyrequirement is that the olivine phenocrysts are not acci-dental or exotic to the lava compositions Support for thisprocedure can be seen in the calculated olivine liquid-line-of-descent shown by the curved arrays with 5 olivineaddition increments (Fig 12c) Fractional crystallization ofolivine from model primary magmas having 85^95wt FeO and 153^175wt MgO will yield arrays of deriva-tive liquids and randomly sorted olivines that in most butnot all cases connect the picrites and high-MgO basalts Anewer version of the PRIMELT software (Herzberg ampAsimow 2008) can perform olivine addition to and sub-traction from lavas with highly variably olivine phenocrystcontents and yields results that converge with those shownin Fig 12
Fig 11 Relationship between abundance of olivine phenocrysts and whole-rock MgO contents for the Keogh Lake picrites (a) Tracings ofolivine grains (gray) in six high-MgO samples made from high resolution (2400 dpi) scans of petrographic thin-sections Scale bars represent10mm sample numbers are indicated at the upper left (b) Modal olivine vs whole-rock MgO Continuous line is the best-fit line for 10 high-MgO samplesThe areal proportion of olivine was calculated from raster images of olivine grains using ImageJ image analysis software whichprovides an acceptable estimation of the modal abundance (eg Chayes 1954)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
25
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
Gray lines = final melting pressure
L + Ol + Opx
07
08
09
Pressure (GPa)
10
3
4
5
6
1
SOLIDUS
01
04
0506
L + Ol
2
03
02
PeridotiteKR-4003
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
Thick black lines = initial melting pressure
Dashed lines = melt fraction
Melt Fraction
2
3 4 5 6
L+Ol+Opx
SolidusSpinel
SolidusGarnet Peridotite
7
L+Ol
0 10 20 30 4040
45
50
55
60
MgO (wt)
SiO2
(wt)
PeridotiteKR-4003
Melt Fraction
0 10 20 305
10
15
CaO(wt)
MgO (wt)
3
4 5 6 7
2
46
2
Peridotite Partial Melts
Thick black line = initial melting pressureGray lines = final melting pressure
Peridotite
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
02
Pressure (GPa)
0302
04
Pressure (GPa)
Picrite
High-MgO basalt
Tholeiitic basalt
(c)
(d)
Karmutsen flood basaltsOlivine addition modelOlivine addition (5 increment)Primary magma
PicriteHigh-MgO basaltTholeiitic basalt
(a)
(b)
800 850 900
910
920
930
940
950
Karmutsen flood basalts
Olivine addition modelOlivine addition (5 increment)Primary magma
Gray lines = Fo contents of olivine
Fig 12 Estimated primary magma compositions for three Keogh Lake picrites and high-MgO basalts (samples 4723A4 4723A135616A7) using the forward and inverse modeling technique of Herzberg et al (2007) Karmutsen compositions and modeling results areoverlain on diagrams provided by C Herzberg (a) Whole-rock FeO vs MgO for Karmutsen samples from this study Total iron estimatedto be FeO is 090 Gray lines show olivine compositions that would precipitate from liquid of a given MgO^FeO composition Black lineswith crosses show results of olivine addition (inverse model) using PRIMELT1 (Herzberg et al 2007) Gray field highlights olivinecompositions with Fo contents of 91^92 predicted to precipitate (b) FeO vs MgO showing Karmutsen lava compositions and results ofa forward model for accumulated fractional melting of fertile peridotite KR-4003 (Kettle River Peridotite Walter 1998) (c) SiO2 vsMgO for Karmtusen lavas and model results (d) CaO vs MgO showing Karmtusen lava compositions and model results To brieflysummarize the technique [see Herzberg et al (2007) for complete description] potential parental magma compositions for the high-MgOlava series were selected (highest MgO and appropriate CaO) and using PRIMELT1 software olivine was incrementally added tothe selected compositions to show an array of potential primary magma compositions (inverse model) Then using PRIMELT1 theresults from the inverse model were compared with a range of accumulated fractional melts for fertile peridotite derived fromparameterization of the experimental results for KR-4003 from Walter (1998) forward model Herzberg amp OrsquoHara (2002) A melt fractionwas sought that was unique to both the inverse and forward models (Herzberg et al 2007) A unique solution was found when there was a com-mon melt fraction for both models in FeO^MgO and CaO^MgO^Al2O3^SiO2 (CMAS) projection space This modeling assumes thatolivine was the only phase crystallizing and ignores chromite precipitation and possible augite fractionation in the mantle (Herzbergamp OrsquoHara 2002) Results are best for a residue of spinel lherzolite (not pyroxenite) The presence of accumulated olivine in samples ofthe Keogh Lake picrites used as starting compositions does not significantly affect the results because we are modeling addition of olivine(see text for further description) The tholeiitic basalts cannot be used for modeling because they are all saturated in plagthorn cpxthornolThick black line for initial melting pressure at 4 GPa is not shown in (c) for clarity Gray area in (a) indicates parental magmas thatwill crystallize olivine with Fo 91^92 at the surface Dashed lines in (c) indicate melt fraction Solidi for spinel and garnet peridotite arelabelled in (c)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
The PRIMELT1 results indicate that if the Karmutsenpicritic lavas were derived from accumulated fractionalmelting of fertile peridotite the primary magmas wouldhave contained 15^17wt MgO and 10wt CaO(Fig 12 Table 6) formed from 23^27 partial meltingand would have crystallized olivine with a Fo content of 91 (Fig 12 Table 6) Calculated mantle temperatures forthe Karmutsen picrites (14908C) indicate melting ofanomalously hot mantle (100^2008C above ambient) com-pared with ambient mantle that produces mid-ocean ridgebasalt (MORB 1280^14008C Herzberg et al 2007)which initiated between 3 and 4 GPa Several picritesappear to be near-primary melts and required very littleaddition of olivine (eg sample 93G171 with measured158wt MgO) These temperature and melting esti-mates of the Karmutsen picrites are consistent withdecompression of hot mantle peridotite in an actively con-vecting plume head (ie plume initiation model) Threesamples (picrite 5614A1 and high-MgO basalts 5614A3and 5614A5) are lower in iron than the other Karmutsenlavas with FeO in the 65^80wt range (Fig 12c) andhave MgO and FeO contents that are similar to primitiveMORB from the Sequeiros fracture zone of the EastPacific Rise (EPR Perfit et al 1996) These samples
however differ from EPR MORB in having higher SiO2
and lower Al2O3 and they are restricted geographicallyto one area (Sara Lake Fig 2)We therefore have evidence for primary magmas with
MgO contents that range from a possible low of 110wt to as high as 174wt with inferred mantle potential tem-peratures from 1340 to 15208C This is similar to the rangedisplayed by lavas from the Ontong Java Plateau deter-mined by Herzberg (2004) who found that primarymagma compositions assuming a peridotite sourcewould contain 17wt MgO and 10wt CaO(Table 6) and that these magmas would have formed by27 melting at a mantle potential temperature of15008C (first melting occurring at 36 GPa) Fitton ampGodard (2004) estimated 30 partial melting for theOntong Java lavas based on the Zr contents of primarymagmas of Kroenke-type basalt calculated by incrementaladdition of equilibrium olivine However if a considerableamount of eclogite was involved in the formation of theOntong Java plateau magmas the excess temperatures esti-mated by Herzberg et al (2007) may not be required(Korenaga 2005) The variability in mantle potential tem-perature for the Keogh Lake picrites is also similar to thewide range of temperatures that have been calculated for
Table 6 Estimated primary magma compositions for Karmutsen basalts and other oceanic plateaus or islands
Sample 4723A4 4723A13 5616A7 93G171 Average OJP Mauna Keay Gorgonaz
Ontong Java primary magma composition for accumulated fraction melting (AFM) from Herzberg (2004)yMauna Kea primary magma composition is average of four samples of Herzberg (2006 table 1)zGorgona primary magma composition for AFM for 1F fertile source from Herzberg amp OrsquoHara (2002 table 4)Total iron estimated to be FeO is 090
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
27
lavas from single volcanoes in the Galapagos Hawaii andelsewhere by Herzberg amp Asimow (2008) Those workersinterpreted the range to reflect the transport of magmasfrom both the cool periphery and the hotter axis of aplume A thermally heterogeneous plume will yield pri-mary magmas with different MgO contents and the rangeof primary magma compositions and their inferred mantlepotential temperatures for Karmutsen lavas may reflectmelting from different parts of the plume
Source of the Karmutsen Formation lavasThe Sr^Nd^Hf^Pb isotopic compositions of theKarmutsen volcanic rocks on Vancouver Island indicatethat they formed from plume-type mantle containing adepleted component that is compositionally similar to butdistinct from other oceanic plateaus that formed in thePacific Ocean (eg Ontong Java and Caribbean Fig 13)Trace element and isotopic constraints can be used to testwhether the depleted component of the KarmutsenFormation was intrinsic to the mantle plume and distinctfrom the source of MORB or the result of mixing or invol-vement of ambient depleted MORB mantle with plume-type mantle The Karmutsen tholeiitic basalts have initialeNd and
87Sr86Sr ratios that fall within the range of basaltsfrom the Caribbean Plateau (eg Kerr et al 1996 1997Hauff et al 2000 Kerr 2003) and a range of Hawaiian iso-tope data (eg Frey et al 2005) and lie within and abovethe range for the Ontong Java Plateau (Tejada et al 2004Fig 13a) Karmutsen tholeiitic basalts with the highest ini-tial eNd and lowest initial 87Sr86Sr lie within and justbelow the field for age-corrected northern EPR MORB at230 Ma (eg Niu et al1999 Regelous et al1999) Initial Hfand Nd isotopic compositions for the Karmutsen samplesare noticeably offset to the low side of the ocean islandbasalt (OIB) array (Vervoort et al 1999) and lie belowand slightly overlap the low eHf end of fields for Hawaiiand the Caribbean Plateau (Fig 13c) the Karmutsen sam-ples mostly fall between and slightly overlap a field forage-corrected Pacific MORB at 230 Ma and Ontong Java(Mahoney et al 1992 1994 Nowell et al 1998 Chauvel ampBlichert-Toft 2001) Karmutsen lava Pb isotope ratios over-lap the broad range for the Caribbean Plateau and thelinear trends for the Karmutsen samples intersect thefields for EPR MORB Hawaii and Ontong Java in208Pb^206Pb space but do not intersect these fields in207Pb^206Pb space (Fig 13d and e) The more radiogenic207Pb204Pb for a given 206Pb204Pb of the Karmutsenbasalts requires high UPb ratios at an ancient time when235U was abundant similar to the Caribbean Plateau andenriched in comparison with EPR MORB Hawaii andOntong JavaThe low eHf^low eNd component of the Karmutsen
basalts that places them below the OIB mantle array andpartly within the field for age-corrected Pacific MORBcan form from low ratios of LuHf and high SmNd over
time (eg Salters amp White 1998) MORB will develop aHf^Nd isotopic signature over time that falls below themantle array Low LuHf can also lead to low 176Hf177Hffrom remelting of residues produced by melting in theabsence of garnet (eg Salters amp Hart 1991 Salters ampWhite 1998) The Hf^Nd isotopic compositions of theKarmutsen Formation indicate the possible involvementof depleted MORB mantle however similar to Iceland(Fitton et al 2003) Nb^Zr^Y variations provide a way todistinguish between a depleted component within theplume and a MORB-like component The Karmutsenbasalts and picrites fall within the Iceland array and donot overlap the MORB field in a logarithmic plot of NbYvs ZrY (Fig 13b) using the fields established by Fittonet al (1997) Similar to the depleted component within theCaribbean Plateau (Thompson et al 2003) the trend ofHf^Nd isotopic compositions towards Pacific MORB-likecompositions is not followed by MORB values in Nb^Zr^Y systematics and this suggests that the depleted compo-nent in the source of the Karmutsen basalts is distinctfrom the source of MORB and probably an intrinsic partof the plume The relatively radiogenic Pb isotopic ratiosand REE patterns distinct from MORB also support thisinterpretationTrace-element characteristics suggest the possible invol-
vement of sub-arc lithospheric mantle in the generation ofthe Karmutsen picrites Lassiter et al (1995) suggested thatmixing of a plume-type source with eNdthorn 6 to thorn 7 witharc material with low NbTh could reproduce the varia-tions observed in the Karmutsen basalts however theyconcluded that the absence of low NbLa ratios in theirsample of tholeiitic basalts restricts the amount of arc litho-sphere involved The study of Lassiter et al (1995) wasbased on major and trace element and Sr Nd and Pb iso-topic data for a suite of 29 samples from Buttle Lake in theStrathcona Provincial Park on central Vancouver Island(Fig 1) Whereas there are no clear HFSE depletions inthe Karmutsen tholeiitic basalts the low NbLa (and TaLa) of the Karmutsen picritic lavas in this study indicatethe possible involvement of Paleozoic arc lithosphere onVancouver Island (Fig 8)
REE modeling dynamic melting andsource mineralogyThe combined isotopic and trace-element variations of thepicritic and tholeiitic Karmutsen lavas indicate that differ-ences in trace elements may have originated from differentmelting histories For example the LuHf ratios of theKarmutsen picritic lavas (mean 008) are considerablyhigher than those in the tholeiitic basalts (mean 002)however the small range of overlapping initial eHf valuesfor the two lava suites suggests that these differences devel-oped during melting within the plume (Fig 9b) Below wemodel the effects of source mineralogy and depth of
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
28
melting on differences in trace-element concentrations inthe tholeiitic basalts and picritesDynamic melting models simulate progressive decom-
pression melting where some of the melt fraction isretained in the residue and only when the degree of partialmelting is greater than the critical mass porosity and the
source becomes permeable is excess melt extracted fromthe residue (Zou1998) For the Karmutsen lavas evolutionof trace element concentrations was simulated using theincongruent dynamic melting model developed by Zou ampReid (2001) Melting of the mantle at pressures greater than1 GPa involves incongruent melting (eg Longhi 2002)
OIB array
PacificMORB(230 Ma)
(0 Ma)
EPR(230 Ma)
-4
-2
0
2
4
8
10
12
14
16
18
20
22
-4 -2 0 2 4 6 8 10 12 14
minus2
0
2
4
6
8
10
12
14
0702 0703 0704 0705 0706
IndianMORB
87Sr 86Sr (initial)
Hf (initial)
(a) (c)
Nd (initial)
Karmutsen tholeiitic basalt
HawaiiOntong JavaCaribbean Plateau
Karmutsen high-MgO lava
high-MgO lavasKarmutsen
1540
1545
1550
1555
1560
1565
175 180 185 190 195 200375
380
385
390
395
175 180 185 190 195 200
(d) (e)
EPR
EPR
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb
Karmutsen Formation (initial)
HawaiiOntong JavaCaribbean Plateau
East Pacific Rise
Karmutsen Formation (measured)
Juan de FucaGorda
ε Nd
(initial)
Karmutsen Formation (different age-correction for 3 picrites)
(0 Ma)
NbY
001
01
1
1 10
ZrY
OIB
NMORB
(b)
Iceland array
6
Fig 13 Comparison of age-corrected (230 Ma) Sr^Nd^Hf^Pb isotopic compositions for Karmutsen flood basalts onVancouver Island withage-corrected OIB and MORB and Nb^Zr^Y variation (a) Initial eNd vs 87Sr86Sr (b) NbYand ZrY variation (after Fitton et al 1997)(c) Initial eHf vs eNd (d) Measured and initial 207Pb204Pb vs 206Pb204Pb (e) Measured and initial 208Pb204Pb vs 206Pb204Pb References fordata sources are listed in Electronic Appendix 4 Most of the compiled data were extracted from the GEOROC database (httpgeorocmpch-mainzgwdgdegeoroc) OIB array line in (c) is fromVervoort et al (1999) EPR East Pacific Rise Several high-MgO samples affected bysecondary alteration were not plotted for clarity in Pb isotope plots Estimation of fields for age-corrected Pacific MORB and EPR at 230 Mawere made using parent^daughter concentrations (in ppm) from Salters amp Stracke (2004) of Lufrac14 0063 Hffrac14 0202 Smfrac14 027 Ndfrac14 071Rbfrac14 0086 and Srfrac14 836 In the NbYvs ZrYplot the normal (N)-MORB and OIB fields not including Icelandic OIB are from data com-pilations of Fitton et al (1997 2003) The OIB field is data from 780 basalts (MgO45wt ) from major ocean islands given by Fitton et al(2003) The N-MORB field is based on data from the Reykjanes Ridge EPR and Southwest Indian Ridge from Fitton et al (1997) The twoparallel gray lines in (b) (Iceland array) are the limits of data for Icelandic rocks
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
29
with olivine being produced during the reactioncpxthornopxthorn spfrac14meltthornol for spinel peridotites The mod-eling used coefficients for incongruent melting reactionsbased on experiments on lherzolite melting (eg Kinzleramp Grove 1992 Longhi 2002) and was separated into (1)melting of spinel lherzolite (2) two combined intervals ofmelting for a garnet lherzolite (gt lherz) and spinel lher-zolite (sp lherz) source (3) melting of garnet lherzoliteThe model solutions are non-unique and a range indegree of melting partition coefficients melt reactioncoefficients and proportion of mixed source componentsis possible to achieve acceptable solutions (see Fig 14 cap-tion for description of modeling parameters anduncertainties)The LREE-depleted high-MgO lavas require melting of
a depleted spinel lherzolite source (Fig 14a) The modelingresults indicate a high degree of melting (22^25) similarto the results from PRIMELT1 based on major-elementvariations (discussed above) of a LREE-depleted sourceThe high-MgO lavas lack a residual garnet signature Theenriched tholeiitic basalts involved melting of both garnetand spinel lherzolite and represent aggregate melts pro-duced from continuous melting throughout the depth ofthe melting column (Fig 14b) Trace-element concentra-tions in the tholeiitic and picritic lavas are not consistentwith low-degree melting of garnet lherzolite alone(Fig 14c) The modeling results indicate that the aggregatemelts that formed the tholeiitic basalts would haveinvolved lesser proportions of enriched small-degreemelts (1^6 melting) generated at depth and greateramounts of depleted high-degree melts (12^25) gener-ated at lower pressure (a ratio of garnet lherzolite tospinel lherzolite melt of 14 provides the best fits seeFig 14b and description in figure caption) The totaldegree of melting for the tholeiitic basalts was high(23^27) similar to the degree of partial melting in thespinel lherzolite facies for the primary melts that eruptedas the Keogh Lake picrites of a source that was also prob-ably depleted in LREE
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
(c)
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
Melting of garnet lherzolite
High-MgO lavas
Melting of spinel lherzolite
source (07DM+03PM)
source (07DM+03PM)
6 melting
30 melting
(b)
(a)
30 melting
Sam
ple
Cho
ndrit
eS
ampl
eC
hond
rite 20 melting
1 melting
Melting of garnet lherzoliteand spinel lherzolite
x gt lherz + 4x sp lherz
5 meltingx gt lherz + 4x sp lherz
Sam
ple
Cho
ndrit
e
Tholeiitic lavas
Tholeiitic lavas
High-MgO lavas
source (07DM+03PM)
Fig 14 Trace-element modeling results for incongruent dynamicmantle melting for picritic and tholeiitic lavas from the KarmutsenFormation Three steps of modeling are shown (a) melting of spinellherzolite compared with high-MgO lavas (b) melting of garnet lher-zolite and spinel lherzolite compared with tholeiitic lavas (c) meltingof garnet lherzolite compared with tholeiitic and picritic lavasShaded fields in each panel for tholeiitic basalts and picrites arelabeled Patterns with symbols in each panel are modeling results(patterns are 5 melting increments in (a) and (b) and 1 incre-ments in (c) PM primitive mantle DM depleted MORB mantle gtlherz garnet lherzolite sp lherz spinel lherzolite In (b) the ratios ofper cent melting for garnet and spinel lherzolite are indicated (x gtlherzthorn 4x sp lherz means that for 5 melting 1 melt in garnetfacies and 4 melt in spinel facies where x is per cent melting in theinterval of modeling) The ratio of the respective melt proportions inthe aggregate melt (x gt lherzthorn 4x sp lherz) was kept constant andthis ratio of melt contribution provided the best fit to the data Arange of proportion of source components (07DMthorn 03PM to
09DMthorn 01PM) can reproduce the variation in the high-MgOlavas Melting modeling uses the formulation of Zou amp Reid (2001)an example calculation is shown in their Appendix PM fromMcDonough amp Sun (1995) and DM from Salters amp Stracke (2004)Melt reaction coefficients of spinel lherzolite from Kinzler amp Grove(1992) and garnet lherzolite fromWalter (1998) Partition coefficientsfrom Salters amp Stracke (2004) and Shaw (2000) were kept constantSource mineralogy for spinel lherzolite (018cpx027opx052ol003sp) from Kinzler (1997) and for garnet lherzolite(034cpx008opx053ol005gt) from Salters amp Stracke (2004) Arange of source compositions for melting of spinel lherzolite(07DMthorn 03PM to 03DMthorn 07PM) reproduce REE patternssimilar to those of the high-MgO lavas with best fits for 22ndash25melting and 07DMthorn 03PM source composition Concentrationsof tholeiitic basalts cannot be reproduced from an entirelyprimitive or depleted source
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
30
The uncertainty in the composition of the source in thespinel lherzolite melting model for the high-MgO lavasprecludes determining whether the source was moredepleted in incompatible trace elements than the source ofthe tholeiitic lavas (see Fig 14 caption for description) It ispossible that early high-pressure low-degree melting left aresidue within parts of the plume (possibly the hotter inte-rior portion) that was depleted in trace elements (egElliott et al 1991) further decompression and high-degreemelting of such depleted regions could then have generatedthe depleted high-MgO Karmutsen lavas Shallow high-degree melting would preferentially sample depletedregions whereas deeper low-degree melting may notsample more refractory depleted regions (eg CaribbeanEscuder-Viruete et al 2007) The picrites lie at the highend of the range of Nd isotopic compositions and havelow NbZr similar to depleted Icelandic basalts (Fittonet al 2003) therefore the picrites may represent the partsof the mantle plume that were the most depleted prior tothe initiation of melting As Fitton et al (2003) suggestedthese depleted melts may only rarely be erupted withoutcombining with more voluminous less depleted melts andthey may be detected only as undifferentiated small-volume flows like the Keogh Lake picrites within theKarmutsen Formation on Vancouver IslandDecompression melting within the mantle plume initiatedwithin the garnet stability field (35^25 GPa) at highmantle potential temperatures (Tp414508C) and pro-ceeded beneath oceanic arc lithosphere within the spinelstability field (25^09 GPa) where more extensivedegrees of melting could occur
Magmatic evolution of Karmutsentholeiitic basaltsPrimary magmas in large igneous provinces leave exten-sive coarse-grained cumulate residues within or below thecrust these mafic and ultramafic plutonic sequences repre-sent a significant proportion of the original magma thatpartially crystallized at depth (eg Farnetani et al 1996)For example seismic and petrological studies of OntongJava (eg Farnetani et al 1996) combined with the use ofMELTS (Ghiorso amp Sack 1995) indicate that the volcanicsequence may represent only 40 of the total magmareaching the Moho Neal et al (1997) estimated that30^45 fractional crystallization took place for theOntong Java magmas and produced cumulate thicknessesof 7^14 km corresponding to 11^22 km of flood basaltsdepending on the presence of pre-existing oceanic crustand the total crustal thickness (25^35 km) Interpretationsof seismic velocity measurements suggest the presence ofpyroxene and gabbroic cumulates 9^16 km thick beneaththe Ontong Java Plateau (eg Hussong et al 1979)Wrangellia is characterized by high crustal velocities in
seismic refraction lines which is indicative of mafic
plutonic rocks extending to depth beneath VancouverIsland (Clowes et al 1995)Wrangellia crust is 25^30 kmthick beneath Vancouver Island and is underlain by astrongly reflective zone of high velocity and density thathas been interpreted as a major shear zone where lowerWrangellia lithosphere was detached (Clowes et al 1995)The vast majority of the Karmutsen flows are evolved tho-leiitic lavas (eg low MgO high FeOMgO) indicating animportant role for partial crystallization of melts at lowpressure and a significant portion of the crust beneathVancouver Island has seismic properties that are consistentwith crystalline residues from partially crystallizedKarmutsen magmas To test the proportion of the primarymagma that fractionated within the crust MELTS(Ghiorso amp Sack 1995) was used to simulate fractionalcrystallization using several estimated primary magmacompositions from the PRIMELT1 modeling results Themajor-element composition of the primary magmas couldnot be estimated for the tholeiitic basalts because of exten-sive plagthorn cpxthornol fractionation and thus the MELTSmodeling of the picrites is used as a proxy for the evolutionof major elements in the volcanic sequence A pressure of 1kbar was used with variable water contents (anhydrousand 02wt H2O) calculated at the quartz^fayalite^magnetite (QFM) oxygen buffer Experimental resultsfrom Ontong Java indicate that most crystallizationoccurs in magma chambers shallower than 6 km deep(52 kbar Sano amp Yamashita 2004) however some crys-tallization may take place at greater depths (3^5 kbarFarnetani et al 1996)A selection of the MELTS results are shown in Fig 15
Olivine and spinel crystallize at high temperature until15^20wt of the liquid mass has fractionated and theresidual magma contains 9^10wt MgO (Fig 15f) At1235^12258C plagioclase begins crystallizing and between1235 and 11908C olivine ceases crystallizing and clinopyr-oxene saturates (Fig 15f) As expected the addition of clin-opyroxene and plagioclase to the crystallization sequencecauses substantial changes in the composition of the resid-ual liquid FeO and TiO2 increase and Al2O3 and CaOdecrease while the decrease in MgO lessens considerably(Fig 15) The near-vertical trends for the Karmutsen tho-leiitic basalts especially for CaO do not match the calcu-lated liquid-line-of-descent very well which misses manyof the high-CaO high-Al2O3 low-FeO basalts A positivecorrelation between Al2O3CaO and SrNd indicates thataccumulation of plagioclase played a major role in the evo-lution of the tholeiitic basalts (Fig 15g) Increasing thecrystallization pressure slightly and the water content ofthe melts does not systematically improve the match ofthe MELTS models to the observed dataThe MELTS results indicate that a significant propor-
tion of the crystallization takes place before plagioclase(15^20 crystallization) and clinopyroxene (35^45
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
31
crystallization) join the crystallization sequence Theseresults suggest that445 crystallization of the primarymagmas occurred (estimated from the residual liquid percent in Fig 15f) and if the original flood basalt stratigra-phy on Vancouver Island was 6 km thick at least anequivalent amount of complementary plutonic rocksshould have crystallized The Karmutsen basalts and theirplutonic residues thus represent a significant addition ofcrust (perhaps 412 km thickness) previously thickenedduring the formation of the underlying Paleozoic arc
Growth of the Wrangellia oceanic plateauon Vancouver IslandThe extrusive part of the Wrangellia oceanic plateau onVancouver Island was constructed in a three-layersequence of mostly tholeiitic basalts with a restrictedrange of composition (Fig 16) The plateau overliesDevonian to Mississippian arc volcanic sequences andmarine sediments of Mississippian to Permian ages thatare intruded by gabbroic rocks related to the flood basaltsThe volcanic stratigraphy of theWrangellia plateau formed
00
05
10
15
20
25
30
35
40
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
16
0 2 4 6 8 10 12 14 16 18 20
10
11
12
13
14
15
16
17
18
19
20
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
0 2 4 6 8 10 12 14 16 18 20
MgO (wt)
0 10 20 30 40 50
percent of total mass
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
) Mineral proportions
Sp (e)
Ol
CpxPlag
PicriteHigh-MgO basaltTholeiitic basalt
TiO2 (wt)
MgO (wt)
Al2O3 (wt) CaO (wt)
FeO (wt)
MgO(a) (b)
(c) (d)
MgO (wt)MgO (wt)
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
)
Residual liquid ()
1220
1190
Ol + Sp
Plag+Ol+Sp Plag+Cpx+Sp
02 wt H2O
no H2O
Residual liquid and crystallizing phases
(f)
0102030405060708090100
plagioclaseaccumulation
plagioclaseaccumulation
plagioclase
accumulatio
n
10
12
14
16
0 10 20 30 40
Al2O3CaO
SrNd
tholeiitic basalts
(g)
Fig 15 Forward fractional crystallization modeling results for major elements from MELTS (Ghiorso amp Sack 1995) compared with picriticand tholeiitic lavas from the Karmutsen Formation (a^d) Major-element compositions of residual liquids compared with Karmutsen lavacompositions For clarity the results are shown only for one starting composition (open circle) using the composition of the estimated primarymagma for the high-MgO lava series from the modeling technique of Herzberg et al (2007) for sample 4723A13 as shown in Fig 12 The esti-mated primary magmas for samples 4723A3 and 93G171 yield similar results A pressure of 1 kbar was used with no H2O and 02wt H2Ocalculated at QFM black line is anhydrous and grey line is 02wt H2OThe end-point for the anhydrous result is at 78 fractionation and11308C and the end-point for the result with 02wt H2O is 75 fractionation and11108C (e) Mineral proportions of crystallized phases fromone MELTrun (sample 4723A13) with no H2O (f) Per cent residual liquid vs temperature labelled with intervals of crystallizing phases (g)Al2O3CaO vs SrNd for the tholeiitic basalts
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
32
in a single eruptive phase (c 231^225 Ma) as an emergentbasalt sequence during a deeper-water shallow-water andsubaerial stage and subsided below sea level relativelysoon after its formation (Fig 16) The deeper-water stagecontains pillowed and massive submarine flows (3000mthick) that built up the submarine volcanic edifice from theseafloor The lowest pillow basalts were emplaced onunconsolidated and lithified fine-grained siliciclastic andcarbonaceous sediments containing Daonella beds(c 235^232 Ma) that are intruded by abundantKarmutsen mafic sills (Fig 16a) The different pillowedand massive submarine flows resulted primarily from dif-ferent effusive rates and local topography Some of themassive submarine flows commonly recognized by radiat-ing columnar jointing represent master tubes for the deliv-ery of lava to distal parts of submarine flow fields Thepillowed flows which commonly have uneven topography
in modern sequences may have served as levees to themassive submarine flows rather than allowing unob-structed spreading and inflation typical of the massive sub-aerial flows that form on shallow slopes in continentalflood basalt provinces (eg Self et al 1997) The rapid suc-cession of flows is evident by the near absence of sedimentbetween submarine flows Minor volumes of picritic pillowlavas occur in the upper part of the pillowed flow sequencein areas of the plateau exposed on northern VancouverIsland (Fig 16 Nixon et al 2008)The growth of the plateau in shallow water led to
emplacement of an increasing proportion of volcaniclastics(pillow breccias and hyaloclastites) that accumulated up to1500m thick in some areas (Fig 16b Nixon et al 2008)Horizons of pillowed flows occur locally within the volca-niclastic stratigraphy Sedimentary structures (eg gradedbedding fluidization structures) and petrographic textures
Late Devonian-Mississippian arc
Daonella beds
Middle Triassic
deeper water
Late Devonian-Mississippian arc
pillowed flows massive submarine
flows
pillow brecciahyaloclastiteshallow water
~4 km
~6 km
Late Triassic limestoneinterflow limestonelenses
Karmutsen Formation
mafic intrusions sediment-sill complex
Karmutsen Formation
(a) (b)
(c) (d)
subaerial flows
Late Devonian-Mississippian arc Late Devonian-Mississippian arc
picritic pillow basalts
Miss to Permian marine sediments
sills
Fig 16 Growth of the Wrangellia oceanic plateau onVancouver Island (a) Deep-water eruptive phase of pillowed and massive submarineflows with almost no intervening sediments onto MiddleTriassic shale containing Daonella beds (c 235^232 Ma) The basement of the plateauis composed of Devonian to Mississippian arc sequences overlain by Mississippian to Permian carbonates and siliciclastic sediments(b) Shallow-water flow emplacement with increasing proportion of volcaniclastics resulted primarily from cooling^contraction granulationPicritic pillow basalts occur primarily near the transition between pillowed submarine flows and volcaniclastic units Pillowed and massiveflows occur locally within the volcaniclastic unit (c) Emergent stage of growth consists of overlapping subaerial flow fields similar to continen-tal flood basalts with pahoehoe flow structures Multi-tiered columnar jointing occurrs locally in response to ponding of flows (eg Bondre et al2004) Interflow carbonate deposits and plagioclase megacrystic flows developed locally during the waning stage of volcanism (d) Post-volcanicsubsidence and deposition of Late Carnian to Norian (c 212^223 Ma) platform carbonates on top of the plateau Units are labeled in the dia-grams Thicknesses of the underlying units are not to scale
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
33
within the volcaniclastic rocks indicate formation primar-ily via cooling^contraction granulation magma^water^steam interaction autobrecciation resedimentation andmass-wasting rather than pyroclastic eruption (Nixonet al 2008)A broad subaerial platform was constructed by emplace-
ment of subaerial lava flows during an emergent subaerialstage (Fig 16c) The subaerial unit predominantly com-prises well-layered sheet flows that formed in a similarway to continental flood basalts as hundred(s) of pahoehoeflow fields erupted on shallow slopes (Self et al 1997) how-ever volcaniclastic and pillowed flow units are locally pre-sent mostly in the upper parts of the stratigraphyLocalized ponding of flows led to the formation of multi-tiered columnar jointing in some subaerial flows (egBondre et al 2004) As volcanism waned local interflowcarbonate deposits developed and plagioclase megacrysticflows were erupted Subsidence of the plateau is recordedby the interflow lenses and deposits of platform carbonates(40^1500m thick) on top of the Karmutsen basalts thatcontain Late Carnian to Early Norian fossils (Fig 16d c216^227 Ma Nixon et al 2008)The evolution of the internal architecture of the
Wrangellia oceanic plateau was primarily related to theconditions of emplacement (ie water depth and eruptionenvironment) There does not appear to have been a sys-tematic coincident evolution or shift of lava compositionsonVancouver Island as has been recognized in most con-tinental flood basalts and parts of the Wrangellia oceanicplateau in Alaska and Yukon (Greene et al 2008 2009)In general with the exception of the picritic pillow lavasin the submarine stratigraphy and local variations betweenhigh-MgO and tholeiitic lavas there are no significantcompositional differences between lavas erupted duringthe different stages of growth of the plateau (ie submarineand subaerial flows) nor are substantial compositional dif-ferences generally found between massive submarine flowsin contact with pillowed flows tholeiitic basalt flows nearthe base of the Karmutsen Formation are compositionallysimilar to flows sampled near the top of the volcanicsequence Where it was possible to sample continuous sec-tions of volcanic stratigraphy greater than several hun-dreds of meters (eg Keogh^Maynard Lake area west ofthe Karmutsen Range Fig 2) there is no clear relationshipbetween stratigraphic position and compositionalvariation
CONCLUSIONSThe Karmutsen Formation covers extensive areas ofVancouver Island (20 000 km2) and forms part of a largeoceanic plateau erupted over a geologically short interval(c 225^231 Ma) The tripartite volcanic stratigraphyexposed on Vancouver Island is at least 6 km thick andcomprises a succession of submarine flows volcaniclastic
deposits and massive subaerial flowsThese volcanologicaldifferences are primarily related to the eruption environ-ment (deep-water shallow-water or subaerial) The rapidgrowth of the plateau prevented significant accumulationsof intervening sediment except in the uppermost stratigra-phy where isolated limestone lenses commonly associatedwith pillow basalt and hyaloclastite deposits preserve arecord of the final stages of subsidence of the plateauThe Wrangellia plateau on Vancouver Island was con-
structed dominantly of tholeiitic basalt with restrictedmajor- and trace-element and isotopic compositional var-iations Minor volumes of picritic pillow basalts eruptedlate in the submarine phase The Karmutsen basalts wereemplaced onto older Paleozoic arc volcanic sequences andlow NbLa for the picrites may indicate a contributionfrom the sub-arc lithospheric mantle Modeling of the pet-rogenesis of the Karmutsen picrites indicates melting ofanomalously hot mantle (15008C) and extensive degreesof partial melting (23^27) and is consistent with aplume initiation model Combined Sr^Nd^Hf^Pb isotopicsystematics indicate that the picritic and tholeiitic lavaswere derived from a depleted plume-type mantle sourcedistinct from the source of MORB that has compositionalsimilarities to the source of the lavas forming theCaribbean Plateau Picrites were generated by melting ofspinel lherzolite whereas the tholeiitic basalts formed asaggregate melts of both garnet and spinel lherzolite Thetholeiitic basalts underwent extensive low-pressure fractio-nation (52^3 kbar) and seismic studies indicate that someof theWrangellia crust beneathVancouver Island may cor-respond to the plutonic residues of the Karmutsen floodbasalts
ACKNOWLEDGEMENTSWe are grateful to Nick Arndt for helping us get this pro-ject started and Nick Massey for insights into VancouverIsland geology We greatly appreciate the assistance withPRIMELT modeling from Claude Herzberg We thankMikkel Schau for his insight and enthusiasm during field-work Jane Barling assisted with analyses by MC-ICP-MSThoughtful and thorough reviews of the manuscript byJohn Mahoney Fred Frey Wendy Bohrson DougalJerram and Teal Riley are greatly appreciated Fundingwas generously provided by the Rocks to Riches Programadministered by the BC amp Yukon Chamber of Mines in2004 by the BC Geological Survey and by NSERCDiscovery Grants to JSS and DW ARG was supportedby a University Graduate Fellowship at UBC
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
34
REFERENCESBarker F Brown A S Budahn J R amp Plafker G (1989) Back-arc
with frontal-arc component origin of Triassic Karmutsen basaltBritish Columbia Canada Chemical Geology 75 81^102
Blichert-Toft JWeis D Maerschalk C Agranier A amp Albare de F(2003) Hawaiian hot spot dynamics as inferred from the Hf and Pbisotope evolution of Mauna Kea volcano Geochemistry GeophysicsGeosystems 4(2) 1^27 doi1010292002GC000340
Bondre N R Duraiswami R A amp Dole G (2004) Morphologyand emplacement of flows from the Deccan Volcanic ProvinceIndia Bulletin of Volcanology 66 29^45
Carlisle D (1963) Pillow breccias and their aquagene tuffs QuadraIsland British Columbia Journal of Geology 71 48^71
Carlisle D (1972) Late Paleozoic to Mid-Triassic sedimentary-volcanic sequence on Northeastern Vancouver Island Geological
Society of Canada Paper Report of Activities 72-1 Part B 24^30Carlisle D amp SuzukiT (1974) Emergent basalt and submergent car-
bonate^clastic sequences including the Upper Triassic Dilleri andWelleri zones on Vancouver Island Canadian Journal of Earth
Sciences 11 254^279Chauvel C amp Blichert-Toft J (2001) A hafnium isotope and trace
element perspective on melting of the depleted mantle Earth and
Planetary Science Letters 190 137^151Chayes F (1954)The theory of thin-section analysis Journal of Geology
62 92^101Cho M Liou J G amp Maruyama S (1987) Transition from the zeo-
lite to prehnite^pumpellyite facies in the Karmutsen metabasitesVancouver Island British Columbia Journal of Petrology 27(2)467^494
Clowes R M Zelt C A Amor J R amp Ellis R M (1995)Lithospheric structure in the southern Canadian Cordillera froma network of seismic refraction lines Canadian Journal of Earth
Sciences 32(10) 1485^1513Coffin M F amp Eldholm O (1994) Large igneous provinces Crustal
structure dimensions and external consequences Reviews of
Geophysics 32(1) 1^36Cox K G (1980) A model for flood basalt volcanism Journal of
Petrology 21 629^650Eldholm O amp Coffin M F (2000) Large igneous provinces
and plate tectonics In Richards M A Gordon R G amp vander Hilst R D (eds) The History and Dynamics of Global
Plate Motions American Geophysical Union Geophysical Monograph 121309^326
Elliott T R Hawkesworth C J amp Grolaquo nvold K (1991) Dynamicmelting of the Iceland plume Nature 351 201^206
Escuder-Viruete J Pecurren rez-Estaucurren n A Contreras F Joubert MWeis D Ullrich T D amp Spadea P (2007) Plume mantle sourceheterogeneity through time Insights from the Duarte ComplexHispaniola northeastern Caribbean Journal of Geophysical Research112(B04203) doi1010292006JB004323
Farnetani C G Richards M A amp Ghiorso M S (1996)Petrological models of magma evolution and deep crustal structurebeneath hotspots and flood basalt provinces Earth and Planetary
Science Letters 143 81^94Fitton J G amp Godard M (2004) Origin and evolution of magmas
on the Ontong Java Plateau In Fitton J G Mahoney J JWallace P J amp Saunders A D (eds) Origin and Evolution of the
Ontong Java Plateau Geological Society London Special Publications 229151^178
Fitton J G Saunders A D Norry M J Hardarson B S ampTaylor R N (1997) Thermal and chemical structure of theIceland plume Earth and Planetary Science Letters 153 197^208
Fitton J G Saunders A D Kempton P D amp Hardarson B S(2003) Does depleted mantle form an intrinsic part of the Icelandplume Geochemistry Geophysics Geosystems 4(3) 1032 doi1010292002GC000424
Frey F A Huang S Blichert-Toft J Regelous M amp Boyet M(2005) Origin of depleted components in basalt related to theHawaiian hotspot Evidence from isotopic and incompatible ele-ment ratios Geochemistry Geophysics Geosystems 6(1) 23 doi1010292004GC000757
Galer S J G amp Abouchami W (1998) Practical application of leadtriple spiking for correction of instrumental mass discriminationMineralogical Magazine 62A 491^492
Ghiorso M S amp Sack R O (1995) Chemical mass transfer in mag-matic processes IV A revised and internally consistent thermody-namic model for the interpolation and extrapolation of liquid^solidequilibria in magmatic systems at elevated temperatures and pres-sures Contributions to Mineralogy and Petrology 119 197^212
Greene A R Scoates J S amp Weis D (2005)WrangelliaTerrane onVancouver Island Distribution of flood basalts with implicationsfor potential Ni^Cu^PGE mineralization In Grant B (ed)Geological Fieldwork 2004 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2005-1 209^220Greene A R Scoates J S Nixon G T amp Weis D (2006) Picritic
lavas and basal sills in the Karmutsen flood basalt provinceWrangellia northern Vancouver Island In Grant B (ed)Geological Fieldwork 2005 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2006-1 39^52Greene A R Scoates J S amp Weis D (2008) Wrangellia flood
basalts in Alaska A record of plume^lithosphere interaction in aLate Triassic accreted oceanic plateau Geochemistry Geophysics
Geosystems 9(Q12004) doi1010292008GC002092Greene A R Scoates J S Weis D amp Israel S (2009)
Geochemistry of triassic flood basalts from the Yukon (Canada)segment of the accreted Wrangellia oceanic plateau Lithosdoi101016jlithos200811010
Hauff F Hoernle K Tilton G Graham D W amp Kerr A C(2000) Large volume recycling of oceanic lithosphere over shorttime scales geochemical constraints from the Caribbean LargeIgneous Province Earth and Planetary Science Letters 174 247^263
Hauff F Hoernle K amp Schmidt A (2003) Sr^Nd^Pb compositionof Mesozoic Pacific oceanic crust (Site 1149 and 801 ODP Leg185)Implications for alteration of ocean crust and the input into theIzu^Bonin^Mariana subduction system Geochemistry Geophysics
Geosystems 4(8) doi1010292002GC000421Herzberg C (2004) Partial melting below the Ontong Java Plateau
In Fitton J G Mahoney J J Wallace P J amp Saunders A D(eds) Origin and Evolution of the Ontong Java Plateau Geological SocietyLondon Special Publications 229 179^183
Herzberg C (2006) Petrology and thermal structure of the Hawaiianplume from Mauna Kea volcano Nature 444(7119) 605
Herzberg C amp Asimow P D (2008) Petrology of some oceanicisland basalts PRIMELT2XLS software for primary magma cal-culation Geochemistry Geophysics Geosystems 9(Q09001) doi1010292008GC002057
Herzberg C amp OrsquoHara M J (2002) Plume-associated ultramaficmagmas of Phanerozoic age Journal of Petrology 43(10) 1857^1883
Herzberg C Asimow P D Arndt N Niu Y Lesher C MFitton J G Cheadle M J amp Saunders A D (2007)Temperatures in ambient mantle and plumes Constraints frombasalts picrites and komatiites Geochemistry Geophysics Geosystems8(Q02006) doi1010292006GC001390
Huang S amp Frey F A (2005) Trace element abundances of MaunaKea basalt from phase 2 of the Hawaii Scientific Drilling Project
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
35
Petrogenetic implications of correlations with major element con-tent and isotopic ratios Geochemistry Geophysics Geosystems 4(6)doi1010292002GC000322
Hussong D M Wipperman L K amp Kroenke L W (1979) Thecrustal structure of the Ontong Java and Manihiki OceanicPlateaus Journal of Geophysical Research 84(B11) 6003^6010
Jerram D A amp Widdowson M (2005) The anatomy of ContinentalFlood Basalt Provinces geological constraints on the processes andproducts of flood volcanism Lithos 79(3^4) 385^405
Jones D L Silberling N J amp Hillhouse J (1977)Wrangellia a dis-placed terrane in northwestern North America CanadianJournal ofEarth Sciences 14(11) 2565^2577
Kerr A C (2003) Oceanic Plateaus In Rudnick R (ed) The Crust
Treatise on GeochemistryVol 3 Oxford Elsevier Science pp 537^565Kerr A C amp Mahoney J J (2007) Oceanic plateaus Problematic
plumes potential paradigms Chemical Geology 241 332^353Kerr A CTarney J Marriner G F Klaver GT Saunders A D
amp Thirlwall M F (1996) The geochemistry and petrogenesis ofthe Late-Cretaceous picrites and basalts of Curacao NetherlandsAntilles a remnant of an oceanic plateau Contributions to
Mineralogy and Petrology 124(1) 29^43Kerr A C Marriner G F Tarney J Nivia A Saunders A D
Thirlwall M F amp Sinton A C (1997) Cretaceous basaltic ter-ranes in Western Colombia Elemental chronological and Sr-Ndisotopic constraints on petrogenesis Journal of Petrology 38(6)677^702
Kinzler R J (1997) Melting of mantle peridotite at pressuresapproaching the spinel to garnet transition Application to mid-ocean ridge basalt petrogenesis Journal of Geophysical Research
102(B1) 853^874Kinzler R J amp Grove T L (1992) Primary magmas of mid-ocean
ridge basalts 1 Experiments and methods Journal of Geophysical
Research 97(B5) 6885^6906Korenaga J (2005) Why did not the Ontong Java Plateau form sub-
aerially Earth and Planetary Science Letters 234 385^399Kuniyoshi S (1972) Petrology of the Karmutsen (Volcanic) Group
northeastern Vancouver Island British Columbia PhD disserta-tion University of California Los Angeles 242 pp
Lassiter J C DePaolo D J amp Mahoney J J (1995) Geochemistry ofthe Wrangellia flood basalt province Implications for the role ofcontinental and oceanic lithosphere in flood basalt genesis Journalof Petrology 36(4) 983^1009
LeBas M J (2000) IUGS reclassification of the high-Mg and picriticvolcanic rocks Journal of Petrology 41(10) 1467^1470
Longhi J (2002) Some phase equilibrium systematics of lherzolitemelting I Geochemistry Geophysics Geosystems 3(3) doi1010292001GC000204
MacDonald G A amp Katsura T (1964) Chemical composition ofHawaiian lavas Journal of Petrology 5 82^133
Mahoney J J (1987) An isotopic survey of Pacific oceanic plateausimplications for their nature and origin In Keating B HFryer P Batiza R amp Boehlert GW (eds) Seamounts Islands andAtolls American Geophysical Union Geophysical Monograph 43 207^220
Mahoney J LeRoex A P Peng Z Fisher R L amp Natland J H(1992) Southwestern limits of Indian Ocean Ridge mantle and theorigin of low 206Pb204Pb mid-ocean ridge basalt isotope systema-tics of the central Southwest Indian Ridge (178^508E) Journal ofGeophysical Research 97(B13) 19771^19790
Mahoney J J Sinton J M Kurz M D MacDougall J DSpencer K J amp Lugmair GW (1994) Isotope and trace elementcharacteristics of a super-fast spreading ridge East Pacific rise13^238S Earth and Planetary Science Letters 121 173^193
Massey N W D (1995a) Geology and mineral resources of theAlberni^Nanaimo Lakes sheet Vancouver Island 92F1W 92F2Eand part of 92F7E British Columbia Ministry of Energy Mines and
Petroleum Resources Paper 1992-2 132 ppMassey N W D (1995b) Geology and mineral resources of the
Cowichan Lake sheet Vancouver Island 92C16 British Columbia
Ministry of Energy Mines and Petroleum Resources Paper 1992-3 112 ppMassey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005a) Digital Geology Map of British Columbia Tile NM9 MidCoast BC British Columbia Ministry of Energy and Mines Geofile
2005-2Massey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005b) Digital Geology Map of British Columbia Tile NM10Southwest BC British Columbia Ministry of Energy and Mines Geofile
2005-3McDonough W F amp Sun S (1995) The composition of the Earth
Chemical Geology 120 223^253Monger JW H amp Journeay J M (1994) Basement geology and tec-
tonic evolution of theVancouver region In Monger JW H (ed)Geology and Geological Hazards of the Vancouver Region Southwestern
British Columbia Geological Survey of Canada Bulletin 481 3^25Muller J E (1977) Geology of Vancouver Island Geological Survey of
Canada Open File Map 463Muller J E (1980)The Paleozoic Sicker Group of Vancouver Island British
Columbia Geological Survey of Canada Paper 79-30 22 ppMuller J E Northcote K E amp Carlisle D (1974) Geology and
mineral deposits of Alert Bay^Cape Scott map area VancouverIsland British Columbia Geological Survey of Canada Paper 74-8 77
Neal C R Mahoney J J Kroenke L W Duncan R A ampPetterson M G (1997) The Ontong^Java Plateau InMahoney J J amp Coffin M F (eds) Large Igneous Provinces
Continental Oceanic and Planetary FloodVolcanism American Geophysical
Union Geophysical Monograph 100 183^216Niu Y Collerson K D Batiza R Wendt J I amp Regelous M
(1999) Origin of enriched-typed mid-ocean ridge basalt at ridgesfar from mantle plumes The East Pacific Rise at 11820rsquoN Journalof Geophysical Research 104 7067^7088
Nixon G T amp Orr A J (2007) Recent revisions to the EarlyMesozoic stratigraphy of Northern Vancouver Island (NTS 102I092L) and metallogenic implicatinos British Columbia InGrant B (ed) Geological Fieldwork 2006 British Columbia Ministry of
Energy Mines and Petroleum Resources Paper 2007-1 163^177Nixon GT Hammack J L KoyanagiV M Payie G J Haggart
JW Orchard M JTozerT Friedman R M Archibald D APalfy J amp Cordey F (2006a) Geology of the Quatsino^PortMcNeill area northernVancouver Island British Columbia Ministry
of Energy Mines and Petroleum Resources Geoscience Map 2006-2 scale150 000
Nixon G T Kelman M C Stevenson D Stokes L A ampJohnston K A (2006b) Preliminary geology of the Nimpkishmap area (NTS 092L07) northern Vancouver Island BritishColumbia In Grant B amp Newell J M (eds) Geological Fieldwork2005 British Columbia Ministry of Energy Mines and Petroleum Resources
Paper 2006-1 135^152Nixon G T Laroque J Pals A Styan J Greene A R amp
Scoates J S (2008) High-Mg lavas in the Karmutsen flood basaltsnorthernVancouver Island (NTS 092L) Stratigraphic setting andmetallogenic significance In Grant B (ed) Geological Fieldwork
2007 British Columbia Ministry of Energy Mines and Petroleum
Resources Paper 2008-1 175^190Nowell G M Kempton P D Noble S R Fitton J G
Saunders A Mahoney J J amp Taylor R N (1998) High precisionHf isotope measurements of MORB and OIB by thermal ionisation
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
36
mass spectrometry insights into the depleted mantle Chemical
Geology 149 211^233Parrish R R amp McNicoll V J (1992) U^Pb age determinations
from the southern Vancouver Island area British Columbia InRadiogenic Age and Isotopic Studies Report 5 Geological Survey of
Canada Paper 91-2 79^86Peate DW (1997) The Parana^Etendeka Province In Coffin M F
amp Mahoney J (eds) Large Igneous Provinces Continental Oceanic andPlanetary Flood Volcanism American Geophysical Union Geophysical
Monograph 100 217^245Perfit M R Fornari D J Ridley W I Kirk P D Casey J
Kastens K A Reynolds J R Edwards M Desonie DShuster R amp Paradis S (1996) Recent volcanism in the Siqueirostransform fault picritic basalts and implications for MORBmagma genesis Earth and Planetary Science Letters 141(1^4) 91^108
Pretorius W Weis D Williams G Hanano D Kieffer B ampScoates J S (2006) Complete trace elemental characterization ofgranitoid (USGSG-2GSP-2) reference materials by high resolutioninductively coupled plasma-mass spectrometry Geostandards and
Geoanalytical Research 30(1) 39^54Regelous M Niu Y Wendt J I Batiza R Greig A amp
Collerson K D (1999) Variations in the geochemistry of magma-tism on the East Pacific Rise at 10830rsquoN since 800 ka Earth and
Planetary Science Letters 168 45^63Recurren villon S Chauvel C Arndt N T Pik R Martineau F
Fourcade S amp Marty B (2002) Heterogeneity of the Caribbeanplateau mantle source Sr O and He isotopic compositions of oli-vine and clinopyroxene from Gorgona Island Earth and Planetary
Science Letters 205(1^2) 91Rhodes J M (1996) Geochemical stratigraphy of lava flows samples
by the Hawaii Scientific Drilling Project Journal of Geophysical
Research 101(B5) 11729^11746Rhodes J M amp Vollinger M J (2004) Composition of basaltic lavas
sampled by phase-2 of the Hawaii Scientific Drilling ProjectGeochemical stratigraphy and magma types Geochemistry
Geophysics Geosystems 5(3) doi1010292002GC000434Richards M A Jones D L Duncan R A amp DePaolo D J (1991)
A mantle plume initiation model for the Wrangellia flood basaltand other oceanic plateaus Science 254 263^267
Salters V J M amp Hart S R (1991) The mantle sources of oceanridges islands and arcs the Hf-isotope connection Earth and
Planetary Science Letters 104 364^380Salters V J M amp Stracke A (2004) Composition of the depleted
SaltersV J amp WhiteW (1998) Hf isotope constraints on mantle evo-lution Chemical Geology 145 447^460
SanoT amp Yamashita S (2004) Experimental petrology of basementlavas from Ocean Drilling Program Leg 192 implications for dif-ferentiation processes in Ontong Java Plateau magmas InFitton J G Mahoney J JWallace P J amp Saunders A D (eds)Origin and Evolution of the Ontong Java Plateau Geological Society
London Special Publications 229 185^218Saunders A D (2005) Large igneous provinces Origin and environ-
mental consequences Elements 1 259^263Self S ThordarsonT amp Keszthely L (1997) Emplacement of conti-
nental flood basalt lava flows In Mahoney J J amp Coffin M F(eds) Large Igneous Provinces Continental Oceanic and Planetary Flood
Volcanism American Geophysical Union Geophysical Monograph 100381^410
Shaw D M (2000) Continuous (dynamic) melting theory revisitedCanadian Mineralogist 38(5) 1041^1063
Sluggett C L (2003) Uranium^lead age and geochemical constraintson Paleozoic and Early Mesozoic magmatism in WrangelliaTerrane Saltspring Island British Columbia BSc thesisUniversity of British ColumbiaVancouver 84 pp
Storey M Mahoney J J amp Saunders A D (1997) Cretaceousbasalts in Madagascar and the transition between plume and con-tinental lithospheric mantle sources In Mahoney J J ampCoffin M F (eds) Large Igneous Provinces Continental Oceanic and
Planetary Flood Volcanism Aamerican Geophysical Union Geophysical
Monograph 100 95^122Surdam R C (1967) Low-grade metamorphism of the Karmutsen
Group PhD dissertation University of California Los Angeles288 pp
Sutherland-Brown A Yorath C J Anderson R G amp Dom K(1986) Geological maps of southern Vancouver IslandLITHOPROBE I 92C10 11 14 16 92F1 2 7 8 Geological Survey ofCanada Open File 1272
Tejada M L G Mahoney J J Castillo P R Ingle S P Sheth HC amp Weis D (2004) Pin-pricking the elephant evidence on theorigin of the Ontong Java Plateau from Pb^Sr^Hf^Nd isotopiccharacteristics of ODP Leg 192 basalts In Fitton J GMahoney J J Wallace P J amp Saunders A D (eds) Origin and
Evolution of the Ontong Java Plateau Geological Society London Special
Publications 229 133^150Thompson P M E Kempton P D amp Kerr A C (2008) Evaluation
of the effects of alteration and leaching on Sm^Nd and Lu^Hf sys-tematics in submarine mafic rocks Lithos 104(1^4) 164^176
Thompson P M E Kempton P D White R V Kerr A CTarney J Saunders A D Fitton J G amp McBirney A (2003)Hf-Nd isotope constraints on the origin of the CretaceousCaribbean plateau and its relationship to the Galapagos plumeEarth and Planetary Science Letters 217(1^2) 59^75
Vervoort J D Patchett P Blichert-Toft J amp Albare de F (1999)Relationships between Lu^Hf and Sm^Nd isotopic systems in theglobal sedimentary system Earth and Planetary Science Letters 16879^99
Walter M J (1998) Melting of garnet peridotite and the origin ofkomatiite and depleted lithosphere Journal of Petrology 39(1) 29^60
Weis D Kieffer B Maerschalk C Barling J de Jong JWilliams G A Hanano D Mattielli N Scoates J SGoolaerts A Friedman R A amp Mahoney J B (2006) High-pre-cision isotopic characterization of USGS reference materials byTIMS and MC-ICP-MS Geochemistry Geophysics Geosystems
7(Q08006) doi1010292006GC001283Weis D Kieffer B Hanano D Silva I N Barling J
Pretorius W Maerschalk C amp Mattielli N (2007) Hf isotopecompositions of US Geological Survey reference materialsGeochemistry Geophysics Geosystems 8(Q06006) doi1010292006GC001473
Wheeler J O amp McFeely P (1991) Tectonic assemblage map of theCanadian Cordillera and adjacent part of the United States ofAmerica Geological Survey of Canada Map 1712A
WhiteW M Albare de F amp Tecurren louk P (2000) High-precision analy-sis of Pb isotope ratios by multi-collector ICP-MS Chemical Geology167 257^270
Yorath C J Sutherland Brown A amp Massey N W D (1999)LITHOPROBE southern Vancouver Island British Columbia Geological
Survey of Canada Bulletin 498 145 ppZou H (1998) Trace element fractionation during modal and
non-model dynamic melting and open-system melting Amathematical treatment Geochimica et Cosmochimica Acta 62(11)1937^1945
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
37
Zou H amp Reid M R (2001) Quantitative modeling of trace elementfractionation during incongruent dynamic melting Geochimica et
Cosmochimica Acta 65(1) 153^162
APPENDIX SAMPLEPREPARAT ION ANDANALYT ICAL METHODSOnly the freshest rocks in the field were sampled and onlythe least altered samples were selected for chemical andisotopic analysis based on thorough petrographic inspec-tion Fifty-nine of the 129 collected samples were crushed(400 g) into small pieces 52mm in diameter in aRocklabs hydraulic piston crusher between WC plates tominimize contamination The coarse-crush was mixedand 100 g was powdered in a planetary mill using agatejars and balls cleaned with quartz sand between samples
ActLabs analytical methodsThe major- and trace-element compositions of the whole-rock powders were determined at Activation LaboratoriesLtd (Actlabs) in Ancaster Ontario Analytical techniquesand detection limits are also available from Actlabs (httpwwwactlabscommethsub_code4erehtm) The analyticalmethod for each of the elements analyzed is indicated inTable 2 For the major elements a 02 g sample was mixedwith a lithium metaborate^lithium tetraborate mixtureand fused in a graphite crucible The molten mixture waspoured into a 5 HNO3 solution and shaken until dis-solved (30min) The samples were analyzed for majoroxides and selected trace elements by inductively coupledplasma optical emission spectrometry (ICP-OES) on acombination simultaneous^sequential Thermo Jarrell-AshEnviro II system Internal calibration was achieved usinga variety of international reference materials (eg W-2BIR-1 DNC-1) and independent control samplesAdditional trace elements were analyzed by both instru-mental neutron activation analysis (INAA) and induc-tively couple plasma mass spectrometry (ICP-MS) ForINAA15^25 g of sample was weighed into small polyeth-ylene vials and irradiated with a control international ref-erence material CANMET WMS-1 and NiCr flux wires ata thermal neutron flux of 71012 neutronscm2 per s in theMcMaster Nuclear Reactor Following a 7 day waitingperiod the samples were measured on an Ortec high-purity Ge detector linked to a Canberra Series 95 multi-channel analyzer Activities for each element weredecay- and weight-corrected and compared with a detectorcalibration developed from multiple international certifiedreference materials For ICP-MS 025 g of sample wasdigested in HF followed by HNO3^HClO4 heated andtaken to dryness The samples were brought back into solu-tion with HCl Samples were analyzed using a PerkinElmer Optima 3000 ICP In-lab standards or certified ref-erence materials (eg W-2 BIR-1 DNC-1) were used for
quality control A total of 15 blind duplicates were ana-lyzed from the same sample powders to assessreproducibility
University of Massachusetts analyticalmethodsFourteen sample duplicate powders taken from the same100 g sample aliquot (high-MgO basalts and picrites)were also analyzed at the Ronald B Gilmore X-RayFluorescence (XRF) Laboratory at the University ofMassachusetts Major elements were measured on a fusedLa-bearing lithium borate glass disc using a SiemensMRS-400 spectrometer Trace element concentrations(Rb Sr Ba Ce Nb Zr Y Pb Zn Ga Ni Cr V) weremeasured on a separate powder pellet using a PhilipsPW2400 sequential spectrometer using a Rh tube LOIand ferrous iron measurements were made as describedby Rhodes amp Vollinger (2004) Precision and accuracy esti-mates for the data have been described by Rhodes (1996)and Rhodes amp Vollinger (2004) Results for each sampleare the average of two separate analyses and are shown inElectronic Appendix 2 for a comparison of analyses ofduplicate powders using XRFand ICP-MS
PCIGR analytical methodsA subset of 19 samples was selected for high-precisiontrace-element analysis and Sr Nd Pb and Hf isotopicanalysis at the Pacific Centre for Isotopic andGeochemical Research (PCIGR) at the University ofBritish Columbia (UBC) Samples were selected from the59 samples analyzed for whole-rock chemistry at ActLabsbased on major- and trace-element chemistry alteration(low LOI and petrographic alteration index) and samplelocation Samples were prepared for trace-element analysisat the PCIGR by the technique described by Pretorius et al(2006) on unleached rock powders Sample powders(100mg) were weighed in 7ml screw-top Savillex bea-kers and dissolved in 1ml 14N HNO3 and 5ml 48 HFon a hotplate for 48 h at 1308C with periodic ultrasonica-tion Samples were dried and redissolved in 6ml 6N HClon a hotplate for 24 h and then dried and redissolved in1ml concentrated HNO3 for 24 h before final dryingTrace element abundances were measured by high-resolution inductively coupled plasma-mass spectrometry(HR-ICP-MS) with aThermo Finnigan Element2 systemfollowing the procedures described by Pretorius et al(2006) within 24 h of redissolution HFSE and LILE weremeasured in medium-resolution mode at 2000 dilutionusing a PFATeflon spray chamber washed with aqua regiafor 3min between samples REE were measured in high-resolution mode and U and Pb in low-resolution modeat 2000 dilution using a glass spray chamber washedwith 2 HNO3 between samples Total procedural blanksand reference materials (BCR-2 BHVO-2) were analyzedwith the batch of samples Indium was used as an internal
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
38
standard in all samples and standard solutionsBackground and standard solutions were analyzed afterevery five samples to detect memory effects and massdrift Results for PCIGR trace-element analyses areshown in Electronic Appendix 3 for the full set of analyzedtrace elementsSample digestion for purification of Sr Nd Hf and Pb
for column chemistry began by weighing each samplepowder (400^500mg) prior to leaching All samples wereleached with 6N HCl and placed in an ultrasonic bath for15min Samples were rinsed twice with 18 M-cm H2Obetween each leaching step (15 total) until the supernatantliquid was clear [following the method of Mahoney (1987)]Samples were then dried on a hotplate for 24 h andweighed again Sample solutions were then prepared bydissolving 100^250mg of the leached powder dissolvedin 1ml 14N HNO3 and 10ml 48 HF on a hotplate for48 h at 1308C with periodic ultrasonication Samples weredried and redissolved in 6ml 6N HCl on a hotplate for24 h and then dried Pb was separated using anion exhangecolumns and the discard was used for Sr REE and Hfseparation Nd was separated from the other REE and Hfrequired two additional purification steps Detailed proce-dures for column chemistry for separating Sr Nd and Pbat the PCIGR have been described byWeis et al (2006) andHf purification has been described byWeis et al (2007) Srand Nd isotope ratios were measured by thermal ioniza-tion mass spectrometry (TIMS) on a Thermo FinniganTriton system in static mode with relay matrix rotation ona single Ta and double Re^Ta filament respectively Fourto five filaments per barrel of 21 were occupied by stan-dards (NBS 987 for Sr and LaJolla for Nd) for each barrelwhere samples were run Sample Sr and Nd isotopic com-positions were corrected for mass fractionation using86Sr88Srfrac14 01194 and 146Nd144Ndfrac14 07219 Each samplewas then normalized using the barrel average of the refer-ence material relative to the values of143Nd144Ndfrac14 0511858 and 87Sr86Srfrac14 0710248 (Weiset al 2006) During the course of the Vancouver Islandanalyses the La Jolla Nd standard gave an average valueof 05118566 (nfrac14 7) and NBS987 standard gave an aver-age of 07102408 (nfrac1411) (2 error is reported as times106) 147Sm144Nd errors are 15 or 0006 USGeological Survey (USGS) reference material BHVO-2was processed with the samples and yielded Sr and Nd iso-topic ratios of 07034607 and 05129786 respectivelyThese are in agreement with the published values of070347920 and 051298411 respectively (Weis et al2006)
Pb and Hf isotopic compositions were analyzed by mul-tiple collector-inductively coupled plasma-mass spectrome-try (MC-ICP-MS) on a static multi-collection on a NuPlasma (Nu Instruments) system The detailed analyticalprocedure for Pb isotopic analyses on the Nu Plasma atthe PCIGR has been described by Weis et al (2006) Theconfiguration for Pb analyses allows for collection of PbTl and Hg togetherTl and Hg are used to monitor instru-mental mass discrimination and isobaric overlap repec-tively All sample solutions were analyzed withapproximately the same PbTl ratio (4) as the referencematerial NIST SRM 981 To accomplish this a small ali-quot of each sample solution from the Pb columns was ana-lyzed on the Element2 to determine the precise amount ofPb available for analysis on the Nu Plasma The SRM 981standard was run after every two samples on the NuPlasma During the time samples were run analyses ofthe SRM 981 Pb reference material gave values of206Pb204Pbfrac1416940322 207Pb204Pbfrac1415495823and 208Pb204Pbfrac14 36713164 (nfrac14 61 2 error is reportedas times 104) these values are in excellent agreement withreported TIMS triple-spike values of Galer amp Abouchami(1998) Fractionation-corrected Pb isotopic ratios were fur-ther corrected by the sample-standard bracketing methodor the ln^ln correction method described by White et al(2000) and Blichert-Toft et al (2003) Leached powders ofUSGS reference material BHVO-2 yielded Pb isotopicratios of 206Pb204Pbfrac14186454 8 207Pb204Pbfrac14 1549105 and 208Pb204Pbfrac14 38222514 and BCR-2yielded 206Pb204Pbfrac141880466 207Pb204Pbfrac14 1562518 and 208Pb204Pbfrac14 3883496 (2 error isreported as times 104) These values are in agreement withthose for leached residues of BHVO-2 and BCR-2 fromWeis et al (2006)Hf isotopic compositions were analyzed following the
procedures detailed byWeis et al (2007) The configurationfor Hf analyses monitored Lu mass 175 andYb mass 172 toallow for interference correction to masses 174 and 176 Hfisotopic ratios were normalized internally for mass fractio-nation to a 179Hf177Hf ratio of 07325 using an exponentialcorrection Standards were run after every two samplesand sample results were normalized to the ratio of the in-run daily average and a 176Hf177Hf ratio for JMC-475 of0282160 During the course of analyses the Hf standardJMC-475 gave an average value 02821679 (nfrac14130)USGS reference materials BCR-2 and BHVO-2 were pro-cessed with the samples and yielded Hf isotopic ratios of02828675 and 02831005 respectively Publishedvalues for BCR-2 and BHVO-2 are 02828717 and0283104 8 respectively (Weis et al 2007)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
1Ni concentrations for these samples were determined by XRFTHOL tholeiitic basalt PIC picrite HI-MG high MgO basalt CG coarse-grained (sill or gabbro) MIN SIL mineralizedsill OUTLIER anomalous pillowed flow in plots MA Mount Arrowsmith SL Schoen Lake KR Karmutsen Range QIQuadra Island Sample locations are given using the Universal Transverse Mercator (UTM) coordinate system (NAD83zones 9 and 10) Analyses were performed at Activation Laboratory (ActLabs) Fe2O3
is total iron expressed as Fe2O3LOI loss on ignition All major elements Sr V and Y were measured by ICP quadrupole OES on solutions of fusedsamples Cu Ni Pb and Zn were measured by total dilution ICP Cs Ga Ge Hf Nb Rb Ta Th U Zr and REE weremeasured by magnetic-sector ICP on solutions of fused samples Co Cr and Sc were measured by INAA Blanks arebelow detection limit See Electronic Appendices 2 and 3 for complete XRF data and PCIGR trace element datarespectively Major elements for sample 93G17 are normalized anhydrous Sample 5615A7 was analyzed by XRFSamples from Quadra Island are not shown in figures
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
17
picritic and high-MgO basalt pillow lavas extend to20wt MgO (Fig 6 and references listed in thecaption)The tholeiitic basalts from the Karmutsen Range dis-
play a tight range of parallel light rare earth element(LREE)-enriched REE patterns (LaYbCNfrac14 21^24mean 2202) whereas the picrites and high-MgObasalts are characterized by sub-parallel LREE-depleted
patterns (LaYbCNfrac14 03^07 mean 06 02) with lowerREE abundances than the tholeiitic basalts (Fig 7)Tholeiitic basalts from the Schoen Lake and MountArrowsmith areas have similar REE patterns tosamples from the Karmutsen Range (Schoen Lake LaYbCNfrac14 21^26 mean 2303 Mount Arrowsmith LaYbCNfrac1419^25 mean 2204) and the coarse-grainedmafic rocks have similar REE patterns to the tholeiitic
Depleted MORB average(Salters amp Stracke 2004)
Schoen LakeSchoen Lake
Mount Arrowsmith
Karmutsen RangeS
ampl
eC
hond
rite
Sam
ple
Cho
ndrit
e
Sam
ple
Prim
itive
Man
tleS
ampl
eP
rimiti
ve M
antle
Sam
ple
Prim
itive
Man
tle
Mount Arrowsmith
Karmutsen Range
(a) (b)
(c) (d)
(e) (f )
Sam
ple
Cho
ndrit
e
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained mafic rock
Pillowed flowMineralized sill
1
10
100
1
10
100
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
01
1
10
100
1
10
100
1
10
100
Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
Fig 7 Whole-rock REE and other incompatible-element concentrations for the Karmutsen Formation (a) (c) and (e) are chondrite-normal-ized REE patterns for samples from each of the three main field areas onVancouver Island (b) (d) and (f) are primitive mantle-normalizedtrace-element patterns for samples from each of the three main field areas All normalization values are from McDonough amp Sun (1995) Theclear distinction between LREE-enriched tholeiitic basalts and LREE-depleted picrites the effects of alteration on the LILE (especially loss ofK and Rb) and the different range for the vertical scale in (b) (extends down to 01) should be noted
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
18
basalts (LaYbCNfrac14 21^29 mean 2508) Two LREE-depleted coarse-grained mafic rocks from the SchoenLake area have similar REE patterns (LaYbCNfrac14 07) tothe picrites from the Keogh Lake area indicating that thisdistinctive suite of rocks is exposed as far south asSchoen Lake that is 75 km away (Fig 7) The anomalouspillowed flow from the Karmutsen Range has lower LaYbCN (13^18) than the main group of tholeiitic basaltsfrom the Karmutsen Range (Fig 7) The mineralized sillfrom the Schoen Lake area is distinctly LREE-enriched(LaYbCNfrac14 33) with the highest REE abundances(LaCNfrac14 940) of all Karmutsen samples this may reflectcontamination by adjacent sediment also evident in thin-section where sulfide blebs are cored by quartz (Greeneet al 2006)The primitive mantle-normalized trace-element pat-
terns (Fig 7) and trace-element variations and ratios(Fig 8) highlight the differences in trace-element
concentrations between the four main groups ofKarmutsen flood basalts onVancouver Island The tholeii-tic basalts from all three areas have relatively smooth par-allel trace-element patterns some samples have smallpositive Sr anomalies relative to Nd and Sm and manysamples have prominent negative K anomalies relative toU and Nb reflecting K loss during alteration (Fig 7) Thelarge ion lithophile elements (LILE) in the tholeiiticbasalts (mainly Rb and K not Cs) are depleted relativeto the high field strength elements (HFSE) and LREELILE segments especially for the Schoen Lake area(Fig 7d) are remarkably parallel The picrites andhigh-MgO basalts form a tight band of parallel trace-element patterns and are depleted in HFSE and LREEwith positive Sr anomalies and relatively low Rb Thecoarse-grained mafic rocks from the Karmutsen Rangehave identical trace-element patterns to the tholeiiticbasalts Samples from the Schoen Lake area have similar
000
025
050
075
100
125
0 1 2 3 4 5
00
04
08
12
16
0 5 10 15 20 25
MgO (wt)
0
5
10
15
0 50 100 150
Hf (ppm)
Th (ppm)
NbLaNb (ppm)
Zr (ppm)
(a)
(c)
(b)
(d)
Th (ppm)
00
02
04
06
08
10
12
00 01 02 03 04 05
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
NbLa=1
4723A4 5615A12
U (ppm)
Fig 8 Whole-rock trace-element concentrations and ratios for the Karmutsen Formation [except (b) which is vs wt MgO] (a) Nb vs Zr(b) NbLa vs MgO (c) Th vs Hf (d) Th vs U There is a clear distinction between the tholeiitic basalts and picrites in Nb and Zr both inconcentration and the slope of each trend
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
19
trace-element patterns to the Karmutsen Range except forthe two LREE-depleted coarse-grained mafic rocks andthe mineralized sill (Fig 7d)
Sr^Nd^Hf^Pb isotopic compositionsA total of 19 samples from the four main groups ofthe Karmutsen Formation were selected for radiogenicisotopic geochemistry on the basis of major- and trace-element variation stratigraphic position and geographicallocation The samples have overlapping age-correctedHf and Nd isotope ratios and distinct ranges of Sr isotopiccompositions (Fig 9) The tholeiitic basalts picriteshigh-MgO basalt and coarse-grained mafic rocks have
initial eHffrac14thorn87 to thorn126 and eNdfrac14thorn66 to thorn88 cor-rected for in situ radioactive decay since 230 Ma (Fig 9Tables 3 and 4) All the Karmutsen samples form a well-defined linear array in a Lu^Hf isochron diagram corre-sponding to an age of 24115 Ma (Fig 9) This is withinerror of the accepted age of the Karmutsen Formationand indicates that the Hf isotope systematics behaved as aclosed system since c 230 Ma The anomalous pillowedflow has similar initial eHf (thorn98) and slightly lower initialeNd (thorn62) compared with the other Karmutsen samplesThe picrites and high-MgO basalt have higher initial87Sr86Sr (070398^070518) than the tholeiitic basalts(initial 87Sr86Srfrac14 070306^070381) and the coarse-grained mafic rocks (initial 87Sr86Srfrac14 070265^070428)
05128
05129
05130
05131
05132
015 017 019 021 023 025
02829
02830
02831
02832
02833
02834
000 002 004 006 008 010
4
6
8
10
0702 0703 0704 0705 07066
8
10
12
14
5 6 7 8 9 10
12
0
1
2
3
4
5
6
7
0702 0703 0704 0705 0706
(d)
LOI (wt )
87Sr 86Sr
4723A2
Nd
143Nd144Nd
147Sm144Nd
87Sr 86Sr (230 Ma)
(230 Ma)
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
(a)
(c)
176Hf177Hf
Age=241 plusmn 15 Ma=+990 plusmn 064
176Lu177Hf
Hf (230 Ma)
Nd (230 Ma)
(e)
Hf (i)
(b)
Fig 9 Whole-rock Sr Nd and Hf isotopic compositions for the Karmutsen Formation (a) 143Nd144Nd vs 147Sm144Nd (b) 176Hf177Hf vs176Lu177Hf The slope of the best-fit line for all samples corresponds to an age of 24115 Ma (c) Initial eNd vs 87Sr86Sr Age-correction to230 Ma (d) LOI vs 87Sr86Sr (e) Initial eHf vs eNd Average 2 error bars are shown in a corner of each panel Complete chemical duplicatesshown inTables 3 and 4 (samples 4720A7 and 4722A4) are circled in each plot
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
overlap the ranges of the picrites and tholeiitic basalts(Fig 9 Table 3)The measured Pb isotopic compositions of the tholeiitic
basalts are more radiogenic than those of the picrites andthe most magnesian Keogh Lake picrites have the leastradiogenic Pb isotopic compositions The range ofinitial Pb isotope ratios for the picrites is206Pb204Pbfrac1417868^18580 207Pb204Pbfrac1415547^15562and 208Pb204Pbfrac14 37873^38257 and the range for thetholeiitic basalts is 206Pb204Pbfrac1418782^19098207Pb204Pbfrac1415570^15584 and 208Pb204Pbfrac14 37312^38587 (Fig 10 Table 5) The coarse-grained mafic rocksoverlap the range of initial Pb isotopic compositions forthe tholeiitic basalts with 206Pb204Pbfrac1418652^19155207Pb204Pbfrac1415568^15588 and 208Pb204Pbfrac14 38153^38735 One high-MgO basalt has the highest initial206Pb204Pb (19242) and 207Pb204Pb (15590) and thelowest initial 208Pb204Pb (37676) The Pb isotopic ratiosfor Karmutsen samples define broadly linearrelationships in Pb isotope plots The age-corrected Pb
isotopic compositions of several picrites have been affectedby U and Pb (andor Th) mobility during alteration(Fig 10)
ALTERAT IONThe Karmutsen basalts have retained most of their origi-nal igneous structures and textures however secondaryalteration and low-grade metamorphism have producedzeolitic- and prehnite^pumpellyite-bearing mineral assem-blages (Table 1 Cho et al 1987) Alteration and metamor-phism have primarily affected the distribution of the LILEand the Sr isotopic systematics The Keogh Lake picriteshave been affected more by alteration than the tholeiiticbasalts and have higher LOI (up to 55wt ) variableLILE and higher measured Sr Nd and Hf and lowermeasured Pb isotope ratios than the tholeiitic basalts Srand Pb isotope compositions for picrites and tholeiiticbasalts show a relationship with LOI (Figs 9 and 10)whereas Nd and Hf isotopic compositions do not correlate
Table 3 Sr and Nd isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Rb Sr 87Sr86Sr 2m87Rb86Sr 87Sr86Srt Sm Nd 143Nd144Nd 2m eNd
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses carried out at the PCIGR thecomplete trace element analyses are given in Electronic Appendix 3
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
21
with LOI (not shown) The relatively high apparent initialSr isotopic compositions for the high-MgO lavas (up to07052) probably resulted from post-magmatic loss of Rbor an increase in 87Sr86Sr through addition of seawater Sr(eg Hauff et al 2003) High-MgO lavas from theCaribbean plateau have similarly high 87Sr86Sr comparedwith basalts (Recurren villon et al 2002)The correction for in situdecay on initial Pb isotopic ratios has been affected bymobilization of U and Th (and Pb) in the whole-rockssince their formation A thorough acid leaching duringsample preparation was used that has been shown to effec-tively remove alteration phases (Weis et al 2006 NobreSilva et al in preparation) Whole-rock SmNd ratiosand age-correction of Nd isotopes may be affected bythe leaching procedure for submarine rocks older than50 Ma (Thompson et al 2008 Fig 9) there ishowever very little variation in Nd isotopic compositionsof the picrites or tholeiitic basalts and this system does notappear to be disturbed by alteration The HFSE abun-dances for tholeiitic and picritic basalts exhibit clear
linear correlations in binary diagrams (Fig 8) whereasplots of LILE vs HFSE are highly scattered (not shown)because of the mobility of some of the alkali and alkalineearth LILE (Cs Rb Ba K) and Sr for most samplesduring alterationThe degree of alteration in the Karmutsen samples does
not appear to be related to eruption environment or depthof burial Pillow rims are compositionally different frompillow cores (Surdam1967 Kuniyoshi1972) and aquagenetuffs are chemically different from isolated pillows withinpillow breccias however submarine and subaerial basaltsexhibit a similar degree of alterationThere is no clear cor-relation between the submarine and subaerial basalts andsome commonly used chemical alteration indices (eg BaRb vs K2OP2O5 Huang amp Frey 2005) There is also nodefinitive correlation between the petrographic alterationindex (Table 1) and chemical alteration indices although10 of 13 tholeiitic basalts with the highest K and LILEabundances have the highest petrographic alterationindex of three (seeTable 1)
Table 4 Hf isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Lu Hf 176Hf177Hf 2m eHf176Lu177Hf 176Hf177Hft eHf(t)
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
OLIV INE ACCUMULAT ION INH IGH-MgO AND PICR IT IC LAVASGeochemical trends and petrographic characteristics indi-cate that accumulation of olivine played an important rolein the formation of the high-MgO basalts and picrites fromthe Keogh Lake area on northern Vancouver Island TheKeogh Lake samples show a strong linear correlation inplots of Al2O3 TiO2 ScYb and Ni vs MgO and many ofthe picrites have abundant clusters of olivine pseudomorphs(Table1 Fig11)There is a strong linear correlationbetween
modal per cent olivine and whole-rock magnesium contentmagnesium contents range between 10 and 19wt MgOand the proportion of olivine phenocrysts in these samplesvaries from 0 to 42 vol (Fig 11)With a few exceptionsboth the size range and median size of the olivine pheno-crysts in most samples are comparable The clear correla-tion between the proportion of olivine phenocrysts andwhole-rock magnesium contents indicates that accumula-tion of olivine was directly responsible for the high MgOcontents (410wt ) in most of the picritic lavas
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
207Pb204Pb
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
208Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
(a) (b)
(c) (e)
4723A24723A2
4723A2
4723A2
4722A4
4723A4
4722A4
4723A4
4723A134723A3
4723A3
4723A13
1556
1558
1560
1562
1564
1566
1568
186 190 194 198 202 206 2101550
1552
1554
1556
1558
1560
178 180 182 184 186 188 190 192 194
380
384
388
392
396
186 190 194 198 202 206 210374
378
382
386
390
178 180 182 184 186 188 190 192 194
206Pb204Pb(d)
LOI (wt )
0
1
2
3
4
5
6
7
186 190 194 198 202 206 210
4723A2
Fig 10 Pb isotopic compositions of leached whole-rock samples measured by MC-ICP-MS for the Karmutsen Formation Error bars are smal-ler than symbols (a) Measured 207Pb204Pb vs 206Pb204Pb (b) Initial 207Pb204Pb vs 206Pb204Pb Age-correction to 230 Ma (c) Measured208Pb204Pb vs 206Pb204Pb (d) LOI vs 206Pb204Pb (e) Initial 208Pb204Pb vs 206Pb204Pb Complete chemical duplicates shown in Table 5(samples 4720A7 and 4722A4) are circled in each plot The dashed lines in (b) and (e) show the differences in age-corrections for two picrites(4723A4 4722A4) using the measured UTh and Pb concentrations for each sample and the age-corrections when concentrations are used fromthe two picrites (4723A3 4723A13) that appear to be least affected by alteration
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
23
DISCUSS IONThe compositional and spatial development of the eruptedlava sequences of the Wrangellia oceanic plateau onVancouver Island provide information about the evolutionof a mantle plume upwelling beneath oceanic lithospherein an analogous way to the interpretation of continentalflood basalts The Caribbean and Ontong Java plateausare the two primary examples of accreted parts of oceanicplateaus that have been studied to date and comparisonscan be made with the Wrangellia oceanic plateauHowever the Wrangellia oceanic plateau is a distinctiveoccurrence of an oceanic plateau because it formed on thecrust of a Devonian to Mississippian intra-oceanic islandarc overlain by Mississippian to Permian limestones andpelagic sediments The nature and thickness of oceaniclithospheric mantle are considered to play a fundamentalrole in the decompression and evolution of a mantleplume and thus the compositions of the erupted lavasequences (Greene et al 2008) The geochemical and
stratigraphic relationships observed in the KarmutsenFormation onVancouver Island and the focus of the subse-quent discussion sections provide constraints on (1) thetemperature and degree of melting required to generatethe primary picritic magmas (2) the composition of themantle source (3) the depth of melting and residual min-eralogy in the source region (4) the low-pressure evolutionof the Karmutsen tholeiitic basalts (5) the growth of theWrangellia oceanic plateau
Melting conditions and major-elementcomposition of the primary magmasThe primary magmas to most flood basalt provinces arebelieved to be picritic (eg Cox 1980) and they have beenfound in many continental and oceanic flood basaltsequences worldwide (eg Siberia Karoo Paranacurren ^Etendeka Caribbean Deccan Saunders 2005) Near-pri-mary picritic lavas with low total alkali abundances
Table 5 Pb isotopic compositions of Karmutsen basaltsVancouver Island BC
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
require high-degree partial melting of the mantle source(eg Herzberg amp OrsquoHara 2002) The Keogh Lake picritesfrom northernVancouver Island are the best candidates foridentifying the least modified partial melts of the mantleplume source despite having accumulated olivine pheno-crysts and can be used to estimate conditions of meltingand the composition of the primary magmas for theKarmutsen FormationPrimary magma compositions mantle potential tem-
peratures and source melt fractions for the picritic lavas ofthe Karmutsen Formation were calculated from primitivewhole-rock compositions using PRIMELT1XLS software(Herzberg et al 2007) and are presented in Fig 12 andTable 6 A detailed discussion of the computationalmethod has been given by Herzberg amp OrsquoHara (2002)and Herzberg et al (2007) key features of the approachare outlined belowPrimary magma composition is calculated for a primi-
tive lava by incremental addition of olivine and compari-son with primary melts of mantle peridotite Lavas thathad experienced plagioclase andor clinopyroxene fractio-nation are excluded from this analysis All calculated pri-mary magma compositions are assumed to be derived byaccumulated fractional melting of fertile peridotite KR-4003 Uncertainties in fertile peridotite composition donot propagate to significant variations in melt fractionand mantle potential temperature melting of depletedperidotite propagates to calculated melt fractions that aretoo high but with a negligible error in mantle potential
temperature PRIMELT1XLS calculates the olivine liqui-dus temperature at 1atm which should be similar to theactual eruption temperature of the primary magmaand is accurate to 308C at the 2 level of confidenceMantle potential temperature is model dependent witha precision that is similar to the accuracy of the olivineliquidus temperature (C Herzberg personal communica-tion 2008)With regard to the evidence for accumulated olivine in
the Keogh Lake picrites it should not matter whether themodeled lava composition is aphyric or laden with accu-mulated olivine phenocrysts PRIMELT1 computes theprimary magma composition by adding olivine to a lavathat has experienced olivine fractionation in this way itinverts the effects of fractional crystallization The onlyrequirement is that the olivine phenocrysts are not acci-dental or exotic to the lava compositions Support for thisprocedure can be seen in the calculated olivine liquid-line-of-descent shown by the curved arrays with 5 olivineaddition increments (Fig 12c) Fractional crystallization ofolivine from model primary magmas having 85^95wt FeO and 153^175wt MgO will yield arrays of deriva-tive liquids and randomly sorted olivines that in most butnot all cases connect the picrites and high-MgO basalts Anewer version of the PRIMELT software (Herzberg ampAsimow 2008) can perform olivine addition to and sub-traction from lavas with highly variably olivine phenocrystcontents and yields results that converge with those shownin Fig 12
Fig 11 Relationship between abundance of olivine phenocrysts and whole-rock MgO contents for the Keogh Lake picrites (a) Tracings ofolivine grains (gray) in six high-MgO samples made from high resolution (2400 dpi) scans of petrographic thin-sections Scale bars represent10mm sample numbers are indicated at the upper left (b) Modal olivine vs whole-rock MgO Continuous line is the best-fit line for 10 high-MgO samplesThe areal proportion of olivine was calculated from raster images of olivine grains using ImageJ image analysis software whichprovides an acceptable estimation of the modal abundance (eg Chayes 1954)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
25
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
Gray lines = final melting pressure
L + Ol + Opx
07
08
09
Pressure (GPa)
10
3
4
5
6
1
SOLIDUS
01
04
0506
L + Ol
2
03
02
PeridotiteKR-4003
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
Thick black lines = initial melting pressure
Dashed lines = melt fraction
Melt Fraction
2
3 4 5 6
L+Ol+Opx
SolidusSpinel
SolidusGarnet Peridotite
7
L+Ol
0 10 20 30 4040
45
50
55
60
MgO (wt)
SiO2
(wt)
PeridotiteKR-4003
Melt Fraction
0 10 20 305
10
15
CaO(wt)
MgO (wt)
3
4 5 6 7
2
46
2
Peridotite Partial Melts
Thick black line = initial melting pressureGray lines = final melting pressure
Peridotite
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
02
Pressure (GPa)
0302
04
Pressure (GPa)
Picrite
High-MgO basalt
Tholeiitic basalt
(c)
(d)
Karmutsen flood basaltsOlivine addition modelOlivine addition (5 increment)Primary magma
PicriteHigh-MgO basaltTholeiitic basalt
(a)
(b)
800 850 900
910
920
930
940
950
Karmutsen flood basalts
Olivine addition modelOlivine addition (5 increment)Primary magma
Gray lines = Fo contents of olivine
Fig 12 Estimated primary magma compositions for three Keogh Lake picrites and high-MgO basalts (samples 4723A4 4723A135616A7) using the forward and inverse modeling technique of Herzberg et al (2007) Karmutsen compositions and modeling results areoverlain on diagrams provided by C Herzberg (a) Whole-rock FeO vs MgO for Karmutsen samples from this study Total iron estimatedto be FeO is 090 Gray lines show olivine compositions that would precipitate from liquid of a given MgO^FeO composition Black lineswith crosses show results of olivine addition (inverse model) using PRIMELT1 (Herzberg et al 2007) Gray field highlights olivinecompositions with Fo contents of 91^92 predicted to precipitate (b) FeO vs MgO showing Karmutsen lava compositions and results ofa forward model for accumulated fractional melting of fertile peridotite KR-4003 (Kettle River Peridotite Walter 1998) (c) SiO2 vsMgO for Karmtusen lavas and model results (d) CaO vs MgO showing Karmtusen lava compositions and model results To brieflysummarize the technique [see Herzberg et al (2007) for complete description] potential parental magma compositions for the high-MgOlava series were selected (highest MgO and appropriate CaO) and using PRIMELT1 software olivine was incrementally added tothe selected compositions to show an array of potential primary magma compositions (inverse model) Then using PRIMELT1 theresults from the inverse model were compared with a range of accumulated fractional melts for fertile peridotite derived fromparameterization of the experimental results for KR-4003 from Walter (1998) forward model Herzberg amp OrsquoHara (2002) A melt fractionwas sought that was unique to both the inverse and forward models (Herzberg et al 2007) A unique solution was found when there was a com-mon melt fraction for both models in FeO^MgO and CaO^MgO^Al2O3^SiO2 (CMAS) projection space This modeling assumes thatolivine was the only phase crystallizing and ignores chromite precipitation and possible augite fractionation in the mantle (Herzbergamp OrsquoHara 2002) Results are best for a residue of spinel lherzolite (not pyroxenite) The presence of accumulated olivine in samples ofthe Keogh Lake picrites used as starting compositions does not significantly affect the results because we are modeling addition of olivine(see text for further description) The tholeiitic basalts cannot be used for modeling because they are all saturated in plagthorn cpxthornolThick black line for initial melting pressure at 4 GPa is not shown in (c) for clarity Gray area in (a) indicates parental magmas thatwill crystallize olivine with Fo 91^92 at the surface Dashed lines in (c) indicate melt fraction Solidi for spinel and garnet peridotite arelabelled in (c)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
The PRIMELT1 results indicate that if the Karmutsenpicritic lavas were derived from accumulated fractionalmelting of fertile peridotite the primary magmas wouldhave contained 15^17wt MgO and 10wt CaO(Fig 12 Table 6) formed from 23^27 partial meltingand would have crystallized olivine with a Fo content of 91 (Fig 12 Table 6) Calculated mantle temperatures forthe Karmutsen picrites (14908C) indicate melting ofanomalously hot mantle (100^2008C above ambient) com-pared with ambient mantle that produces mid-ocean ridgebasalt (MORB 1280^14008C Herzberg et al 2007)which initiated between 3 and 4 GPa Several picritesappear to be near-primary melts and required very littleaddition of olivine (eg sample 93G171 with measured158wt MgO) These temperature and melting esti-mates of the Karmutsen picrites are consistent withdecompression of hot mantle peridotite in an actively con-vecting plume head (ie plume initiation model) Threesamples (picrite 5614A1 and high-MgO basalts 5614A3and 5614A5) are lower in iron than the other Karmutsenlavas with FeO in the 65^80wt range (Fig 12c) andhave MgO and FeO contents that are similar to primitiveMORB from the Sequeiros fracture zone of the EastPacific Rise (EPR Perfit et al 1996) These samples
however differ from EPR MORB in having higher SiO2
and lower Al2O3 and they are restricted geographicallyto one area (Sara Lake Fig 2)We therefore have evidence for primary magmas with
MgO contents that range from a possible low of 110wt to as high as 174wt with inferred mantle potential tem-peratures from 1340 to 15208C This is similar to the rangedisplayed by lavas from the Ontong Java Plateau deter-mined by Herzberg (2004) who found that primarymagma compositions assuming a peridotite sourcewould contain 17wt MgO and 10wt CaO(Table 6) and that these magmas would have formed by27 melting at a mantle potential temperature of15008C (first melting occurring at 36 GPa) Fitton ampGodard (2004) estimated 30 partial melting for theOntong Java lavas based on the Zr contents of primarymagmas of Kroenke-type basalt calculated by incrementaladdition of equilibrium olivine However if a considerableamount of eclogite was involved in the formation of theOntong Java plateau magmas the excess temperatures esti-mated by Herzberg et al (2007) may not be required(Korenaga 2005) The variability in mantle potential tem-perature for the Keogh Lake picrites is also similar to thewide range of temperatures that have been calculated for
Table 6 Estimated primary magma compositions for Karmutsen basalts and other oceanic plateaus or islands
Sample 4723A4 4723A13 5616A7 93G171 Average OJP Mauna Keay Gorgonaz
Ontong Java primary magma composition for accumulated fraction melting (AFM) from Herzberg (2004)yMauna Kea primary magma composition is average of four samples of Herzberg (2006 table 1)zGorgona primary magma composition for AFM for 1F fertile source from Herzberg amp OrsquoHara (2002 table 4)Total iron estimated to be FeO is 090
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
27
lavas from single volcanoes in the Galapagos Hawaii andelsewhere by Herzberg amp Asimow (2008) Those workersinterpreted the range to reflect the transport of magmasfrom both the cool periphery and the hotter axis of aplume A thermally heterogeneous plume will yield pri-mary magmas with different MgO contents and the rangeof primary magma compositions and their inferred mantlepotential temperatures for Karmutsen lavas may reflectmelting from different parts of the plume
Source of the Karmutsen Formation lavasThe Sr^Nd^Hf^Pb isotopic compositions of theKarmutsen volcanic rocks on Vancouver Island indicatethat they formed from plume-type mantle containing adepleted component that is compositionally similar to butdistinct from other oceanic plateaus that formed in thePacific Ocean (eg Ontong Java and Caribbean Fig 13)Trace element and isotopic constraints can be used to testwhether the depleted component of the KarmutsenFormation was intrinsic to the mantle plume and distinctfrom the source of MORB or the result of mixing or invol-vement of ambient depleted MORB mantle with plume-type mantle The Karmutsen tholeiitic basalts have initialeNd and
87Sr86Sr ratios that fall within the range of basaltsfrom the Caribbean Plateau (eg Kerr et al 1996 1997Hauff et al 2000 Kerr 2003) and a range of Hawaiian iso-tope data (eg Frey et al 2005) and lie within and abovethe range for the Ontong Java Plateau (Tejada et al 2004Fig 13a) Karmutsen tholeiitic basalts with the highest ini-tial eNd and lowest initial 87Sr86Sr lie within and justbelow the field for age-corrected northern EPR MORB at230 Ma (eg Niu et al1999 Regelous et al1999) Initial Hfand Nd isotopic compositions for the Karmutsen samplesare noticeably offset to the low side of the ocean islandbasalt (OIB) array (Vervoort et al 1999) and lie belowand slightly overlap the low eHf end of fields for Hawaiiand the Caribbean Plateau (Fig 13c) the Karmutsen sam-ples mostly fall between and slightly overlap a field forage-corrected Pacific MORB at 230 Ma and Ontong Java(Mahoney et al 1992 1994 Nowell et al 1998 Chauvel ampBlichert-Toft 2001) Karmutsen lava Pb isotope ratios over-lap the broad range for the Caribbean Plateau and thelinear trends for the Karmutsen samples intersect thefields for EPR MORB Hawaii and Ontong Java in208Pb^206Pb space but do not intersect these fields in207Pb^206Pb space (Fig 13d and e) The more radiogenic207Pb204Pb for a given 206Pb204Pb of the Karmutsenbasalts requires high UPb ratios at an ancient time when235U was abundant similar to the Caribbean Plateau andenriched in comparison with EPR MORB Hawaii andOntong JavaThe low eHf^low eNd component of the Karmutsen
basalts that places them below the OIB mantle array andpartly within the field for age-corrected Pacific MORBcan form from low ratios of LuHf and high SmNd over
time (eg Salters amp White 1998) MORB will develop aHf^Nd isotopic signature over time that falls below themantle array Low LuHf can also lead to low 176Hf177Hffrom remelting of residues produced by melting in theabsence of garnet (eg Salters amp Hart 1991 Salters ampWhite 1998) The Hf^Nd isotopic compositions of theKarmutsen Formation indicate the possible involvementof depleted MORB mantle however similar to Iceland(Fitton et al 2003) Nb^Zr^Y variations provide a way todistinguish between a depleted component within theplume and a MORB-like component The Karmutsenbasalts and picrites fall within the Iceland array and donot overlap the MORB field in a logarithmic plot of NbYvs ZrY (Fig 13b) using the fields established by Fittonet al (1997) Similar to the depleted component within theCaribbean Plateau (Thompson et al 2003) the trend ofHf^Nd isotopic compositions towards Pacific MORB-likecompositions is not followed by MORB values in Nb^Zr^Y systematics and this suggests that the depleted compo-nent in the source of the Karmutsen basalts is distinctfrom the source of MORB and probably an intrinsic partof the plume The relatively radiogenic Pb isotopic ratiosand REE patterns distinct from MORB also support thisinterpretationTrace-element characteristics suggest the possible invol-
vement of sub-arc lithospheric mantle in the generation ofthe Karmutsen picrites Lassiter et al (1995) suggested thatmixing of a plume-type source with eNdthorn 6 to thorn 7 witharc material with low NbTh could reproduce the varia-tions observed in the Karmutsen basalts however theyconcluded that the absence of low NbLa ratios in theirsample of tholeiitic basalts restricts the amount of arc litho-sphere involved The study of Lassiter et al (1995) wasbased on major and trace element and Sr Nd and Pb iso-topic data for a suite of 29 samples from Buttle Lake in theStrathcona Provincial Park on central Vancouver Island(Fig 1) Whereas there are no clear HFSE depletions inthe Karmutsen tholeiitic basalts the low NbLa (and TaLa) of the Karmutsen picritic lavas in this study indicatethe possible involvement of Paleozoic arc lithosphere onVancouver Island (Fig 8)
REE modeling dynamic melting andsource mineralogyThe combined isotopic and trace-element variations of thepicritic and tholeiitic Karmutsen lavas indicate that differ-ences in trace elements may have originated from differentmelting histories For example the LuHf ratios of theKarmutsen picritic lavas (mean 008) are considerablyhigher than those in the tholeiitic basalts (mean 002)however the small range of overlapping initial eHf valuesfor the two lava suites suggests that these differences devel-oped during melting within the plume (Fig 9b) Below wemodel the effects of source mineralogy and depth of
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
28
melting on differences in trace-element concentrations inthe tholeiitic basalts and picritesDynamic melting models simulate progressive decom-
pression melting where some of the melt fraction isretained in the residue and only when the degree of partialmelting is greater than the critical mass porosity and the
source becomes permeable is excess melt extracted fromthe residue (Zou1998) For the Karmutsen lavas evolutionof trace element concentrations was simulated using theincongruent dynamic melting model developed by Zou ampReid (2001) Melting of the mantle at pressures greater than1 GPa involves incongruent melting (eg Longhi 2002)
OIB array
PacificMORB(230 Ma)
(0 Ma)
EPR(230 Ma)
-4
-2
0
2
4
8
10
12
14
16
18
20
22
-4 -2 0 2 4 6 8 10 12 14
minus2
0
2
4
6
8
10
12
14
0702 0703 0704 0705 0706
IndianMORB
87Sr 86Sr (initial)
Hf (initial)
(a) (c)
Nd (initial)
Karmutsen tholeiitic basalt
HawaiiOntong JavaCaribbean Plateau
Karmutsen high-MgO lava
high-MgO lavasKarmutsen
1540
1545
1550
1555
1560
1565
175 180 185 190 195 200375
380
385
390
395
175 180 185 190 195 200
(d) (e)
EPR
EPR
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb
Karmutsen Formation (initial)
HawaiiOntong JavaCaribbean Plateau
East Pacific Rise
Karmutsen Formation (measured)
Juan de FucaGorda
ε Nd
(initial)
Karmutsen Formation (different age-correction for 3 picrites)
(0 Ma)
NbY
001
01
1
1 10
ZrY
OIB
NMORB
(b)
Iceland array
6
Fig 13 Comparison of age-corrected (230 Ma) Sr^Nd^Hf^Pb isotopic compositions for Karmutsen flood basalts onVancouver Island withage-corrected OIB and MORB and Nb^Zr^Y variation (a) Initial eNd vs 87Sr86Sr (b) NbYand ZrY variation (after Fitton et al 1997)(c) Initial eHf vs eNd (d) Measured and initial 207Pb204Pb vs 206Pb204Pb (e) Measured and initial 208Pb204Pb vs 206Pb204Pb References fordata sources are listed in Electronic Appendix 4 Most of the compiled data were extracted from the GEOROC database (httpgeorocmpch-mainzgwdgdegeoroc) OIB array line in (c) is fromVervoort et al (1999) EPR East Pacific Rise Several high-MgO samples affected bysecondary alteration were not plotted for clarity in Pb isotope plots Estimation of fields for age-corrected Pacific MORB and EPR at 230 Mawere made using parent^daughter concentrations (in ppm) from Salters amp Stracke (2004) of Lufrac14 0063 Hffrac14 0202 Smfrac14 027 Ndfrac14 071Rbfrac14 0086 and Srfrac14 836 In the NbYvs ZrYplot the normal (N)-MORB and OIB fields not including Icelandic OIB are from data com-pilations of Fitton et al (1997 2003) The OIB field is data from 780 basalts (MgO45wt ) from major ocean islands given by Fitton et al(2003) The N-MORB field is based on data from the Reykjanes Ridge EPR and Southwest Indian Ridge from Fitton et al (1997) The twoparallel gray lines in (b) (Iceland array) are the limits of data for Icelandic rocks
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
29
with olivine being produced during the reactioncpxthornopxthorn spfrac14meltthornol for spinel peridotites The mod-eling used coefficients for incongruent melting reactionsbased on experiments on lherzolite melting (eg Kinzleramp Grove 1992 Longhi 2002) and was separated into (1)melting of spinel lherzolite (2) two combined intervals ofmelting for a garnet lherzolite (gt lherz) and spinel lher-zolite (sp lherz) source (3) melting of garnet lherzoliteThe model solutions are non-unique and a range indegree of melting partition coefficients melt reactioncoefficients and proportion of mixed source componentsis possible to achieve acceptable solutions (see Fig 14 cap-tion for description of modeling parameters anduncertainties)The LREE-depleted high-MgO lavas require melting of
a depleted spinel lherzolite source (Fig 14a) The modelingresults indicate a high degree of melting (22^25) similarto the results from PRIMELT1 based on major-elementvariations (discussed above) of a LREE-depleted sourceThe high-MgO lavas lack a residual garnet signature Theenriched tholeiitic basalts involved melting of both garnetand spinel lherzolite and represent aggregate melts pro-duced from continuous melting throughout the depth ofthe melting column (Fig 14b) Trace-element concentra-tions in the tholeiitic and picritic lavas are not consistentwith low-degree melting of garnet lherzolite alone(Fig 14c) The modeling results indicate that the aggregatemelts that formed the tholeiitic basalts would haveinvolved lesser proportions of enriched small-degreemelts (1^6 melting) generated at depth and greateramounts of depleted high-degree melts (12^25) gener-ated at lower pressure (a ratio of garnet lherzolite tospinel lherzolite melt of 14 provides the best fits seeFig 14b and description in figure caption) The totaldegree of melting for the tholeiitic basalts was high(23^27) similar to the degree of partial melting in thespinel lherzolite facies for the primary melts that eruptedas the Keogh Lake picrites of a source that was also prob-ably depleted in LREE
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
(c)
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
Melting of garnet lherzolite
High-MgO lavas
Melting of spinel lherzolite
source (07DM+03PM)
source (07DM+03PM)
6 melting
30 melting
(b)
(a)
30 melting
Sam
ple
Cho
ndrit
eS
ampl
eC
hond
rite 20 melting
1 melting
Melting of garnet lherzoliteand spinel lherzolite
x gt lherz + 4x sp lherz
5 meltingx gt lherz + 4x sp lherz
Sam
ple
Cho
ndrit
e
Tholeiitic lavas
Tholeiitic lavas
High-MgO lavas
source (07DM+03PM)
Fig 14 Trace-element modeling results for incongruent dynamicmantle melting for picritic and tholeiitic lavas from the KarmutsenFormation Three steps of modeling are shown (a) melting of spinellherzolite compared with high-MgO lavas (b) melting of garnet lher-zolite and spinel lherzolite compared with tholeiitic lavas (c) meltingof garnet lherzolite compared with tholeiitic and picritic lavasShaded fields in each panel for tholeiitic basalts and picrites arelabeled Patterns with symbols in each panel are modeling results(patterns are 5 melting increments in (a) and (b) and 1 incre-ments in (c) PM primitive mantle DM depleted MORB mantle gtlherz garnet lherzolite sp lherz spinel lherzolite In (b) the ratios ofper cent melting for garnet and spinel lherzolite are indicated (x gtlherzthorn 4x sp lherz means that for 5 melting 1 melt in garnetfacies and 4 melt in spinel facies where x is per cent melting in theinterval of modeling) The ratio of the respective melt proportions inthe aggregate melt (x gt lherzthorn 4x sp lherz) was kept constant andthis ratio of melt contribution provided the best fit to the data Arange of proportion of source components (07DMthorn 03PM to
09DMthorn 01PM) can reproduce the variation in the high-MgOlavas Melting modeling uses the formulation of Zou amp Reid (2001)an example calculation is shown in their Appendix PM fromMcDonough amp Sun (1995) and DM from Salters amp Stracke (2004)Melt reaction coefficients of spinel lherzolite from Kinzler amp Grove(1992) and garnet lherzolite fromWalter (1998) Partition coefficientsfrom Salters amp Stracke (2004) and Shaw (2000) were kept constantSource mineralogy for spinel lherzolite (018cpx027opx052ol003sp) from Kinzler (1997) and for garnet lherzolite(034cpx008opx053ol005gt) from Salters amp Stracke (2004) Arange of source compositions for melting of spinel lherzolite(07DMthorn 03PM to 03DMthorn 07PM) reproduce REE patternssimilar to those of the high-MgO lavas with best fits for 22ndash25melting and 07DMthorn 03PM source composition Concentrationsof tholeiitic basalts cannot be reproduced from an entirelyprimitive or depleted source
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
30
The uncertainty in the composition of the source in thespinel lherzolite melting model for the high-MgO lavasprecludes determining whether the source was moredepleted in incompatible trace elements than the source ofthe tholeiitic lavas (see Fig 14 caption for description) It ispossible that early high-pressure low-degree melting left aresidue within parts of the plume (possibly the hotter inte-rior portion) that was depleted in trace elements (egElliott et al 1991) further decompression and high-degreemelting of such depleted regions could then have generatedthe depleted high-MgO Karmutsen lavas Shallow high-degree melting would preferentially sample depletedregions whereas deeper low-degree melting may notsample more refractory depleted regions (eg CaribbeanEscuder-Viruete et al 2007) The picrites lie at the highend of the range of Nd isotopic compositions and havelow NbZr similar to depleted Icelandic basalts (Fittonet al 2003) therefore the picrites may represent the partsof the mantle plume that were the most depleted prior tothe initiation of melting As Fitton et al (2003) suggestedthese depleted melts may only rarely be erupted withoutcombining with more voluminous less depleted melts andthey may be detected only as undifferentiated small-volume flows like the Keogh Lake picrites within theKarmutsen Formation on Vancouver IslandDecompression melting within the mantle plume initiatedwithin the garnet stability field (35^25 GPa) at highmantle potential temperatures (Tp414508C) and pro-ceeded beneath oceanic arc lithosphere within the spinelstability field (25^09 GPa) where more extensivedegrees of melting could occur
Magmatic evolution of Karmutsentholeiitic basaltsPrimary magmas in large igneous provinces leave exten-sive coarse-grained cumulate residues within or below thecrust these mafic and ultramafic plutonic sequences repre-sent a significant proportion of the original magma thatpartially crystallized at depth (eg Farnetani et al 1996)For example seismic and petrological studies of OntongJava (eg Farnetani et al 1996) combined with the use ofMELTS (Ghiorso amp Sack 1995) indicate that the volcanicsequence may represent only 40 of the total magmareaching the Moho Neal et al (1997) estimated that30^45 fractional crystallization took place for theOntong Java magmas and produced cumulate thicknessesof 7^14 km corresponding to 11^22 km of flood basaltsdepending on the presence of pre-existing oceanic crustand the total crustal thickness (25^35 km) Interpretationsof seismic velocity measurements suggest the presence ofpyroxene and gabbroic cumulates 9^16 km thick beneaththe Ontong Java Plateau (eg Hussong et al 1979)Wrangellia is characterized by high crustal velocities in
seismic refraction lines which is indicative of mafic
plutonic rocks extending to depth beneath VancouverIsland (Clowes et al 1995)Wrangellia crust is 25^30 kmthick beneath Vancouver Island and is underlain by astrongly reflective zone of high velocity and density thathas been interpreted as a major shear zone where lowerWrangellia lithosphere was detached (Clowes et al 1995)The vast majority of the Karmutsen flows are evolved tho-leiitic lavas (eg low MgO high FeOMgO) indicating animportant role for partial crystallization of melts at lowpressure and a significant portion of the crust beneathVancouver Island has seismic properties that are consistentwith crystalline residues from partially crystallizedKarmutsen magmas To test the proportion of the primarymagma that fractionated within the crust MELTS(Ghiorso amp Sack 1995) was used to simulate fractionalcrystallization using several estimated primary magmacompositions from the PRIMELT1 modeling results Themajor-element composition of the primary magmas couldnot be estimated for the tholeiitic basalts because of exten-sive plagthorn cpxthornol fractionation and thus the MELTSmodeling of the picrites is used as a proxy for the evolutionof major elements in the volcanic sequence A pressure of 1kbar was used with variable water contents (anhydrousand 02wt H2O) calculated at the quartz^fayalite^magnetite (QFM) oxygen buffer Experimental resultsfrom Ontong Java indicate that most crystallizationoccurs in magma chambers shallower than 6 km deep(52 kbar Sano amp Yamashita 2004) however some crys-tallization may take place at greater depths (3^5 kbarFarnetani et al 1996)A selection of the MELTS results are shown in Fig 15
Olivine and spinel crystallize at high temperature until15^20wt of the liquid mass has fractionated and theresidual magma contains 9^10wt MgO (Fig 15f) At1235^12258C plagioclase begins crystallizing and between1235 and 11908C olivine ceases crystallizing and clinopyr-oxene saturates (Fig 15f) As expected the addition of clin-opyroxene and plagioclase to the crystallization sequencecauses substantial changes in the composition of the resid-ual liquid FeO and TiO2 increase and Al2O3 and CaOdecrease while the decrease in MgO lessens considerably(Fig 15) The near-vertical trends for the Karmutsen tho-leiitic basalts especially for CaO do not match the calcu-lated liquid-line-of-descent very well which misses manyof the high-CaO high-Al2O3 low-FeO basalts A positivecorrelation between Al2O3CaO and SrNd indicates thataccumulation of plagioclase played a major role in the evo-lution of the tholeiitic basalts (Fig 15g) Increasing thecrystallization pressure slightly and the water content ofthe melts does not systematically improve the match ofthe MELTS models to the observed dataThe MELTS results indicate that a significant propor-
tion of the crystallization takes place before plagioclase(15^20 crystallization) and clinopyroxene (35^45
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
31
crystallization) join the crystallization sequence Theseresults suggest that445 crystallization of the primarymagmas occurred (estimated from the residual liquid percent in Fig 15f) and if the original flood basalt stratigra-phy on Vancouver Island was 6 km thick at least anequivalent amount of complementary plutonic rocksshould have crystallized The Karmutsen basalts and theirplutonic residues thus represent a significant addition ofcrust (perhaps 412 km thickness) previously thickenedduring the formation of the underlying Paleozoic arc
Growth of the Wrangellia oceanic plateauon Vancouver IslandThe extrusive part of the Wrangellia oceanic plateau onVancouver Island was constructed in a three-layersequence of mostly tholeiitic basalts with a restrictedrange of composition (Fig 16) The plateau overliesDevonian to Mississippian arc volcanic sequences andmarine sediments of Mississippian to Permian ages thatare intruded by gabbroic rocks related to the flood basaltsThe volcanic stratigraphy of theWrangellia plateau formed
00
05
10
15
20
25
30
35
40
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
16
0 2 4 6 8 10 12 14 16 18 20
10
11
12
13
14
15
16
17
18
19
20
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
0 2 4 6 8 10 12 14 16 18 20
MgO (wt)
0 10 20 30 40 50
percent of total mass
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
) Mineral proportions
Sp (e)
Ol
CpxPlag
PicriteHigh-MgO basaltTholeiitic basalt
TiO2 (wt)
MgO (wt)
Al2O3 (wt) CaO (wt)
FeO (wt)
MgO(a) (b)
(c) (d)
MgO (wt)MgO (wt)
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
)
Residual liquid ()
1220
1190
Ol + Sp
Plag+Ol+Sp Plag+Cpx+Sp
02 wt H2O
no H2O
Residual liquid and crystallizing phases
(f)
0102030405060708090100
plagioclaseaccumulation
plagioclaseaccumulation
plagioclase
accumulatio
n
10
12
14
16
0 10 20 30 40
Al2O3CaO
SrNd
tholeiitic basalts
(g)
Fig 15 Forward fractional crystallization modeling results for major elements from MELTS (Ghiorso amp Sack 1995) compared with picriticand tholeiitic lavas from the Karmutsen Formation (a^d) Major-element compositions of residual liquids compared with Karmutsen lavacompositions For clarity the results are shown only for one starting composition (open circle) using the composition of the estimated primarymagma for the high-MgO lava series from the modeling technique of Herzberg et al (2007) for sample 4723A13 as shown in Fig 12 The esti-mated primary magmas for samples 4723A3 and 93G171 yield similar results A pressure of 1 kbar was used with no H2O and 02wt H2Ocalculated at QFM black line is anhydrous and grey line is 02wt H2OThe end-point for the anhydrous result is at 78 fractionation and11308C and the end-point for the result with 02wt H2O is 75 fractionation and11108C (e) Mineral proportions of crystallized phases fromone MELTrun (sample 4723A13) with no H2O (f) Per cent residual liquid vs temperature labelled with intervals of crystallizing phases (g)Al2O3CaO vs SrNd for the tholeiitic basalts
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
32
in a single eruptive phase (c 231^225 Ma) as an emergentbasalt sequence during a deeper-water shallow-water andsubaerial stage and subsided below sea level relativelysoon after its formation (Fig 16) The deeper-water stagecontains pillowed and massive submarine flows (3000mthick) that built up the submarine volcanic edifice from theseafloor The lowest pillow basalts were emplaced onunconsolidated and lithified fine-grained siliciclastic andcarbonaceous sediments containing Daonella beds(c 235^232 Ma) that are intruded by abundantKarmutsen mafic sills (Fig 16a) The different pillowedand massive submarine flows resulted primarily from dif-ferent effusive rates and local topography Some of themassive submarine flows commonly recognized by radiat-ing columnar jointing represent master tubes for the deliv-ery of lava to distal parts of submarine flow fields Thepillowed flows which commonly have uneven topography
in modern sequences may have served as levees to themassive submarine flows rather than allowing unob-structed spreading and inflation typical of the massive sub-aerial flows that form on shallow slopes in continentalflood basalt provinces (eg Self et al 1997) The rapid suc-cession of flows is evident by the near absence of sedimentbetween submarine flows Minor volumes of picritic pillowlavas occur in the upper part of the pillowed flow sequencein areas of the plateau exposed on northern VancouverIsland (Fig 16 Nixon et al 2008)The growth of the plateau in shallow water led to
emplacement of an increasing proportion of volcaniclastics(pillow breccias and hyaloclastites) that accumulated up to1500m thick in some areas (Fig 16b Nixon et al 2008)Horizons of pillowed flows occur locally within the volca-niclastic stratigraphy Sedimentary structures (eg gradedbedding fluidization structures) and petrographic textures
Late Devonian-Mississippian arc
Daonella beds
Middle Triassic
deeper water
Late Devonian-Mississippian arc
pillowed flows massive submarine
flows
pillow brecciahyaloclastiteshallow water
~4 km
~6 km
Late Triassic limestoneinterflow limestonelenses
Karmutsen Formation
mafic intrusions sediment-sill complex
Karmutsen Formation
(a) (b)
(c) (d)
subaerial flows
Late Devonian-Mississippian arc Late Devonian-Mississippian arc
picritic pillow basalts
Miss to Permian marine sediments
sills
Fig 16 Growth of the Wrangellia oceanic plateau onVancouver Island (a) Deep-water eruptive phase of pillowed and massive submarineflows with almost no intervening sediments onto MiddleTriassic shale containing Daonella beds (c 235^232 Ma) The basement of the plateauis composed of Devonian to Mississippian arc sequences overlain by Mississippian to Permian carbonates and siliciclastic sediments(b) Shallow-water flow emplacement with increasing proportion of volcaniclastics resulted primarily from cooling^contraction granulationPicritic pillow basalts occur primarily near the transition between pillowed submarine flows and volcaniclastic units Pillowed and massiveflows occur locally within the volcaniclastic unit (c) Emergent stage of growth consists of overlapping subaerial flow fields similar to continen-tal flood basalts with pahoehoe flow structures Multi-tiered columnar jointing occurrs locally in response to ponding of flows (eg Bondre et al2004) Interflow carbonate deposits and plagioclase megacrystic flows developed locally during the waning stage of volcanism (d) Post-volcanicsubsidence and deposition of Late Carnian to Norian (c 212^223 Ma) platform carbonates on top of the plateau Units are labeled in the dia-grams Thicknesses of the underlying units are not to scale
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
33
within the volcaniclastic rocks indicate formation primar-ily via cooling^contraction granulation magma^water^steam interaction autobrecciation resedimentation andmass-wasting rather than pyroclastic eruption (Nixonet al 2008)A broad subaerial platform was constructed by emplace-
ment of subaerial lava flows during an emergent subaerialstage (Fig 16c) The subaerial unit predominantly com-prises well-layered sheet flows that formed in a similarway to continental flood basalts as hundred(s) of pahoehoeflow fields erupted on shallow slopes (Self et al 1997) how-ever volcaniclastic and pillowed flow units are locally pre-sent mostly in the upper parts of the stratigraphyLocalized ponding of flows led to the formation of multi-tiered columnar jointing in some subaerial flows (egBondre et al 2004) As volcanism waned local interflowcarbonate deposits developed and plagioclase megacrysticflows were erupted Subsidence of the plateau is recordedby the interflow lenses and deposits of platform carbonates(40^1500m thick) on top of the Karmutsen basalts thatcontain Late Carnian to Early Norian fossils (Fig 16d c216^227 Ma Nixon et al 2008)The evolution of the internal architecture of the
Wrangellia oceanic plateau was primarily related to theconditions of emplacement (ie water depth and eruptionenvironment) There does not appear to have been a sys-tematic coincident evolution or shift of lava compositionsonVancouver Island as has been recognized in most con-tinental flood basalts and parts of the Wrangellia oceanicplateau in Alaska and Yukon (Greene et al 2008 2009)In general with the exception of the picritic pillow lavasin the submarine stratigraphy and local variations betweenhigh-MgO and tholeiitic lavas there are no significantcompositional differences between lavas erupted duringthe different stages of growth of the plateau (ie submarineand subaerial flows) nor are substantial compositional dif-ferences generally found between massive submarine flowsin contact with pillowed flows tholeiitic basalt flows nearthe base of the Karmutsen Formation are compositionallysimilar to flows sampled near the top of the volcanicsequence Where it was possible to sample continuous sec-tions of volcanic stratigraphy greater than several hun-dreds of meters (eg Keogh^Maynard Lake area west ofthe Karmutsen Range Fig 2) there is no clear relationshipbetween stratigraphic position and compositionalvariation
CONCLUSIONSThe Karmutsen Formation covers extensive areas ofVancouver Island (20 000 km2) and forms part of a largeoceanic plateau erupted over a geologically short interval(c 225^231 Ma) The tripartite volcanic stratigraphyexposed on Vancouver Island is at least 6 km thick andcomprises a succession of submarine flows volcaniclastic
deposits and massive subaerial flowsThese volcanologicaldifferences are primarily related to the eruption environ-ment (deep-water shallow-water or subaerial) The rapidgrowth of the plateau prevented significant accumulationsof intervening sediment except in the uppermost stratigra-phy where isolated limestone lenses commonly associatedwith pillow basalt and hyaloclastite deposits preserve arecord of the final stages of subsidence of the plateauThe Wrangellia plateau on Vancouver Island was con-
structed dominantly of tholeiitic basalt with restrictedmajor- and trace-element and isotopic compositional var-iations Minor volumes of picritic pillow basalts eruptedlate in the submarine phase The Karmutsen basalts wereemplaced onto older Paleozoic arc volcanic sequences andlow NbLa for the picrites may indicate a contributionfrom the sub-arc lithospheric mantle Modeling of the pet-rogenesis of the Karmutsen picrites indicates melting ofanomalously hot mantle (15008C) and extensive degreesof partial melting (23^27) and is consistent with aplume initiation model Combined Sr^Nd^Hf^Pb isotopicsystematics indicate that the picritic and tholeiitic lavaswere derived from a depleted plume-type mantle sourcedistinct from the source of MORB that has compositionalsimilarities to the source of the lavas forming theCaribbean Plateau Picrites were generated by melting ofspinel lherzolite whereas the tholeiitic basalts formed asaggregate melts of both garnet and spinel lherzolite Thetholeiitic basalts underwent extensive low-pressure fractio-nation (52^3 kbar) and seismic studies indicate that someof theWrangellia crust beneathVancouver Island may cor-respond to the plutonic residues of the Karmutsen floodbasalts
ACKNOWLEDGEMENTSWe are grateful to Nick Arndt for helping us get this pro-ject started and Nick Massey for insights into VancouverIsland geology We greatly appreciate the assistance withPRIMELT modeling from Claude Herzberg We thankMikkel Schau for his insight and enthusiasm during field-work Jane Barling assisted with analyses by MC-ICP-MSThoughtful and thorough reviews of the manuscript byJohn Mahoney Fred Frey Wendy Bohrson DougalJerram and Teal Riley are greatly appreciated Fundingwas generously provided by the Rocks to Riches Programadministered by the BC amp Yukon Chamber of Mines in2004 by the BC Geological Survey and by NSERCDiscovery Grants to JSS and DW ARG was supportedby a University Graduate Fellowship at UBC
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
34
REFERENCESBarker F Brown A S Budahn J R amp Plafker G (1989) Back-arc
with frontal-arc component origin of Triassic Karmutsen basaltBritish Columbia Canada Chemical Geology 75 81^102
Blichert-Toft JWeis D Maerschalk C Agranier A amp Albare de F(2003) Hawaiian hot spot dynamics as inferred from the Hf and Pbisotope evolution of Mauna Kea volcano Geochemistry GeophysicsGeosystems 4(2) 1^27 doi1010292002GC000340
Bondre N R Duraiswami R A amp Dole G (2004) Morphologyand emplacement of flows from the Deccan Volcanic ProvinceIndia Bulletin of Volcanology 66 29^45
Carlisle D (1963) Pillow breccias and their aquagene tuffs QuadraIsland British Columbia Journal of Geology 71 48^71
Carlisle D (1972) Late Paleozoic to Mid-Triassic sedimentary-volcanic sequence on Northeastern Vancouver Island Geological
Society of Canada Paper Report of Activities 72-1 Part B 24^30Carlisle D amp SuzukiT (1974) Emergent basalt and submergent car-
bonate^clastic sequences including the Upper Triassic Dilleri andWelleri zones on Vancouver Island Canadian Journal of Earth
Sciences 11 254^279Chauvel C amp Blichert-Toft J (2001) A hafnium isotope and trace
element perspective on melting of the depleted mantle Earth and
Planetary Science Letters 190 137^151Chayes F (1954)The theory of thin-section analysis Journal of Geology
62 92^101Cho M Liou J G amp Maruyama S (1987) Transition from the zeo-
lite to prehnite^pumpellyite facies in the Karmutsen metabasitesVancouver Island British Columbia Journal of Petrology 27(2)467^494
Clowes R M Zelt C A Amor J R amp Ellis R M (1995)Lithospheric structure in the southern Canadian Cordillera froma network of seismic refraction lines Canadian Journal of Earth
Sciences 32(10) 1485^1513Coffin M F amp Eldholm O (1994) Large igneous provinces Crustal
structure dimensions and external consequences Reviews of
Geophysics 32(1) 1^36Cox K G (1980) A model for flood basalt volcanism Journal of
Petrology 21 629^650Eldholm O amp Coffin M F (2000) Large igneous provinces
and plate tectonics In Richards M A Gordon R G amp vander Hilst R D (eds) The History and Dynamics of Global
Plate Motions American Geophysical Union Geophysical Monograph 121309^326
Elliott T R Hawkesworth C J amp Grolaquo nvold K (1991) Dynamicmelting of the Iceland plume Nature 351 201^206
Escuder-Viruete J Pecurren rez-Estaucurren n A Contreras F Joubert MWeis D Ullrich T D amp Spadea P (2007) Plume mantle sourceheterogeneity through time Insights from the Duarte ComplexHispaniola northeastern Caribbean Journal of Geophysical Research112(B04203) doi1010292006JB004323
Farnetani C G Richards M A amp Ghiorso M S (1996)Petrological models of magma evolution and deep crustal structurebeneath hotspots and flood basalt provinces Earth and Planetary
Science Letters 143 81^94Fitton J G amp Godard M (2004) Origin and evolution of magmas
on the Ontong Java Plateau In Fitton J G Mahoney J JWallace P J amp Saunders A D (eds) Origin and Evolution of the
Ontong Java Plateau Geological Society London Special Publications 229151^178
Fitton J G Saunders A D Norry M J Hardarson B S ampTaylor R N (1997) Thermal and chemical structure of theIceland plume Earth and Planetary Science Letters 153 197^208
Fitton J G Saunders A D Kempton P D amp Hardarson B S(2003) Does depleted mantle form an intrinsic part of the Icelandplume Geochemistry Geophysics Geosystems 4(3) 1032 doi1010292002GC000424
Frey F A Huang S Blichert-Toft J Regelous M amp Boyet M(2005) Origin of depleted components in basalt related to theHawaiian hotspot Evidence from isotopic and incompatible ele-ment ratios Geochemistry Geophysics Geosystems 6(1) 23 doi1010292004GC000757
Galer S J G amp Abouchami W (1998) Practical application of leadtriple spiking for correction of instrumental mass discriminationMineralogical Magazine 62A 491^492
Ghiorso M S amp Sack R O (1995) Chemical mass transfer in mag-matic processes IV A revised and internally consistent thermody-namic model for the interpolation and extrapolation of liquid^solidequilibria in magmatic systems at elevated temperatures and pres-sures Contributions to Mineralogy and Petrology 119 197^212
Greene A R Scoates J S amp Weis D (2005)WrangelliaTerrane onVancouver Island Distribution of flood basalts with implicationsfor potential Ni^Cu^PGE mineralization In Grant B (ed)Geological Fieldwork 2004 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2005-1 209^220Greene A R Scoates J S Nixon G T amp Weis D (2006) Picritic
lavas and basal sills in the Karmutsen flood basalt provinceWrangellia northern Vancouver Island In Grant B (ed)Geological Fieldwork 2005 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2006-1 39^52Greene A R Scoates J S amp Weis D (2008) Wrangellia flood
basalts in Alaska A record of plume^lithosphere interaction in aLate Triassic accreted oceanic plateau Geochemistry Geophysics
Geosystems 9(Q12004) doi1010292008GC002092Greene A R Scoates J S Weis D amp Israel S (2009)
Geochemistry of triassic flood basalts from the Yukon (Canada)segment of the accreted Wrangellia oceanic plateau Lithosdoi101016jlithos200811010
Hauff F Hoernle K Tilton G Graham D W amp Kerr A C(2000) Large volume recycling of oceanic lithosphere over shorttime scales geochemical constraints from the Caribbean LargeIgneous Province Earth and Planetary Science Letters 174 247^263
Hauff F Hoernle K amp Schmidt A (2003) Sr^Nd^Pb compositionof Mesozoic Pacific oceanic crust (Site 1149 and 801 ODP Leg185)Implications for alteration of ocean crust and the input into theIzu^Bonin^Mariana subduction system Geochemistry Geophysics
Geosystems 4(8) doi1010292002GC000421Herzberg C (2004) Partial melting below the Ontong Java Plateau
In Fitton J G Mahoney J J Wallace P J amp Saunders A D(eds) Origin and Evolution of the Ontong Java Plateau Geological SocietyLondon Special Publications 229 179^183
Herzberg C (2006) Petrology and thermal structure of the Hawaiianplume from Mauna Kea volcano Nature 444(7119) 605
Herzberg C amp Asimow P D (2008) Petrology of some oceanicisland basalts PRIMELT2XLS software for primary magma cal-culation Geochemistry Geophysics Geosystems 9(Q09001) doi1010292008GC002057
Herzberg C amp OrsquoHara M J (2002) Plume-associated ultramaficmagmas of Phanerozoic age Journal of Petrology 43(10) 1857^1883
Herzberg C Asimow P D Arndt N Niu Y Lesher C MFitton J G Cheadle M J amp Saunders A D (2007)Temperatures in ambient mantle and plumes Constraints frombasalts picrites and komatiites Geochemistry Geophysics Geosystems8(Q02006) doi1010292006GC001390
Huang S amp Frey F A (2005) Trace element abundances of MaunaKea basalt from phase 2 of the Hawaii Scientific Drilling Project
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
35
Petrogenetic implications of correlations with major element con-tent and isotopic ratios Geochemistry Geophysics Geosystems 4(6)doi1010292002GC000322
Hussong D M Wipperman L K amp Kroenke L W (1979) Thecrustal structure of the Ontong Java and Manihiki OceanicPlateaus Journal of Geophysical Research 84(B11) 6003^6010
Jerram D A amp Widdowson M (2005) The anatomy of ContinentalFlood Basalt Provinces geological constraints on the processes andproducts of flood volcanism Lithos 79(3^4) 385^405
Jones D L Silberling N J amp Hillhouse J (1977)Wrangellia a dis-placed terrane in northwestern North America CanadianJournal ofEarth Sciences 14(11) 2565^2577
Kerr A C (2003) Oceanic Plateaus In Rudnick R (ed) The Crust
Treatise on GeochemistryVol 3 Oxford Elsevier Science pp 537^565Kerr A C amp Mahoney J J (2007) Oceanic plateaus Problematic
plumes potential paradigms Chemical Geology 241 332^353Kerr A CTarney J Marriner G F Klaver GT Saunders A D
amp Thirlwall M F (1996) The geochemistry and petrogenesis ofthe Late-Cretaceous picrites and basalts of Curacao NetherlandsAntilles a remnant of an oceanic plateau Contributions to
Mineralogy and Petrology 124(1) 29^43Kerr A C Marriner G F Tarney J Nivia A Saunders A D
Thirlwall M F amp Sinton A C (1997) Cretaceous basaltic ter-ranes in Western Colombia Elemental chronological and Sr-Ndisotopic constraints on petrogenesis Journal of Petrology 38(6)677^702
Kinzler R J (1997) Melting of mantle peridotite at pressuresapproaching the spinel to garnet transition Application to mid-ocean ridge basalt petrogenesis Journal of Geophysical Research
102(B1) 853^874Kinzler R J amp Grove T L (1992) Primary magmas of mid-ocean
ridge basalts 1 Experiments and methods Journal of Geophysical
Research 97(B5) 6885^6906Korenaga J (2005) Why did not the Ontong Java Plateau form sub-
aerially Earth and Planetary Science Letters 234 385^399Kuniyoshi S (1972) Petrology of the Karmutsen (Volcanic) Group
northeastern Vancouver Island British Columbia PhD disserta-tion University of California Los Angeles 242 pp
Lassiter J C DePaolo D J amp Mahoney J J (1995) Geochemistry ofthe Wrangellia flood basalt province Implications for the role ofcontinental and oceanic lithosphere in flood basalt genesis Journalof Petrology 36(4) 983^1009
LeBas M J (2000) IUGS reclassification of the high-Mg and picriticvolcanic rocks Journal of Petrology 41(10) 1467^1470
Longhi J (2002) Some phase equilibrium systematics of lherzolitemelting I Geochemistry Geophysics Geosystems 3(3) doi1010292001GC000204
MacDonald G A amp Katsura T (1964) Chemical composition ofHawaiian lavas Journal of Petrology 5 82^133
Mahoney J J (1987) An isotopic survey of Pacific oceanic plateausimplications for their nature and origin In Keating B HFryer P Batiza R amp Boehlert GW (eds) Seamounts Islands andAtolls American Geophysical Union Geophysical Monograph 43 207^220
Mahoney J LeRoex A P Peng Z Fisher R L amp Natland J H(1992) Southwestern limits of Indian Ocean Ridge mantle and theorigin of low 206Pb204Pb mid-ocean ridge basalt isotope systema-tics of the central Southwest Indian Ridge (178^508E) Journal ofGeophysical Research 97(B13) 19771^19790
Mahoney J J Sinton J M Kurz M D MacDougall J DSpencer K J amp Lugmair GW (1994) Isotope and trace elementcharacteristics of a super-fast spreading ridge East Pacific rise13^238S Earth and Planetary Science Letters 121 173^193
Massey N W D (1995a) Geology and mineral resources of theAlberni^Nanaimo Lakes sheet Vancouver Island 92F1W 92F2Eand part of 92F7E British Columbia Ministry of Energy Mines and
Petroleum Resources Paper 1992-2 132 ppMassey N W D (1995b) Geology and mineral resources of the
Cowichan Lake sheet Vancouver Island 92C16 British Columbia
Ministry of Energy Mines and Petroleum Resources Paper 1992-3 112 ppMassey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005a) Digital Geology Map of British Columbia Tile NM9 MidCoast BC British Columbia Ministry of Energy and Mines Geofile
2005-2Massey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005b) Digital Geology Map of British Columbia Tile NM10Southwest BC British Columbia Ministry of Energy and Mines Geofile
2005-3McDonough W F amp Sun S (1995) The composition of the Earth
Chemical Geology 120 223^253Monger JW H amp Journeay J M (1994) Basement geology and tec-
tonic evolution of theVancouver region In Monger JW H (ed)Geology and Geological Hazards of the Vancouver Region Southwestern
British Columbia Geological Survey of Canada Bulletin 481 3^25Muller J E (1977) Geology of Vancouver Island Geological Survey of
Canada Open File Map 463Muller J E (1980)The Paleozoic Sicker Group of Vancouver Island British
Columbia Geological Survey of Canada Paper 79-30 22 ppMuller J E Northcote K E amp Carlisle D (1974) Geology and
mineral deposits of Alert Bay^Cape Scott map area VancouverIsland British Columbia Geological Survey of Canada Paper 74-8 77
Neal C R Mahoney J J Kroenke L W Duncan R A ampPetterson M G (1997) The Ontong^Java Plateau InMahoney J J amp Coffin M F (eds) Large Igneous Provinces
Continental Oceanic and Planetary FloodVolcanism American Geophysical
Union Geophysical Monograph 100 183^216Niu Y Collerson K D Batiza R Wendt J I amp Regelous M
(1999) Origin of enriched-typed mid-ocean ridge basalt at ridgesfar from mantle plumes The East Pacific Rise at 11820rsquoN Journalof Geophysical Research 104 7067^7088
Nixon G T amp Orr A J (2007) Recent revisions to the EarlyMesozoic stratigraphy of Northern Vancouver Island (NTS 102I092L) and metallogenic implicatinos British Columbia InGrant B (ed) Geological Fieldwork 2006 British Columbia Ministry of
Energy Mines and Petroleum Resources Paper 2007-1 163^177Nixon GT Hammack J L KoyanagiV M Payie G J Haggart
JW Orchard M JTozerT Friedman R M Archibald D APalfy J amp Cordey F (2006a) Geology of the Quatsino^PortMcNeill area northernVancouver Island British Columbia Ministry
of Energy Mines and Petroleum Resources Geoscience Map 2006-2 scale150 000
Nixon G T Kelman M C Stevenson D Stokes L A ampJohnston K A (2006b) Preliminary geology of the Nimpkishmap area (NTS 092L07) northern Vancouver Island BritishColumbia In Grant B amp Newell J M (eds) Geological Fieldwork2005 British Columbia Ministry of Energy Mines and Petroleum Resources
Paper 2006-1 135^152Nixon G T Laroque J Pals A Styan J Greene A R amp
Scoates J S (2008) High-Mg lavas in the Karmutsen flood basaltsnorthernVancouver Island (NTS 092L) Stratigraphic setting andmetallogenic significance In Grant B (ed) Geological Fieldwork
2007 British Columbia Ministry of Energy Mines and Petroleum
Resources Paper 2008-1 175^190Nowell G M Kempton P D Noble S R Fitton J G
Saunders A Mahoney J J amp Taylor R N (1998) High precisionHf isotope measurements of MORB and OIB by thermal ionisation
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
36
mass spectrometry insights into the depleted mantle Chemical
Geology 149 211^233Parrish R R amp McNicoll V J (1992) U^Pb age determinations
from the southern Vancouver Island area British Columbia InRadiogenic Age and Isotopic Studies Report 5 Geological Survey of
Canada Paper 91-2 79^86Peate DW (1997) The Parana^Etendeka Province In Coffin M F
amp Mahoney J (eds) Large Igneous Provinces Continental Oceanic andPlanetary Flood Volcanism American Geophysical Union Geophysical
Monograph 100 217^245Perfit M R Fornari D J Ridley W I Kirk P D Casey J
Kastens K A Reynolds J R Edwards M Desonie DShuster R amp Paradis S (1996) Recent volcanism in the Siqueirostransform fault picritic basalts and implications for MORBmagma genesis Earth and Planetary Science Letters 141(1^4) 91^108
Pretorius W Weis D Williams G Hanano D Kieffer B ampScoates J S (2006) Complete trace elemental characterization ofgranitoid (USGSG-2GSP-2) reference materials by high resolutioninductively coupled plasma-mass spectrometry Geostandards and
Geoanalytical Research 30(1) 39^54Regelous M Niu Y Wendt J I Batiza R Greig A amp
Collerson K D (1999) Variations in the geochemistry of magma-tism on the East Pacific Rise at 10830rsquoN since 800 ka Earth and
Planetary Science Letters 168 45^63Recurren villon S Chauvel C Arndt N T Pik R Martineau F
Fourcade S amp Marty B (2002) Heterogeneity of the Caribbeanplateau mantle source Sr O and He isotopic compositions of oli-vine and clinopyroxene from Gorgona Island Earth and Planetary
Science Letters 205(1^2) 91Rhodes J M (1996) Geochemical stratigraphy of lava flows samples
by the Hawaii Scientific Drilling Project Journal of Geophysical
Research 101(B5) 11729^11746Rhodes J M amp Vollinger M J (2004) Composition of basaltic lavas
sampled by phase-2 of the Hawaii Scientific Drilling ProjectGeochemical stratigraphy and magma types Geochemistry
Geophysics Geosystems 5(3) doi1010292002GC000434Richards M A Jones D L Duncan R A amp DePaolo D J (1991)
A mantle plume initiation model for the Wrangellia flood basaltand other oceanic plateaus Science 254 263^267
Salters V J M amp Hart S R (1991) The mantle sources of oceanridges islands and arcs the Hf-isotope connection Earth and
Planetary Science Letters 104 364^380Salters V J M amp Stracke A (2004) Composition of the depleted
SaltersV J amp WhiteW (1998) Hf isotope constraints on mantle evo-lution Chemical Geology 145 447^460
SanoT amp Yamashita S (2004) Experimental petrology of basementlavas from Ocean Drilling Program Leg 192 implications for dif-ferentiation processes in Ontong Java Plateau magmas InFitton J G Mahoney J JWallace P J amp Saunders A D (eds)Origin and Evolution of the Ontong Java Plateau Geological Society
London Special Publications 229 185^218Saunders A D (2005) Large igneous provinces Origin and environ-
mental consequences Elements 1 259^263Self S ThordarsonT amp Keszthely L (1997) Emplacement of conti-
nental flood basalt lava flows In Mahoney J J amp Coffin M F(eds) Large Igneous Provinces Continental Oceanic and Planetary Flood
Volcanism American Geophysical Union Geophysical Monograph 100381^410
Shaw D M (2000) Continuous (dynamic) melting theory revisitedCanadian Mineralogist 38(5) 1041^1063
Sluggett C L (2003) Uranium^lead age and geochemical constraintson Paleozoic and Early Mesozoic magmatism in WrangelliaTerrane Saltspring Island British Columbia BSc thesisUniversity of British ColumbiaVancouver 84 pp
Storey M Mahoney J J amp Saunders A D (1997) Cretaceousbasalts in Madagascar and the transition between plume and con-tinental lithospheric mantle sources In Mahoney J J ampCoffin M F (eds) Large Igneous Provinces Continental Oceanic and
Planetary Flood Volcanism Aamerican Geophysical Union Geophysical
Monograph 100 95^122Surdam R C (1967) Low-grade metamorphism of the Karmutsen
Group PhD dissertation University of California Los Angeles288 pp
Sutherland-Brown A Yorath C J Anderson R G amp Dom K(1986) Geological maps of southern Vancouver IslandLITHOPROBE I 92C10 11 14 16 92F1 2 7 8 Geological Survey ofCanada Open File 1272
Tejada M L G Mahoney J J Castillo P R Ingle S P Sheth HC amp Weis D (2004) Pin-pricking the elephant evidence on theorigin of the Ontong Java Plateau from Pb^Sr^Hf^Nd isotopiccharacteristics of ODP Leg 192 basalts In Fitton J GMahoney J J Wallace P J amp Saunders A D (eds) Origin and
Evolution of the Ontong Java Plateau Geological Society London Special
Publications 229 133^150Thompson P M E Kempton P D amp Kerr A C (2008) Evaluation
of the effects of alteration and leaching on Sm^Nd and Lu^Hf sys-tematics in submarine mafic rocks Lithos 104(1^4) 164^176
Thompson P M E Kempton P D White R V Kerr A CTarney J Saunders A D Fitton J G amp McBirney A (2003)Hf-Nd isotope constraints on the origin of the CretaceousCaribbean plateau and its relationship to the Galapagos plumeEarth and Planetary Science Letters 217(1^2) 59^75
Vervoort J D Patchett P Blichert-Toft J amp Albare de F (1999)Relationships between Lu^Hf and Sm^Nd isotopic systems in theglobal sedimentary system Earth and Planetary Science Letters 16879^99
Walter M J (1998) Melting of garnet peridotite and the origin ofkomatiite and depleted lithosphere Journal of Petrology 39(1) 29^60
Weis D Kieffer B Maerschalk C Barling J de Jong JWilliams G A Hanano D Mattielli N Scoates J SGoolaerts A Friedman R A amp Mahoney J B (2006) High-pre-cision isotopic characterization of USGS reference materials byTIMS and MC-ICP-MS Geochemistry Geophysics Geosystems
7(Q08006) doi1010292006GC001283Weis D Kieffer B Hanano D Silva I N Barling J
Pretorius W Maerschalk C amp Mattielli N (2007) Hf isotopecompositions of US Geological Survey reference materialsGeochemistry Geophysics Geosystems 8(Q06006) doi1010292006GC001473
Wheeler J O amp McFeely P (1991) Tectonic assemblage map of theCanadian Cordillera and adjacent part of the United States ofAmerica Geological Survey of Canada Map 1712A
WhiteW M Albare de F amp Tecurren louk P (2000) High-precision analy-sis of Pb isotope ratios by multi-collector ICP-MS Chemical Geology167 257^270
Yorath C J Sutherland Brown A amp Massey N W D (1999)LITHOPROBE southern Vancouver Island British Columbia Geological
Survey of Canada Bulletin 498 145 ppZou H (1998) Trace element fractionation during modal and
non-model dynamic melting and open-system melting Amathematical treatment Geochimica et Cosmochimica Acta 62(11)1937^1945
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
37
Zou H amp Reid M R (2001) Quantitative modeling of trace elementfractionation during incongruent dynamic melting Geochimica et
Cosmochimica Acta 65(1) 153^162
APPENDIX SAMPLEPREPARAT ION ANDANALYT ICAL METHODSOnly the freshest rocks in the field were sampled and onlythe least altered samples were selected for chemical andisotopic analysis based on thorough petrographic inspec-tion Fifty-nine of the 129 collected samples were crushed(400 g) into small pieces 52mm in diameter in aRocklabs hydraulic piston crusher between WC plates tominimize contamination The coarse-crush was mixedand 100 g was powdered in a planetary mill using agatejars and balls cleaned with quartz sand between samples
ActLabs analytical methodsThe major- and trace-element compositions of the whole-rock powders were determined at Activation LaboratoriesLtd (Actlabs) in Ancaster Ontario Analytical techniquesand detection limits are also available from Actlabs (httpwwwactlabscommethsub_code4erehtm) The analyticalmethod for each of the elements analyzed is indicated inTable 2 For the major elements a 02 g sample was mixedwith a lithium metaborate^lithium tetraborate mixtureand fused in a graphite crucible The molten mixture waspoured into a 5 HNO3 solution and shaken until dis-solved (30min) The samples were analyzed for majoroxides and selected trace elements by inductively coupledplasma optical emission spectrometry (ICP-OES) on acombination simultaneous^sequential Thermo Jarrell-AshEnviro II system Internal calibration was achieved usinga variety of international reference materials (eg W-2BIR-1 DNC-1) and independent control samplesAdditional trace elements were analyzed by both instru-mental neutron activation analysis (INAA) and induc-tively couple plasma mass spectrometry (ICP-MS) ForINAA15^25 g of sample was weighed into small polyeth-ylene vials and irradiated with a control international ref-erence material CANMET WMS-1 and NiCr flux wires ata thermal neutron flux of 71012 neutronscm2 per s in theMcMaster Nuclear Reactor Following a 7 day waitingperiod the samples were measured on an Ortec high-purity Ge detector linked to a Canberra Series 95 multi-channel analyzer Activities for each element weredecay- and weight-corrected and compared with a detectorcalibration developed from multiple international certifiedreference materials For ICP-MS 025 g of sample wasdigested in HF followed by HNO3^HClO4 heated andtaken to dryness The samples were brought back into solu-tion with HCl Samples were analyzed using a PerkinElmer Optima 3000 ICP In-lab standards or certified ref-erence materials (eg W-2 BIR-1 DNC-1) were used for
quality control A total of 15 blind duplicates were ana-lyzed from the same sample powders to assessreproducibility
University of Massachusetts analyticalmethodsFourteen sample duplicate powders taken from the same100 g sample aliquot (high-MgO basalts and picrites)were also analyzed at the Ronald B Gilmore X-RayFluorescence (XRF) Laboratory at the University ofMassachusetts Major elements were measured on a fusedLa-bearing lithium borate glass disc using a SiemensMRS-400 spectrometer Trace element concentrations(Rb Sr Ba Ce Nb Zr Y Pb Zn Ga Ni Cr V) weremeasured on a separate powder pellet using a PhilipsPW2400 sequential spectrometer using a Rh tube LOIand ferrous iron measurements were made as describedby Rhodes amp Vollinger (2004) Precision and accuracy esti-mates for the data have been described by Rhodes (1996)and Rhodes amp Vollinger (2004) Results for each sampleare the average of two separate analyses and are shown inElectronic Appendix 2 for a comparison of analyses ofduplicate powders using XRFand ICP-MS
PCIGR analytical methodsA subset of 19 samples was selected for high-precisiontrace-element analysis and Sr Nd Pb and Hf isotopicanalysis at the Pacific Centre for Isotopic andGeochemical Research (PCIGR) at the University ofBritish Columbia (UBC) Samples were selected from the59 samples analyzed for whole-rock chemistry at ActLabsbased on major- and trace-element chemistry alteration(low LOI and petrographic alteration index) and samplelocation Samples were prepared for trace-element analysisat the PCIGR by the technique described by Pretorius et al(2006) on unleached rock powders Sample powders(100mg) were weighed in 7ml screw-top Savillex bea-kers and dissolved in 1ml 14N HNO3 and 5ml 48 HFon a hotplate for 48 h at 1308C with periodic ultrasonica-tion Samples were dried and redissolved in 6ml 6N HClon a hotplate for 24 h and then dried and redissolved in1ml concentrated HNO3 for 24 h before final dryingTrace element abundances were measured by high-resolution inductively coupled plasma-mass spectrometry(HR-ICP-MS) with aThermo Finnigan Element2 systemfollowing the procedures described by Pretorius et al(2006) within 24 h of redissolution HFSE and LILE weremeasured in medium-resolution mode at 2000 dilutionusing a PFATeflon spray chamber washed with aqua regiafor 3min between samples REE were measured in high-resolution mode and U and Pb in low-resolution modeat 2000 dilution using a glass spray chamber washedwith 2 HNO3 between samples Total procedural blanksand reference materials (BCR-2 BHVO-2) were analyzedwith the batch of samples Indium was used as an internal
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
38
standard in all samples and standard solutionsBackground and standard solutions were analyzed afterevery five samples to detect memory effects and massdrift Results for PCIGR trace-element analyses areshown in Electronic Appendix 3 for the full set of analyzedtrace elementsSample digestion for purification of Sr Nd Hf and Pb
for column chemistry began by weighing each samplepowder (400^500mg) prior to leaching All samples wereleached with 6N HCl and placed in an ultrasonic bath for15min Samples were rinsed twice with 18 M-cm H2Obetween each leaching step (15 total) until the supernatantliquid was clear [following the method of Mahoney (1987)]Samples were then dried on a hotplate for 24 h andweighed again Sample solutions were then prepared bydissolving 100^250mg of the leached powder dissolvedin 1ml 14N HNO3 and 10ml 48 HF on a hotplate for48 h at 1308C with periodic ultrasonication Samples weredried and redissolved in 6ml 6N HCl on a hotplate for24 h and then dried Pb was separated using anion exhangecolumns and the discard was used for Sr REE and Hfseparation Nd was separated from the other REE and Hfrequired two additional purification steps Detailed proce-dures for column chemistry for separating Sr Nd and Pbat the PCIGR have been described byWeis et al (2006) andHf purification has been described byWeis et al (2007) Srand Nd isotope ratios were measured by thermal ioniza-tion mass spectrometry (TIMS) on a Thermo FinniganTriton system in static mode with relay matrix rotation ona single Ta and double Re^Ta filament respectively Fourto five filaments per barrel of 21 were occupied by stan-dards (NBS 987 for Sr and LaJolla for Nd) for each barrelwhere samples were run Sample Sr and Nd isotopic com-positions were corrected for mass fractionation using86Sr88Srfrac14 01194 and 146Nd144Ndfrac14 07219 Each samplewas then normalized using the barrel average of the refer-ence material relative to the values of143Nd144Ndfrac14 0511858 and 87Sr86Srfrac14 0710248 (Weiset al 2006) During the course of the Vancouver Islandanalyses the La Jolla Nd standard gave an average valueof 05118566 (nfrac14 7) and NBS987 standard gave an aver-age of 07102408 (nfrac1411) (2 error is reported as times106) 147Sm144Nd errors are 15 or 0006 USGeological Survey (USGS) reference material BHVO-2was processed with the samples and yielded Sr and Nd iso-topic ratios of 07034607 and 05129786 respectivelyThese are in agreement with the published values of070347920 and 051298411 respectively (Weis et al2006)
Pb and Hf isotopic compositions were analyzed by mul-tiple collector-inductively coupled plasma-mass spectrome-try (MC-ICP-MS) on a static multi-collection on a NuPlasma (Nu Instruments) system The detailed analyticalprocedure for Pb isotopic analyses on the Nu Plasma atthe PCIGR has been described by Weis et al (2006) Theconfiguration for Pb analyses allows for collection of PbTl and Hg togetherTl and Hg are used to monitor instru-mental mass discrimination and isobaric overlap repec-tively All sample solutions were analyzed withapproximately the same PbTl ratio (4) as the referencematerial NIST SRM 981 To accomplish this a small ali-quot of each sample solution from the Pb columns was ana-lyzed on the Element2 to determine the precise amount ofPb available for analysis on the Nu Plasma The SRM 981standard was run after every two samples on the NuPlasma During the time samples were run analyses ofthe SRM 981 Pb reference material gave values of206Pb204Pbfrac1416940322 207Pb204Pbfrac1415495823and 208Pb204Pbfrac14 36713164 (nfrac14 61 2 error is reportedas times 104) these values are in excellent agreement withreported TIMS triple-spike values of Galer amp Abouchami(1998) Fractionation-corrected Pb isotopic ratios were fur-ther corrected by the sample-standard bracketing methodor the ln^ln correction method described by White et al(2000) and Blichert-Toft et al (2003) Leached powders ofUSGS reference material BHVO-2 yielded Pb isotopicratios of 206Pb204Pbfrac14186454 8 207Pb204Pbfrac14 1549105 and 208Pb204Pbfrac14 38222514 and BCR-2yielded 206Pb204Pbfrac141880466 207Pb204Pbfrac14 1562518 and 208Pb204Pbfrac14 3883496 (2 error isreported as times 104) These values are in agreement withthose for leached residues of BHVO-2 and BCR-2 fromWeis et al (2006)Hf isotopic compositions were analyzed following the
procedures detailed byWeis et al (2007) The configurationfor Hf analyses monitored Lu mass 175 andYb mass 172 toallow for interference correction to masses 174 and 176 Hfisotopic ratios were normalized internally for mass fractio-nation to a 179Hf177Hf ratio of 07325 using an exponentialcorrection Standards were run after every two samplesand sample results were normalized to the ratio of the in-run daily average and a 176Hf177Hf ratio for JMC-475 of0282160 During the course of analyses the Hf standardJMC-475 gave an average value 02821679 (nfrac14130)USGS reference materials BCR-2 and BHVO-2 were pro-cessed with the samples and yielded Hf isotopic ratios of02828675 and 02831005 respectively Publishedvalues for BCR-2 and BHVO-2 are 02828717 and0283104 8 respectively (Weis et al 2007)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
1Ni concentrations for these samples were determined by XRFTHOL tholeiitic basalt PIC picrite HI-MG high MgO basalt CG coarse-grained (sill or gabbro) MIN SIL mineralizedsill OUTLIER anomalous pillowed flow in plots MA Mount Arrowsmith SL Schoen Lake KR Karmutsen Range QIQuadra Island Sample locations are given using the Universal Transverse Mercator (UTM) coordinate system (NAD83zones 9 and 10) Analyses were performed at Activation Laboratory (ActLabs) Fe2O3
is total iron expressed as Fe2O3LOI loss on ignition All major elements Sr V and Y were measured by ICP quadrupole OES on solutions of fusedsamples Cu Ni Pb and Zn were measured by total dilution ICP Cs Ga Ge Hf Nb Rb Ta Th U Zr and REE weremeasured by magnetic-sector ICP on solutions of fused samples Co Cr and Sc were measured by INAA Blanks arebelow detection limit See Electronic Appendices 2 and 3 for complete XRF data and PCIGR trace element datarespectively Major elements for sample 93G17 are normalized anhydrous Sample 5615A7 was analyzed by XRFSamples from Quadra Island are not shown in figures
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
17
picritic and high-MgO basalt pillow lavas extend to20wt MgO (Fig 6 and references listed in thecaption)The tholeiitic basalts from the Karmutsen Range dis-
play a tight range of parallel light rare earth element(LREE)-enriched REE patterns (LaYbCNfrac14 21^24mean 2202) whereas the picrites and high-MgObasalts are characterized by sub-parallel LREE-depleted
patterns (LaYbCNfrac14 03^07 mean 06 02) with lowerREE abundances than the tholeiitic basalts (Fig 7)Tholeiitic basalts from the Schoen Lake and MountArrowsmith areas have similar REE patterns tosamples from the Karmutsen Range (Schoen Lake LaYbCNfrac14 21^26 mean 2303 Mount Arrowsmith LaYbCNfrac1419^25 mean 2204) and the coarse-grainedmafic rocks have similar REE patterns to the tholeiitic
Depleted MORB average(Salters amp Stracke 2004)
Schoen LakeSchoen Lake
Mount Arrowsmith
Karmutsen RangeS
ampl
eC
hond
rite
Sam
ple
Cho
ndrit
e
Sam
ple
Prim
itive
Man
tleS
ampl
eP
rimiti
ve M
antle
Sam
ple
Prim
itive
Man
tle
Mount Arrowsmith
Karmutsen Range
(a) (b)
(c) (d)
(e) (f )
Sam
ple
Cho
ndrit
e
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained mafic rock
Pillowed flowMineralized sill
1
10
100
1
10
100
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
01
1
10
100
1
10
100
1
10
100
Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
Fig 7 Whole-rock REE and other incompatible-element concentrations for the Karmutsen Formation (a) (c) and (e) are chondrite-normal-ized REE patterns for samples from each of the three main field areas onVancouver Island (b) (d) and (f) are primitive mantle-normalizedtrace-element patterns for samples from each of the three main field areas All normalization values are from McDonough amp Sun (1995) Theclear distinction between LREE-enriched tholeiitic basalts and LREE-depleted picrites the effects of alteration on the LILE (especially loss ofK and Rb) and the different range for the vertical scale in (b) (extends down to 01) should be noted
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
18
basalts (LaYbCNfrac14 21^29 mean 2508) Two LREE-depleted coarse-grained mafic rocks from the SchoenLake area have similar REE patterns (LaYbCNfrac14 07) tothe picrites from the Keogh Lake area indicating that thisdistinctive suite of rocks is exposed as far south asSchoen Lake that is 75 km away (Fig 7) The anomalouspillowed flow from the Karmutsen Range has lower LaYbCN (13^18) than the main group of tholeiitic basaltsfrom the Karmutsen Range (Fig 7) The mineralized sillfrom the Schoen Lake area is distinctly LREE-enriched(LaYbCNfrac14 33) with the highest REE abundances(LaCNfrac14 940) of all Karmutsen samples this may reflectcontamination by adjacent sediment also evident in thin-section where sulfide blebs are cored by quartz (Greeneet al 2006)The primitive mantle-normalized trace-element pat-
terns (Fig 7) and trace-element variations and ratios(Fig 8) highlight the differences in trace-element
concentrations between the four main groups ofKarmutsen flood basalts onVancouver Island The tholeii-tic basalts from all three areas have relatively smooth par-allel trace-element patterns some samples have smallpositive Sr anomalies relative to Nd and Sm and manysamples have prominent negative K anomalies relative toU and Nb reflecting K loss during alteration (Fig 7) Thelarge ion lithophile elements (LILE) in the tholeiiticbasalts (mainly Rb and K not Cs) are depleted relativeto the high field strength elements (HFSE) and LREELILE segments especially for the Schoen Lake area(Fig 7d) are remarkably parallel The picrites andhigh-MgO basalts form a tight band of parallel trace-element patterns and are depleted in HFSE and LREEwith positive Sr anomalies and relatively low Rb Thecoarse-grained mafic rocks from the Karmutsen Rangehave identical trace-element patterns to the tholeiiticbasalts Samples from the Schoen Lake area have similar
000
025
050
075
100
125
0 1 2 3 4 5
00
04
08
12
16
0 5 10 15 20 25
MgO (wt)
0
5
10
15
0 50 100 150
Hf (ppm)
Th (ppm)
NbLaNb (ppm)
Zr (ppm)
(a)
(c)
(b)
(d)
Th (ppm)
00
02
04
06
08
10
12
00 01 02 03 04 05
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
NbLa=1
4723A4 5615A12
U (ppm)
Fig 8 Whole-rock trace-element concentrations and ratios for the Karmutsen Formation [except (b) which is vs wt MgO] (a) Nb vs Zr(b) NbLa vs MgO (c) Th vs Hf (d) Th vs U There is a clear distinction between the tholeiitic basalts and picrites in Nb and Zr both inconcentration and the slope of each trend
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
19
trace-element patterns to the Karmutsen Range except forthe two LREE-depleted coarse-grained mafic rocks andthe mineralized sill (Fig 7d)
Sr^Nd^Hf^Pb isotopic compositionsA total of 19 samples from the four main groups ofthe Karmutsen Formation were selected for radiogenicisotopic geochemistry on the basis of major- and trace-element variation stratigraphic position and geographicallocation The samples have overlapping age-correctedHf and Nd isotope ratios and distinct ranges of Sr isotopiccompositions (Fig 9) The tholeiitic basalts picriteshigh-MgO basalt and coarse-grained mafic rocks have
initial eHffrac14thorn87 to thorn126 and eNdfrac14thorn66 to thorn88 cor-rected for in situ radioactive decay since 230 Ma (Fig 9Tables 3 and 4) All the Karmutsen samples form a well-defined linear array in a Lu^Hf isochron diagram corre-sponding to an age of 24115 Ma (Fig 9) This is withinerror of the accepted age of the Karmutsen Formationand indicates that the Hf isotope systematics behaved as aclosed system since c 230 Ma The anomalous pillowedflow has similar initial eHf (thorn98) and slightly lower initialeNd (thorn62) compared with the other Karmutsen samplesThe picrites and high-MgO basalt have higher initial87Sr86Sr (070398^070518) than the tholeiitic basalts(initial 87Sr86Srfrac14 070306^070381) and the coarse-grained mafic rocks (initial 87Sr86Srfrac14 070265^070428)
05128
05129
05130
05131
05132
015 017 019 021 023 025
02829
02830
02831
02832
02833
02834
000 002 004 006 008 010
4
6
8
10
0702 0703 0704 0705 07066
8
10
12
14
5 6 7 8 9 10
12
0
1
2
3
4
5
6
7
0702 0703 0704 0705 0706
(d)
LOI (wt )
87Sr 86Sr
4723A2
Nd
143Nd144Nd
147Sm144Nd
87Sr 86Sr (230 Ma)
(230 Ma)
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
(a)
(c)
176Hf177Hf
Age=241 plusmn 15 Ma=+990 plusmn 064
176Lu177Hf
Hf (230 Ma)
Nd (230 Ma)
(e)
Hf (i)
(b)
Fig 9 Whole-rock Sr Nd and Hf isotopic compositions for the Karmutsen Formation (a) 143Nd144Nd vs 147Sm144Nd (b) 176Hf177Hf vs176Lu177Hf The slope of the best-fit line for all samples corresponds to an age of 24115 Ma (c) Initial eNd vs 87Sr86Sr Age-correction to230 Ma (d) LOI vs 87Sr86Sr (e) Initial eHf vs eNd Average 2 error bars are shown in a corner of each panel Complete chemical duplicatesshown inTables 3 and 4 (samples 4720A7 and 4722A4) are circled in each plot
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
overlap the ranges of the picrites and tholeiitic basalts(Fig 9 Table 3)The measured Pb isotopic compositions of the tholeiitic
basalts are more radiogenic than those of the picrites andthe most magnesian Keogh Lake picrites have the leastradiogenic Pb isotopic compositions The range ofinitial Pb isotope ratios for the picrites is206Pb204Pbfrac1417868^18580 207Pb204Pbfrac1415547^15562and 208Pb204Pbfrac14 37873^38257 and the range for thetholeiitic basalts is 206Pb204Pbfrac1418782^19098207Pb204Pbfrac1415570^15584 and 208Pb204Pbfrac14 37312^38587 (Fig 10 Table 5) The coarse-grained mafic rocksoverlap the range of initial Pb isotopic compositions forthe tholeiitic basalts with 206Pb204Pbfrac1418652^19155207Pb204Pbfrac1415568^15588 and 208Pb204Pbfrac14 38153^38735 One high-MgO basalt has the highest initial206Pb204Pb (19242) and 207Pb204Pb (15590) and thelowest initial 208Pb204Pb (37676) The Pb isotopic ratiosfor Karmutsen samples define broadly linearrelationships in Pb isotope plots The age-corrected Pb
isotopic compositions of several picrites have been affectedby U and Pb (andor Th) mobility during alteration(Fig 10)
ALTERAT IONThe Karmutsen basalts have retained most of their origi-nal igneous structures and textures however secondaryalteration and low-grade metamorphism have producedzeolitic- and prehnite^pumpellyite-bearing mineral assem-blages (Table 1 Cho et al 1987) Alteration and metamor-phism have primarily affected the distribution of the LILEand the Sr isotopic systematics The Keogh Lake picriteshave been affected more by alteration than the tholeiiticbasalts and have higher LOI (up to 55wt ) variableLILE and higher measured Sr Nd and Hf and lowermeasured Pb isotope ratios than the tholeiitic basalts Srand Pb isotope compositions for picrites and tholeiiticbasalts show a relationship with LOI (Figs 9 and 10)whereas Nd and Hf isotopic compositions do not correlate
Table 3 Sr and Nd isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Rb Sr 87Sr86Sr 2m87Rb86Sr 87Sr86Srt Sm Nd 143Nd144Nd 2m eNd
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses carried out at the PCIGR thecomplete trace element analyses are given in Electronic Appendix 3
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
21
with LOI (not shown) The relatively high apparent initialSr isotopic compositions for the high-MgO lavas (up to07052) probably resulted from post-magmatic loss of Rbor an increase in 87Sr86Sr through addition of seawater Sr(eg Hauff et al 2003) High-MgO lavas from theCaribbean plateau have similarly high 87Sr86Sr comparedwith basalts (Recurren villon et al 2002)The correction for in situdecay on initial Pb isotopic ratios has been affected bymobilization of U and Th (and Pb) in the whole-rockssince their formation A thorough acid leaching duringsample preparation was used that has been shown to effec-tively remove alteration phases (Weis et al 2006 NobreSilva et al in preparation) Whole-rock SmNd ratiosand age-correction of Nd isotopes may be affected bythe leaching procedure for submarine rocks older than50 Ma (Thompson et al 2008 Fig 9) there ishowever very little variation in Nd isotopic compositionsof the picrites or tholeiitic basalts and this system does notappear to be disturbed by alteration The HFSE abun-dances for tholeiitic and picritic basalts exhibit clear
linear correlations in binary diagrams (Fig 8) whereasplots of LILE vs HFSE are highly scattered (not shown)because of the mobility of some of the alkali and alkalineearth LILE (Cs Rb Ba K) and Sr for most samplesduring alterationThe degree of alteration in the Karmutsen samples does
not appear to be related to eruption environment or depthof burial Pillow rims are compositionally different frompillow cores (Surdam1967 Kuniyoshi1972) and aquagenetuffs are chemically different from isolated pillows withinpillow breccias however submarine and subaerial basaltsexhibit a similar degree of alterationThere is no clear cor-relation between the submarine and subaerial basalts andsome commonly used chemical alteration indices (eg BaRb vs K2OP2O5 Huang amp Frey 2005) There is also nodefinitive correlation between the petrographic alterationindex (Table 1) and chemical alteration indices although10 of 13 tholeiitic basalts with the highest K and LILEabundances have the highest petrographic alterationindex of three (seeTable 1)
Table 4 Hf isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Lu Hf 176Hf177Hf 2m eHf176Lu177Hf 176Hf177Hft eHf(t)
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
OLIV INE ACCUMULAT ION INH IGH-MgO AND PICR IT IC LAVASGeochemical trends and petrographic characteristics indi-cate that accumulation of olivine played an important rolein the formation of the high-MgO basalts and picrites fromthe Keogh Lake area on northern Vancouver Island TheKeogh Lake samples show a strong linear correlation inplots of Al2O3 TiO2 ScYb and Ni vs MgO and many ofthe picrites have abundant clusters of olivine pseudomorphs(Table1 Fig11)There is a strong linear correlationbetween
modal per cent olivine and whole-rock magnesium contentmagnesium contents range between 10 and 19wt MgOand the proportion of olivine phenocrysts in these samplesvaries from 0 to 42 vol (Fig 11)With a few exceptionsboth the size range and median size of the olivine pheno-crysts in most samples are comparable The clear correla-tion between the proportion of olivine phenocrysts andwhole-rock magnesium contents indicates that accumula-tion of olivine was directly responsible for the high MgOcontents (410wt ) in most of the picritic lavas
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
207Pb204Pb
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
208Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
(a) (b)
(c) (e)
4723A24723A2
4723A2
4723A2
4722A4
4723A4
4722A4
4723A4
4723A134723A3
4723A3
4723A13
1556
1558
1560
1562
1564
1566
1568
186 190 194 198 202 206 2101550
1552
1554
1556
1558
1560
178 180 182 184 186 188 190 192 194
380
384
388
392
396
186 190 194 198 202 206 210374
378
382
386
390
178 180 182 184 186 188 190 192 194
206Pb204Pb(d)
LOI (wt )
0
1
2
3
4
5
6
7
186 190 194 198 202 206 210
4723A2
Fig 10 Pb isotopic compositions of leached whole-rock samples measured by MC-ICP-MS for the Karmutsen Formation Error bars are smal-ler than symbols (a) Measured 207Pb204Pb vs 206Pb204Pb (b) Initial 207Pb204Pb vs 206Pb204Pb Age-correction to 230 Ma (c) Measured208Pb204Pb vs 206Pb204Pb (d) LOI vs 206Pb204Pb (e) Initial 208Pb204Pb vs 206Pb204Pb Complete chemical duplicates shown in Table 5(samples 4720A7 and 4722A4) are circled in each plot The dashed lines in (b) and (e) show the differences in age-corrections for two picrites(4723A4 4722A4) using the measured UTh and Pb concentrations for each sample and the age-corrections when concentrations are used fromthe two picrites (4723A3 4723A13) that appear to be least affected by alteration
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
23
DISCUSS IONThe compositional and spatial development of the eruptedlava sequences of the Wrangellia oceanic plateau onVancouver Island provide information about the evolutionof a mantle plume upwelling beneath oceanic lithospherein an analogous way to the interpretation of continentalflood basalts The Caribbean and Ontong Java plateausare the two primary examples of accreted parts of oceanicplateaus that have been studied to date and comparisonscan be made with the Wrangellia oceanic plateauHowever the Wrangellia oceanic plateau is a distinctiveoccurrence of an oceanic plateau because it formed on thecrust of a Devonian to Mississippian intra-oceanic islandarc overlain by Mississippian to Permian limestones andpelagic sediments The nature and thickness of oceaniclithospheric mantle are considered to play a fundamentalrole in the decompression and evolution of a mantleplume and thus the compositions of the erupted lavasequences (Greene et al 2008) The geochemical and
stratigraphic relationships observed in the KarmutsenFormation onVancouver Island and the focus of the subse-quent discussion sections provide constraints on (1) thetemperature and degree of melting required to generatethe primary picritic magmas (2) the composition of themantle source (3) the depth of melting and residual min-eralogy in the source region (4) the low-pressure evolutionof the Karmutsen tholeiitic basalts (5) the growth of theWrangellia oceanic plateau
Melting conditions and major-elementcomposition of the primary magmasThe primary magmas to most flood basalt provinces arebelieved to be picritic (eg Cox 1980) and they have beenfound in many continental and oceanic flood basaltsequences worldwide (eg Siberia Karoo Paranacurren ^Etendeka Caribbean Deccan Saunders 2005) Near-pri-mary picritic lavas with low total alkali abundances
Table 5 Pb isotopic compositions of Karmutsen basaltsVancouver Island BC
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
require high-degree partial melting of the mantle source(eg Herzberg amp OrsquoHara 2002) The Keogh Lake picritesfrom northernVancouver Island are the best candidates foridentifying the least modified partial melts of the mantleplume source despite having accumulated olivine pheno-crysts and can be used to estimate conditions of meltingand the composition of the primary magmas for theKarmutsen FormationPrimary magma compositions mantle potential tem-
peratures and source melt fractions for the picritic lavas ofthe Karmutsen Formation were calculated from primitivewhole-rock compositions using PRIMELT1XLS software(Herzberg et al 2007) and are presented in Fig 12 andTable 6 A detailed discussion of the computationalmethod has been given by Herzberg amp OrsquoHara (2002)and Herzberg et al (2007) key features of the approachare outlined belowPrimary magma composition is calculated for a primi-
tive lava by incremental addition of olivine and compari-son with primary melts of mantle peridotite Lavas thathad experienced plagioclase andor clinopyroxene fractio-nation are excluded from this analysis All calculated pri-mary magma compositions are assumed to be derived byaccumulated fractional melting of fertile peridotite KR-4003 Uncertainties in fertile peridotite composition donot propagate to significant variations in melt fractionand mantle potential temperature melting of depletedperidotite propagates to calculated melt fractions that aretoo high but with a negligible error in mantle potential
temperature PRIMELT1XLS calculates the olivine liqui-dus temperature at 1atm which should be similar to theactual eruption temperature of the primary magmaand is accurate to 308C at the 2 level of confidenceMantle potential temperature is model dependent witha precision that is similar to the accuracy of the olivineliquidus temperature (C Herzberg personal communica-tion 2008)With regard to the evidence for accumulated olivine in
the Keogh Lake picrites it should not matter whether themodeled lava composition is aphyric or laden with accu-mulated olivine phenocrysts PRIMELT1 computes theprimary magma composition by adding olivine to a lavathat has experienced olivine fractionation in this way itinverts the effects of fractional crystallization The onlyrequirement is that the olivine phenocrysts are not acci-dental or exotic to the lava compositions Support for thisprocedure can be seen in the calculated olivine liquid-line-of-descent shown by the curved arrays with 5 olivineaddition increments (Fig 12c) Fractional crystallization ofolivine from model primary magmas having 85^95wt FeO and 153^175wt MgO will yield arrays of deriva-tive liquids and randomly sorted olivines that in most butnot all cases connect the picrites and high-MgO basalts Anewer version of the PRIMELT software (Herzberg ampAsimow 2008) can perform olivine addition to and sub-traction from lavas with highly variably olivine phenocrystcontents and yields results that converge with those shownin Fig 12
Fig 11 Relationship between abundance of olivine phenocrysts and whole-rock MgO contents for the Keogh Lake picrites (a) Tracings ofolivine grains (gray) in six high-MgO samples made from high resolution (2400 dpi) scans of petrographic thin-sections Scale bars represent10mm sample numbers are indicated at the upper left (b) Modal olivine vs whole-rock MgO Continuous line is the best-fit line for 10 high-MgO samplesThe areal proportion of olivine was calculated from raster images of olivine grains using ImageJ image analysis software whichprovides an acceptable estimation of the modal abundance (eg Chayes 1954)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
25
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
Gray lines = final melting pressure
L + Ol + Opx
07
08
09
Pressure (GPa)
10
3
4
5
6
1
SOLIDUS
01
04
0506
L + Ol
2
03
02
PeridotiteKR-4003
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
Thick black lines = initial melting pressure
Dashed lines = melt fraction
Melt Fraction
2
3 4 5 6
L+Ol+Opx
SolidusSpinel
SolidusGarnet Peridotite
7
L+Ol
0 10 20 30 4040
45
50
55
60
MgO (wt)
SiO2
(wt)
PeridotiteKR-4003
Melt Fraction
0 10 20 305
10
15
CaO(wt)
MgO (wt)
3
4 5 6 7
2
46
2
Peridotite Partial Melts
Thick black line = initial melting pressureGray lines = final melting pressure
Peridotite
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
02
Pressure (GPa)
0302
04
Pressure (GPa)
Picrite
High-MgO basalt
Tholeiitic basalt
(c)
(d)
Karmutsen flood basaltsOlivine addition modelOlivine addition (5 increment)Primary magma
PicriteHigh-MgO basaltTholeiitic basalt
(a)
(b)
800 850 900
910
920
930
940
950
Karmutsen flood basalts
Olivine addition modelOlivine addition (5 increment)Primary magma
Gray lines = Fo contents of olivine
Fig 12 Estimated primary magma compositions for three Keogh Lake picrites and high-MgO basalts (samples 4723A4 4723A135616A7) using the forward and inverse modeling technique of Herzberg et al (2007) Karmutsen compositions and modeling results areoverlain on diagrams provided by C Herzberg (a) Whole-rock FeO vs MgO for Karmutsen samples from this study Total iron estimatedto be FeO is 090 Gray lines show olivine compositions that would precipitate from liquid of a given MgO^FeO composition Black lineswith crosses show results of olivine addition (inverse model) using PRIMELT1 (Herzberg et al 2007) Gray field highlights olivinecompositions with Fo contents of 91^92 predicted to precipitate (b) FeO vs MgO showing Karmutsen lava compositions and results ofa forward model for accumulated fractional melting of fertile peridotite KR-4003 (Kettle River Peridotite Walter 1998) (c) SiO2 vsMgO for Karmtusen lavas and model results (d) CaO vs MgO showing Karmtusen lava compositions and model results To brieflysummarize the technique [see Herzberg et al (2007) for complete description] potential parental magma compositions for the high-MgOlava series were selected (highest MgO and appropriate CaO) and using PRIMELT1 software olivine was incrementally added tothe selected compositions to show an array of potential primary magma compositions (inverse model) Then using PRIMELT1 theresults from the inverse model were compared with a range of accumulated fractional melts for fertile peridotite derived fromparameterization of the experimental results for KR-4003 from Walter (1998) forward model Herzberg amp OrsquoHara (2002) A melt fractionwas sought that was unique to both the inverse and forward models (Herzberg et al 2007) A unique solution was found when there was a com-mon melt fraction for both models in FeO^MgO and CaO^MgO^Al2O3^SiO2 (CMAS) projection space This modeling assumes thatolivine was the only phase crystallizing and ignores chromite precipitation and possible augite fractionation in the mantle (Herzbergamp OrsquoHara 2002) Results are best for a residue of spinel lherzolite (not pyroxenite) The presence of accumulated olivine in samples ofthe Keogh Lake picrites used as starting compositions does not significantly affect the results because we are modeling addition of olivine(see text for further description) The tholeiitic basalts cannot be used for modeling because they are all saturated in plagthorn cpxthornolThick black line for initial melting pressure at 4 GPa is not shown in (c) for clarity Gray area in (a) indicates parental magmas thatwill crystallize olivine with Fo 91^92 at the surface Dashed lines in (c) indicate melt fraction Solidi for spinel and garnet peridotite arelabelled in (c)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
The PRIMELT1 results indicate that if the Karmutsenpicritic lavas were derived from accumulated fractionalmelting of fertile peridotite the primary magmas wouldhave contained 15^17wt MgO and 10wt CaO(Fig 12 Table 6) formed from 23^27 partial meltingand would have crystallized olivine with a Fo content of 91 (Fig 12 Table 6) Calculated mantle temperatures forthe Karmutsen picrites (14908C) indicate melting ofanomalously hot mantle (100^2008C above ambient) com-pared with ambient mantle that produces mid-ocean ridgebasalt (MORB 1280^14008C Herzberg et al 2007)which initiated between 3 and 4 GPa Several picritesappear to be near-primary melts and required very littleaddition of olivine (eg sample 93G171 with measured158wt MgO) These temperature and melting esti-mates of the Karmutsen picrites are consistent withdecompression of hot mantle peridotite in an actively con-vecting plume head (ie plume initiation model) Threesamples (picrite 5614A1 and high-MgO basalts 5614A3and 5614A5) are lower in iron than the other Karmutsenlavas with FeO in the 65^80wt range (Fig 12c) andhave MgO and FeO contents that are similar to primitiveMORB from the Sequeiros fracture zone of the EastPacific Rise (EPR Perfit et al 1996) These samples
however differ from EPR MORB in having higher SiO2
and lower Al2O3 and they are restricted geographicallyto one area (Sara Lake Fig 2)We therefore have evidence for primary magmas with
MgO contents that range from a possible low of 110wt to as high as 174wt with inferred mantle potential tem-peratures from 1340 to 15208C This is similar to the rangedisplayed by lavas from the Ontong Java Plateau deter-mined by Herzberg (2004) who found that primarymagma compositions assuming a peridotite sourcewould contain 17wt MgO and 10wt CaO(Table 6) and that these magmas would have formed by27 melting at a mantle potential temperature of15008C (first melting occurring at 36 GPa) Fitton ampGodard (2004) estimated 30 partial melting for theOntong Java lavas based on the Zr contents of primarymagmas of Kroenke-type basalt calculated by incrementaladdition of equilibrium olivine However if a considerableamount of eclogite was involved in the formation of theOntong Java plateau magmas the excess temperatures esti-mated by Herzberg et al (2007) may not be required(Korenaga 2005) The variability in mantle potential tem-perature for the Keogh Lake picrites is also similar to thewide range of temperatures that have been calculated for
Table 6 Estimated primary magma compositions for Karmutsen basalts and other oceanic plateaus or islands
Sample 4723A4 4723A13 5616A7 93G171 Average OJP Mauna Keay Gorgonaz
Ontong Java primary magma composition for accumulated fraction melting (AFM) from Herzberg (2004)yMauna Kea primary magma composition is average of four samples of Herzberg (2006 table 1)zGorgona primary magma composition for AFM for 1F fertile source from Herzberg amp OrsquoHara (2002 table 4)Total iron estimated to be FeO is 090
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
27
lavas from single volcanoes in the Galapagos Hawaii andelsewhere by Herzberg amp Asimow (2008) Those workersinterpreted the range to reflect the transport of magmasfrom both the cool periphery and the hotter axis of aplume A thermally heterogeneous plume will yield pri-mary magmas with different MgO contents and the rangeof primary magma compositions and their inferred mantlepotential temperatures for Karmutsen lavas may reflectmelting from different parts of the plume
Source of the Karmutsen Formation lavasThe Sr^Nd^Hf^Pb isotopic compositions of theKarmutsen volcanic rocks on Vancouver Island indicatethat they formed from plume-type mantle containing adepleted component that is compositionally similar to butdistinct from other oceanic plateaus that formed in thePacific Ocean (eg Ontong Java and Caribbean Fig 13)Trace element and isotopic constraints can be used to testwhether the depleted component of the KarmutsenFormation was intrinsic to the mantle plume and distinctfrom the source of MORB or the result of mixing or invol-vement of ambient depleted MORB mantle with plume-type mantle The Karmutsen tholeiitic basalts have initialeNd and
87Sr86Sr ratios that fall within the range of basaltsfrom the Caribbean Plateau (eg Kerr et al 1996 1997Hauff et al 2000 Kerr 2003) and a range of Hawaiian iso-tope data (eg Frey et al 2005) and lie within and abovethe range for the Ontong Java Plateau (Tejada et al 2004Fig 13a) Karmutsen tholeiitic basalts with the highest ini-tial eNd and lowest initial 87Sr86Sr lie within and justbelow the field for age-corrected northern EPR MORB at230 Ma (eg Niu et al1999 Regelous et al1999) Initial Hfand Nd isotopic compositions for the Karmutsen samplesare noticeably offset to the low side of the ocean islandbasalt (OIB) array (Vervoort et al 1999) and lie belowand slightly overlap the low eHf end of fields for Hawaiiand the Caribbean Plateau (Fig 13c) the Karmutsen sam-ples mostly fall between and slightly overlap a field forage-corrected Pacific MORB at 230 Ma and Ontong Java(Mahoney et al 1992 1994 Nowell et al 1998 Chauvel ampBlichert-Toft 2001) Karmutsen lava Pb isotope ratios over-lap the broad range for the Caribbean Plateau and thelinear trends for the Karmutsen samples intersect thefields for EPR MORB Hawaii and Ontong Java in208Pb^206Pb space but do not intersect these fields in207Pb^206Pb space (Fig 13d and e) The more radiogenic207Pb204Pb for a given 206Pb204Pb of the Karmutsenbasalts requires high UPb ratios at an ancient time when235U was abundant similar to the Caribbean Plateau andenriched in comparison with EPR MORB Hawaii andOntong JavaThe low eHf^low eNd component of the Karmutsen
basalts that places them below the OIB mantle array andpartly within the field for age-corrected Pacific MORBcan form from low ratios of LuHf and high SmNd over
time (eg Salters amp White 1998) MORB will develop aHf^Nd isotopic signature over time that falls below themantle array Low LuHf can also lead to low 176Hf177Hffrom remelting of residues produced by melting in theabsence of garnet (eg Salters amp Hart 1991 Salters ampWhite 1998) The Hf^Nd isotopic compositions of theKarmutsen Formation indicate the possible involvementof depleted MORB mantle however similar to Iceland(Fitton et al 2003) Nb^Zr^Y variations provide a way todistinguish between a depleted component within theplume and a MORB-like component The Karmutsenbasalts and picrites fall within the Iceland array and donot overlap the MORB field in a logarithmic plot of NbYvs ZrY (Fig 13b) using the fields established by Fittonet al (1997) Similar to the depleted component within theCaribbean Plateau (Thompson et al 2003) the trend ofHf^Nd isotopic compositions towards Pacific MORB-likecompositions is not followed by MORB values in Nb^Zr^Y systematics and this suggests that the depleted compo-nent in the source of the Karmutsen basalts is distinctfrom the source of MORB and probably an intrinsic partof the plume The relatively radiogenic Pb isotopic ratiosand REE patterns distinct from MORB also support thisinterpretationTrace-element characteristics suggest the possible invol-
vement of sub-arc lithospheric mantle in the generation ofthe Karmutsen picrites Lassiter et al (1995) suggested thatmixing of a plume-type source with eNdthorn 6 to thorn 7 witharc material with low NbTh could reproduce the varia-tions observed in the Karmutsen basalts however theyconcluded that the absence of low NbLa ratios in theirsample of tholeiitic basalts restricts the amount of arc litho-sphere involved The study of Lassiter et al (1995) wasbased on major and trace element and Sr Nd and Pb iso-topic data for a suite of 29 samples from Buttle Lake in theStrathcona Provincial Park on central Vancouver Island(Fig 1) Whereas there are no clear HFSE depletions inthe Karmutsen tholeiitic basalts the low NbLa (and TaLa) of the Karmutsen picritic lavas in this study indicatethe possible involvement of Paleozoic arc lithosphere onVancouver Island (Fig 8)
REE modeling dynamic melting andsource mineralogyThe combined isotopic and trace-element variations of thepicritic and tholeiitic Karmutsen lavas indicate that differ-ences in trace elements may have originated from differentmelting histories For example the LuHf ratios of theKarmutsen picritic lavas (mean 008) are considerablyhigher than those in the tholeiitic basalts (mean 002)however the small range of overlapping initial eHf valuesfor the two lava suites suggests that these differences devel-oped during melting within the plume (Fig 9b) Below wemodel the effects of source mineralogy and depth of
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
28
melting on differences in trace-element concentrations inthe tholeiitic basalts and picritesDynamic melting models simulate progressive decom-
pression melting where some of the melt fraction isretained in the residue and only when the degree of partialmelting is greater than the critical mass porosity and the
source becomes permeable is excess melt extracted fromthe residue (Zou1998) For the Karmutsen lavas evolutionof trace element concentrations was simulated using theincongruent dynamic melting model developed by Zou ampReid (2001) Melting of the mantle at pressures greater than1 GPa involves incongruent melting (eg Longhi 2002)
OIB array
PacificMORB(230 Ma)
(0 Ma)
EPR(230 Ma)
-4
-2
0
2
4
8
10
12
14
16
18
20
22
-4 -2 0 2 4 6 8 10 12 14
minus2
0
2
4
6
8
10
12
14
0702 0703 0704 0705 0706
IndianMORB
87Sr 86Sr (initial)
Hf (initial)
(a) (c)
Nd (initial)
Karmutsen tholeiitic basalt
HawaiiOntong JavaCaribbean Plateau
Karmutsen high-MgO lava
high-MgO lavasKarmutsen
1540
1545
1550
1555
1560
1565
175 180 185 190 195 200375
380
385
390
395
175 180 185 190 195 200
(d) (e)
EPR
EPR
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb
Karmutsen Formation (initial)
HawaiiOntong JavaCaribbean Plateau
East Pacific Rise
Karmutsen Formation (measured)
Juan de FucaGorda
ε Nd
(initial)
Karmutsen Formation (different age-correction for 3 picrites)
(0 Ma)
NbY
001
01
1
1 10
ZrY
OIB
NMORB
(b)
Iceland array
6
Fig 13 Comparison of age-corrected (230 Ma) Sr^Nd^Hf^Pb isotopic compositions for Karmutsen flood basalts onVancouver Island withage-corrected OIB and MORB and Nb^Zr^Y variation (a) Initial eNd vs 87Sr86Sr (b) NbYand ZrY variation (after Fitton et al 1997)(c) Initial eHf vs eNd (d) Measured and initial 207Pb204Pb vs 206Pb204Pb (e) Measured and initial 208Pb204Pb vs 206Pb204Pb References fordata sources are listed in Electronic Appendix 4 Most of the compiled data were extracted from the GEOROC database (httpgeorocmpch-mainzgwdgdegeoroc) OIB array line in (c) is fromVervoort et al (1999) EPR East Pacific Rise Several high-MgO samples affected bysecondary alteration were not plotted for clarity in Pb isotope plots Estimation of fields for age-corrected Pacific MORB and EPR at 230 Mawere made using parent^daughter concentrations (in ppm) from Salters amp Stracke (2004) of Lufrac14 0063 Hffrac14 0202 Smfrac14 027 Ndfrac14 071Rbfrac14 0086 and Srfrac14 836 In the NbYvs ZrYplot the normal (N)-MORB and OIB fields not including Icelandic OIB are from data com-pilations of Fitton et al (1997 2003) The OIB field is data from 780 basalts (MgO45wt ) from major ocean islands given by Fitton et al(2003) The N-MORB field is based on data from the Reykjanes Ridge EPR and Southwest Indian Ridge from Fitton et al (1997) The twoparallel gray lines in (b) (Iceland array) are the limits of data for Icelandic rocks
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
29
with olivine being produced during the reactioncpxthornopxthorn spfrac14meltthornol for spinel peridotites The mod-eling used coefficients for incongruent melting reactionsbased on experiments on lherzolite melting (eg Kinzleramp Grove 1992 Longhi 2002) and was separated into (1)melting of spinel lherzolite (2) two combined intervals ofmelting for a garnet lherzolite (gt lherz) and spinel lher-zolite (sp lherz) source (3) melting of garnet lherzoliteThe model solutions are non-unique and a range indegree of melting partition coefficients melt reactioncoefficients and proportion of mixed source componentsis possible to achieve acceptable solutions (see Fig 14 cap-tion for description of modeling parameters anduncertainties)The LREE-depleted high-MgO lavas require melting of
a depleted spinel lherzolite source (Fig 14a) The modelingresults indicate a high degree of melting (22^25) similarto the results from PRIMELT1 based on major-elementvariations (discussed above) of a LREE-depleted sourceThe high-MgO lavas lack a residual garnet signature Theenriched tholeiitic basalts involved melting of both garnetand spinel lherzolite and represent aggregate melts pro-duced from continuous melting throughout the depth ofthe melting column (Fig 14b) Trace-element concentra-tions in the tholeiitic and picritic lavas are not consistentwith low-degree melting of garnet lherzolite alone(Fig 14c) The modeling results indicate that the aggregatemelts that formed the tholeiitic basalts would haveinvolved lesser proportions of enriched small-degreemelts (1^6 melting) generated at depth and greateramounts of depleted high-degree melts (12^25) gener-ated at lower pressure (a ratio of garnet lherzolite tospinel lherzolite melt of 14 provides the best fits seeFig 14b and description in figure caption) The totaldegree of melting for the tholeiitic basalts was high(23^27) similar to the degree of partial melting in thespinel lherzolite facies for the primary melts that eruptedas the Keogh Lake picrites of a source that was also prob-ably depleted in LREE
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
(c)
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
Melting of garnet lherzolite
High-MgO lavas
Melting of spinel lherzolite
source (07DM+03PM)
source (07DM+03PM)
6 melting
30 melting
(b)
(a)
30 melting
Sam
ple
Cho
ndrit
eS
ampl
eC
hond
rite 20 melting
1 melting
Melting of garnet lherzoliteand spinel lherzolite
x gt lherz + 4x sp lherz
5 meltingx gt lherz + 4x sp lherz
Sam
ple
Cho
ndrit
e
Tholeiitic lavas
Tholeiitic lavas
High-MgO lavas
source (07DM+03PM)
Fig 14 Trace-element modeling results for incongruent dynamicmantle melting for picritic and tholeiitic lavas from the KarmutsenFormation Three steps of modeling are shown (a) melting of spinellherzolite compared with high-MgO lavas (b) melting of garnet lher-zolite and spinel lherzolite compared with tholeiitic lavas (c) meltingof garnet lherzolite compared with tholeiitic and picritic lavasShaded fields in each panel for tholeiitic basalts and picrites arelabeled Patterns with symbols in each panel are modeling results(patterns are 5 melting increments in (a) and (b) and 1 incre-ments in (c) PM primitive mantle DM depleted MORB mantle gtlherz garnet lherzolite sp lherz spinel lherzolite In (b) the ratios ofper cent melting for garnet and spinel lherzolite are indicated (x gtlherzthorn 4x sp lherz means that for 5 melting 1 melt in garnetfacies and 4 melt in spinel facies where x is per cent melting in theinterval of modeling) The ratio of the respective melt proportions inthe aggregate melt (x gt lherzthorn 4x sp lherz) was kept constant andthis ratio of melt contribution provided the best fit to the data Arange of proportion of source components (07DMthorn 03PM to
09DMthorn 01PM) can reproduce the variation in the high-MgOlavas Melting modeling uses the formulation of Zou amp Reid (2001)an example calculation is shown in their Appendix PM fromMcDonough amp Sun (1995) and DM from Salters amp Stracke (2004)Melt reaction coefficients of spinel lherzolite from Kinzler amp Grove(1992) and garnet lherzolite fromWalter (1998) Partition coefficientsfrom Salters amp Stracke (2004) and Shaw (2000) were kept constantSource mineralogy for spinel lherzolite (018cpx027opx052ol003sp) from Kinzler (1997) and for garnet lherzolite(034cpx008opx053ol005gt) from Salters amp Stracke (2004) Arange of source compositions for melting of spinel lherzolite(07DMthorn 03PM to 03DMthorn 07PM) reproduce REE patternssimilar to those of the high-MgO lavas with best fits for 22ndash25melting and 07DMthorn 03PM source composition Concentrationsof tholeiitic basalts cannot be reproduced from an entirelyprimitive or depleted source
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
30
The uncertainty in the composition of the source in thespinel lherzolite melting model for the high-MgO lavasprecludes determining whether the source was moredepleted in incompatible trace elements than the source ofthe tholeiitic lavas (see Fig 14 caption for description) It ispossible that early high-pressure low-degree melting left aresidue within parts of the plume (possibly the hotter inte-rior portion) that was depleted in trace elements (egElliott et al 1991) further decompression and high-degreemelting of such depleted regions could then have generatedthe depleted high-MgO Karmutsen lavas Shallow high-degree melting would preferentially sample depletedregions whereas deeper low-degree melting may notsample more refractory depleted regions (eg CaribbeanEscuder-Viruete et al 2007) The picrites lie at the highend of the range of Nd isotopic compositions and havelow NbZr similar to depleted Icelandic basalts (Fittonet al 2003) therefore the picrites may represent the partsof the mantle plume that were the most depleted prior tothe initiation of melting As Fitton et al (2003) suggestedthese depleted melts may only rarely be erupted withoutcombining with more voluminous less depleted melts andthey may be detected only as undifferentiated small-volume flows like the Keogh Lake picrites within theKarmutsen Formation on Vancouver IslandDecompression melting within the mantle plume initiatedwithin the garnet stability field (35^25 GPa) at highmantle potential temperatures (Tp414508C) and pro-ceeded beneath oceanic arc lithosphere within the spinelstability field (25^09 GPa) where more extensivedegrees of melting could occur
Magmatic evolution of Karmutsentholeiitic basaltsPrimary magmas in large igneous provinces leave exten-sive coarse-grained cumulate residues within or below thecrust these mafic and ultramafic plutonic sequences repre-sent a significant proportion of the original magma thatpartially crystallized at depth (eg Farnetani et al 1996)For example seismic and petrological studies of OntongJava (eg Farnetani et al 1996) combined with the use ofMELTS (Ghiorso amp Sack 1995) indicate that the volcanicsequence may represent only 40 of the total magmareaching the Moho Neal et al (1997) estimated that30^45 fractional crystallization took place for theOntong Java magmas and produced cumulate thicknessesof 7^14 km corresponding to 11^22 km of flood basaltsdepending on the presence of pre-existing oceanic crustand the total crustal thickness (25^35 km) Interpretationsof seismic velocity measurements suggest the presence ofpyroxene and gabbroic cumulates 9^16 km thick beneaththe Ontong Java Plateau (eg Hussong et al 1979)Wrangellia is characterized by high crustal velocities in
seismic refraction lines which is indicative of mafic
plutonic rocks extending to depth beneath VancouverIsland (Clowes et al 1995)Wrangellia crust is 25^30 kmthick beneath Vancouver Island and is underlain by astrongly reflective zone of high velocity and density thathas been interpreted as a major shear zone where lowerWrangellia lithosphere was detached (Clowes et al 1995)The vast majority of the Karmutsen flows are evolved tho-leiitic lavas (eg low MgO high FeOMgO) indicating animportant role for partial crystallization of melts at lowpressure and a significant portion of the crust beneathVancouver Island has seismic properties that are consistentwith crystalline residues from partially crystallizedKarmutsen magmas To test the proportion of the primarymagma that fractionated within the crust MELTS(Ghiorso amp Sack 1995) was used to simulate fractionalcrystallization using several estimated primary magmacompositions from the PRIMELT1 modeling results Themajor-element composition of the primary magmas couldnot be estimated for the tholeiitic basalts because of exten-sive plagthorn cpxthornol fractionation and thus the MELTSmodeling of the picrites is used as a proxy for the evolutionof major elements in the volcanic sequence A pressure of 1kbar was used with variable water contents (anhydrousand 02wt H2O) calculated at the quartz^fayalite^magnetite (QFM) oxygen buffer Experimental resultsfrom Ontong Java indicate that most crystallizationoccurs in magma chambers shallower than 6 km deep(52 kbar Sano amp Yamashita 2004) however some crys-tallization may take place at greater depths (3^5 kbarFarnetani et al 1996)A selection of the MELTS results are shown in Fig 15
Olivine and spinel crystallize at high temperature until15^20wt of the liquid mass has fractionated and theresidual magma contains 9^10wt MgO (Fig 15f) At1235^12258C plagioclase begins crystallizing and between1235 and 11908C olivine ceases crystallizing and clinopyr-oxene saturates (Fig 15f) As expected the addition of clin-opyroxene and plagioclase to the crystallization sequencecauses substantial changes in the composition of the resid-ual liquid FeO and TiO2 increase and Al2O3 and CaOdecrease while the decrease in MgO lessens considerably(Fig 15) The near-vertical trends for the Karmutsen tho-leiitic basalts especially for CaO do not match the calcu-lated liquid-line-of-descent very well which misses manyof the high-CaO high-Al2O3 low-FeO basalts A positivecorrelation between Al2O3CaO and SrNd indicates thataccumulation of plagioclase played a major role in the evo-lution of the tholeiitic basalts (Fig 15g) Increasing thecrystallization pressure slightly and the water content ofthe melts does not systematically improve the match ofthe MELTS models to the observed dataThe MELTS results indicate that a significant propor-
tion of the crystallization takes place before plagioclase(15^20 crystallization) and clinopyroxene (35^45
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
31
crystallization) join the crystallization sequence Theseresults suggest that445 crystallization of the primarymagmas occurred (estimated from the residual liquid percent in Fig 15f) and if the original flood basalt stratigra-phy on Vancouver Island was 6 km thick at least anequivalent amount of complementary plutonic rocksshould have crystallized The Karmutsen basalts and theirplutonic residues thus represent a significant addition ofcrust (perhaps 412 km thickness) previously thickenedduring the formation of the underlying Paleozoic arc
Growth of the Wrangellia oceanic plateauon Vancouver IslandThe extrusive part of the Wrangellia oceanic plateau onVancouver Island was constructed in a three-layersequence of mostly tholeiitic basalts with a restrictedrange of composition (Fig 16) The plateau overliesDevonian to Mississippian arc volcanic sequences andmarine sediments of Mississippian to Permian ages thatare intruded by gabbroic rocks related to the flood basaltsThe volcanic stratigraphy of theWrangellia plateau formed
00
05
10
15
20
25
30
35
40
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
16
0 2 4 6 8 10 12 14 16 18 20
10
11
12
13
14
15
16
17
18
19
20
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
0 2 4 6 8 10 12 14 16 18 20
MgO (wt)
0 10 20 30 40 50
percent of total mass
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
) Mineral proportions
Sp (e)
Ol
CpxPlag
PicriteHigh-MgO basaltTholeiitic basalt
TiO2 (wt)
MgO (wt)
Al2O3 (wt) CaO (wt)
FeO (wt)
MgO(a) (b)
(c) (d)
MgO (wt)MgO (wt)
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
)
Residual liquid ()
1220
1190
Ol + Sp
Plag+Ol+Sp Plag+Cpx+Sp
02 wt H2O
no H2O
Residual liquid and crystallizing phases
(f)
0102030405060708090100
plagioclaseaccumulation
plagioclaseaccumulation
plagioclase
accumulatio
n
10
12
14
16
0 10 20 30 40
Al2O3CaO
SrNd
tholeiitic basalts
(g)
Fig 15 Forward fractional crystallization modeling results for major elements from MELTS (Ghiorso amp Sack 1995) compared with picriticand tholeiitic lavas from the Karmutsen Formation (a^d) Major-element compositions of residual liquids compared with Karmutsen lavacompositions For clarity the results are shown only for one starting composition (open circle) using the composition of the estimated primarymagma for the high-MgO lava series from the modeling technique of Herzberg et al (2007) for sample 4723A13 as shown in Fig 12 The esti-mated primary magmas for samples 4723A3 and 93G171 yield similar results A pressure of 1 kbar was used with no H2O and 02wt H2Ocalculated at QFM black line is anhydrous and grey line is 02wt H2OThe end-point for the anhydrous result is at 78 fractionation and11308C and the end-point for the result with 02wt H2O is 75 fractionation and11108C (e) Mineral proportions of crystallized phases fromone MELTrun (sample 4723A13) with no H2O (f) Per cent residual liquid vs temperature labelled with intervals of crystallizing phases (g)Al2O3CaO vs SrNd for the tholeiitic basalts
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
32
in a single eruptive phase (c 231^225 Ma) as an emergentbasalt sequence during a deeper-water shallow-water andsubaerial stage and subsided below sea level relativelysoon after its formation (Fig 16) The deeper-water stagecontains pillowed and massive submarine flows (3000mthick) that built up the submarine volcanic edifice from theseafloor The lowest pillow basalts were emplaced onunconsolidated and lithified fine-grained siliciclastic andcarbonaceous sediments containing Daonella beds(c 235^232 Ma) that are intruded by abundantKarmutsen mafic sills (Fig 16a) The different pillowedand massive submarine flows resulted primarily from dif-ferent effusive rates and local topography Some of themassive submarine flows commonly recognized by radiat-ing columnar jointing represent master tubes for the deliv-ery of lava to distal parts of submarine flow fields Thepillowed flows which commonly have uneven topography
in modern sequences may have served as levees to themassive submarine flows rather than allowing unob-structed spreading and inflation typical of the massive sub-aerial flows that form on shallow slopes in continentalflood basalt provinces (eg Self et al 1997) The rapid suc-cession of flows is evident by the near absence of sedimentbetween submarine flows Minor volumes of picritic pillowlavas occur in the upper part of the pillowed flow sequencein areas of the plateau exposed on northern VancouverIsland (Fig 16 Nixon et al 2008)The growth of the plateau in shallow water led to
emplacement of an increasing proportion of volcaniclastics(pillow breccias and hyaloclastites) that accumulated up to1500m thick in some areas (Fig 16b Nixon et al 2008)Horizons of pillowed flows occur locally within the volca-niclastic stratigraphy Sedimentary structures (eg gradedbedding fluidization structures) and petrographic textures
Late Devonian-Mississippian arc
Daonella beds
Middle Triassic
deeper water
Late Devonian-Mississippian arc
pillowed flows massive submarine
flows
pillow brecciahyaloclastiteshallow water
~4 km
~6 km
Late Triassic limestoneinterflow limestonelenses
Karmutsen Formation
mafic intrusions sediment-sill complex
Karmutsen Formation
(a) (b)
(c) (d)
subaerial flows
Late Devonian-Mississippian arc Late Devonian-Mississippian arc
picritic pillow basalts
Miss to Permian marine sediments
sills
Fig 16 Growth of the Wrangellia oceanic plateau onVancouver Island (a) Deep-water eruptive phase of pillowed and massive submarineflows with almost no intervening sediments onto MiddleTriassic shale containing Daonella beds (c 235^232 Ma) The basement of the plateauis composed of Devonian to Mississippian arc sequences overlain by Mississippian to Permian carbonates and siliciclastic sediments(b) Shallow-water flow emplacement with increasing proportion of volcaniclastics resulted primarily from cooling^contraction granulationPicritic pillow basalts occur primarily near the transition between pillowed submarine flows and volcaniclastic units Pillowed and massiveflows occur locally within the volcaniclastic unit (c) Emergent stage of growth consists of overlapping subaerial flow fields similar to continen-tal flood basalts with pahoehoe flow structures Multi-tiered columnar jointing occurrs locally in response to ponding of flows (eg Bondre et al2004) Interflow carbonate deposits and plagioclase megacrystic flows developed locally during the waning stage of volcanism (d) Post-volcanicsubsidence and deposition of Late Carnian to Norian (c 212^223 Ma) platform carbonates on top of the plateau Units are labeled in the dia-grams Thicknesses of the underlying units are not to scale
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
33
within the volcaniclastic rocks indicate formation primar-ily via cooling^contraction granulation magma^water^steam interaction autobrecciation resedimentation andmass-wasting rather than pyroclastic eruption (Nixonet al 2008)A broad subaerial platform was constructed by emplace-
ment of subaerial lava flows during an emergent subaerialstage (Fig 16c) The subaerial unit predominantly com-prises well-layered sheet flows that formed in a similarway to continental flood basalts as hundred(s) of pahoehoeflow fields erupted on shallow slopes (Self et al 1997) how-ever volcaniclastic and pillowed flow units are locally pre-sent mostly in the upper parts of the stratigraphyLocalized ponding of flows led to the formation of multi-tiered columnar jointing in some subaerial flows (egBondre et al 2004) As volcanism waned local interflowcarbonate deposits developed and plagioclase megacrysticflows were erupted Subsidence of the plateau is recordedby the interflow lenses and deposits of platform carbonates(40^1500m thick) on top of the Karmutsen basalts thatcontain Late Carnian to Early Norian fossils (Fig 16d c216^227 Ma Nixon et al 2008)The evolution of the internal architecture of the
Wrangellia oceanic plateau was primarily related to theconditions of emplacement (ie water depth and eruptionenvironment) There does not appear to have been a sys-tematic coincident evolution or shift of lava compositionsonVancouver Island as has been recognized in most con-tinental flood basalts and parts of the Wrangellia oceanicplateau in Alaska and Yukon (Greene et al 2008 2009)In general with the exception of the picritic pillow lavasin the submarine stratigraphy and local variations betweenhigh-MgO and tholeiitic lavas there are no significantcompositional differences between lavas erupted duringthe different stages of growth of the plateau (ie submarineand subaerial flows) nor are substantial compositional dif-ferences generally found between massive submarine flowsin contact with pillowed flows tholeiitic basalt flows nearthe base of the Karmutsen Formation are compositionallysimilar to flows sampled near the top of the volcanicsequence Where it was possible to sample continuous sec-tions of volcanic stratigraphy greater than several hun-dreds of meters (eg Keogh^Maynard Lake area west ofthe Karmutsen Range Fig 2) there is no clear relationshipbetween stratigraphic position and compositionalvariation
CONCLUSIONSThe Karmutsen Formation covers extensive areas ofVancouver Island (20 000 km2) and forms part of a largeoceanic plateau erupted over a geologically short interval(c 225^231 Ma) The tripartite volcanic stratigraphyexposed on Vancouver Island is at least 6 km thick andcomprises a succession of submarine flows volcaniclastic
deposits and massive subaerial flowsThese volcanologicaldifferences are primarily related to the eruption environ-ment (deep-water shallow-water or subaerial) The rapidgrowth of the plateau prevented significant accumulationsof intervening sediment except in the uppermost stratigra-phy where isolated limestone lenses commonly associatedwith pillow basalt and hyaloclastite deposits preserve arecord of the final stages of subsidence of the plateauThe Wrangellia plateau on Vancouver Island was con-
structed dominantly of tholeiitic basalt with restrictedmajor- and trace-element and isotopic compositional var-iations Minor volumes of picritic pillow basalts eruptedlate in the submarine phase The Karmutsen basalts wereemplaced onto older Paleozoic arc volcanic sequences andlow NbLa for the picrites may indicate a contributionfrom the sub-arc lithospheric mantle Modeling of the pet-rogenesis of the Karmutsen picrites indicates melting ofanomalously hot mantle (15008C) and extensive degreesof partial melting (23^27) and is consistent with aplume initiation model Combined Sr^Nd^Hf^Pb isotopicsystematics indicate that the picritic and tholeiitic lavaswere derived from a depleted plume-type mantle sourcedistinct from the source of MORB that has compositionalsimilarities to the source of the lavas forming theCaribbean Plateau Picrites were generated by melting ofspinel lherzolite whereas the tholeiitic basalts formed asaggregate melts of both garnet and spinel lherzolite Thetholeiitic basalts underwent extensive low-pressure fractio-nation (52^3 kbar) and seismic studies indicate that someof theWrangellia crust beneathVancouver Island may cor-respond to the plutonic residues of the Karmutsen floodbasalts
ACKNOWLEDGEMENTSWe are grateful to Nick Arndt for helping us get this pro-ject started and Nick Massey for insights into VancouverIsland geology We greatly appreciate the assistance withPRIMELT modeling from Claude Herzberg We thankMikkel Schau for his insight and enthusiasm during field-work Jane Barling assisted with analyses by MC-ICP-MSThoughtful and thorough reviews of the manuscript byJohn Mahoney Fred Frey Wendy Bohrson DougalJerram and Teal Riley are greatly appreciated Fundingwas generously provided by the Rocks to Riches Programadministered by the BC amp Yukon Chamber of Mines in2004 by the BC Geological Survey and by NSERCDiscovery Grants to JSS and DW ARG was supportedby a University Graduate Fellowship at UBC
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
34
REFERENCESBarker F Brown A S Budahn J R amp Plafker G (1989) Back-arc
with frontal-arc component origin of Triassic Karmutsen basaltBritish Columbia Canada Chemical Geology 75 81^102
Blichert-Toft JWeis D Maerschalk C Agranier A amp Albare de F(2003) Hawaiian hot spot dynamics as inferred from the Hf and Pbisotope evolution of Mauna Kea volcano Geochemistry GeophysicsGeosystems 4(2) 1^27 doi1010292002GC000340
Bondre N R Duraiswami R A amp Dole G (2004) Morphologyand emplacement of flows from the Deccan Volcanic ProvinceIndia Bulletin of Volcanology 66 29^45
Carlisle D (1963) Pillow breccias and their aquagene tuffs QuadraIsland British Columbia Journal of Geology 71 48^71
Carlisle D (1972) Late Paleozoic to Mid-Triassic sedimentary-volcanic sequence on Northeastern Vancouver Island Geological
Society of Canada Paper Report of Activities 72-1 Part B 24^30Carlisle D amp SuzukiT (1974) Emergent basalt and submergent car-
bonate^clastic sequences including the Upper Triassic Dilleri andWelleri zones on Vancouver Island Canadian Journal of Earth
Sciences 11 254^279Chauvel C amp Blichert-Toft J (2001) A hafnium isotope and trace
element perspective on melting of the depleted mantle Earth and
Planetary Science Letters 190 137^151Chayes F (1954)The theory of thin-section analysis Journal of Geology
62 92^101Cho M Liou J G amp Maruyama S (1987) Transition from the zeo-
lite to prehnite^pumpellyite facies in the Karmutsen metabasitesVancouver Island British Columbia Journal of Petrology 27(2)467^494
Clowes R M Zelt C A Amor J R amp Ellis R M (1995)Lithospheric structure in the southern Canadian Cordillera froma network of seismic refraction lines Canadian Journal of Earth
Sciences 32(10) 1485^1513Coffin M F amp Eldholm O (1994) Large igneous provinces Crustal
structure dimensions and external consequences Reviews of
Geophysics 32(1) 1^36Cox K G (1980) A model for flood basalt volcanism Journal of
Petrology 21 629^650Eldholm O amp Coffin M F (2000) Large igneous provinces
and plate tectonics In Richards M A Gordon R G amp vander Hilst R D (eds) The History and Dynamics of Global
Plate Motions American Geophysical Union Geophysical Monograph 121309^326
Elliott T R Hawkesworth C J amp Grolaquo nvold K (1991) Dynamicmelting of the Iceland plume Nature 351 201^206
Escuder-Viruete J Pecurren rez-Estaucurren n A Contreras F Joubert MWeis D Ullrich T D amp Spadea P (2007) Plume mantle sourceheterogeneity through time Insights from the Duarte ComplexHispaniola northeastern Caribbean Journal of Geophysical Research112(B04203) doi1010292006JB004323
Farnetani C G Richards M A amp Ghiorso M S (1996)Petrological models of magma evolution and deep crustal structurebeneath hotspots and flood basalt provinces Earth and Planetary
Science Letters 143 81^94Fitton J G amp Godard M (2004) Origin and evolution of magmas
on the Ontong Java Plateau In Fitton J G Mahoney J JWallace P J amp Saunders A D (eds) Origin and Evolution of the
Ontong Java Plateau Geological Society London Special Publications 229151^178
Fitton J G Saunders A D Norry M J Hardarson B S ampTaylor R N (1997) Thermal and chemical structure of theIceland plume Earth and Planetary Science Letters 153 197^208
Fitton J G Saunders A D Kempton P D amp Hardarson B S(2003) Does depleted mantle form an intrinsic part of the Icelandplume Geochemistry Geophysics Geosystems 4(3) 1032 doi1010292002GC000424
Frey F A Huang S Blichert-Toft J Regelous M amp Boyet M(2005) Origin of depleted components in basalt related to theHawaiian hotspot Evidence from isotopic and incompatible ele-ment ratios Geochemistry Geophysics Geosystems 6(1) 23 doi1010292004GC000757
Galer S J G amp Abouchami W (1998) Practical application of leadtriple spiking for correction of instrumental mass discriminationMineralogical Magazine 62A 491^492
Ghiorso M S amp Sack R O (1995) Chemical mass transfer in mag-matic processes IV A revised and internally consistent thermody-namic model for the interpolation and extrapolation of liquid^solidequilibria in magmatic systems at elevated temperatures and pres-sures Contributions to Mineralogy and Petrology 119 197^212
Greene A R Scoates J S amp Weis D (2005)WrangelliaTerrane onVancouver Island Distribution of flood basalts with implicationsfor potential Ni^Cu^PGE mineralization In Grant B (ed)Geological Fieldwork 2004 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2005-1 209^220Greene A R Scoates J S Nixon G T amp Weis D (2006) Picritic
lavas and basal sills in the Karmutsen flood basalt provinceWrangellia northern Vancouver Island In Grant B (ed)Geological Fieldwork 2005 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2006-1 39^52Greene A R Scoates J S amp Weis D (2008) Wrangellia flood
basalts in Alaska A record of plume^lithosphere interaction in aLate Triassic accreted oceanic plateau Geochemistry Geophysics
Geosystems 9(Q12004) doi1010292008GC002092Greene A R Scoates J S Weis D amp Israel S (2009)
Geochemistry of triassic flood basalts from the Yukon (Canada)segment of the accreted Wrangellia oceanic plateau Lithosdoi101016jlithos200811010
Hauff F Hoernle K Tilton G Graham D W amp Kerr A C(2000) Large volume recycling of oceanic lithosphere over shorttime scales geochemical constraints from the Caribbean LargeIgneous Province Earth and Planetary Science Letters 174 247^263
Hauff F Hoernle K amp Schmidt A (2003) Sr^Nd^Pb compositionof Mesozoic Pacific oceanic crust (Site 1149 and 801 ODP Leg185)Implications for alteration of ocean crust and the input into theIzu^Bonin^Mariana subduction system Geochemistry Geophysics
Geosystems 4(8) doi1010292002GC000421Herzberg C (2004) Partial melting below the Ontong Java Plateau
In Fitton J G Mahoney J J Wallace P J amp Saunders A D(eds) Origin and Evolution of the Ontong Java Plateau Geological SocietyLondon Special Publications 229 179^183
Herzberg C (2006) Petrology and thermal structure of the Hawaiianplume from Mauna Kea volcano Nature 444(7119) 605
Herzberg C amp Asimow P D (2008) Petrology of some oceanicisland basalts PRIMELT2XLS software for primary magma cal-culation Geochemistry Geophysics Geosystems 9(Q09001) doi1010292008GC002057
Herzberg C amp OrsquoHara M J (2002) Plume-associated ultramaficmagmas of Phanerozoic age Journal of Petrology 43(10) 1857^1883
Herzberg C Asimow P D Arndt N Niu Y Lesher C MFitton J G Cheadle M J amp Saunders A D (2007)Temperatures in ambient mantle and plumes Constraints frombasalts picrites and komatiites Geochemistry Geophysics Geosystems8(Q02006) doi1010292006GC001390
Huang S amp Frey F A (2005) Trace element abundances of MaunaKea basalt from phase 2 of the Hawaii Scientific Drilling Project
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
35
Petrogenetic implications of correlations with major element con-tent and isotopic ratios Geochemistry Geophysics Geosystems 4(6)doi1010292002GC000322
Hussong D M Wipperman L K amp Kroenke L W (1979) Thecrustal structure of the Ontong Java and Manihiki OceanicPlateaus Journal of Geophysical Research 84(B11) 6003^6010
Jerram D A amp Widdowson M (2005) The anatomy of ContinentalFlood Basalt Provinces geological constraints on the processes andproducts of flood volcanism Lithos 79(3^4) 385^405
Jones D L Silberling N J amp Hillhouse J (1977)Wrangellia a dis-placed terrane in northwestern North America CanadianJournal ofEarth Sciences 14(11) 2565^2577
Kerr A C (2003) Oceanic Plateaus In Rudnick R (ed) The Crust
Treatise on GeochemistryVol 3 Oxford Elsevier Science pp 537^565Kerr A C amp Mahoney J J (2007) Oceanic plateaus Problematic
plumes potential paradigms Chemical Geology 241 332^353Kerr A CTarney J Marriner G F Klaver GT Saunders A D
amp Thirlwall M F (1996) The geochemistry and petrogenesis ofthe Late-Cretaceous picrites and basalts of Curacao NetherlandsAntilles a remnant of an oceanic plateau Contributions to
Mineralogy and Petrology 124(1) 29^43Kerr A C Marriner G F Tarney J Nivia A Saunders A D
Thirlwall M F amp Sinton A C (1997) Cretaceous basaltic ter-ranes in Western Colombia Elemental chronological and Sr-Ndisotopic constraints on petrogenesis Journal of Petrology 38(6)677^702
Kinzler R J (1997) Melting of mantle peridotite at pressuresapproaching the spinel to garnet transition Application to mid-ocean ridge basalt petrogenesis Journal of Geophysical Research
102(B1) 853^874Kinzler R J amp Grove T L (1992) Primary magmas of mid-ocean
ridge basalts 1 Experiments and methods Journal of Geophysical
Research 97(B5) 6885^6906Korenaga J (2005) Why did not the Ontong Java Plateau form sub-
aerially Earth and Planetary Science Letters 234 385^399Kuniyoshi S (1972) Petrology of the Karmutsen (Volcanic) Group
northeastern Vancouver Island British Columbia PhD disserta-tion University of California Los Angeles 242 pp
Lassiter J C DePaolo D J amp Mahoney J J (1995) Geochemistry ofthe Wrangellia flood basalt province Implications for the role ofcontinental and oceanic lithosphere in flood basalt genesis Journalof Petrology 36(4) 983^1009
LeBas M J (2000) IUGS reclassification of the high-Mg and picriticvolcanic rocks Journal of Petrology 41(10) 1467^1470
Longhi J (2002) Some phase equilibrium systematics of lherzolitemelting I Geochemistry Geophysics Geosystems 3(3) doi1010292001GC000204
MacDonald G A amp Katsura T (1964) Chemical composition ofHawaiian lavas Journal of Petrology 5 82^133
Mahoney J J (1987) An isotopic survey of Pacific oceanic plateausimplications for their nature and origin In Keating B HFryer P Batiza R amp Boehlert GW (eds) Seamounts Islands andAtolls American Geophysical Union Geophysical Monograph 43 207^220
Mahoney J LeRoex A P Peng Z Fisher R L amp Natland J H(1992) Southwestern limits of Indian Ocean Ridge mantle and theorigin of low 206Pb204Pb mid-ocean ridge basalt isotope systema-tics of the central Southwest Indian Ridge (178^508E) Journal ofGeophysical Research 97(B13) 19771^19790
Mahoney J J Sinton J M Kurz M D MacDougall J DSpencer K J amp Lugmair GW (1994) Isotope and trace elementcharacteristics of a super-fast spreading ridge East Pacific rise13^238S Earth and Planetary Science Letters 121 173^193
Massey N W D (1995a) Geology and mineral resources of theAlberni^Nanaimo Lakes sheet Vancouver Island 92F1W 92F2Eand part of 92F7E British Columbia Ministry of Energy Mines and
Petroleum Resources Paper 1992-2 132 ppMassey N W D (1995b) Geology and mineral resources of the
Cowichan Lake sheet Vancouver Island 92C16 British Columbia
Ministry of Energy Mines and Petroleum Resources Paper 1992-3 112 ppMassey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005a) Digital Geology Map of British Columbia Tile NM9 MidCoast BC British Columbia Ministry of Energy and Mines Geofile
2005-2Massey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005b) Digital Geology Map of British Columbia Tile NM10Southwest BC British Columbia Ministry of Energy and Mines Geofile
2005-3McDonough W F amp Sun S (1995) The composition of the Earth
Chemical Geology 120 223^253Monger JW H amp Journeay J M (1994) Basement geology and tec-
tonic evolution of theVancouver region In Monger JW H (ed)Geology and Geological Hazards of the Vancouver Region Southwestern
British Columbia Geological Survey of Canada Bulletin 481 3^25Muller J E (1977) Geology of Vancouver Island Geological Survey of
Canada Open File Map 463Muller J E (1980)The Paleozoic Sicker Group of Vancouver Island British
Columbia Geological Survey of Canada Paper 79-30 22 ppMuller J E Northcote K E amp Carlisle D (1974) Geology and
mineral deposits of Alert Bay^Cape Scott map area VancouverIsland British Columbia Geological Survey of Canada Paper 74-8 77
Neal C R Mahoney J J Kroenke L W Duncan R A ampPetterson M G (1997) The Ontong^Java Plateau InMahoney J J amp Coffin M F (eds) Large Igneous Provinces
Continental Oceanic and Planetary FloodVolcanism American Geophysical
Union Geophysical Monograph 100 183^216Niu Y Collerson K D Batiza R Wendt J I amp Regelous M
(1999) Origin of enriched-typed mid-ocean ridge basalt at ridgesfar from mantle plumes The East Pacific Rise at 11820rsquoN Journalof Geophysical Research 104 7067^7088
Nixon G T amp Orr A J (2007) Recent revisions to the EarlyMesozoic stratigraphy of Northern Vancouver Island (NTS 102I092L) and metallogenic implicatinos British Columbia InGrant B (ed) Geological Fieldwork 2006 British Columbia Ministry of
Energy Mines and Petroleum Resources Paper 2007-1 163^177Nixon GT Hammack J L KoyanagiV M Payie G J Haggart
JW Orchard M JTozerT Friedman R M Archibald D APalfy J amp Cordey F (2006a) Geology of the Quatsino^PortMcNeill area northernVancouver Island British Columbia Ministry
of Energy Mines and Petroleum Resources Geoscience Map 2006-2 scale150 000
Nixon G T Kelman M C Stevenson D Stokes L A ampJohnston K A (2006b) Preliminary geology of the Nimpkishmap area (NTS 092L07) northern Vancouver Island BritishColumbia In Grant B amp Newell J M (eds) Geological Fieldwork2005 British Columbia Ministry of Energy Mines and Petroleum Resources
Paper 2006-1 135^152Nixon G T Laroque J Pals A Styan J Greene A R amp
Scoates J S (2008) High-Mg lavas in the Karmutsen flood basaltsnorthernVancouver Island (NTS 092L) Stratigraphic setting andmetallogenic significance In Grant B (ed) Geological Fieldwork
2007 British Columbia Ministry of Energy Mines and Petroleum
Resources Paper 2008-1 175^190Nowell G M Kempton P D Noble S R Fitton J G
Saunders A Mahoney J J amp Taylor R N (1998) High precisionHf isotope measurements of MORB and OIB by thermal ionisation
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
36
mass spectrometry insights into the depleted mantle Chemical
Geology 149 211^233Parrish R R amp McNicoll V J (1992) U^Pb age determinations
from the southern Vancouver Island area British Columbia InRadiogenic Age and Isotopic Studies Report 5 Geological Survey of
Canada Paper 91-2 79^86Peate DW (1997) The Parana^Etendeka Province In Coffin M F
amp Mahoney J (eds) Large Igneous Provinces Continental Oceanic andPlanetary Flood Volcanism American Geophysical Union Geophysical
Monograph 100 217^245Perfit M R Fornari D J Ridley W I Kirk P D Casey J
Kastens K A Reynolds J R Edwards M Desonie DShuster R amp Paradis S (1996) Recent volcanism in the Siqueirostransform fault picritic basalts and implications for MORBmagma genesis Earth and Planetary Science Letters 141(1^4) 91^108
Pretorius W Weis D Williams G Hanano D Kieffer B ampScoates J S (2006) Complete trace elemental characterization ofgranitoid (USGSG-2GSP-2) reference materials by high resolutioninductively coupled plasma-mass spectrometry Geostandards and
Geoanalytical Research 30(1) 39^54Regelous M Niu Y Wendt J I Batiza R Greig A amp
Collerson K D (1999) Variations in the geochemistry of magma-tism on the East Pacific Rise at 10830rsquoN since 800 ka Earth and
Planetary Science Letters 168 45^63Recurren villon S Chauvel C Arndt N T Pik R Martineau F
Fourcade S amp Marty B (2002) Heterogeneity of the Caribbeanplateau mantle source Sr O and He isotopic compositions of oli-vine and clinopyroxene from Gorgona Island Earth and Planetary
Science Letters 205(1^2) 91Rhodes J M (1996) Geochemical stratigraphy of lava flows samples
by the Hawaii Scientific Drilling Project Journal of Geophysical
Research 101(B5) 11729^11746Rhodes J M amp Vollinger M J (2004) Composition of basaltic lavas
sampled by phase-2 of the Hawaii Scientific Drilling ProjectGeochemical stratigraphy and magma types Geochemistry
Geophysics Geosystems 5(3) doi1010292002GC000434Richards M A Jones D L Duncan R A amp DePaolo D J (1991)
A mantle plume initiation model for the Wrangellia flood basaltand other oceanic plateaus Science 254 263^267
Salters V J M amp Hart S R (1991) The mantle sources of oceanridges islands and arcs the Hf-isotope connection Earth and
Planetary Science Letters 104 364^380Salters V J M amp Stracke A (2004) Composition of the depleted
SaltersV J amp WhiteW (1998) Hf isotope constraints on mantle evo-lution Chemical Geology 145 447^460
SanoT amp Yamashita S (2004) Experimental petrology of basementlavas from Ocean Drilling Program Leg 192 implications for dif-ferentiation processes in Ontong Java Plateau magmas InFitton J G Mahoney J JWallace P J amp Saunders A D (eds)Origin and Evolution of the Ontong Java Plateau Geological Society
London Special Publications 229 185^218Saunders A D (2005) Large igneous provinces Origin and environ-
mental consequences Elements 1 259^263Self S ThordarsonT amp Keszthely L (1997) Emplacement of conti-
nental flood basalt lava flows In Mahoney J J amp Coffin M F(eds) Large Igneous Provinces Continental Oceanic and Planetary Flood
Volcanism American Geophysical Union Geophysical Monograph 100381^410
Shaw D M (2000) Continuous (dynamic) melting theory revisitedCanadian Mineralogist 38(5) 1041^1063
Sluggett C L (2003) Uranium^lead age and geochemical constraintson Paleozoic and Early Mesozoic magmatism in WrangelliaTerrane Saltspring Island British Columbia BSc thesisUniversity of British ColumbiaVancouver 84 pp
Storey M Mahoney J J amp Saunders A D (1997) Cretaceousbasalts in Madagascar and the transition between plume and con-tinental lithospheric mantle sources In Mahoney J J ampCoffin M F (eds) Large Igneous Provinces Continental Oceanic and
Planetary Flood Volcanism Aamerican Geophysical Union Geophysical
Monograph 100 95^122Surdam R C (1967) Low-grade metamorphism of the Karmutsen
Group PhD dissertation University of California Los Angeles288 pp
Sutherland-Brown A Yorath C J Anderson R G amp Dom K(1986) Geological maps of southern Vancouver IslandLITHOPROBE I 92C10 11 14 16 92F1 2 7 8 Geological Survey ofCanada Open File 1272
Tejada M L G Mahoney J J Castillo P R Ingle S P Sheth HC amp Weis D (2004) Pin-pricking the elephant evidence on theorigin of the Ontong Java Plateau from Pb^Sr^Hf^Nd isotopiccharacteristics of ODP Leg 192 basalts In Fitton J GMahoney J J Wallace P J amp Saunders A D (eds) Origin and
Evolution of the Ontong Java Plateau Geological Society London Special
Publications 229 133^150Thompson P M E Kempton P D amp Kerr A C (2008) Evaluation
of the effects of alteration and leaching on Sm^Nd and Lu^Hf sys-tematics in submarine mafic rocks Lithos 104(1^4) 164^176
Thompson P M E Kempton P D White R V Kerr A CTarney J Saunders A D Fitton J G amp McBirney A (2003)Hf-Nd isotope constraints on the origin of the CretaceousCaribbean plateau and its relationship to the Galapagos plumeEarth and Planetary Science Letters 217(1^2) 59^75
Vervoort J D Patchett P Blichert-Toft J amp Albare de F (1999)Relationships between Lu^Hf and Sm^Nd isotopic systems in theglobal sedimentary system Earth and Planetary Science Letters 16879^99
Walter M J (1998) Melting of garnet peridotite and the origin ofkomatiite and depleted lithosphere Journal of Petrology 39(1) 29^60
Weis D Kieffer B Maerschalk C Barling J de Jong JWilliams G A Hanano D Mattielli N Scoates J SGoolaerts A Friedman R A amp Mahoney J B (2006) High-pre-cision isotopic characterization of USGS reference materials byTIMS and MC-ICP-MS Geochemistry Geophysics Geosystems
7(Q08006) doi1010292006GC001283Weis D Kieffer B Hanano D Silva I N Barling J
Pretorius W Maerschalk C amp Mattielli N (2007) Hf isotopecompositions of US Geological Survey reference materialsGeochemistry Geophysics Geosystems 8(Q06006) doi1010292006GC001473
Wheeler J O amp McFeely P (1991) Tectonic assemblage map of theCanadian Cordillera and adjacent part of the United States ofAmerica Geological Survey of Canada Map 1712A
WhiteW M Albare de F amp Tecurren louk P (2000) High-precision analy-sis of Pb isotope ratios by multi-collector ICP-MS Chemical Geology167 257^270
Yorath C J Sutherland Brown A amp Massey N W D (1999)LITHOPROBE southern Vancouver Island British Columbia Geological
Survey of Canada Bulletin 498 145 ppZou H (1998) Trace element fractionation during modal and
non-model dynamic melting and open-system melting Amathematical treatment Geochimica et Cosmochimica Acta 62(11)1937^1945
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
37
Zou H amp Reid M R (2001) Quantitative modeling of trace elementfractionation during incongruent dynamic melting Geochimica et
Cosmochimica Acta 65(1) 153^162
APPENDIX SAMPLEPREPARAT ION ANDANALYT ICAL METHODSOnly the freshest rocks in the field were sampled and onlythe least altered samples were selected for chemical andisotopic analysis based on thorough petrographic inspec-tion Fifty-nine of the 129 collected samples were crushed(400 g) into small pieces 52mm in diameter in aRocklabs hydraulic piston crusher between WC plates tominimize contamination The coarse-crush was mixedand 100 g was powdered in a planetary mill using agatejars and balls cleaned with quartz sand between samples
ActLabs analytical methodsThe major- and trace-element compositions of the whole-rock powders were determined at Activation LaboratoriesLtd (Actlabs) in Ancaster Ontario Analytical techniquesand detection limits are also available from Actlabs (httpwwwactlabscommethsub_code4erehtm) The analyticalmethod for each of the elements analyzed is indicated inTable 2 For the major elements a 02 g sample was mixedwith a lithium metaborate^lithium tetraborate mixtureand fused in a graphite crucible The molten mixture waspoured into a 5 HNO3 solution and shaken until dis-solved (30min) The samples were analyzed for majoroxides and selected trace elements by inductively coupledplasma optical emission spectrometry (ICP-OES) on acombination simultaneous^sequential Thermo Jarrell-AshEnviro II system Internal calibration was achieved usinga variety of international reference materials (eg W-2BIR-1 DNC-1) and independent control samplesAdditional trace elements were analyzed by both instru-mental neutron activation analysis (INAA) and induc-tively couple plasma mass spectrometry (ICP-MS) ForINAA15^25 g of sample was weighed into small polyeth-ylene vials and irradiated with a control international ref-erence material CANMET WMS-1 and NiCr flux wires ata thermal neutron flux of 71012 neutronscm2 per s in theMcMaster Nuclear Reactor Following a 7 day waitingperiod the samples were measured on an Ortec high-purity Ge detector linked to a Canberra Series 95 multi-channel analyzer Activities for each element weredecay- and weight-corrected and compared with a detectorcalibration developed from multiple international certifiedreference materials For ICP-MS 025 g of sample wasdigested in HF followed by HNO3^HClO4 heated andtaken to dryness The samples were brought back into solu-tion with HCl Samples were analyzed using a PerkinElmer Optima 3000 ICP In-lab standards or certified ref-erence materials (eg W-2 BIR-1 DNC-1) were used for
quality control A total of 15 blind duplicates were ana-lyzed from the same sample powders to assessreproducibility
University of Massachusetts analyticalmethodsFourteen sample duplicate powders taken from the same100 g sample aliquot (high-MgO basalts and picrites)were also analyzed at the Ronald B Gilmore X-RayFluorescence (XRF) Laboratory at the University ofMassachusetts Major elements were measured on a fusedLa-bearing lithium borate glass disc using a SiemensMRS-400 spectrometer Trace element concentrations(Rb Sr Ba Ce Nb Zr Y Pb Zn Ga Ni Cr V) weremeasured on a separate powder pellet using a PhilipsPW2400 sequential spectrometer using a Rh tube LOIand ferrous iron measurements were made as describedby Rhodes amp Vollinger (2004) Precision and accuracy esti-mates for the data have been described by Rhodes (1996)and Rhodes amp Vollinger (2004) Results for each sampleare the average of two separate analyses and are shown inElectronic Appendix 2 for a comparison of analyses ofduplicate powders using XRFand ICP-MS
PCIGR analytical methodsA subset of 19 samples was selected for high-precisiontrace-element analysis and Sr Nd Pb and Hf isotopicanalysis at the Pacific Centre for Isotopic andGeochemical Research (PCIGR) at the University ofBritish Columbia (UBC) Samples were selected from the59 samples analyzed for whole-rock chemistry at ActLabsbased on major- and trace-element chemistry alteration(low LOI and petrographic alteration index) and samplelocation Samples were prepared for trace-element analysisat the PCIGR by the technique described by Pretorius et al(2006) on unleached rock powders Sample powders(100mg) were weighed in 7ml screw-top Savillex bea-kers and dissolved in 1ml 14N HNO3 and 5ml 48 HFon a hotplate for 48 h at 1308C with periodic ultrasonica-tion Samples were dried and redissolved in 6ml 6N HClon a hotplate for 24 h and then dried and redissolved in1ml concentrated HNO3 for 24 h before final dryingTrace element abundances were measured by high-resolution inductively coupled plasma-mass spectrometry(HR-ICP-MS) with aThermo Finnigan Element2 systemfollowing the procedures described by Pretorius et al(2006) within 24 h of redissolution HFSE and LILE weremeasured in medium-resolution mode at 2000 dilutionusing a PFATeflon spray chamber washed with aqua regiafor 3min between samples REE were measured in high-resolution mode and U and Pb in low-resolution modeat 2000 dilution using a glass spray chamber washedwith 2 HNO3 between samples Total procedural blanksand reference materials (BCR-2 BHVO-2) were analyzedwith the batch of samples Indium was used as an internal
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
38
standard in all samples and standard solutionsBackground and standard solutions were analyzed afterevery five samples to detect memory effects and massdrift Results for PCIGR trace-element analyses areshown in Electronic Appendix 3 for the full set of analyzedtrace elementsSample digestion for purification of Sr Nd Hf and Pb
for column chemistry began by weighing each samplepowder (400^500mg) prior to leaching All samples wereleached with 6N HCl and placed in an ultrasonic bath for15min Samples were rinsed twice with 18 M-cm H2Obetween each leaching step (15 total) until the supernatantliquid was clear [following the method of Mahoney (1987)]Samples were then dried on a hotplate for 24 h andweighed again Sample solutions were then prepared bydissolving 100^250mg of the leached powder dissolvedin 1ml 14N HNO3 and 10ml 48 HF on a hotplate for48 h at 1308C with periodic ultrasonication Samples weredried and redissolved in 6ml 6N HCl on a hotplate for24 h and then dried Pb was separated using anion exhangecolumns and the discard was used for Sr REE and Hfseparation Nd was separated from the other REE and Hfrequired two additional purification steps Detailed proce-dures for column chemistry for separating Sr Nd and Pbat the PCIGR have been described byWeis et al (2006) andHf purification has been described byWeis et al (2007) Srand Nd isotope ratios were measured by thermal ioniza-tion mass spectrometry (TIMS) on a Thermo FinniganTriton system in static mode with relay matrix rotation ona single Ta and double Re^Ta filament respectively Fourto five filaments per barrel of 21 were occupied by stan-dards (NBS 987 for Sr and LaJolla for Nd) for each barrelwhere samples were run Sample Sr and Nd isotopic com-positions were corrected for mass fractionation using86Sr88Srfrac14 01194 and 146Nd144Ndfrac14 07219 Each samplewas then normalized using the barrel average of the refer-ence material relative to the values of143Nd144Ndfrac14 0511858 and 87Sr86Srfrac14 0710248 (Weiset al 2006) During the course of the Vancouver Islandanalyses the La Jolla Nd standard gave an average valueof 05118566 (nfrac14 7) and NBS987 standard gave an aver-age of 07102408 (nfrac1411) (2 error is reported as times106) 147Sm144Nd errors are 15 or 0006 USGeological Survey (USGS) reference material BHVO-2was processed with the samples and yielded Sr and Nd iso-topic ratios of 07034607 and 05129786 respectivelyThese are in agreement with the published values of070347920 and 051298411 respectively (Weis et al2006)
Pb and Hf isotopic compositions were analyzed by mul-tiple collector-inductively coupled plasma-mass spectrome-try (MC-ICP-MS) on a static multi-collection on a NuPlasma (Nu Instruments) system The detailed analyticalprocedure for Pb isotopic analyses on the Nu Plasma atthe PCIGR has been described by Weis et al (2006) Theconfiguration for Pb analyses allows for collection of PbTl and Hg togetherTl and Hg are used to monitor instru-mental mass discrimination and isobaric overlap repec-tively All sample solutions were analyzed withapproximately the same PbTl ratio (4) as the referencematerial NIST SRM 981 To accomplish this a small ali-quot of each sample solution from the Pb columns was ana-lyzed on the Element2 to determine the precise amount ofPb available for analysis on the Nu Plasma The SRM 981standard was run after every two samples on the NuPlasma During the time samples were run analyses ofthe SRM 981 Pb reference material gave values of206Pb204Pbfrac1416940322 207Pb204Pbfrac1415495823and 208Pb204Pbfrac14 36713164 (nfrac14 61 2 error is reportedas times 104) these values are in excellent agreement withreported TIMS triple-spike values of Galer amp Abouchami(1998) Fractionation-corrected Pb isotopic ratios were fur-ther corrected by the sample-standard bracketing methodor the ln^ln correction method described by White et al(2000) and Blichert-Toft et al (2003) Leached powders ofUSGS reference material BHVO-2 yielded Pb isotopicratios of 206Pb204Pbfrac14186454 8 207Pb204Pbfrac14 1549105 and 208Pb204Pbfrac14 38222514 and BCR-2yielded 206Pb204Pbfrac141880466 207Pb204Pbfrac14 1562518 and 208Pb204Pbfrac14 3883496 (2 error isreported as times 104) These values are in agreement withthose for leached residues of BHVO-2 and BCR-2 fromWeis et al (2006)Hf isotopic compositions were analyzed following the
procedures detailed byWeis et al (2007) The configurationfor Hf analyses monitored Lu mass 175 andYb mass 172 toallow for interference correction to masses 174 and 176 Hfisotopic ratios were normalized internally for mass fractio-nation to a 179Hf177Hf ratio of 07325 using an exponentialcorrection Standards were run after every two samplesand sample results were normalized to the ratio of the in-run daily average and a 176Hf177Hf ratio for JMC-475 of0282160 During the course of analyses the Hf standardJMC-475 gave an average value 02821679 (nfrac14130)USGS reference materials BCR-2 and BHVO-2 were pro-cessed with the samples and yielded Hf isotopic ratios of02828675 and 02831005 respectively Publishedvalues for BCR-2 and BHVO-2 are 02828717 and0283104 8 respectively (Weis et al 2007)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
39
picritic and high-MgO basalt pillow lavas extend to20wt MgO (Fig 6 and references listed in thecaption)The tholeiitic basalts from the Karmutsen Range dis-
play a tight range of parallel light rare earth element(LREE)-enriched REE patterns (LaYbCNfrac14 21^24mean 2202) whereas the picrites and high-MgObasalts are characterized by sub-parallel LREE-depleted
patterns (LaYbCNfrac14 03^07 mean 06 02) with lowerREE abundances than the tholeiitic basalts (Fig 7)Tholeiitic basalts from the Schoen Lake and MountArrowsmith areas have similar REE patterns tosamples from the Karmutsen Range (Schoen Lake LaYbCNfrac14 21^26 mean 2303 Mount Arrowsmith LaYbCNfrac1419^25 mean 2204) and the coarse-grainedmafic rocks have similar REE patterns to the tholeiitic
Depleted MORB average(Salters amp Stracke 2004)
Schoen LakeSchoen Lake
Mount Arrowsmith
Karmutsen RangeS
ampl
eC
hond
rite
Sam
ple
Cho
ndrit
e
Sam
ple
Prim
itive
Man
tleS
ampl
eP
rimiti
ve M
antle
Sam
ple
Prim
itive
Man
tle
Mount Arrowsmith
Karmutsen Range
(a) (b)
(c) (d)
(e) (f )
Sam
ple
Cho
ndrit
e
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained mafic rock
Pillowed flowMineralized sill
1
10
100
1
10
100
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
01
1
10
100
1
10
100
1
10
100
Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Cs RbBa Th U K Nb Ta La Ce Pr Nd Sr Sm Zr Hf Ti Eu Gd Tb Dy Ho Y Er Yb Lu
Fig 7 Whole-rock REE and other incompatible-element concentrations for the Karmutsen Formation (a) (c) and (e) are chondrite-normal-ized REE patterns for samples from each of the three main field areas onVancouver Island (b) (d) and (f) are primitive mantle-normalizedtrace-element patterns for samples from each of the three main field areas All normalization values are from McDonough amp Sun (1995) Theclear distinction between LREE-enriched tholeiitic basalts and LREE-depleted picrites the effects of alteration on the LILE (especially loss ofK and Rb) and the different range for the vertical scale in (b) (extends down to 01) should be noted
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
18
basalts (LaYbCNfrac14 21^29 mean 2508) Two LREE-depleted coarse-grained mafic rocks from the SchoenLake area have similar REE patterns (LaYbCNfrac14 07) tothe picrites from the Keogh Lake area indicating that thisdistinctive suite of rocks is exposed as far south asSchoen Lake that is 75 km away (Fig 7) The anomalouspillowed flow from the Karmutsen Range has lower LaYbCN (13^18) than the main group of tholeiitic basaltsfrom the Karmutsen Range (Fig 7) The mineralized sillfrom the Schoen Lake area is distinctly LREE-enriched(LaYbCNfrac14 33) with the highest REE abundances(LaCNfrac14 940) of all Karmutsen samples this may reflectcontamination by adjacent sediment also evident in thin-section where sulfide blebs are cored by quartz (Greeneet al 2006)The primitive mantle-normalized trace-element pat-
terns (Fig 7) and trace-element variations and ratios(Fig 8) highlight the differences in trace-element
concentrations between the four main groups ofKarmutsen flood basalts onVancouver Island The tholeii-tic basalts from all three areas have relatively smooth par-allel trace-element patterns some samples have smallpositive Sr anomalies relative to Nd and Sm and manysamples have prominent negative K anomalies relative toU and Nb reflecting K loss during alteration (Fig 7) Thelarge ion lithophile elements (LILE) in the tholeiiticbasalts (mainly Rb and K not Cs) are depleted relativeto the high field strength elements (HFSE) and LREELILE segments especially for the Schoen Lake area(Fig 7d) are remarkably parallel The picrites andhigh-MgO basalts form a tight band of parallel trace-element patterns and are depleted in HFSE and LREEwith positive Sr anomalies and relatively low Rb Thecoarse-grained mafic rocks from the Karmutsen Rangehave identical trace-element patterns to the tholeiiticbasalts Samples from the Schoen Lake area have similar
000
025
050
075
100
125
0 1 2 3 4 5
00
04
08
12
16
0 5 10 15 20 25
MgO (wt)
0
5
10
15
0 50 100 150
Hf (ppm)
Th (ppm)
NbLaNb (ppm)
Zr (ppm)
(a)
(c)
(b)
(d)
Th (ppm)
00
02
04
06
08
10
12
00 01 02 03 04 05
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
NbLa=1
4723A4 5615A12
U (ppm)
Fig 8 Whole-rock trace-element concentrations and ratios for the Karmutsen Formation [except (b) which is vs wt MgO] (a) Nb vs Zr(b) NbLa vs MgO (c) Th vs Hf (d) Th vs U There is a clear distinction between the tholeiitic basalts and picrites in Nb and Zr both inconcentration and the slope of each trend
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
19
trace-element patterns to the Karmutsen Range except forthe two LREE-depleted coarse-grained mafic rocks andthe mineralized sill (Fig 7d)
Sr^Nd^Hf^Pb isotopic compositionsA total of 19 samples from the four main groups ofthe Karmutsen Formation were selected for radiogenicisotopic geochemistry on the basis of major- and trace-element variation stratigraphic position and geographicallocation The samples have overlapping age-correctedHf and Nd isotope ratios and distinct ranges of Sr isotopiccompositions (Fig 9) The tholeiitic basalts picriteshigh-MgO basalt and coarse-grained mafic rocks have
initial eHffrac14thorn87 to thorn126 and eNdfrac14thorn66 to thorn88 cor-rected for in situ radioactive decay since 230 Ma (Fig 9Tables 3 and 4) All the Karmutsen samples form a well-defined linear array in a Lu^Hf isochron diagram corre-sponding to an age of 24115 Ma (Fig 9) This is withinerror of the accepted age of the Karmutsen Formationand indicates that the Hf isotope systematics behaved as aclosed system since c 230 Ma The anomalous pillowedflow has similar initial eHf (thorn98) and slightly lower initialeNd (thorn62) compared with the other Karmutsen samplesThe picrites and high-MgO basalt have higher initial87Sr86Sr (070398^070518) than the tholeiitic basalts(initial 87Sr86Srfrac14 070306^070381) and the coarse-grained mafic rocks (initial 87Sr86Srfrac14 070265^070428)
05128
05129
05130
05131
05132
015 017 019 021 023 025
02829
02830
02831
02832
02833
02834
000 002 004 006 008 010
4
6
8
10
0702 0703 0704 0705 07066
8
10
12
14
5 6 7 8 9 10
12
0
1
2
3
4
5
6
7
0702 0703 0704 0705 0706
(d)
LOI (wt )
87Sr 86Sr
4723A2
Nd
143Nd144Nd
147Sm144Nd
87Sr 86Sr (230 Ma)
(230 Ma)
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
(a)
(c)
176Hf177Hf
Age=241 plusmn 15 Ma=+990 plusmn 064
176Lu177Hf
Hf (230 Ma)
Nd (230 Ma)
(e)
Hf (i)
(b)
Fig 9 Whole-rock Sr Nd and Hf isotopic compositions for the Karmutsen Formation (a) 143Nd144Nd vs 147Sm144Nd (b) 176Hf177Hf vs176Lu177Hf The slope of the best-fit line for all samples corresponds to an age of 24115 Ma (c) Initial eNd vs 87Sr86Sr Age-correction to230 Ma (d) LOI vs 87Sr86Sr (e) Initial eHf vs eNd Average 2 error bars are shown in a corner of each panel Complete chemical duplicatesshown inTables 3 and 4 (samples 4720A7 and 4722A4) are circled in each plot
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
overlap the ranges of the picrites and tholeiitic basalts(Fig 9 Table 3)The measured Pb isotopic compositions of the tholeiitic
basalts are more radiogenic than those of the picrites andthe most magnesian Keogh Lake picrites have the leastradiogenic Pb isotopic compositions The range ofinitial Pb isotope ratios for the picrites is206Pb204Pbfrac1417868^18580 207Pb204Pbfrac1415547^15562and 208Pb204Pbfrac14 37873^38257 and the range for thetholeiitic basalts is 206Pb204Pbfrac1418782^19098207Pb204Pbfrac1415570^15584 and 208Pb204Pbfrac14 37312^38587 (Fig 10 Table 5) The coarse-grained mafic rocksoverlap the range of initial Pb isotopic compositions forthe tholeiitic basalts with 206Pb204Pbfrac1418652^19155207Pb204Pbfrac1415568^15588 and 208Pb204Pbfrac14 38153^38735 One high-MgO basalt has the highest initial206Pb204Pb (19242) and 207Pb204Pb (15590) and thelowest initial 208Pb204Pb (37676) The Pb isotopic ratiosfor Karmutsen samples define broadly linearrelationships in Pb isotope plots The age-corrected Pb
isotopic compositions of several picrites have been affectedby U and Pb (andor Th) mobility during alteration(Fig 10)
ALTERAT IONThe Karmutsen basalts have retained most of their origi-nal igneous structures and textures however secondaryalteration and low-grade metamorphism have producedzeolitic- and prehnite^pumpellyite-bearing mineral assem-blages (Table 1 Cho et al 1987) Alteration and metamor-phism have primarily affected the distribution of the LILEand the Sr isotopic systematics The Keogh Lake picriteshave been affected more by alteration than the tholeiiticbasalts and have higher LOI (up to 55wt ) variableLILE and higher measured Sr Nd and Hf and lowermeasured Pb isotope ratios than the tholeiitic basalts Srand Pb isotope compositions for picrites and tholeiiticbasalts show a relationship with LOI (Figs 9 and 10)whereas Nd and Hf isotopic compositions do not correlate
Table 3 Sr and Nd isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Rb Sr 87Sr86Sr 2m87Rb86Sr 87Sr86Srt Sm Nd 143Nd144Nd 2m eNd
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses carried out at the PCIGR thecomplete trace element analyses are given in Electronic Appendix 3
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
21
with LOI (not shown) The relatively high apparent initialSr isotopic compositions for the high-MgO lavas (up to07052) probably resulted from post-magmatic loss of Rbor an increase in 87Sr86Sr through addition of seawater Sr(eg Hauff et al 2003) High-MgO lavas from theCaribbean plateau have similarly high 87Sr86Sr comparedwith basalts (Recurren villon et al 2002)The correction for in situdecay on initial Pb isotopic ratios has been affected bymobilization of U and Th (and Pb) in the whole-rockssince their formation A thorough acid leaching duringsample preparation was used that has been shown to effec-tively remove alteration phases (Weis et al 2006 NobreSilva et al in preparation) Whole-rock SmNd ratiosand age-correction of Nd isotopes may be affected bythe leaching procedure for submarine rocks older than50 Ma (Thompson et al 2008 Fig 9) there ishowever very little variation in Nd isotopic compositionsof the picrites or tholeiitic basalts and this system does notappear to be disturbed by alteration The HFSE abun-dances for tholeiitic and picritic basalts exhibit clear
linear correlations in binary diagrams (Fig 8) whereasplots of LILE vs HFSE are highly scattered (not shown)because of the mobility of some of the alkali and alkalineearth LILE (Cs Rb Ba K) and Sr for most samplesduring alterationThe degree of alteration in the Karmutsen samples does
not appear to be related to eruption environment or depthof burial Pillow rims are compositionally different frompillow cores (Surdam1967 Kuniyoshi1972) and aquagenetuffs are chemically different from isolated pillows withinpillow breccias however submarine and subaerial basaltsexhibit a similar degree of alterationThere is no clear cor-relation between the submarine and subaerial basalts andsome commonly used chemical alteration indices (eg BaRb vs K2OP2O5 Huang amp Frey 2005) There is also nodefinitive correlation between the petrographic alterationindex (Table 1) and chemical alteration indices although10 of 13 tholeiitic basalts with the highest K and LILEabundances have the highest petrographic alterationindex of three (seeTable 1)
Table 4 Hf isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Lu Hf 176Hf177Hf 2m eHf176Lu177Hf 176Hf177Hft eHf(t)
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
OLIV INE ACCUMULAT ION INH IGH-MgO AND PICR IT IC LAVASGeochemical trends and petrographic characteristics indi-cate that accumulation of olivine played an important rolein the formation of the high-MgO basalts and picrites fromthe Keogh Lake area on northern Vancouver Island TheKeogh Lake samples show a strong linear correlation inplots of Al2O3 TiO2 ScYb and Ni vs MgO and many ofthe picrites have abundant clusters of olivine pseudomorphs(Table1 Fig11)There is a strong linear correlationbetween
modal per cent olivine and whole-rock magnesium contentmagnesium contents range between 10 and 19wt MgOand the proportion of olivine phenocrysts in these samplesvaries from 0 to 42 vol (Fig 11)With a few exceptionsboth the size range and median size of the olivine pheno-crysts in most samples are comparable The clear correla-tion between the proportion of olivine phenocrysts andwhole-rock magnesium contents indicates that accumula-tion of olivine was directly responsible for the high MgOcontents (410wt ) in most of the picritic lavas
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
207Pb204Pb
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
208Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
(a) (b)
(c) (e)
4723A24723A2
4723A2
4723A2
4722A4
4723A4
4722A4
4723A4
4723A134723A3
4723A3
4723A13
1556
1558
1560
1562
1564
1566
1568
186 190 194 198 202 206 2101550
1552
1554
1556
1558
1560
178 180 182 184 186 188 190 192 194
380
384
388
392
396
186 190 194 198 202 206 210374
378
382
386
390
178 180 182 184 186 188 190 192 194
206Pb204Pb(d)
LOI (wt )
0
1
2
3
4
5
6
7
186 190 194 198 202 206 210
4723A2
Fig 10 Pb isotopic compositions of leached whole-rock samples measured by MC-ICP-MS for the Karmutsen Formation Error bars are smal-ler than symbols (a) Measured 207Pb204Pb vs 206Pb204Pb (b) Initial 207Pb204Pb vs 206Pb204Pb Age-correction to 230 Ma (c) Measured208Pb204Pb vs 206Pb204Pb (d) LOI vs 206Pb204Pb (e) Initial 208Pb204Pb vs 206Pb204Pb Complete chemical duplicates shown in Table 5(samples 4720A7 and 4722A4) are circled in each plot The dashed lines in (b) and (e) show the differences in age-corrections for two picrites(4723A4 4722A4) using the measured UTh and Pb concentrations for each sample and the age-corrections when concentrations are used fromthe two picrites (4723A3 4723A13) that appear to be least affected by alteration
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
23
DISCUSS IONThe compositional and spatial development of the eruptedlava sequences of the Wrangellia oceanic plateau onVancouver Island provide information about the evolutionof a mantle plume upwelling beneath oceanic lithospherein an analogous way to the interpretation of continentalflood basalts The Caribbean and Ontong Java plateausare the two primary examples of accreted parts of oceanicplateaus that have been studied to date and comparisonscan be made with the Wrangellia oceanic plateauHowever the Wrangellia oceanic plateau is a distinctiveoccurrence of an oceanic plateau because it formed on thecrust of a Devonian to Mississippian intra-oceanic islandarc overlain by Mississippian to Permian limestones andpelagic sediments The nature and thickness of oceaniclithospheric mantle are considered to play a fundamentalrole in the decompression and evolution of a mantleplume and thus the compositions of the erupted lavasequences (Greene et al 2008) The geochemical and
stratigraphic relationships observed in the KarmutsenFormation onVancouver Island and the focus of the subse-quent discussion sections provide constraints on (1) thetemperature and degree of melting required to generatethe primary picritic magmas (2) the composition of themantle source (3) the depth of melting and residual min-eralogy in the source region (4) the low-pressure evolutionof the Karmutsen tholeiitic basalts (5) the growth of theWrangellia oceanic plateau
Melting conditions and major-elementcomposition of the primary magmasThe primary magmas to most flood basalt provinces arebelieved to be picritic (eg Cox 1980) and they have beenfound in many continental and oceanic flood basaltsequences worldwide (eg Siberia Karoo Paranacurren ^Etendeka Caribbean Deccan Saunders 2005) Near-pri-mary picritic lavas with low total alkali abundances
Table 5 Pb isotopic compositions of Karmutsen basaltsVancouver Island BC
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
require high-degree partial melting of the mantle source(eg Herzberg amp OrsquoHara 2002) The Keogh Lake picritesfrom northernVancouver Island are the best candidates foridentifying the least modified partial melts of the mantleplume source despite having accumulated olivine pheno-crysts and can be used to estimate conditions of meltingand the composition of the primary magmas for theKarmutsen FormationPrimary magma compositions mantle potential tem-
peratures and source melt fractions for the picritic lavas ofthe Karmutsen Formation were calculated from primitivewhole-rock compositions using PRIMELT1XLS software(Herzberg et al 2007) and are presented in Fig 12 andTable 6 A detailed discussion of the computationalmethod has been given by Herzberg amp OrsquoHara (2002)and Herzberg et al (2007) key features of the approachare outlined belowPrimary magma composition is calculated for a primi-
tive lava by incremental addition of olivine and compari-son with primary melts of mantle peridotite Lavas thathad experienced plagioclase andor clinopyroxene fractio-nation are excluded from this analysis All calculated pri-mary magma compositions are assumed to be derived byaccumulated fractional melting of fertile peridotite KR-4003 Uncertainties in fertile peridotite composition donot propagate to significant variations in melt fractionand mantle potential temperature melting of depletedperidotite propagates to calculated melt fractions that aretoo high but with a negligible error in mantle potential
temperature PRIMELT1XLS calculates the olivine liqui-dus temperature at 1atm which should be similar to theactual eruption temperature of the primary magmaand is accurate to 308C at the 2 level of confidenceMantle potential temperature is model dependent witha precision that is similar to the accuracy of the olivineliquidus temperature (C Herzberg personal communica-tion 2008)With regard to the evidence for accumulated olivine in
the Keogh Lake picrites it should not matter whether themodeled lava composition is aphyric or laden with accu-mulated olivine phenocrysts PRIMELT1 computes theprimary magma composition by adding olivine to a lavathat has experienced olivine fractionation in this way itinverts the effects of fractional crystallization The onlyrequirement is that the olivine phenocrysts are not acci-dental or exotic to the lava compositions Support for thisprocedure can be seen in the calculated olivine liquid-line-of-descent shown by the curved arrays with 5 olivineaddition increments (Fig 12c) Fractional crystallization ofolivine from model primary magmas having 85^95wt FeO and 153^175wt MgO will yield arrays of deriva-tive liquids and randomly sorted olivines that in most butnot all cases connect the picrites and high-MgO basalts Anewer version of the PRIMELT software (Herzberg ampAsimow 2008) can perform olivine addition to and sub-traction from lavas with highly variably olivine phenocrystcontents and yields results that converge with those shownin Fig 12
Fig 11 Relationship between abundance of olivine phenocrysts and whole-rock MgO contents for the Keogh Lake picrites (a) Tracings ofolivine grains (gray) in six high-MgO samples made from high resolution (2400 dpi) scans of petrographic thin-sections Scale bars represent10mm sample numbers are indicated at the upper left (b) Modal olivine vs whole-rock MgO Continuous line is the best-fit line for 10 high-MgO samplesThe areal proportion of olivine was calculated from raster images of olivine grains using ImageJ image analysis software whichprovides an acceptable estimation of the modal abundance (eg Chayes 1954)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
25
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
Gray lines = final melting pressure
L + Ol + Opx
07
08
09
Pressure (GPa)
10
3
4
5
6
1
SOLIDUS
01
04
0506
L + Ol
2
03
02
PeridotiteKR-4003
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
Thick black lines = initial melting pressure
Dashed lines = melt fraction
Melt Fraction
2
3 4 5 6
L+Ol+Opx
SolidusSpinel
SolidusGarnet Peridotite
7
L+Ol
0 10 20 30 4040
45
50
55
60
MgO (wt)
SiO2
(wt)
PeridotiteKR-4003
Melt Fraction
0 10 20 305
10
15
CaO(wt)
MgO (wt)
3
4 5 6 7
2
46
2
Peridotite Partial Melts
Thick black line = initial melting pressureGray lines = final melting pressure
Peridotite
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
02
Pressure (GPa)
0302
04
Pressure (GPa)
Picrite
High-MgO basalt
Tholeiitic basalt
(c)
(d)
Karmutsen flood basaltsOlivine addition modelOlivine addition (5 increment)Primary magma
PicriteHigh-MgO basaltTholeiitic basalt
(a)
(b)
800 850 900
910
920
930
940
950
Karmutsen flood basalts
Olivine addition modelOlivine addition (5 increment)Primary magma
Gray lines = Fo contents of olivine
Fig 12 Estimated primary magma compositions for three Keogh Lake picrites and high-MgO basalts (samples 4723A4 4723A135616A7) using the forward and inverse modeling technique of Herzberg et al (2007) Karmutsen compositions and modeling results areoverlain on diagrams provided by C Herzberg (a) Whole-rock FeO vs MgO for Karmutsen samples from this study Total iron estimatedto be FeO is 090 Gray lines show olivine compositions that would precipitate from liquid of a given MgO^FeO composition Black lineswith crosses show results of olivine addition (inverse model) using PRIMELT1 (Herzberg et al 2007) Gray field highlights olivinecompositions with Fo contents of 91^92 predicted to precipitate (b) FeO vs MgO showing Karmutsen lava compositions and results ofa forward model for accumulated fractional melting of fertile peridotite KR-4003 (Kettle River Peridotite Walter 1998) (c) SiO2 vsMgO for Karmtusen lavas and model results (d) CaO vs MgO showing Karmtusen lava compositions and model results To brieflysummarize the technique [see Herzberg et al (2007) for complete description] potential parental magma compositions for the high-MgOlava series were selected (highest MgO and appropriate CaO) and using PRIMELT1 software olivine was incrementally added tothe selected compositions to show an array of potential primary magma compositions (inverse model) Then using PRIMELT1 theresults from the inverse model were compared with a range of accumulated fractional melts for fertile peridotite derived fromparameterization of the experimental results for KR-4003 from Walter (1998) forward model Herzberg amp OrsquoHara (2002) A melt fractionwas sought that was unique to both the inverse and forward models (Herzberg et al 2007) A unique solution was found when there was a com-mon melt fraction for both models in FeO^MgO and CaO^MgO^Al2O3^SiO2 (CMAS) projection space This modeling assumes thatolivine was the only phase crystallizing and ignores chromite precipitation and possible augite fractionation in the mantle (Herzbergamp OrsquoHara 2002) Results are best for a residue of spinel lherzolite (not pyroxenite) The presence of accumulated olivine in samples ofthe Keogh Lake picrites used as starting compositions does not significantly affect the results because we are modeling addition of olivine(see text for further description) The tholeiitic basalts cannot be used for modeling because they are all saturated in plagthorn cpxthornolThick black line for initial melting pressure at 4 GPa is not shown in (c) for clarity Gray area in (a) indicates parental magmas thatwill crystallize olivine with Fo 91^92 at the surface Dashed lines in (c) indicate melt fraction Solidi for spinel and garnet peridotite arelabelled in (c)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
The PRIMELT1 results indicate that if the Karmutsenpicritic lavas were derived from accumulated fractionalmelting of fertile peridotite the primary magmas wouldhave contained 15^17wt MgO and 10wt CaO(Fig 12 Table 6) formed from 23^27 partial meltingand would have crystallized olivine with a Fo content of 91 (Fig 12 Table 6) Calculated mantle temperatures forthe Karmutsen picrites (14908C) indicate melting ofanomalously hot mantle (100^2008C above ambient) com-pared with ambient mantle that produces mid-ocean ridgebasalt (MORB 1280^14008C Herzberg et al 2007)which initiated between 3 and 4 GPa Several picritesappear to be near-primary melts and required very littleaddition of olivine (eg sample 93G171 with measured158wt MgO) These temperature and melting esti-mates of the Karmutsen picrites are consistent withdecompression of hot mantle peridotite in an actively con-vecting plume head (ie plume initiation model) Threesamples (picrite 5614A1 and high-MgO basalts 5614A3and 5614A5) are lower in iron than the other Karmutsenlavas with FeO in the 65^80wt range (Fig 12c) andhave MgO and FeO contents that are similar to primitiveMORB from the Sequeiros fracture zone of the EastPacific Rise (EPR Perfit et al 1996) These samples
however differ from EPR MORB in having higher SiO2
and lower Al2O3 and they are restricted geographicallyto one area (Sara Lake Fig 2)We therefore have evidence for primary magmas with
MgO contents that range from a possible low of 110wt to as high as 174wt with inferred mantle potential tem-peratures from 1340 to 15208C This is similar to the rangedisplayed by lavas from the Ontong Java Plateau deter-mined by Herzberg (2004) who found that primarymagma compositions assuming a peridotite sourcewould contain 17wt MgO and 10wt CaO(Table 6) and that these magmas would have formed by27 melting at a mantle potential temperature of15008C (first melting occurring at 36 GPa) Fitton ampGodard (2004) estimated 30 partial melting for theOntong Java lavas based on the Zr contents of primarymagmas of Kroenke-type basalt calculated by incrementaladdition of equilibrium olivine However if a considerableamount of eclogite was involved in the formation of theOntong Java plateau magmas the excess temperatures esti-mated by Herzberg et al (2007) may not be required(Korenaga 2005) The variability in mantle potential tem-perature for the Keogh Lake picrites is also similar to thewide range of temperatures that have been calculated for
Table 6 Estimated primary magma compositions for Karmutsen basalts and other oceanic plateaus or islands
Sample 4723A4 4723A13 5616A7 93G171 Average OJP Mauna Keay Gorgonaz
Ontong Java primary magma composition for accumulated fraction melting (AFM) from Herzberg (2004)yMauna Kea primary magma composition is average of four samples of Herzberg (2006 table 1)zGorgona primary magma composition for AFM for 1F fertile source from Herzberg amp OrsquoHara (2002 table 4)Total iron estimated to be FeO is 090
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
27
lavas from single volcanoes in the Galapagos Hawaii andelsewhere by Herzberg amp Asimow (2008) Those workersinterpreted the range to reflect the transport of magmasfrom both the cool periphery and the hotter axis of aplume A thermally heterogeneous plume will yield pri-mary magmas with different MgO contents and the rangeof primary magma compositions and their inferred mantlepotential temperatures for Karmutsen lavas may reflectmelting from different parts of the plume
Source of the Karmutsen Formation lavasThe Sr^Nd^Hf^Pb isotopic compositions of theKarmutsen volcanic rocks on Vancouver Island indicatethat they formed from plume-type mantle containing adepleted component that is compositionally similar to butdistinct from other oceanic plateaus that formed in thePacific Ocean (eg Ontong Java and Caribbean Fig 13)Trace element and isotopic constraints can be used to testwhether the depleted component of the KarmutsenFormation was intrinsic to the mantle plume and distinctfrom the source of MORB or the result of mixing or invol-vement of ambient depleted MORB mantle with plume-type mantle The Karmutsen tholeiitic basalts have initialeNd and
87Sr86Sr ratios that fall within the range of basaltsfrom the Caribbean Plateau (eg Kerr et al 1996 1997Hauff et al 2000 Kerr 2003) and a range of Hawaiian iso-tope data (eg Frey et al 2005) and lie within and abovethe range for the Ontong Java Plateau (Tejada et al 2004Fig 13a) Karmutsen tholeiitic basalts with the highest ini-tial eNd and lowest initial 87Sr86Sr lie within and justbelow the field for age-corrected northern EPR MORB at230 Ma (eg Niu et al1999 Regelous et al1999) Initial Hfand Nd isotopic compositions for the Karmutsen samplesare noticeably offset to the low side of the ocean islandbasalt (OIB) array (Vervoort et al 1999) and lie belowand slightly overlap the low eHf end of fields for Hawaiiand the Caribbean Plateau (Fig 13c) the Karmutsen sam-ples mostly fall between and slightly overlap a field forage-corrected Pacific MORB at 230 Ma and Ontong Java(Mahoney et al 1992 1994 Nowell et al 1998 Chauvel ampBlichert-Toft 2001) Karmutsen lava Pb isotope ratios over-lap the broad range for the Caribbean Plateau and thelinear trends for the Karmutsen samples intersect thefields for EPR MORB Hawaii and Ontong Java in208Pb^206Pb space but do not intersect these fields in207Pb^206Pb space (Fig 13d and e) The more radiogenic207Pb204Pb for a given 206Pb204Pb of the Karmutsenbasalts requires high UPb ratios at an ancient time when235U was abundant similar to the Caribbean Plateau andenriched in comparison with EPR MORB Hawaii andOntong JavaThe low eHf^low eNd component of the Karmutsen
basalts that places them below the OIB mantle array andpartly within the field for age-corrected Pacific MORBcan form from low ratios of LuHf and high SmNd over
time (eg Salters amp White 1998) MORB will develop aHf^Nd isotopic signature over time that falls below themantle array Low LuHf can also lead to low 176Hf177Hffrom remelting of residues produced by melting in theabsence of garnet (eg Salters amp Hart 1991 Salters ampWhite 1998) The Hf^Nd isotopic compositions of theKarmutsen Formation indicate the possible involvementof depleted MORB mantle however similar to Iceland(Fitton et al 2003) Nb^Zr^Y variations provide a way todistinguish between a depleted component within theplume and a MORB-like component The Karmutsenbasalts and picrites fall within the Iceland array and donot overlap the MORB field in a logarithmic plot of NbYvs ZrY (Fig 13b) using the fields established by Fittonet al (1997) Similar to the depleted component within theCaribbean Plateau (Thompson et al 2003) the trend ofHf^Nd isotopic compositions towards Pacific MORB-likecompositions is not followed by MORB values in Nb^Zr^Y systematics and this suggests that the depleted compo-nent in the source of the Karmutsen basalts is distinctfrom the source of MORB and probably an intrinsic partof the plume The relatively radiogenic Pb isotopic ratiosand REE patterns distinct from MORB also support thisinterpretationTrace-element characteristics suggest the possible invol-
vement of sub-arc lithospheric mantle in the generation ofthe Karmutsen picrites Lassiter et al (1995) suggested thatmixing of a plume-type source with eNdthorn 6 to thorn 7 witharc material with low NbTh could reproduce the varia-tions observed in the Karmutsen basalts however theyconcluded that the absence of low NbLa ratios in theirsample of tholeiitic basalts restricts the amount of arc litho-sphere involved The study of Lassiter et al (1995) wasbased on major and trace element and Sr Nd and Pb iso-topic data for a suite of 29 samples from Buttle Lake in theStrathcona Provincial Park on central Vancouver Island(Fig 1) Whereas there are no clear HFSE depletions inthe Karmutsen tholeiitic basalts the low NbLa (and TaLa) of the Karmutsen picritic lavas in this study indicatethe possible involvement of Paleozoic arc lithosphere onVancouver Island (Fig 8)
REE modeling dynamic melting andsource mineralogyThe combined isotopic and trace-element variations of thepicritic and tholeiitic Karmutsen lavas indicate that differ-ences in trace elements may have originated from differentmelting histories For example the LuHf ratios of theKarmutsen picritic lavas (mean 008) are considerablyhigher than those in the tholeiitic basalts (mean 002)however the small range of overlapping initial eHf valuesfor the two lava suites suggests that these differences devel-oped during melting within the plume (Fig 9b) Below wemodel the effects of source mineralogy and depth of
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
28
melting on differences in trace-element concentrations inthe tholeiitic basalts and picritesDynamic melting models simulate progressive decom-
pression melting where some of the melt fraction isretained in the residue and only when the degree of partialmelting is greater than the critical mass porosity and the
source becomes permeable is excess melt extracted fromthe residue (Zou1998) For the Karmutsen lavas evolutionof trace element concentrations was simulated using theincongruent dynamic melting model developed by Zou ampReid (2001) Melting of the mantle at pressures greater than1 GPa involves incongruent melting (eg Longhi 2002)
OIB array
PacificMORB(230 Ma)
(0 Ma)
EPR(230 Ma)
-4
-2
0
2
4
8
10
12
14
16
18
20
22
-4 -2 0 2 4 6 8 10 12 14
minus2
0
2
4
6
8
10
12
14
0702 0703 0704 0705 0706
IndianMORB
87Sr 86Sr (initial)
Hf (initial)
(a) (c)
Nd (initial)
Karmutsen tholeiitic basalt
HawaiiOntong JavaCaribbean Plateau
Karmutsen high-MgO lava
high-MgO lavasKarmutsen
1540
1545
1550
1555
1560
1565
175 180 185 190 195 200375
380
385
390
395
175 180 185 190 195 200
(d) (e)
EPR
EPR
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb
Karmutsen Formation (initial)
HawaiiOntong JavaCaribbean Plateau
East Pacific Rise
Karmutsen Formation (measured)
Juan de FucaGorda
ε Nd
(initial)
Karmutsen Formation (different age-correction for 3 picrites)
(0 Ma)
NbY
001
01
1
1 10
ZrY
OIB
NMORB
(b)
Iceland array
6
Fig 13 Comparison of age-corrected (230 Ma) Sr^Nd^Hf^Pb isotopic compositions for Karmutsen flood basalts onVancouver Island withage-corrected OIB and MORB and Nb^Zr^Y variation (a) Initial eNd vs 87Sr86Sr (b) NbYand ZrY variation (after Fitton et al 1997)(c) Initial eHf vs eNd (d) Measured and initial 207Pb204Pb vs 206Pb204Pb (e) Measured and initial 208Pb204Pb vs 206Pb204Pb References fordata sources are listed in Electronic Appendix 4 Most of the compiled data were extracted from the GEOROC database (httpgeorocmpch-mainzgwdgdegeoroc) OIB array line in (c) is fromVervoort et al (1999) EPR East Pacific Rise Several high-MgO samples affected bysecondary alteration were not plotted for clarity in Pb isotope plots Estimation of fields for age-corrected Pacific MORB and EPR at 230 Mawere made using parent^daughter concentrations (in ppm) from Salters amp Stracke (2004) of Lufrac14 0063 Hffrac14 0202 Smfrac14 027 Ndfrac14 071Rbfrac14 0086 and Srfrac14 836 In the NbYvs ZrYplot the normal (N)-MORB and OIB fields not including Icelandic OIB are from data com-pilations of Fitton et al (1997 2003) The OIB field is data from 780 basalts (MgO45wt ) from major ocean islands given by Fitton et al(2003) The N-MORB field is based on data from the Reykjanes Ridge EPR and Southwest Indian Ridge from Fitton et al (1997) The twoparallel gray lines in (b) (Iceland array) are the limits of data for Icelandic rocks
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
29
with olivine being produced during the reactioncpxthornopxthorn spfrac14meltthornol for spinel peridotites The mod-eling used coefficients for incongruent melting reactionsbased on experiments on lherzolite melting (eg Kinzleramp Grove 1992 Longhi 2002) and was separated into (1)melting of spinel lherzolite (2) two combined intervals ofmelting for a garnet lherzolite (gt lherz) and spinel lher-zolite (sp lherz) source (3) melting of garnet lherzoliteThe model solutions are non-unique and a range indegree of melting partition coefficients melt reactioncoefficients and proportion of mixed source componentsis possible to achieve acceptable solutions (see Fig 14 cap-tion for description of modeling parameters anduncertainties)The LREE-depleted high-MgO lavas require melting of
a depleted spinel lherzolite source (Fig 14a) The modelingresults indicate a high degree of melting (22^25) similarto the results from PRIMELT1 based on major-elementvariations (discussed above) of a LREE-depleted sourceThe high-MgO lavas lack a residual garnet signature Theenriched tholeiitic basalts involved melting of both garnetand spinel lherzolite and represent aggregate melts pro-duced from continuous melting throughout the depth ofthe melting column (Fig 14b) Trace-element concentra-tions in the tholeiitic and picritic lavas are not consistentwith low-degree melting of garnet lherzolite alone(Fig 14c) The modeling results indicate that the aggregatemelts that formed the tholeiitic basalts would haveinvolved lesser proportions of enriched small-degreemelts (1^6 melting) generated at depth and greateramounts of depleted high-degree melts (12^25) gener-ated at lower pressure (a ratio of garnet lherzolite tospinel lherzolite melt of 14 provides the best fits seeFig 14b and description in figure caption) The totaldegree of melting for the tholeiitic basalts was high(23^27) similar to the degree of partial melting in thespinel lherzolite facies for the primary melts that eruptedas the Keogh Lake picrites of a source that was also prob-ably depleted in LREE
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
(c)
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
Melting of garnet lherzolite
High-MgO lavas
Melting of spinel lherzolite
source (07DM+03PM)
source (07DM+03PM)
6 melting
30 melting
(b)
(a)
30 melting
Sam
ple
Cho
ndrit
eS
ampl
eC
hond
rite 20 melting
1 melting
Melting of garnet lherzoliteand spinel lherzolite
x gt lherz + 4x sp lherz
5 meltingx gt lherz + 4x sp lherz
Sam
ple
Cho
ndrit
e
Tholeiitic lavas
Tholeiitic lavas
High-MgO lavas
source (07DM+03PM)
Fig 14 Trace-element modeling results for incongruent dynamicmantle melting for picritic and tholeiitic lavas from the KarmutsenFormation Three steps of modeling are shown (a) melting of spinellherzolite compared with high-MgO lavas (b) melting of garnet lher-zolite and spinel lherzolite compared with tholeiitic lavas (c) meltingof garnet lherzolite compared with tholeiitic and picritic lavasShaded fields in each panel for tholeiitic basalts and picrites arelabeled Patterns with symbols in each panel are modeling results(patterns are 5 melting increments in (a) and (b) and 1 incre-ments in (c) PM primitive mantle DM depleted MORB mantle gtlherz garnet lherzolite sp lherz spinel lherzolite In (b) the ratios ofper cent melting for garnet and spinel lherzolite are indicated (x gtlherzthorn 4x sp lherz means that for 5 melting 1 melt in garnetfacies and 4 melt in spinel facies where x is per cent melting in theinterval of modeling) The ratio of the respective melt proportions inthe aggregate melt (x gt lherzthorn 4x sp lherz) was kept constant andthis ratio of melt contribution provided the best fit to the data Arange of proportion of source components (07DMthorn 03PM to
09DMthorn 01PM) can reproduce the variation in the high-MgOlavas Melting modeling uses the formulation of Zou amp Reid (2001)an example calculation is shown in their Appendix PM fromMcDonough amp Sun (1995) and DM from Salters amp Stracke (2004)Melt reaction coefficients of spinel lherzolite from Kinzler amp Grove(1992) and garnet lherzolite fromWalter (1998) Partition coefficientsfrom Salters amp Stracke (2004) and Shaw (2000) were kept constantSource mineralogy for spinel lherzolite (018cpx027opx052ol003sp) from Kinzler (1997) and for garnet lherzolite(034cpx008opx053ol005gt) from Salters amp Stracke (2004) Arange of source compositions for melting of spinel lherzolite(07DMthorn 03PM to 03DMthorn 07PM) reproduce REE patternssimilar to those of the high-MgO lavas with best fits for 22ndash25melting and 07DMthorn 03PM source composition Concentrationsof tholeiitic basalts cannot be reproduced from an entirelyprimitive or depleted source
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
30
The uncertainty in the composition of the source in thespinel lherzolite melting model for the high-MgO lavasprecludes determining whether the source was moredepleted in incompatible trace elements than the source ofthe tholeiitic lavas (see Fig 14 caption for description) It ispossible that early high-pressure low-degree melting left aresidue within parts of the plume (possibly the hotter inte-rior portion) that was depleted in trace elements (egElliott et al 1991) further decompression and high-degreemelting of such depleted regions could then have generatedthe depleted high-MgO Karmutsen lavas Shallow high-degree melting would preferentially sample depletedregions whereas deeper low-degree melting may notsample more refractory depleted regions (eg CaribbeanEscuder-Viruete et al 2007) The picrites lie at the highend of the range of Nd isotopic compositions and havelow NbZr similar to depleted Icelandic basalts (Fittonet al 2003) therefore the picrites may represent the partsof the mantle plume that were the most depleted prior tothe initiation of melting As Fitton et al (2003) suggestedthese depleted melts may only rarely be erupted withoutcombining with more voluminous less depleted melts andthey may be detected only as undifferentiated small-volume flows like the Keogh Lake picrites within theKarmutsen Formation on Vancouver IslandDecompression melting within the mantle plume initiatedwithin the garnet stability field (35^25 GPa) at highmantle potential temperatures (Tp414508C) and pro-ceeded beneath oceanic arc lithosphere within the spinelstability field (25^09 GPa) where more extensivedegrees of melting could occur
Magmatic evolution of Karmutsentholeiitic basaltsPrimary magmas in large igneous provinces leave exten-sive coarse-grained cumulate residues within or below thecrust these mafic and ultramafic plutonic sequences repre-sent a significant proportion of the original magma thatpartially crystallized at depth (eg Farnetani et al 1996)For example seismic and petrological studies of OntongJava (eg Farnetani et al 1996) combined with the use ofMELTS (Ghiorso amp Sack 1995) indicate that the volcanicsequence may represent only 40 of the total magmareaching the Moho Neal et al (1997) estimated that30^45 fractional crystallization took place for theOntong Java magmas and produced cumulate thicknessesof 7^14 km corresponding to 11^22 km of flood basaltsdepending on the presence of pre-existing oceanic crustand the total crustal thickness (25^35 km) Interpretationsof seismic velocity measurements suggest the presence ofpyroxene and gabbroic cumulates 9^16 km thick beneaththe Ontong Java Plateau (eg Hussong et al 1979)Wrangellia is characterized by high crustal velocities in
seismic refraction lines which is indicative of mafic
plutonic rocks extending to depth beneath VancouverIsland (Clowes et al 1995)Wrangellia crust is 25^30 kmthick beneath Vancouver Island and is underlain by astrongly reflective zone of high velocity and density thathas been interpreted as a major shear zone where lowerWrangellia lithosphere was detached (Clowes et al 1995)The vast majority of the Karmutsen flows are evolved tho-leiitic lavas (eg low MgO high FeOMgO) indicating animportant role for partial crystallization of melts at lowpressure and a significant portion of the crust beneathVancouver Island has seismic properties that are consistentwith crystalline residues from partially crystallizedKarmutsen magmas To test the proportion of the primarymagma that fractionated within the crust MELTS(Ghiorso amp Sack 1995) was used to simulate fractionalcrystallization using several estimated primary magmacompositions from the PRIMELT1 modeling results Themajor-element composition of the primary magmas couldnot be estimated for the tholeiitic basalts because of exten-sive plagthorn cpxthornol fractionation and thus the MELTSmodeling of the picrites is used as a proxy for the evolutionof major elements in the volcanic sequence A pressure of 1kbar was used with variable water contents (anhydrousand 02wt H2O) calculated at the quartz^fayalite^magnetite (QFM) oxygen buffer Experimental resultsfrom Ontong Java indicate that most crystallizationoccurs in magma chambers shallower than 6 km deep(52 kbar Sano amp Yamashita 2004) however some crys-tallization may take place at greater depths (3^5 kbarFarnetani et al 1996)A selection of the MELTS results are shown in Fig 15
Olivine and spinel crystallize at high temperature until15^20wt of the liquid mass has fractionated and theresidual magma contains 9^10wt MgO (Fig 15f) At1235^12258C plagioclase begins crystallizing and between1235 and 11908C olivine ceases crystallizing and clinopyr-oxene saturates (Fig 15f) As expected the addition of clin-opyroxene and plagioclase to the crystallization sequencecauses substantial changes in the composition of the resid-ual liquid FeO and TiO2 increase and Al2O3 and CaOdecrease while the decrease in MgO lessens considerably(Fig 15) The near-vertical trends for the Karmutsen tho-leiitic basalts especially for CaO do not match the calcu-lated liquid-line-of-descent very well which misses manyof the high-CaO high-Al2O3 low-FeO basalts A positivecorrelation between Al2O3CaO and SrNd indicates thataccumulation of plagioclase played a major role in the evo-lution of the tholeiitic basalts (Fig 15g) Increasing thecrystallization pressure slightly and the water content ofthe melts does not systematically improve the match ofthe MELTS models to the observed dataThe MELTS results indicate that a significant propor-
tion of the crystallization takes place before plagioclase(15^20 crystallization) and clinopyroxene (35^45
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
31
crystallization) join the crystallization sequence Theseresults suggest that445 crystallization of the primarymagmas occurred (estimated from the residual liquid percent in Fig 15f) and if the original flood basalt stratigra-phy on Vancouver Island was 6 km thick at least anequivalent amount of complementary plutonic rocksshould have crystallized The Karmutsen basalts and theirplutonic residues thus represent a significant addition ofcrust (perhaps 412 km thickness) previously thickenedduring the formation of the underlying Paleozoic arc
Growth of the Wrangellia oceanic plateauon Vancouver IslandThe extrusive part of the Wrangellia oceanic plateau onVancouver Island was constructed in a three-layersequence of mostly tholeiitic basalts with a restrictedrange of composition (Fig 16) The plateau overliesDevonian to Mississippian arc volcanic sequences andmarine sediments of Mississippian to Permian ages thatare intruded by gabbroic rocks related to the flood basaltsThe volcanic stratigraphy of theWrangellia plateau formed
00
05
10
15
20
25
30
35
40
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
16
0 2 4 6 8 10 12 14 16 18 20
10
11
12
13
14
15
16
17
18
19
20
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
0 2 4 6 8 10 12 14 16 18 20
MgO (wt)
0 10 20 30 40 50
percent of total mass
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
) Mineral proportions
Sp (e)
Ol
CpxPlag
PicriteHigh-MgO basaltTholeiitic basalt
TiO2 (wt)
MgO (wt)
Al2O3 (wt) CaO (wt)
FeO (wt)
MgO(a) (b)
(c) (d)
MgO (wt)MgO (wt)
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
)
Residual liquid ()
1220
1190
Ol + Sp
Plag+Ol+Sp Plag+Cpx+Sp
02 wt H2O
no H2O
Residual liquid and crystallizing phases
(f)
0102030405060708090100
plagioclaseaccumulation
plagioclaseaccumulation
plagioclase
accumulatio
n
10
12
14
16
0 10 20 30 40
Al2O3CaO
SrNd
tholeiitic basalts
(g)
Fig 15 Forward fractional crystallization modeling results for major elements from MELTS (Ghiorso amp Sack 1995) compared with picriticand tholeiitic lavas from the Karmutsen Formation (a^d) Major-element compositions of residual liquids compared with Karmutsen lavacompositions For clarity the results are shown only for one starting composition (open circle) using the composition of the estimated primarymagma for the high-MgO lava series from the modeling technique of Herzberg et al (2007) for sample 4723A13 as shown in Fig 12 The esti-mated primary magmas for samples 4723A3 and 93G171 yield similar results A pressure of 1 kbar was used with no H2O and 02wt H2Ocalculated at QFM black line is anhydrous and grey line is 02wt H2OThe end-point for the anhydrous result is at 78 fractionation and11308C and the end-point for the result with 02wt H2O is 75 fractionation and11108C (e) Mineral proportions of crystallized phases fromone MELTrun (sample 4723A13) with no H2O (f) Per cent residual liquid vs temperature labelled with intervals of crystallizing phases (g)Al2O3CaO vs SrNd for the tholeiitic basalts
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
32
in a single eruptive phase (c 231^225 Ma) as an emergentbasalt sequence during a deeper-water shallow-water andsubaerial stage and subsided below sea level relativelysoon after its formation (Fig 16) The deeper-water stagecontains pillowed and massive submarine flows (3000mthick) that built up the submarine volcanic edifice from theseafloor The lowest pillow basalts were emplaced onunconsolidated and lithified fine-grained siliciclastic andcarbonaceous sediments containing Daonella beds(c 235^232 Ma) that are intruded by abundantKarmutsen mafic sills (Fig 16a) The different pillowedand massive submarine flows resulted primarily from dif-ferent effusive rates and local topography Some of themassive submarine flows commonly recognized by radiat-ing columnar jointing represent master tubes for the deliv-ery of lava to distal parts of submarine flow fields Thepillowed flows which commonly have uneven topography
in modern sequences may have served as levees to themassive submarine flows rather than allowing unob-structed spreading and inflation typical of the massive sub-aerial flows that form on shallow slopes in continentalflood basalt provinces (eg Self et al 1997) The rapid suc-cession of flows is evident by the near absence of sedimentbetween submarine flows Minor volumes of picritic pillowlavas occur in the upper part of the pillowed flow sequencein areas of the plateau exposed on northern VancouverIsland (Fig 16 Nixon et al 2008)The growth of the plateau in shallow water led to
emplacement of an increasing proportion of volcaniclastics(pillow breccias and hyaloclastites) that accumulated up to1500m thick in some areas (Fig 16b Nixon et al 2008)Horizons of pillowed flows occur locally within the volca-niclastic stratigraphy Sedimentary structures (eg gradedbedding fluidization structures) and petrographic textures
Late Devonian-Mississippian arc
Daonella beds
Middle Triassic
deeper water
Late Devonian-Mississippian arc
pillowed flows massive submarine
flows
pillow brecciahyaloclastiteshallow water
~4 km
~6 km
Late Triassic limestoneinterflow limestonelenses
Karmutsen Formation
mafic intrusions sediment-sill complex
Karmutsen Formation
(a) (b)
(c) (d)
subaerial flows
Late Devonian-Mississippian arc Late Devonian-Mississippian arc
picritic pillow basalts
Miss to Permian marine sediments
sills
Fig 16 Growth of the Wrangellia oceanic plateau onVancouver Island (a) Deep-water eruptive phase of pillowed and massive submarineflows with almost no intervening sediments onto MiddleTriassic shale containing Daonella beds (c 235^232 Ma) The basement of the plateauis composed of Devonian to Mississippian arc sequences overlain by Mississippian to Permian carbonates and siliciclastic sediments(b) Shallow-water flow emplacement with increasing proportion of volcaniclastics resulted primarily from cooling^contraction granulationPicritic pillow basalts occur primarily near the transition between pillowed submarine flows and volcaniclastic units Pillowed and massiveflows occur locally within the volcaniclastic unit (c) Emergent stage of growth consists of overlapping subaerial flow fields similar to continen-tal flood basalts with pahoehoe flow structures Multi-tiered columnar jointing occurrs locally in response to ponding of flows (eg Bondre et al2004) Interflow carbonate deposits and plagioclase megacrystic flows developed locally during the waning stage of volcanism (d) Post-volcanicsubsidence and deposition of Late Carnian to Norian (c 212^223 Ma) platform carbonates on top of the plateau Units are labeled in the dia-grams Thicknesses of the underlying units are not to scale
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
33
within the volcaniclastic rocks indicate formation primar-ily via cooling^contraction granulation magma^water^steam interaction autobrecciation resedimentation andmass-wasting rather than pyroclastic eruption (Nixonet al 2008)A broad subaerial platform was constructed by emplace-
ment of subaerial lava flows during an emergent subaerialstage (Fig 16c) The subaerial unit predominantly com-prises well-layered sheet flows that formed in a similarway to continental flood basalts as hundred(s) of pahoehoeflow fields erupted on shallow slopes (Self et al 1997) how-ever volcaniclastic and pillowed flow units are locally pre-sent mostly in the upper parts of the stratigraphyLocalized ponding of flows led to the formation of multi-tiered columnar jointing in some subaerial flows (egBondre et al 2004) As volcanism waned local interflowcarbonate deposits developed and plagioclase megacrysticflows were erupted Subsidence of the plateau is recordedby the interflow lenses and deposits of platform carbonates(40^1500m thick) on top of the Karmutsen basalts thatcontain Late Carnian to Early Norian fossils (Fig 16d c216^227 Ma Nixon et al 2008)The evolution of the internal architecture of the
Wrangellia oceanic plateau was primarily related to theconditions of emplacement (ie water depth and eruptionenvironment) There does not appear to have been a sys-tematic coincident evolution or shift of lava compositionsonVancouver Island as has been recognized in most con-tinental flood basalts and parts of the Wrangellia oceanicplateau in Alaska and Yukon (Greene et al 2008 2009)In general with the exception of the picritic pillow lavasin the submarine stratigraphy and local variations betweenhigh-MgO and tholeiitic lavas there are no significantcompositional differences between lavas erupted duringthe different stages of growth of the plateau (ie submarineand subaerial flows) nor are substantial compositional dif-ferences generally found between massive submarine flowsin contact with pillowed flows tholeiitic basalt flows nearthe base of the Karmutsen Formation are compositionallysimilar to flows sampled near the top of the volcanicsequence Where it was possible to sample continuous sec-tions of volcanic stratigraphy greater than several hun-dreds of meters (eg Keogh^Maynard Lake area west ofthe Karmutsen Range Fig 2) there is no clear relationshipbetween stratigraphic position and compositionalvariation
CONCLUSIONSThe Karmutsen Formation covers extensive areas ofVancouver Island (20 000 km2) and forms part of a largeoceanic plateau erupted over a geologically short interval(c 225^231 Ma) The tripartite volcanic stratigraphyexposed on Vancouver Island is at least 6 km thick andcomprises a succession of submarine flows volcaniclastic
deposits and massive subaerial flowsThese volcanologicaldifferences are primarily related to the eruption environ-ment (deep-water shallow-water or subaerial) The rapidgrowth of the plateau prevented significant accumulationsof intervening sediment except in the uppermost stratigra-phy where isolated limestone lenses commonly associatedwith pillow basalt and hyaloclastite deposits preserve arecord of the final stages of subsidence of the plateauThe Wrangellia plateau on Vancouver Island was con-
structed dominantly of tholeiitic basalt with restrictedmajor- and trace-element and isotopic compositional var-iations Minor volumes of picritic pillow basalts eruptedlate in the submarine phase The Karmutsen basalts wereemplaced onto older Paleozoic arc volcanic sequences andlow NbLa for the picrites may indicate a contributionfrom the sub-arc lithospheric mantle Modeling of the pet-rogenesis of the Karmutsen picrites indicates melting ofanomalously hot mantle (15008C) and extensive degreesof partial melting (23^27) and is consistent with aplume initiation model Combined Sr^Nd^Hf^Pb isotopicsystematics indicate that the picritic and tholeiitic lavaswere derived from a depleted plume-type mantle sourcedistinct from the source of MORB that has compositionalsimilarities to the source of the lavas forming theCaribbean Plateau Picrites were generated by melting ofspinel lherzolite whereas the tholeiitic basalts formed asaggregate melts of both garnet and spinel lherzolite Thetholeiitic basalts underwent extensive low-pressure fractio-nation (52^3 kbar) and seismic studies indicate that someof theWrangellia crust beneathVancouver Island may cor-respond to the plutonic residues of the Karmutsen floodbasalts
ACKNOWLEDGEMENTSWe are grateful to Nick Arndt for helping us get this pro-ject started and Nick Massey for insights into VancouverIsland geology We greatly appreciate the assistance withPRIMELT modeling from Claude Herzberg We thankMikkel Schau for his insight and enthusiasm during field-work Jane Barling assisted with analyses by MC-ICP-MSThoughtful and thorough reviews of the manuscript byJohn Mahoney Fred Frey Wendy Bohrson DougalJerram and Teal Riley are greatly appreciated Fundingwas generously provided by the Rocks to Riches Programadministered by the BC amp Yukon Chamber of Mines in2004 by the BC Geological Survey and by NSERCDiscovery Grants to JSS and DW ARG was supportedby a University Graduate Fellowship at UBC
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
34
REFERENCESBarker F Brown A S Budahn J R amp Plafker G (1989) Back-arc
with frontal-arc component origin of Triassic Karmutsen basaltBritish Columbia Canada Chemical Geology 75 81^102
Blichert-Toft JWeis D Maerschalk C Agranier A amp Albare de F(2003) Hawaiian hot spot dynamics as inferred from the Hf and Pbisotope evolution of Mauna Kea volcano Geochemistry GeophysicsGeosystems 4(2) 1^27 doi1010292002GC000340
Bondre N R Duraiswami R A amp Dole G (2004) Morphologyand emplacement of flows from the Deccan Volcanic ProvinceIndia Bulletin of Volcanology 66 29^45
Carlisle D (1963) Pillow breccias and their aquagene tuffs QuadraIsland British Columbia Journal of Geology 71 48^71
Carlisle D (1972) Late Paleozoic to Mid-Triassic sedimentary-volcanic sequence on Northeastern Vancouver Island Geological
Society of Canada Paper Report of Activities 72-1 Part B 24^30Carlisle D amp SuzukiT (1974) Emergent basalt and submergent car-
bonate^clastic sequences including the Upper Triassic Dilleri andWelleri zones on Vancouver Island Canadian Journal of Earth
Sciences 11 254^279Chauvel C amp Blichert-Toft J (2001) A hafnium isotope and trace
element perspective on melting of the depleted mantle Earth and
Planetary Science Letters 190 137^151Chayes F (1954)The theory of thin-section analysis Journal of Geology
62 92^101Cho M Liou J G amp Maruyama S (1987) Transition from the zeo-
lite to prehnite^pumpellyite facies in the Karmutsen metabasitesVancouver Island British Columbia Journal of Petrology 27(2)467^494
Clowes R M Zelt C A Amor J R amp Ellis R M (1995)Lithospheric structure in the southern Canadian Cordillera froma network of seismic refraction lines Canadian Journal of Earth
Sciences 32(10) 1485^1513Coffin M F amp Eldholm O (1994) Large igneous provinces Crustal
structure dimensions and external consequences Reviews of
Geophysics 32(1) 1^36Cox K G (1980) A model for flood basalt volcanism Journal of
Petrology 21 629^650Eldholm O amp Coffin M F (2000) Large igneous provinces
and plate tectonics In Richards M A Gordon R G amp vander Hilst R D (eds) The History and Dynamics of Global
Plate Motions American Geophysical Union Geophysical Monograph 121309^326
Elliott T R Hawkesworth C J amp Grolaquo nvold K (1991) Dynamicmelting of the Iceland plume Nature 351 201^206
Escuder-Viruete J Pecurren rez-Estaucurren n A Contreras F Joubert MWeis D Ullrich T D amp Spadea P (2007) Plume mantle sourceheterogeneity through time Insights from the Duarte ComplexHispaniola northeastern Caribbean Journal of Geophysical Research112(B04203) doi1010292006JB004323
Farnetani C G Richards M A amp Ghiorso M S (1996)Petrological models of magma evolution and deep crustal structurebeneath hotspots and flood basalt provinces Earth and Planetary
Science Letters 143 81^94Fitton J G amp Godard M (2004) Origin and evolution of magmas
on the Ontong Java Plateau In Fitton J G Mahoney J JWallace P J amp Saunders A D (eds) Origin and Evolution of the
Ontong Java Plateau Geological Society London Special Publications 229151^178
Fitton J G Saunders A D Norry M J Hardarson B S ampTaylor R N (1997) Thermal and chemical structure of theIceland plume Earth and Planetary Science Letters 153 197^208
Fitton J G Saunders A D Kempton P D amp Hardarson B S(2003) Does depleted mantle form an intrinsic part of the Icelandplume Geochemistry Geophysics Geosystems 4(3) 1032 doi1010292002GC000424
Frey F A Huang S Blichert-Toft J Regelous M amp Boyet M(2005) Origin of depleted components in basalt related to theHawaiian hotspot Evidence from isotopic and incompatible ele-ment ratios Geochemistry Geophysics Geosystems 6(1) 23 doi1010292004GC000757
Galer S J G amp Abouchami W (1998) Practical application of leadtriple spiking for correction of instrumental mass discriminationMineralogical Magazine 62A 491^492
Ghiorso M S amp Sack R O (1995) Chemical mass transfer in mag-matic processes IV A revised and internally consistent thermody-namic model for the interpolation and extrapolation of liquid^solidequilibria in magmatic systems at elevated temperatures and pres-sures Contributions to Mineralogy and Petrology 119 197^212
Greene A R Scoates J S amp Weis D (2005)WrangelliaTerrane onVancouver Island Distribution of flood basalts with implicationsfor potential Ni^Cu^PGE mineralization In Grant B (ed)Geological Fieldwork 2004 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2005-1 209^220Greene A R Scoates J S Nixon G T amp Weis D (2006) Picritic
lavas and basal sills in the Karmutsen flood basalt provinceWrangellia northern Vancouver Island In Grant B (ed)Geological Fieldwork 2005 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2006-1 39^52Greene A R Scoates J S amp Weis D (2008) Wrangellia flood
basalts in Alaska A record of plume^lithosphere interaction in aLate Triassic accreted oceanic plateau Geochemistry Geophysics
Geosystems 9(Q12004) doi1010292008GC002092Greene A R Scoates J S Weis D amp Israel S (2009)
Geochemistry of triassic flood basalts from the Yukon (Canada)segment of the accreted Wrangellia oceanic plateau Lithosdoi101016jlithos200811010
Hauff F Hoernle K Tilton G Graham D W amp Kerr A C(2000) Large volume recycling of oceanic lithosphere over shorttime scales geochemical constraints from the Caribbean LargeIgneous Province Earth and Planetary Science Letters 174 247^263
Hauff F Hoernle K amp Schmidt A (2003) Sr^Nd^Pb compositionof Mesozoic Pacific oceanic crust (Site 1149 and 801 ODP Leg185)Implications for alteration of ocean crust and the input into theIzu^Bonin^Mariana subduction system Geochemistry Geophysics
Geosystems 4(8) doi1010292002GC000421Herzberg C (2004) Partial melting below the Ontong Java Plateau
In Fitton J G Mahoney J J Wallace P J amp Saunders A D(eds) Origin and Evolution of the Ontong Java Plateau Geological SocietyLondon Special Publications 229 179^183
Herzberg C (2006) Petrology and thermal structure of the Hawaiianplume from Mauna Kea volcano Nature 444(7119) 605
Herzberg C amp Asimow P D (2008) Petrology of some oceanicisland basalts PRIMELT2XLS software for primary magma cal-culation Geochemistry Geophysics Geosystems 9(Q09001) doi1010292008GC002057
Herzberg C amp OrsquoHara M J (2002) Plume-associated ultramaficmagmas of Phanerozoic age Journal of Petrology 43(10) 1857^1883
Herzberg C Asimow P D Arndt N Niu Y Lesher C MFitton J G Cheadle M J amp Saunders A D (2007)Temperatures in ambient mantle and plumes Constraints frombasalts picrites and komatiites Geochemistry Geophysics Geosystems8(Q02006) doi1010292006GC001390
Huang S amp Frey F A (2005) Trace element abundances of MaunaKea basalt from phase 2 of the Hawaii Scientific Drilling Project
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
35
Petrogenetic implications of correlations with major element con-tent and isotopic ratios Geochemistry Geophysics Geosystems 4(6)doi1010292002GC000322
Hussong D M Wipperman L K amp Kroenke L W (1979) Thecrustal structure of the Ontong Java and Manihiki OceanicPlateaus Journal of Geophysical Research 84(B11) 6003^6010
Jerram D A amp Widdowson M (2005) The anatomy of ContinentalFlood Basalt Provinces geological constraints on the processes andproducts of flood volcanism Lithos 79(3^4) 385^405
Jones D L Silberling N J amp Hillhouse J (1977)Wrangellia a dis-placed terrane in northwestern North America CanadianJournal ofEarth Sciences 14(11) 2565^2577
Kerr A C (2003) Oceanic Plateaus In Rudnick R (ed) The Crust
Treatise on GeochemistryVol 3 Oxford Elsevier Science pp 537^565Kerr A C amp Mahoney J J (2007) Oceanic plateaus Problematic
plumes potential paradigms Chemical Geology 241 332^353Kerr A CTarney J Marriner G F Klaver GT Saunders A D
amp Thirlwall M F (1996) The geochemistry and petrogenesis ofthe Late-Cretaceous picrites and basalts of Curacao NetherlandsAntilles a remnant of an oceanic plateau Contributions to
Mineralogy and Petrology 124(1) 29^43Kerr A C Marriner G F Tarney J Nivia A Saunders A D
Thirlwall M F amp Sinton A C (1997) Cretaceous basaltic ter-ranes in Western Colombia Elemental chronological and Sr-Ndisotopic constraints on petrogenesis Journal of Petrology 38(6)677^702
Kinzler R J (1997) Melting of mantle peridotite at pressuresapproaching the spinel to garnet transition Application to mid-ocean ridge basalt petrogenesis Journal of Geophysical Research
102(B1) 853^874Kinzler R J amp Grove T L (1992) Primary magmas of mid-ocean
ridge basalts 1 Experiments and methods Journal of Geophysical
Research 97(B5) 6885^6906Korenaga J (2005) Why did not the Ontong Java Plateau form sub-
aerially Earth and Planetary Science Letters 234 385^399Kuniyoshi S (1972) Petrology of the Karmutsen (Volcanic) Group
northeastern Vancouver Island British Columbia PhD disserta-tion University of California Los Angeles 242 pp
Lassiter J C DePaolo D J amp Mahoney J J (1995) Geochemistry ofthe Wrangellia flood basalt province Implications for the role ofcontinental and oceanic lithosphere in flood basalt genesis Journalof Petrology 36(4) 983^1009
LeBas M J (2000) IUGS reclassification of the high-Mg and picriticvolcanic rocks Journal of Petrology 41(10) 1467^1470
Longhi J (2002) Some phase equilibrium systematics of lherzolitemelting I Geochemistry Geophysics Geosystems 3(3) doi1010292001GC000204
MacDonald G A amp Katsura T (1964) Chemical composition ofHawaiian lavas Journal of Petrology 5 82^133
Mahoney J J (1987) An isotopic survey of Pacific oceanic plateausimplications for their nature and origin In Keating B HFryer P Batiza R amp Boehlert GW (eds) Seamounts Islands andAtolls American Geophysical Union Geophysical Monograph 43 207^220
Mahoney J LeRoex A P Peng Z Fisher R L amp Natland J H(1992) Southwestern limits of Indian Ocean Ridge mantle and theorigin of low 206Pb204Pb mid-ocean ridge basalt isotope systema-tics of the central Southwest Indian Ridge (178^508E) Journal ofGeophysical Research 97(B13) 19771^19790
Mahoney J J Sinton J M Kurz M D MacDougall J DSpencer K J amp Lugmair GW (1994) Isotope and trace elementcharacteristics of a super-fast spreading ridge East Pacific rise13^238S Earth and Planetary Science Letters 121 173^193
Massey N W D (1995a) Geology and mineral resources of theAlberni^Nanaimo Lakes sheet Vancouver Island 92F1W 92F2Eand part of 92F7E British Columbia Ministry of Energy Mines and
Petroleum Resources Paper 1992-2 132 ppMassey N W D (1995b) Geology and mineral resources of the
Cowichan Lake sheet Vancouver Island 92C16 British Columbia
Ministry of Energy Mines and Petroleum Resources Paper 1992-3 112 ppMassey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005a) Digital Geology Map of British Columbia Tile NM9 MidCoast BC British Columbia Ministry of Energy and Mines Geofile
2005-2Massey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005b) Digital Geology Map of British Columbia Tile NM10Southwest BC British Columbia Ministry of Energy and Mines Geofile
2005-3McDonough W F amp Sun S (1995) The composition of the Earth
Chemical Geology 120 223^253Monger JW H amp Journeay J M (1994) Basement geology and tec-
tonic evolution of theVancouver region In Monger JW H (ed)Geology and Geological Hazards of the Vancouver Region Southwestern
British Columbia Geological Survey of Canada Bulletin 481 3^25Muller J E (1977) Geology of Vancouver Island Geological Survey of
Canada Open File Map 463Muller J E (1980)The Paleozoic Sicker Group of Vancouver Island British
Columbia Geological Survey of Canada Paper 79-30 22 ppMuller J E Northcote K E amp Carlisle D (1974) Geology and
mineral deposits of Alert Bay^Cape Scott map area VancouverIsland British Columbia Geological Survey of Canada Paper 74-8 77
Neal C R Mahoney J J Kroenke L W Duncan R A ampPetterson M G (1997) The Ontong^Java Plateau InMahoney J J amp Coffin M F (eds) Large Igneous Provinces
Continental Oceanic and Planetary FloodVolcanism American Geophysical
Union Geophysical Monograph 100 183^216Niu Y Collerson K D Batiza R Wendt J I amp Regelous M
(1999) Origin of enriched-typed mid-ocean ridge basalt at ridgesfar from mantle plumes The East Pacific Rise at 11820rsquoN Journalof Geophysical Research 104 7067^7088
Nixon G T amp Orr A J (2007) Recent revisions to the EarlyMesozoic stratigraphy of Northern Vancouver Island (NTS 102I092L) and metallogenic implicatinos British Columbia InGrant B (ed) Geological Fieldwork 2006 British Columbia Ministry of
Energy Mines and Petroleum Resources Paper 2007-1 163^177Nixon GT Hammack J L KoyanagiV M Payie G J Haggart
JW Orchard M JTozerT Friedman R M Archibald D APalfy J amp Cordey F (2006a) Geology of the Quatsino^PortMcNeill area northernVancouver Island British Columbia Ministry
of Energy Mines and Petroleum Resources Geoscience Map 2006-2 scale150 000
Nixon G T Kelman M C Stevenson D Stokes L A ampJohnston K A (2006b) Preliminary geology of the Nimpkishmap area (NTS 092L07) northern Vancouver Island BritishColumbia In Grant B amp Newell J M (eds) Geological Fieldwork2005 British Columbia Ministry of Energy Mines and Petroleum Resources
Paper 2006-1 135^152Nixon G T Laroque J Pals A Styan J Greene A R amp
Scoates J S (2008) High-Mg lavas in the Karmutsen flood basaltsnorthernVancouver Island (NTS 092L) Stratigraphic setting andmetallogenic significance In Grant B (ed) Geological Fieldwork
2007 British Columbia Ministry of Energy Mines and Petroleum
Resources Paper 2008-1 175^190Nowell G M Kempton P D Noble S R Fitton J G
Saunders A Mahoney J J amp Taylor R N (1998) High precisionHf isotope measurements of MORB and OIB by thermal ionisation
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
36
mass spectrometry insights into the depleted mantle Chemical
Geology 149 211^233Parrish R R amp McNicoll V J (1992) U^Pb age determinations
from the southern Vancouver Island area British Columbia InRadiogenic Age and Isotopic Studies Report 5 Geological Survey of
Canada Paper 91-2 79^86Peate DW (1997) The Parana^Etendeka Province In Coffin M F
amp Mahoney J (eds) Large Igneous Provinces Continental Oceanic andPlanetary Flood Volcanism American Geophysical Union Geophysical
Monograph 100 217^245Perfit M R Fornari D J Ridley W I Kirk P D Casey J
Kastens K A Reynolds J R Edwards M Desonie DShuster R amp Paradis S (1996) Recent volcanism in the Siqueirostransform fault picritic basalts and implications for MORBmagma genesis Earth and Planetary Science Letters 141(1^4) 91^108
Pretorius W Weis D Williams G Hanano D Kieffer B ampScoates J S (2006) Complete trace elemental characterization ofgranitoid (USGSG-2GSP-2) reference materials by high resolutioninductively coupled plasma-mass spectrometry Geostandards and
Geoanalytical Research 30(1) 39^54Regelous M Niu Y Wendt J I Batiza R Greig A amp
Collerson K D (1999) Variations in the geochemistry of magma-tism on the East Pacific Rise at 10830rsquoN since 800 ka Earth and
Planetary Science Letters 168 45^63Recurren villon S Chauvel C Arndt N T Pik R Martineau F
Fourcade S amp Marty B (2002) Heterogeneity of the Caribbeanplateau mantle source Sr O and He isotopic compositions of oli-vine and clinopyroxene from Gorgona Island Earth and Planetary
Science Letters 205(1^2) 91Rhodes J M (1996) Geochemical stratigraphy of lava flows samples
by the Hawaii Scientific Drilling Project Journal of Geophysical
Research 101(B5) 11729^11746Rhodes J M amp Vollinger M J (2004) Composition of basaltic lavas
sampled by phase-2 of the Hawaii Scientific Drilling ProjectGeochemical stratigraphy and magma types Geochemistry
Geophysics Geosystems 5(3) doi1010292002GC000434Richards M A Jones D L Duncan R A amp DePaolo D J (1991)
A mantle plume initiation model for the Wrangellia flood basaltand other oceanic plateaus Science 254 263^267
Salters V J M amp Hart S R (1991) The mantle sources of oceanridges islands and arcs the Hf-isotope connection Earth and
Planetary Science Letters 104 364^380Salters V J M amp Stracke A (2004) Composition of the depleted
SaltersV J amp WhiteW (1998) Hf isotope constraints on mantle evo-lution Chemical Geology 145 447^460
SanoT amp Yamashita S (2004) Experimental petrology of basementlavas from Ocean Drilling Program Leg 192 implications for dif-ferentiation processes in Ontong Java Plateau magmas InFitton J G Mahoney J JWallace P J amp Saunders A D (eds)Origin and Evolution of the Ontong Java Plateau Geological Society
London Special Publications 229 185^218Saunders A D (2005) Large igneous provinces Origin and environ-
mental consequences Elements 1 259^263Self S ThordarsonT amp Keszthely L (1997) Emplacement of conti-
nental flood basalt lava flows In Mahoney J J amp Coffin M F(eds) Large Igneous Provinces Continental Oceanic and Planetary Flood
Volcanism American Geophysical Union Geophysical Monograph 100381^410
Shaw D M (2000) Continuous (dynamic) melting theory revisitedCanadian Mineralogist 38(5) 1041^1063
Sluggett C L (2003) Uranium^lead age and geochemical constraintson Paleozoic and Early Mesozoic magmatism in WrangelliaTerrane Saltspring Island British Columbia BSc thesisUniversity of British ColumbiaVancouver 84 pp
Storey M Mahoney J J amp Saunders A D (1997) Cretaceousbasalts in Madagascar and the transition between plume and con-tinental lithospheric mantle sources In Mahoney J J ampCoffin M F (eds) Large Igneous Provinces Continental Oceanic and
Planetary Flood Volcanism Aamerican Geophysical Union Geophysical
Monograph 100 95^122Surdam R C (1967) Low-grade metamorphism of the Karmutsen
Group PhD dissertation University of California Los Angeles288 pp
Sutherland-Brown A Yorath C J Anderson R G amp Dom K(1986) Geological maps of southern Vancouver IslandLITHOPROBE I 92C10 11 14 16 92F1 2 7 8 Geological Survey ofCanada Open File 1272
Tejada M L G Mahoney J J Castillo P R Ingle S P Sheth HC amp Weis D (2004) Pin-pricking the elephant evidence on theorigin of the Ontong Java Plateau from Pb^Sr^Hf^Nd isotopiccharacteristics of ODP Leg 192 basalts In Fitton J GMahoney J J Wallace P J amp Saunders A D (eds) Origin and
Evolution of the Ontong Java Plateau Geological Society London Special
Publications 229 133^150Thompson P M E Kempton P D amp Kerr A C (2008) Evaluation
of the effects of alteration and leaching on Sm^Nd and Lu^Hf sys-tematics in submarine mafic rocks Lithos 104(1^4) 164^176
Thompson P M E Kempton P D White R V Kerr A CTarney J Saunders A D Fitton J G amp McBirney A (2003)Hf-Nd isotope constraints on the origin of the CretaceousCaribbean plateau and its relationship to the Galapagos plumeEarth and Planetary Science Letters 217(1^2) 59^75
Vervoort J D Patchett P Blichert-Toft J amp Albare de F (1999)Relationships between Lu^Hf and Sm^Nd isotopic systems in theglobal sedimentary system Earth and Planetary Science Letters 16879^99
Walter M J (1998) Melting of garnet peridotite and the origin ofkomatiite and depleted lithosphere Journal of Petrology 39(1) 29^60
Weis D Kieffer B Maerschalk C Barling J de Jong JWilliams G A Hanano D Mattielli N Scoates J SGoolaerts A Friedman R A amp Mahoney J B (2006) High-pre-cision isotopic characterization of USGS reference materials byTIMS and MC-ICP-MS Geochemistry Geophysics Geosystems
7(Q08006) doi1010292006GC001283Weis D Kieffer B Hanano D Silva I N Barling J
Pretorius W Maerschalk C amp Mattielli N (2007) Hf isotopecompositions of US Geological Survey reference materialsGeochemistry Geophysics Geosystems 8(Q06006) doi1010292006GC001473
Wheeler J O amp McFeely P (1991) Tectonic assemblage map of theCanadian Cordillera and adjacent part of the United States ofAmerica Geological Survey of Canada Map 1712A
WhiteW M Albare de F amp Tecurren louk P (2000) High-precision analy-sis of Pb isotope ratios by multi-collector ICP-MS Chemical Geology167 257^270
Yorath C J Sutherland Brown A amp Massey N W D (1999)LITHOPROBE southern Vancouver Island British Columbia Geological
Survey of Canada Bulletin 498 145 ppZou H (1998) Trace element fractionation during modal and
non-model dynamic melting and open-system melting Amathematical treatment Geochimica et Cosmochimica Acta 62(11)1937^1945
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
37
Zou H amp Reid M R (2001) Quantitative modeling of trace elementfractionation during incongruent dynamic melting Geochimica et
Cosmochimica Acta 65(1) 153^162
APPENDIX SAMPLEPREPARAT ION ANDANALYT ICAL METHODSOnly the freshest rocks in the field were sampled and onlythe least altered samples were selected for chemical andisotopic analysis based on thorough petrographic inspec-tion Fifty-nine of the 129 collected samples were crushed(400 g) into small pieces 52mm in diameter in aRocklabs hydraulic piston crusher between WC plates tominimize contamination The coarse-crush was mixedand 100 g was powdered in a planetary mill using agatejars and balls cleaned with quartz sand between samples
ActLabs analytical methodsThe major- and trace-element compositions of the whole-rock powders were determined at Activation LaboratoriesLtd (Actlabs) in Ancaster Ontario Analytical techniquesand detection limits are also available from Actlabs (httpwwwactlabscommethsub_code4erehtm) The analyticalmethod for each of the elements analyzed is indicated inTable 2 For the major elements a 02 g sample was mixedwith a lithium metaborate^lithium tetraborate mixtureand fused in a graphite crucible The molten mixture waspoured into a 5 HNO3 solution and shaken until dis-solved (30min) The samples were analyzed for majoroxides and selected trace elements by inductively coupledplasma optical emission spectrometry (ICP-OES) on acombination simultaneous^sequential Thermo Jarrell-AshEnviro II system Internal calibration was achieved usinga variety of international reference materials (eg W-2BIR-1 DNC-1) and independent control samplesAdditional trace elements were analyzed by both instru-mental neutron activation analysis (INAA) and induc-tively couple plasma mass spectrometry (ICP-MS) ForINAA15^25 g of sample was weighed into small polyeth-ylene vials and irradiated with a control international ref-erence material CANMET WMS-1 and NiCr flux wires ata thermal neutron flux of 71012 neutronscm2 per s in theMcMaster Nuclear Reactor Following a 7 day waitingperiod the samples were measured on an Ortec high-purity Ge detector linked to a Canberra Series 95 multi-channel analyzer Activities for each element weredecay- and weight-corrected and compared with a detectorcalibration developed from multiple international certifiedreference materials For ICP-MS 025 g of sample wasdigested in HF followed by HNO3^HClO4 heated andtaken to dryness The samples were brought back into solu-tion with HCl Samples were analyzed using a PerkinElmer Optima 3000 ICP In-lab standards or certified ref-erence materials (eg W-2 BIR-1 DNC-1) were used for
quality control A total of 15 blind duplicates were ana-lyzed from the same sample powders to assessreproducibility
University of Massachusetts analyticalmethodsFourteen sample duplicate powders taken from the same100 g sample aliquot (high-MgO basalts and picrites)were also analyzed at the Ronald B Gilmore X-RayFluorescence (XRF) Laboratory at the University ofMassachusetts Major elements were measured on a fusedLa-bearing lithium borate glass disc using a SiemensMRS-400 spectrometer Trace element concentrations(Rb Sr Ba Ce Nb Zr Y Pb Zn Ga Ni Cr V) weremeasured on a separate powder pellet using a PhilipsPW2400 sequential spectrometer using a Rh tube LOIand ferrous iron measurements were made as describedby Rhodes amp Vollinger (2004) Precision and accuracy esti-mates for the data have been described by Rhodes (1996)and Rhodes amp Vollinger (2004) Results for each sampleare the average of two separate analyses and are shown inElectronic Appendix 2 for a comparison of analyses ofduplicate powders using XRFand ICP-MS
PCIGR analytical methodsA subset of 19 samples was selected for high-precisiontrace-element analysis and Sr Nd Pb and Hf isotopicanalysis at the Pacific Centre for Isotopic andGeochemical Research (PCIGR) at the University ofBritish Columbia (UBC) Samples were selected from the59 samples analyzed for whole-rock chemistry at ActLabsbased on major- and trace-element chemistry alteration(low LOI and petrographic alteration index) and samplelocation Samples were prepared for trace-element analysisat the PCIGR by the technique described by Pretorius et al(2006) on unleached rock powders Sample powders(100mg) were weighed in 7ml screw-top Savillex bea-kers and dissolved in 1ml 14N HNO3 and 5ml 48 HFon a hotplate for 48 h at 1308C with periodic ultrasonica-tion Samples were dried and redissolved in 6ml 6N HClon a hotplate for 24 h and then dried and redissolved in1ml concentrated HNO3 for 24 h before final dryingTrace element abundances were measured by high-resolution inductively coupled plasma-mass spectrometry(HR-ICP-MS) with aThermo Finnigan Element2 systemfollowing the procedures described by Pretorius et al(2006) within 24 h of redissolution HFSE and LILE weremeasured in medium-resolution mode at 2000 dilutionusing a PFATeflon spray chamber washed with aqua regiafor 3min between samples REE were measured in high-resolution mode and U and Pb in low-resolution modeat 2000 dilution using a glass spray chamber washedwith 2 HNO3 between samples Total procedural blanksand reference materials (BCR-2 BHVO-2) were analyzedwith the batch of samples Indium was used as an internal
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
38
standard in all samples and standard solutionsBackground and standard solutions were analyzed afterevery five samples to detect memory effects and massdrift Results for PCIGR trace-element analyses areshown in Electronic Appendix 3 for the full set of analyzedtrace elementsSample digestion for purification of Sr Nd Hf and Pb
for column chemistry began by weighing each samplepowder (400^500mg) prior to leaching All samples wereleached with 6N HCl and placed in an ultrasonic bath for15min Samples were rinsed twice with 18 M-cm H2Obetween each leaching step (15 total) until the supernatantliquid was clear [following the method of Mahoney (1987)]Samples were then dried on a hotplate for 24 h andweighed again Sample solutions were then prepared bydissolving 100^250mg of the leached powder dissolvedin 1ml 14N HNO3 and 10ml 48 HF on a hotplate for48 h at 1308C with periodic ultrasonication Samples weredried and redissolved in 6ml 6N HCl on a hotplate for24 h and then dried Pb was separated using anion exhangecolumns and the discard was used for Sr REE and Hfseparation Nd was separated from the other REE and Hfrequired two additional purification steps Detailed proce-dures for column chemistry for separating Sr Nd and Pbat the PCIGR have been described byWeis et al (2006) andHf purification has been described byWeis et al (2007) Srand Nd isotope ratios were measured by thermal ioniza-tion mass spectrometry (TIMS) on a Thermo FinniganTriton system in static mode with relay matrix rotation ona single Ta and double Re^Ta filament respectively Fourto five filaments per barrel of 21 were occupied by stan-dards (NBS 987 for Sr and LaJolla for Nd) for each barrelwhere samples were run Sample Sr and Nd isotopic com-positions were corrected for mass fractionation using86Sr88Srfrac14 01194 and 146Nd144Ndfrac14 07219 Each samplewas then normalized using the barrel average of the refer-ence material relative to the values of143Nd144Ndfrac14 0511858 and 87Sr86Srfrac14 0710248 (Weiset al 2006) During the course of the Vancouver Islandanalyses the La Jolla Nd standard gave an average valueof 05118566 (nfrac14 7) and NBS987 standard gave an aver-age of 07102408 (nfrac1411) (2 error is reported as times106) 147Sm144Nd errors are 15 or 0006 USGeological Survey (USGS) reference material BHVO-2was processed with the samples and yielded Sr and Nd iso-topic ratios of 07034607 and 05129786 respectivelyThese are in agreement with the published values of070347920 and 051298411 respectively (Weis et al2006)
Pb and Hf isotopic compositions were analyzed by mul-tiple collector-inductively coupled plasma-mass spectrome-try (MC-ICP-MS) on a static multi-collection on a NuPlasma (Nu Instruments) system The detailed analyticalprocedure for Pb isotopic analyses on the Nu Plasma atthe PCIGR has been described by Weis et al (2006) Theconfiguration for Pb analyses allows for collection of PbTl and Hg togetherTl and Hg are used to monitor instru-mental mass discrimination and isobaric overlap repec-tively All sample solutions were analyzed withapproximately the same PbTl ratio (4) as the referencematerial NIST SRM 981 To accomplish this a small ali-quot of each sample solution from the Pb columns was ana-lyzed on the Element2 to determine the precise amount ofPb available for analysis on the Nu Plasma The SRM 981standard was run after every two samples on the NuPlasma During the time samples were run analyses ofthe SRM 981 Pb reference material gave values of206Pb204Pbfrac1416940322 207Pb204Pbfrac1415495823and 208Pb204Pbfrac14 36713164 (nfrac14 61 2 error is reportedas times 104) these values are in excellent agreement withreported TIMS triple-spike values of Galer amp Abouchami(1998) Fractionation-corrected Pb isotopic ratios were fur-ther corrected by the sample-standard bracketing methodor the ln^ln correction method described by White et al(2000) and Blichert-Toft et al (2003) Leached powders ofUSGS reference material BHVO-2 yielded Pb isotopicratios of 206Pb204Pbfrac14186454 8 207Pb204Pbfrac14 1549105 and 208Pb204Pbfrac14 38222514 and BCR-2yielded 206Pb204Pbfrac141880466 207Pb204Pbfrac14 1562518 and 208Pb204Pbfrac14 3883496 (2 error isreported as times 104) These values are in agreement withthose for leached residues of BHVO-2 and BCR-2 fromWeis et al (2006)Hf isotopic compositions were analyzed following the
procedures detailed byWeis et al (2007) The configurationfor Hf analyses monitored Lu mass 175 andYb mass 172 toallow for interference correction to masses 174 and 176 Hfisotopic ratios were normalized internally for mass fractio-nation to a 179Hf177Hf ratio of 07325 using an exponentialcorrection Standards were run after every two samplesand sample results were normalized to the ratio of the in-run daily average and a 176Hf177Hf ratio for JMC-475 of0282160 During the course of analyses the Hf standardJMC-475 gave an average value 02821679 (nfrac14130)USGS reference materials BCR-2 and BHVO-2 were pro-cessed with the samples and yielded Hf isotopic ratios of02828675 and 02831005 respectively Publishedvalues for BCR-2 and BHVO-2 are 02828717 and0283104 8 respectively (Weis et al 2007)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
39
basalts (LaYbCNfrac14 21^29 mean 2508) Two LREE-depleted coarse-grained mafic rocks from the SchoenLake area have similar REE patterns (LaYbCNfrac14 07) tothe picrites from the Keogh Lake area indicating that thisdistinctive suite of rocks is exposed as far south asSchoen Lake that is 75 km away (Fig 7) The anomalouspillowed flow from the Karmutsen Range has lower LaYbCN (13^18) than the main group of tholeiitic basaltsfrom the Karmutsen Range (Fig 7) The mineralized sillfrom the Schoen Lake area is distinctly LREE-enriched(LaYbCNfrac14 33) with the highest REE abundances(LaCNfrac14 940) of all Karmutsen samples this may reflectcontamination by adjacent sediment also evident in thin-section where sulfide blebs are cored by quartz (Greeneet al 2006)The primitive mantle-normalized trace-element pat-
terns (Fig 7) and trace-element variations and ratios(Fig 8) highlight the differences in trace-element
concentrations between the four main groups ofKarmutsen flood basalts onVancouver Island The tholeii-tic basalts from all three areas have relatively smooth par-allel trace-element patterns some samples have smallpositive Sr anomalies relative to Nd and Sm and manysamples have prominent negative K anomalies relative toU and Nb reflecting K loss during alteration (Fig 7) Thelarge ion lithophile elements (LILE) in the tholeiiticbasalts (mainly Rb and K not Cs) are depleted relativeto the high field strength elements (HFSE) and LREELILE segments especially for the Schoen Lake area(Fig 7d) are remarkably parallel The picrites andhigh-MgO basalts form a tight band of parallel trace-element patterns and are depleted in HFSE and LREEwith positive Sr anomalies and relatively low Rb Thecoarse-grained mafic rocks from the Karmutsen Rangehave identical trace-element patterns to the tholeiiticbasalts Samples from the Schoen Lake area have similar
000
025
050
075
100
125
0 1 2 3 4 5
00
04
08
12
16
0 5 10 15 20 25
MgO (wt)
0
5
10
15
0 50 100 150
Hf (ppm)
Th (ppm)
NbLaNb (ppm)
Zr (ppm)
(a)
(c)
(b)
(d)
Th (ppm)
00
02
04
06
08
10
12
00 01 02 03 04 05
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
NbLa=1
4723A4 5615A12
U (ppm)
Fig 8 Whole-rock trace-element concentrations and ratios for the Karmutsen Formation [except (b) which is vs wt MgO] (a) Nb vs Zr(b) NbLa vs MgO (c) Th vs Hf (d) Th vs U There is a clear distinction between the tholeiitic basalts and picrites in Nb and Zr both inconcentration and the slope of each trend
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
19
trace-element patterns to the Karmutsen Range except forthe two LREE-depleted coarse-grained mafic rocks andthe mineralized sill (Fig 7d)
Sr^Nd^Hf^Pb isotopic compositionsA total of 19 samples from the four main groups ofthe Karmutsen Formation were selected for radiogenicisotopic geochemistry on the basis of major- and trace-element variation stratigraphic position and geographicallocation The samples have overlapping age-correctedHf and Nd isotope ratios and distinct ranges of Sr isotopiccompositions (Fig 9) The tholeiitic basalts picriteshigh-MgO basalt and coarse-grained mafic rocks have
initial eHffrac14thorn87 to thorn126 and eNdfrac14thorn66 to thorn88 cor-rected for in situ radioactive decay since 230 Ma (Fig 9Tables 3 and 4) All the Karmutsen samples form a well-defined linear array in a Lu^Hf isochron diagram corre-sponding to an age of 24115 Ma (Fig 9) This is withinerror of the accepted age of the Karmutsen Formationand indicates that the Hf isotope systematics behaved as aclosed system since c 230 Ma The anomalous pillowedflow has similar initial eHf (thorn98) and slightly lower initialeNd (thorn62) compared with the other Karmutsen samplesThe picrites and high-MgO basalt have higher initial87Sr86Sr (070398^070518) than the tholeiitic basalts(initial 87Sr86Srfrac14 070306^070381) and the coarse-grained mafic rocks (initial 87Sr86Srfrac14 070265^070428)
05128
05129
05130
05131
05132
015 017 019 021 023 025
02829
02830
02831
02832
02833
02834
000 002 004 006 008 010
4
6
8
10
0702 0703 0704 0705 07066
8
10
12
14
5 6 7 8 9 10
12
0
1
2
3
4
5
6
7
0702 0703 0704 0705 0706
(d)
LOI (wt )
87Sr 86Sr
4723A2
Nd
143Nd144Nd
147Sm144Nd
87Sr 86Sr (230 Ma)
(230 Ma)
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
(a)
(c)
176Hf177Hf
Age=241 plusmn 15 Ma=+990 plusmn 064
176Lu177Hf
Hf (230 Ma)
Nd (230 Ma)
(e)
Hf (i)
(b)
Fig 9 Whole-rock Sr Nd and Hf isotopic compositions for the Karmutsen Formation (a) 143Nd144Nd vs 147Sm144Nd (b) 176Hf177Hf vs176Lu177Hf The slope of the best-fit line for all samples corresponds to an age of 24115 Ma (c) Initial eNd vs 87Sr86Sr Age-correction to230 Ma (d) LOI vs 87Sr86Sr (e) Initial eHf vs eNd Average 2 error bars are shown in a corner of each panel Complete chemical duplicatesshown inTables 3 and 4 (samples 4720A7 and 4722A4) are circled in each plot
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
overlap the ranges of the picrites and tholeiitic basalts(Fig 9 Table 3)The measured Pb isotopic compositions of the tholeiitic
basalts are more radiogenic than those of the picrites andthe most magnesian Keogh Lake picrites have the leastradiogenic Pb isotopic compositions The range ofinitial Pb isotope ratios for the picrites is206Pb204Pbfrac1417868^18580 207Pb204Pbfrac1415547^15562and 208Pb204Pbfrac14 37873^38257 and the range for thetholeiitic basalts is 206Pb204Pbfrac1418782^19098207Pb204Pbfrac1415570^15584 and 208Pb204Pbfrac14 37312^38587 (Fig 10 Table 5) The coarse-grained mafic rocksoverlap the range of initial Pb isotopic compositions forthe tholeiitic basalts with 206Pb204Pbfrac1418652^19155207Pb204Pbfrac1415568^15588 and 208Pb204Pbfrac14 38153^38735 One high-MgO basalt has the highest initial206Pb204Pb (19242) and 207Pb204Pb (15590) and thelowest initial 208Pb204Pb (37676) The Pb isotopic ratiosfor Karmutsen samples define broadly linearrelationships in Pb isotope plots The age-corrected Pb
isotopic compositions of several picrites have been affectedby U and Pb (andor Th) mobility during alteration(Fig 10)
ALTERAT IONThe Karmutsen basalts have retained most of their origi-nal igneous structures and textures however secondaryalteration and low-grade metamorphism have producedzeolitic- and prehnite^pumpellyite-bearing mineral assem-blages (Table 1 Cho et al 1987) Alteration and metamor-phism have primarily affected the distribution of the LILEand the Sr isotopic systematics The Keogh Lake picriteshave been affected more by alteration than the tholeiiticbasalts and have higher LOI (up to 55wt ) variableLILE and higher measured Sr Nd and Hf and lowermeasured Pb isotope ratios than the tholeiitic basalts Srand Pb isotope compositions for picrites and tholeiiticbasalts show a relationship with LOI (Figs 9 and 10)whereas Nd and Hf isotopic compositions do not correlate
Table 3 Sr and Nd isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Rb Sr 87Sr86Sr 2m87Rb86Sr 87Sr86Srt Sm Nd 143Nd144Nd 2m eNd
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses carried out at the PCIGR thecomplete trace element analyses are given in Electronic Appendix 3
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
21
with LOI (not shown) The relatively high apparent initialSr isotopic compositions for the high-MgO lavas (up to07052) probably resulted from post-magmatic loss of Rbor an increase in 87Sr86Sr through addition of seawater Sr(eg Hauff et al 2003) High-MgO lavas from theCaribbean plateau have similarly high 87Sr86Sr comparedwith basalts (Recurren villon et al 2002)The correction for in situdecay on initial Pb isotopic ratios has been affected bymobilization of U and Th (and Pb) in the whole-rockssince their formation A thorough acid leaching duringsample preparation was used that has been shown to effec-tively remove alteration phases (Weis et al 2006 NobreSilva et al in preparation) Whole-rock SmNd ratiosand age-correction of Nd isotopes may be affected bythe leaching procedure for submarine rocks older than50 Ma (Thompson et al 2008 Fig 9) there ishowever very little variation in Nd isotopic compositionsof the picrites or tholeiitic basalts and this system does notappear to be disturbed by alteration The HFSE abun-dances for tholeiitic and picritic basalts exhibit clear
linear correlations in binary diagrams (Fig 8) whereasplots of LILE vs HFSE are highly scattered (not shown)because of the mobility of some of the alkali and alkalineearth LILE (Cs Rb Ba K) and Sr for most samplesduring alterationThe degree of alteration in the Karmutsen samples does
not appear to be related to eruption environment or depthof burial Pillow rims are compositionally different frompillow cores (Surdam1967 Kuniyoshi1972) and aquagenetuffs are chemically different from isolated pillows withinpillow breccias however submarine and subaerial basaltsexhibit a similar degree of alterationThere is no clear cor-relation between the submarine and subaerial basalts andsome commonly used chemical alteration indices (eg BaRb vs K2OP2O5 Huang amp Frey 2005) There is also nodefinitive correlation between the petrographic alterationindex (Table 1) and chemical alteration indices although10 of 13 tholeiitic basalts with the highest K and LILEabundances have the highest petrographic alterationindex of three (seeTable 1)
Table 4 Hf isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Lu Hf 176Hf177Hf 2m eHf176Lu177Hf 176Hf177Hft eHf(t)
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
OLIV INE ACCUMULAT ION INH IGH-MgO AND PICR IT IC LAVASGeochemical trends and petrographic characteristics indi-cate that accumulation of olivine played an important rolein the formation of the high-MgO basalts and picrites fromthe Keogh Lake area on northern Vancouver Island TheKeogh Lake samples show a strong linear correlation inplots of Al2O3 TiO2 ScYb and Ni vs MgO and many ofthe picrites have abundant clusters of olivine pseudomorphs(Table1 Fig11)There is a strong linear correlationbetween
modal per cent olivine and whole-rock magnesium contentmagnesium contents range between 10 and 19wt MgOand the proportion of olivine phenocrysts in these samplesvaries from 0 to 42 vol (Fig 11)With a few exceptionsboth the size range and median size of the olivine pheno-crysts in most samples are comparable The clear correla-tion between the proportion of olivine phenocrysts andwhole-rock magnesium contents indicates that accumula-tion of olivine was directly responsible for the high MgOcontents (410wt ) in most of the picritic lavas
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
207Pb204Pb
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
208Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
(a) (b)
(c) (e)
4723A24723A2
4723A2
4723A2
4722A4
4723A4
4722A4
4723A4
4723A134723A3
4723A3
4723A13
1556
1558
1560
1562
1564
1566
1568
186 190 194 198 202 206 2101550
1552
1554
1556
1558
1560
178 180 182 184 186 188 190 192 194
380
384
388
392
396
186 190 194 198 202 206 210374
378
382
386
390
178 180 182 184 186 188 190 192 194
206Pb204Pb(d)
LOI (wt )
0
1
2
3
4
5
6
7
186 190 194 198 202 206 210
4723A2
Fig 10 Pb isotopic compositions of leached whole-rock samples measured by MC-ICP-MS for the Karmutsen Formation Error bars are smal-ler than symbols (a) Measured 207Pb204Pb vs 206Pb204Pb (b) Initial 207Pb204Pb vs 206Pb204Pb Age-correction to 230 Ma (c) Measured208Pb204Pb vs 206Pb204Pb (d) LOI vs 206Pb204Pb (e) Initial 208Pb204Pb vs 206Pb204Pb Complete chemical duplicates shown in Table 5(samples 4720A7 and 4722A4) are circled in each plot The dashed lines in (b) and (e) show the differences in age-corrections for two picrites(4723A4 4722A4) using the measured UTh and Pb concentrations for each sample and the age-corrections when concentrations are used fromthe two picrites (4723A3 4723A13) that appear to be least affected by alteration
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
23
DISCUSS IONThe compositional and spatial development of the eruptedlava sequences of the Wrangellia oceanic plateau onVancouver Island provide information about the evolutionof a mantle plume upwelling beneath oceanic lithospherein an analogous way to the interpretation of continentalflood basalts The Caribbean and Ontong Java plateausare the two primary examples of accreted parts of oceanicplateaus that have been studied to date and comparisonscan be made with the Wrangellia oceanic plateauHowever the Wrangellia oceanic plateau is a distinctiveoccurrence of an oceanic plateau because it formed on thecrust of a Devonian to Mississippian intra-oceanic islandarc overlain by Mississippian to Permian limestones andpelagic sediments The nature and thickness of oceaniclithospheric mantle are considered to play a fundamentalrole in the decompression and evolution of a mantleplume and thus the compositions of the erupted lavasequences (Greene et al 2008) The geochemical and
stratigraphic relationships observed in the KarmutsenFormation onVancouver Island and the focus of the subse-quent discussion sections provide constraints on (1) thetemperature and degree of melting required to generatethe primary picritic magmas (2) the composition of themantle source (3) the depth of melting and residual min-eralogy in the source region (4) the low-pressure evolutionof the Karmutsen tholeiitic basalts (5) the growth of theWrangellia oceanic plateau
Melting conditions and major-elementcomposition of the primary magmasThe primary magmas to most flood basalt provinces arebelieved to be picritic (eg Cox 1980) and they have beenfound in many continental and oceanic flood basaltsequences worldwide (eg Siberia Karoo Paranacurren ^Etendeka Caribbean Deccan Saunders 2005) Near-pri-mary picritic lavas with low total alkali abundances
Table 5 Pb isotopic compositions of Karmutsen basaltsVancouver Island BC
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
require high-degree partial melting of the mantle source(eg Herzberg amp OrsquoHara 2002) The Keogh Lake picritesfrom northernVancouver Island are the best candidates foridentifying the least modified partial melts of the mantleplume source despite having accumulated olivine pheno-crysts and can be used to estimate conditions of meltingand the composition of the primary magmas for theKarmutsen FormationPrimary magma compositions mantle potential tem-
peratures and source melt fractions for the picritic lavas ofthe Karmutsen Formation were calculated from primitivewhole-rock compositions using PRIMELT1XLS software(Herzberg et al 2007) and are presented in Fig 12 andTable 6 A detailed discussion of the computationalmethod has been given by Herzberg amp OrsquoHara (2002)and Herzberg et al (2007) key features of the approachare outlined belowPrimary magma composition is calculated for a primi-
tive lava by incremental addition of olivine and compari-son with primary melts of mantle peridotite Lavas thathad experienced plagioclase andor clinopyroxene fractio-nation are excluded from this analysis All calculated pri-mary magma compositions are assumed to be derived byaccumulated fractional melting of fertile peridotite KR-4003 Uncertainties in fertile peridotite composition donot propagate to significant variations in melt fractionand mantle potential temperature melting of depletedperidotite propagates to calculated melt fractions that aretoo high but with a negligible error in mantle potential
temperature PRIMELT1XLS calculates the olivine liqui-dus temperature at 1atm which should be similar to theactual eruption temperature of the primary magmaand is accurate to 308C at the 2 level of confidenceMantle potential temperature is model dependent witha precision that is similar to the accuracy of the olivineliquidus temperature (C Herzberg personal communica-tion 2008)With regard to the evidence for accumulated olivine in
the Keogh Lake picrites it should not matter whether themodeled lava composition is aphyric or laden with accu-mulated olivine phenocrysts PRIMELT1 computes theprimary magma composition by adding olivine to a lavathat has experienced olivine fractionation in this way itinverts the effects of fractional crystallization The onlyrequirement is that the olivine phenocrysts are not acci-dental or exotic to the lava compositions Support for thisprocedure can be seen in the calculated olivine liquid-line-of-descent shown by the curved arrays with 5 olivineaddition increments (Fig 12c) Fractional crystallization ofolivine from model primary magmas having 85^95wt FeO and 153^175wt MgO will yield arrays of deriva-tive liquids and randomly sorted olivines that in most butnot all cases connect the picrites and high-MgO basalts Anewer version of the PRIMELT software (Herzberg ampAsimow 2008) can perform olivine addition to and sub-traction from lavas with highly variably olivine phenocrystcontents and yields results that converge with those shownin Fig 12
Fig 11 Relationship between abundance of olivine phenocrysts and whole-rock MgO contents for the Keogh Lake picrites (a) Tracings ofolivine grains (gray) in six high-MgO samples made from high resolution (2400 dpi) scans of petrographic thin-sections Scale bars represent10mm sample numbers are indicated at the upper left (b) Modal olivine vs whole-rock MgO Continuous line is the best-fit line for 10 high-MgO samplesThe areal proportion of olivine was calculated from raster images of olivine grains using ImageJ image analysis software whichprovides an acceptable estimation of the modal abundance (eg Chayes 1954)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
25
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
Gray lines = final melting pressure
L + Ol + Opx
07
08
09
Pressure (GPa)
10
3
4
5
6
1
SOLIDUS
01
04
0506
L + Ol
2
03
02
PeridotiteKR-4003
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
Thick black lines = initial melting pressure
Dashed lines = melt fraction
Melt Fraction
2
3 4 5 6
L+Ol+Opx
SolidusSpinel
SolidusGarnet Peridotite
7
L+Ol
0 10 20 30 4040
45
50
55
60
MgO (wt)
SiO2
(wt)
PeridotiteKR-4003
Melt Fraction
0 10 20 305
10
15
CaO(wt)
MgO (wt)
3
4 5 6 7
2
46
2
Peridotite Partial Melts
Thick black line = initial melting pressureGray lines = final melting pressure
Peridotite
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
02
Pressure (GPa)
0302
04
Pressure (GPa)
Picrite
High-MgO basalt
Tholeiitic basalt
(c)
(d)
Karmutsen flood basaltsOlivine addition modelOlivine addition (5 increment)Primary magma
PicriteHigh-MgO basaltTholeiitic basalt
(a)
(b)
800 850 900
910
920
930
940
950
Karmutsen flood basalts
Olivine addition modelOlivine addition (5 increment)Primary magma
Gray lines = Fo contents of olivine
Fig 12 Estimated primary magma compositions for three Keogh Lake picrites and high-MgO basalts (samples 4723A4 4723A135616A7) using the forward and inverse modeling technique of Herzberg et al (2007) Karmutsen compositions and modeling results areoverlain on diagrams provided by C Herzberg (a) Whole-rock FeO vs MgO for Karmutsen samples from this study Total iron estimatedto be FeO is 090 Gray lines show olivine compositions that would precipitate from liquid of a given MgO^FeO composition Black lineswith crosses show results of olivine addition (inverse model) using PRIMELT1 (Herzberg et al 2007) Gray field highlights olivinecompositions with Fo contents of 91^92 predicted to precipitate (b) FeO vs MgO showing Karmutsen lava compositions and results ofa forward model for accumulated fractional melting of fertile peridotite KR-4003 (Kettle River Peridotite Walter 1998) (c) SiO2 vsMgO for Karmtusen lavas and model results (d) CaO vs MgO showing Karmtusen lava compositions and model results To brieflysummarize the technique [see Herzberg et al (2007) for complete description] potential parental magma compositions for the high-MgOlava series were selected (highest MgO and appropriate CaO) and using PRIMELT1 software olivine was incrementally added tothe selected compositions to show an array of potential primary magma compositions (inverse model) Then using PRIMELT1 theresults from the inverse model were compared with a range of accumulated fractional melts for fertile peridotite derived fromparameterization of the experimental results for KR-4003 from Walter (1998) forward model Herzberg amp OrsquoHara (2002) A melt fractionwas sought that was unique to both the inverse and forward models (Herzberg et al 2007) A unique solution was found when there was a com-mon melt fraction for both models in FeO^MgO and CaO^MgO^Al2O3^SiO2 (CMAS) projection space This modeling assumes thatolivine was the only phase crystallizing and ignores chromite precipitation and possible augite fractionation in the mantle (Herzbergamp OrsquoHara 2002) Results are best for a residue of spinel lherzolite (not pyroxenite) The presence of accumulated olivine in samples ofthe Keogh Lake picrites used as starting compositions does not significantly affect the results because we are modeling addition of olivine(see text for further description) The tholeiitic basalts cannot be used for modeling because they are all saturated in plagthorn cpxthornolThick black line for initial melting pressure at 4 GPa is not shown in (c) for clarity Gray area in (a) indicates parental magmas thatwill crystallize olivine with Fo 91^92 at the surface Dashed lines in (c) indicate melt fraction Solidi for spinel and garnet peridotite arelabelled in (c)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
The PRIMELT1 results indicate that if the Karmutsenpicritic lavas were derived from accumulated fractionalmelting of fertile peridotite the primary magmas wouldhave contained 15^17wt MgO and 10wt CaO(Fig 12 Table 6) formed from 23^27 partial meltingand would have crystallized olivine with a Fo content of 91 (Fig 12 Table 6) Calculated mantle temperatures forthe Karmutsen picrites (14908C) indicate melting ofanomalously hot mantle (100^2008C above ambient) com-pared with ambient mantle that produces mid-ocean ridgebasalt (MORB 1280^14008C Herzberg et al 2007)which initiated between 3 and 4 GPa Several picritesappear to be near-primary melts and required very littleaddition of olivine (eg sample 93G171 with measured158wt MgO) These temperature and melting esti-mates of the Karmutsen picrites are consistent withdecompression of hot mantle peridotite in an actively con-vecting plume head (ie plume initiation model) Threesamples (picrite 5614A1 and high-MgO basalts 5614A3and 5614A5) are lower in iron than the other Karmutsenlavas with FeO in the 65^80wt range (Fig 12c) andhave MgO and FeO contents that are similar to primitiveMORB from the Sequeiros fracture zone of the EastPacific Rise (EPR Perfit et al 1996) These samples
however differ from EPR MORB in having higher SiO2
and lower Al2O3 and they are restricted geographicallyto one area (Sara Lake Fig 2)We therefore have evidence for primary magmas with
MgO contents that range from a possible low of 110wt to as high as 174wt with inferred mantle potential tem-peratures from 1340 to 15208C This is similar to the rangedisplayed by lavas from the Ontong Java Plateau deter-mined by Herzberg (2004) who found that primarymagma compositions assuming a peridotite sourcewould contain 17wt MgO and 10wt CaO(Table 6) and that these magmas would have formed by27 melting at a mantle potential temperature of15008C (first melting occurring at 36 GPa) Fitton ampGodard (2004) estimated 30 partial melting for theOntong Java lavas based on the Zr contents of primarymagmas of Kroenke-type basalt calculated by incrementaladdition of equilibrium olivine However if a considerableamount of eclogite was involved in the formation of theOntong Java plateau magmas the excess temperatures esti-mated by Herzberg et al (2007) may not be required(Korenaga 2005) The variability in mantle potential tem-perature for the Keogh Lake picrites is also similar to thewide range of temperatures that have been calculated for
Table 6 Estimated primary magma compositions for Karmutsen basalts and other oceanic plateaus or islands
Sample 4723A4 4723A13 5616A7 93G171 Average OJP Mauna Keay Gorgonaz
Ontong Java primary magma composition for accumulated fraction melting (AFM) from Herzberg (2004)yMauna Kea primary magma composition is average of four samples of Herzberg (2006 table 1)zGorgona primary magma composition for AFM for 1F fertile source from Herzberg amp OrsquoHara (2002 table 4)Total iron estimated to be FeO is 090
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
27
lavas from single volcanoes in the Galapagos Hawaii andelsewhere by Herzberg amp Asimow (2008) Those workersinterpreted the range to reflect the transport of magmasfrom both the cool periphery and the hotter axis of aplume A thermally heterogeneous plume will yield pri-mary magmas with different MgO contents and the rangeof primary magma compositions and their inferred mantlepotential temperatures for Karmutsen lavas may reflectmelting from different parts of the plume
Source of the Karmutsen Formation lavasThe Sr^Nd^Hf^Pb isotopic compositions of theKarmutsen volcanic rocks on Vancouver Island indicatethat they formed from plume-type mantle containing adepleted component that is compositionally similar to butdistinct from other oceanic plateaus that formed in thePacific Ocean (eg Ontong Java and Caribbean Fig 13)Trace element and isotopic constraints can be used to testwhether the depleted component of the KarmutsenFormation was intrinsic to the mantle plume and distinctfrom the source of MORB or the result of mixing or invol-vement of ambient depleted MORB mantle with plume-type mantle The Karmutsen tholeiitic basalts have initialeNd and
87Sr86Sr ratios that fall within the range of basaltsfrom the Caribbean Plateau (eg Kerr et al 1996 1997Hauff et al 2000 Kerr 2003) and a range of Hawaiian iso-tope data (eg Frey et al 2005) and lie within and abovethe range for the Ontong Java Plateau (Tejada et al 2004Fig 13a) Karmutsen tholeiitic basalts with the highest ini-tial eNd and lowest initial 87Sr86Sr lie within and justbelow the field for age-corrected northern EPR MORB at230 Ma (eg Niu et al1999 Regelous et al1999) Initial Hfand Nd isotopic compositions for the Karmutsen samplesare noticeably offset to the low side of the ocean islandbasalt (OIB) array (Vervoort et al 1999) and lie belowand slightly overlap the low eHf end of fields for Hawaiiand the Caribbean Plateau (Fig 13c) the Karmutsen sam-ples mostly fall between and slightly overlap a field forage-corrected Pacific MORB at 230 Ma and Ontong Java(Mahoney et al 1992 1994 Nowell et al 1998 Chauvel ampBlichert-Toft 2001) Karmutsen lava Pb isotope ratios over-lap the broad range for the Caribbean Plateau and thelinear trends for the Karmutsen samples intersect thefields for EPR MORB Hawaii and Ontong Java in208Pb^206Pb space but do not intersect these fields in207Pb^206Pb space (Fig 13d and e) The more radiogenic207Pb204Pb for a given 206Pb204Pb of the Karmutsenbasalts requires high UPb ratios at an ancient time when235U was abundant similar to the Caribbean Plateau andenriched in comparison with EPR MORB Hawaii andOntong JavaThe low eHf^low eNd component of the Karmutsen
basalts that places them below the OIB mantle array andpartly within the field for age-corrected Pacific MORBcan form from low ratios of LuHf and high SmNd over
time (eg Salters amp White 1998) MORB will develop aHf^Nd isotopic signature over time that falls below themantle array Low LuHf can also lead to low 176Hf177Hffrom remelting of residues produced by melting in theabsence of garnet (eg Salters amp Hart 1991 Salters ampWhite 1998) The Hf^Nd isotopic compositions of theKarmutsen Formation indicate the possible involvementof depleted MORB mantle however similar to Iceland(Fitton et al 2003) Nb^Zr^Y variations provide a way todistinguish between a depleted component within theplume and a MORB-like component The Karmutsenbasalts and picrites fall within the Iceland array and donot overlap the MORB field in a logarithmic plot of NbYvs ZrY (Fig 13b) using the fields established by Fittonet al (1997) Similar to the depleted component within theCaribbean Plateau (Thompson et al 2003) the trend ofHf^Nd isotopic compositions towards Pacific MORB-likecompositions is not followed by MORB values in Nb^Zr^Y systematics and this suggests that the depleted compo-nent in the source of the Karmutsen basalts is distinctfrom the source of MORB and probably an intrinsic partof the plume The relatively radiogenic Pb isotopic ratiosand REE patterns distinct from MORB also support thisinterpretationTrace-element characteristics suggest the possible invol-
vement of sub-arc lithospheric mantle in the generation ofthe Karmutsen picrites Lassiter et al (1995) suggested thatmixing of a plume-type source with eNdthorn 6 to thorn 7 witharc material with low NbTh could reproduce the varia-tions observed in the Karmutsen basalts however theyconcluded that the absence of low NbLa ratios in theirsample of tholeiitic basalts restricts the amount of arc litho-sphere involved The study of Lassiter et al (1995) wasbased on major and trace element and Sr Nd and Pb iso-topic data for a suite of 29 samples from Buttle Lake in theStrathcona Provincial Park on central Vancouver Island(Fig 1) Whereas there are no clear HFSE depletions inthe Karmutsen tholeiitic basalts the low NbLa (and TaLa) of the Karmutsen picritic lavas in this study indicatethe possible involvement of Paleozoic arc lithosphere onVancouver Island (Fig 8)
REE modeling dynamic melting andsource mineralogyThe combined isotopic and trace-element variations of thepicritic and tholeiitic Karmutsen lavas indicate that differ-ences in trace elements may have originated from differentmelting histories For example the LuHf ratios of theKarmutsen picritic lavas (mean 008) are considerablyhigher than those in the tholeiitic basalts (mean 002)however the small range of overlapping initial eHf valuesfor the two lava suites suggests that these differences devel-oped during melting within the plume (Fig 9b) Below wemodel the effects of source mineralogy and depth of
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
28
melting on differences in trace-element concentrations inthe tholeiitic basalts and picritesDynamic melting models simulate progressive decom-
pression melting where some of the melt fraction isretained in the residue and only when the degree of partialmelting is greater than the critical mass porosity and the
source becomes permeable is excess melt extracted fromthe residue (Zou1998) For the Karmutsen lavas evolutionof trace element concentrations was simulated using theincongruent dynamic melting model developed by Zou ampReid (2001) Melting of the mantle at pressures greater than1 GPa involves incongruent melting (eg Longhi 2002)
OIB array
PacificMORB(230 Ma)
(0 Ma)
EPR(230 Ma)
-4
-2
0
2
4
8
10
12
14
16
18
20
22
-4 -2 0 2 4 6 8 10 12 14
minus2
0
2
4
6
8
10
12
14
0702 0703 0704 0705 0706
IndianMORB
87Sr 86Sr (initial)
Hf (initial)
(a) (c)
Nd (initial)
Karmutsen tholeiitic basalt
HawaiiOntong JavaCaribbean Plateau
Karmutsen high-MgO lava
high-MgO lavasKarmutsen
1540
1545
1550
1555
1560
1565
175 180 185 190 195 200375
380
385
390
395
175 180 185 190 195 200
(d) (e)
EPR
EPR
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb
Karmutsen Formation (initial)
HawaiiOntong JavaCaribbean Plateau
East Pacific Rise
Karmutsen Formation (measured)
Juan de FucaGorda
ε Nd
(initial)
Karmutsen Formation (different age-correction for 3 picrites)
(0 Ma)
NbY
001
01
1
1 10
ZrY
OIB
NMORB
(b)
Iceland array
6
Fig 13 Comparison of age-corrected (230 Ma) Sr^Nd^Hf^Pb isotopic compositions for Karmutsen flood basalts onVancouver Island withage-corrected OIB and MORB and Nb^Zr^Y variation (a) Initial eNd vs 87Sr86Sr (b) NbYand ZrY variation (after Fitton et al 1997)(c) Initial eHf vs eNd (d) Measured and initial 207Pb204Pb vs 206Pb204Pb (e) Measured and initial 208Pb204Pb vs 206Pb204Pb References fordata sources are listed in Electronic Appendix 4 Most of the compiled data were extracted from the GEOROC database (httpgeorocmpch-mainzgwdgdegeoroc) OIB array line in (c) is fromVervoort et al (1999) EPR East Pacific Rise Several high-MgO samples affected bysecondary alteration were not plotted for clarity in Pb isotope plots Estimation of fields for age-corrected Pacific MORB and EPR at 230 Mawere made using parent^daughter concentrations (in ppm) from Salters amp Stracke (2004) of Lufrac14 0063 Hffrac14 0202 Smfrac14 027 Ndfrac14 071Rbfrac14 0086 and Srfrac14 836 In the NbYvs ZrYplot the normal (N)-MORB and OIB fields not including Icelandic OIB are from data com-pilations of Fitton et al (1997 2003) The OIB field is data from 780 basalts (MgO45wt ) from major ocean islands given by Fitton et al(2003) The N-MORB field is based on data from the Reykjanes Ridge EPR and Southwest Indian Ridge from Fitton et al (1997) The twoparallel gray lines in (b) (Iceland array) are the limits of data for Icelandic rocks
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
29
with olivine being produced during the reactioncpxthornopxthorn spfrac14meltthornol for spinel peridotites The mod-eling used coefficients for incongruent melting reactionsbased on experiments on lherzolite melting (eg Kinzleramp Grove 1992 Longhi 2002) and was separated into (1)melting of spinel lherzolite (2) two combined intervals ofmelting for a garnet lherzolite (gt lherz) and spinel lher-zolite (sp lherz) source (3) melting of garnet lherzoliteThe model solutions are non-unique and a range indegree of melting partition coefficients melt reactioncoefficients and proportion of mixed source componentsis possible to achieve acceptable solutions (see Fig 14 cap-tion for description of modeling parameters anduncertainties)The LREE-depleted high-MgO lavas require melting of
a depleted spinel lherzolite source (Fig 14a) The modelingresults indicate a high degree of melting (22^25) similarto the results from PRIMELT1 based on major-elementvariations (discussed above) of a LREE-depleted sourceThe high-MgO lavas lack a residual garnet signature Theenriched tholeiitic basalts involved melting of both garnetand spinel lherzolite and represent aggregate melts pro-duced from continuous melting throughout the depth ofthe melting column (Fig 14b) Trace-element concentra-tions in the tholeiitic and picritic lavas are not consistentwith low-degree melting of garnet lherzolite alone(Fig 14c) The modeling results indicate that the aggregatemelts that formed the tholeiitic basalts would haveinvolved lesser proportions of enriched small-degreemelts (1^6 melting) generated at depth and greateramounts of depleted high-degree melts (12^25) gener-ated at lower pressure (a ratio of garnet lherzolite tospinel lherzolite melt of 14 provides the best fits seeFig 14b and description in figure caption) The totaldegree of melting for the tholeiitic basalts was high(23^27) similar to the degree of partial melting in thespinel lherzolite facies for the primary melts that eruptedas the Keogh Lake picrites of a source that was also prob-ably depleted in LREE
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
(c)
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
Melting of garnet lherzolite
High-MgO lavas
Melting of spinel lherzolite
source (07DM+03PM)
source (07DM+03PM)
6 melting
30 melting
(b)
(a)
30 melting
Sam
ple
Cho
ndrit
eS
ampl
eC
hond
rite 20 melting
1 melting
Melting of garnet lherzoliteand spinel lherzolite
x gt lherz + 4x sp lherz
5 meltingx gt lherz + 4x sp lherz
Sam
ple
Cho
ndrit
e
Tholeiitic lavas
Tholeiitic lavas
High-MgO lavas
source (07DM+03PM)
Fig 14 Trace-element modeling results for incongruent dynamicmantle melting for picritic and tholeiitic lavas from the KarmutsenFormation Three steps of modeling are shown (a) melting of spinellherzolite compared with high-MgO lavas (b) melting of garnet lher-zolite and spinel lherzolite compared with tholeiitic lavas (c) meltingof garnet lherzolite compared with tholeiitic and picritic lavasShaded fields in each panel for tholeiitic basalts and picrites arelabeled Patterns with symbols in each panel are modeling results(patterns are 5 melting increments in (a) and (b) and 1 incre-ments in (c) PM primitive mantle DM depleted MORB mantle gtlherz garnet lherzolite sp lherz spinel lherzolite In (b) the ratios ofper cent melting for garnet and spinel lherzolite are indicated (x gtlherzthorn 4x sp lherz means that for 5 melting 1 melt in garnetfacies and 4 melt in spinel facies where x is per cent melting in theinterval of modeling) The ratio of the respective melt proportions inthe aggregate melt (x gt lherzthorn 4x sp lherz) was kept constant andthis ratio of melt contribution provided the best fit to the data Arange of proportion of source components (07DMthorn 03PM to
09DMthorn 01PM) can reproduce the variation in the high-MgOlavas Melting modeling uses the formulation of Zou amp Reid (2001)an example calculation is shown in their Appendix PM fromMcDonough amp Sun (1995) and DM from Salters amp Stracke (2004)Melt reaction coefficients of spinel lherzolite from Kinzler amp Grove(1992) and garnet lherzolite fromWalter (1998) Partition coefficientsfrom Salters amp Stracke (2004) and Shaw (2000) were kept constantSource mineralogy for spinel lherzolite (018cpx027opx052ol003sp) from Kinzler (1997) and for garnet lherzolite(034cpx008opx053ol005gt) from Salters amp Stracke (2004) Arange of source compositions for melting of spinel lherzolite(07DMthorn 03PM to 03DMthorn 07PM) reproduce REE patternssimilar to those of the high-MgO lavas with best fits for 22ndash25melting and 07DMthorn 03PM source composition Concentrationsof tholeiitic basalts cannot be reproduced from an entirelyprimitive or depleted source
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
30
The uncertainty in the composition of the source in thespinel lherzolite melting model for the high-MgO lavasprecludes determining whether the source was moredepleted in incompatible trace elements than the source ofthe tholeiitic lavas (see Fig 14 caption for description) It ispossible that early high-pressure low-degree melting left aresidue within parts of the plume (possibly the hotter inte-rior portion) that was depleted in trace elements (egElliott et al 1991) further decompression and high-degreemelting of such depleted regions could then have generatedthe depleted high-MgO Karmutsen lavas Shallow high-degree melting would preferentially sample depletedregions whereas deeper low-degree melting may notsample more refractory depleted regions (eg CaribbeanEscuder-Viruete et al 2007) The picrites lie at the highend of the range of Nd isotopic compositions and havelow NbZr similar to depleted Icelandic basalts (Fittonet al 2003) therefore the picrites may represent the partsof the mantle plume that were the most depleted prior tothe initiation of melting As Fitton et al (2003) suggestedthese depleted melts may only rarely be erupted withoutcombining with more voluminous less depleted melts andthey may be detected only as undifferentiated small-volume flows like the Keogh Lake picrites within theKarmutsen Formation on Vancouver IslandDecompression melting within the mantle plume initiatedwithin the garnet stability field (35^25 GPa) at highmantle potential temperatures (Tp414508C) and pro-ceeded beneath oceanic arc lithosphere within the spinelstability field (25^09 GPa) where more extensivedegrees of melting could occur
Magmatic evolution of Karmutsentholeiitic basaltsPrimary magmas in large igneous provinces leave exten-sive coarse-grained cumulate residues within or below thecrust these mafic and ultramafic plutonic sequences repre-sent a significant proportion of the original magma thatpartially crystallized at depth (eg Farnetani et al 1996)For example seismic and petrological studies of OntongJava (eg Farnetani et al 1996) combined with the use ofMELTS (Ghiorso amp Sack 1995) indicate that the volcanicsequence may represent only 40 of the total magmareaching the Moho Neal et al (1997) estimated that30^45 fractional crystallization took place for theOntong Java magmas and produced cumulate thicknessesof 7^14 km corresponding to 11^22 km of flood basaltsdepending on the presence of pre-existing oceanic crustand the total crustal thickness (25^35 km) Interpretationsof seismic velocity measurements suggest the presence ofpyroxene and gabbroic cumulates 9^16 km thick beneaththe Ontong Java Plateau (eg Hussong et al 1979)Wrangellia is characterized by high crustal velocities in
seismic refraction lines which is indicative of mafic
plutonic rocks extending to depth beneath VancouverIsland (Clowes et al 1995)Wrangellia crust is 25^30 kmthick beneath Vancouver Island and is underlain by astrongly reflective zone of high velocity and density thathas been interpreted as a major shear zone where lowerWrangellia lithosphere was detached (Clowes et al 1995)The vast majority of the Karmutsen flows are evolved tho-leiitic lavas (eg low MgO high FeOMgO) indicating animportant role for partial crystallization of melts at lowpressure and a significant portion of the crust beneathVancouver Island has seismic properties that are consistentwith crystalline residues from partially crystallizedKarmutsen magmas To test the proportion of the primarymagma that fractionated within the crust MELTS(Ghiorso amp Sack 1995) was used to simulate fractionalcrystallization using several estimated primary magmacompositions from the PRIMELT1 modeling results Themajor-element composition of the primary magmas couldnot be estimated for the tholeiitic basalts because of exten-sive plagthorn cpxthornol fractionation and thus the MELTSmodeling of the picrites is used as a proxy for the evolutionof major elements in the volcanic sequence A pressure of 1kbar was used with variable water contents (anhydrousand 02wt H2O) calculated at the quartz^fayalite^magnetite (QFM) oxygen buffer Experimental resultsfrom Ontong Java indicate that most crystallizationoccurs in magma chambers shallower than 6 km deep(52 kbar Sano amp Yamashita 2004) however some crys-tallization may take place at greater depths (3^5 kbarFarnetani et al 1996)A selection of the MELTS results are shown in Fig 15
Olivine and spinel crystallize at high temperature until15^20wt of the liquid mass has fractionated and theresidual magma contains 9^10wt MgO (Fig 15f) At1235^12258C plagioclase begins crystallizing and between1235 and 11908C olivine ceases crystallizing and clinopyr-oxene saturates (Fig 15f) As expected the addition of clin-opyroxene and plagioclase to the crystallization sequencecauses substantial changes in the composition of the resid-ual liquid FeO and TiO2 increase and Al2O3 and CaOdecrease while the decrease in MgO lessens considerably(Fig 15) The near-vertical trends for the Karmutsen tho-leiitic basalts especially for CaO do not match the calcu-lated liquid-line-of-descent very well which misses manyof the high-CaO high-Al2O3 low-FeO basalts A positivecorrelation between Al2O3CaO and SrNd indicates thataccumulation of plagioclase played a major role in the evo-lution of the tholeiitic basalts (Fig 15g) Increasing thecrystallization pressure slightly and the water content ofthe melts does not systematically improve the match ofthe MELTS models to the observed dataThe MELTS results indicate that a significant propor-
tion of the crystallization takes place before plagioclase(15^20 crystallization) and clinopyroxene (35^45
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
31
crystallization) join the crystallization sequence Theseresults suggest that445 crystallization of the primarymagmas occurred (estimated from the residual liquid percent in Fig 15f) and if the original flood basalt stratigra-phy on Vancouver Island was 6 km thick at least anequivalent amount of complementary plutonic rocksshould have crystallized The Karmutsen basalts and theirplutonic residues thus represent a significant addition ofcrust (perhaps 412 km thickness) previously thickenedduring the formation of the underlying Paleozoic arc
Growth of the Wrangellia oceanic plateauon Vancouver IslandThe extrusive part of the Wrangellia oceanic plateau onVancouver Island was constructed in a three-layersequence of mostly tholeiitic basalts with a restrictedrange of composition (Fig 16) The plateau overliesDevonian to Mississippian arc volcanic sequences andmarine sediments of Mississippian to Permian ages thatare intruded by gabbroic rocks related to the flood basaltsThe volcanic stratigraphy of theWrangellia plateau formed
00
05
10
15
20
25
30
35
40
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
16
0 2 4 6 8 10 12 14 16 18 20
10
11
12
13
14
15
16
17
18
19
20
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
0 2 4 6 8 10 12 14 16 18 20
MgO (wt)
0 10 20 30 40 50
percent of total mass
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
) Mineral proportions
Sp (e)
Ol
CpxPlag
PicriteHigh-MgO basaltTholeiitic basalt
TiO2 (wt)
MgO (wt)
Al2O3 (wt) CaO (wt)
FeO (wt)
MgO(a) (b)
(c) (d)
MgO (wt)MgO (wt)
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
)
Residual liquid ()
1220
1190
Ol + Sp
Plag+Ol+Sp Plag+Cpx+Sp
02 wt H2O
no H2O
Residual liquid and crystallizing phases
(f)
0102030405060708090100
plagioclaseaccumulation
plagioclaseaccumulation
plagioclase
accumulatio
n
10
12
14
16
0 10 20 30 40
Al2O3CaO
SrNd
tholeiitic basalts
(g)
Fig 15 Forward fractional crystallization modeling results for major elements from MELTS (Ghiorso amp Sack 1995) compared with picriticand tholeiitic lavas from the Karmutsen Formation (a^d) Major-element compositions of residual liquids compared with Karmutsen lavacompositions For clarity the results are shown only for one starting composition (open circle) using the composition of the estimated primarymagma for the high-MgO lava series from the modeling technique of Herzberg et al (2007) for sample 4723A13 as shown in Fig 12 The esti-mated primary magmas for samples 4723A3 and 93G171 yield similar results A pressure of 1 kbar was used with no H2O and 02wt H2Ocalculated at QFM black line is anhydrous and grey line is 02wt H2OThe end-point for the anhydrous result is at 78 fractionation and11308C and the end-point for the result with 02wt H2O is 75 fractionation and11108C (e) Mineral proportions of crystallized phases fromone MELTrun (sample 4723A13) with no H2O (f) Per cent residual liquid vs temperature labelled with intervals of crystallizing phases (g)Al2O3CaO vs SrNd for the tholeiitic basalts
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
32
in a single eruptive phase (c 231^225 Ma) as an emergentbasalt sequence during a deeper-water shallow-water andsubaerial stage and subsided below sea level relativelysoon after its formation (Fig 16) The deeper-water stagecontains pillowed and massive submarine flows (3000mthick) that built up the submarine volcanic edifice from theseafloor The lowest pillow basalts were emplaced onunconsolidated and lithified fine-grained siliciclastic andcarbonaceous sediments containing Daonella beds(c 235^232 Ma) that are intruded by abundantKarmutsen mafic sills (Fig 16a) The different pillowedand massive submarine flows resulted primarily from dif-ferent effusive rates and local topography Some of themassive submarine flows commonly recognized by radiat-ing columnar jointing represent master tubes for the deliv-ery of lava to distal parts of submarine flow fields Thepillowed flows which commonly have uneven topography
in modern sequences may have served as levees to themassive submarine flows rather than allowing unob-structed spreading and inflation typical of the massive sub-aerial flows that form on shallow slopes in continentalflood basalt provinces (eg Self et al 1997) The rapid suc-cession of flows is evident by the near absence of sedimentbetween submarine flows Minor volumes of picritic pillowlavas occur in the upper part of the pillowed flow sequencein areas of the plateau exposed on northern VancouverIsland (Fig 16 Nixon et al 2008)The growth of the plateau in shallow water led to
emplacement of an increasing proportion of volcaniclastics(pillow breccias and hyaloclastites) that accumulated up to1500m thick in some areas (Fig 16b Nixon et al 2008)Horizons of pillowed flows occur locally within the volca-niclastic stratigraphy Sedimentary structures (eg gradedbedding fluidization structures) and petrographic textures
Late Devonian-Mississippian arc
Daonella beds
Middle Triassic
deeper water
Late Devonian-Mississippian arc
pillowed flows massive submarine
flows
pillow brecciahyaloclastiteshallow water
~4 km
~6 km
Late Triassic limestoneinterflow limestonelenses
Karmutsen Formation
mafic intrusions sediment-sill complex
Karmutsen Formation
(a) (b)
(c) (d)
subaerial flows
Late Devonian-Mississippian arc Late Devonian-Mississippian arc
picritic pillow basalts
Miss to Permian marine sediments
sills
Fig 16 Growth of the Wrangellia oceanic plateau onVancouver Island (a) Deep-water eruptive phase of pillowed and massive submarineflows with almost no intervening sediments onto MiddleTriassic shale containing Daonella beds (c 235^232 Ma) The basement of the plateauis composed of Devonian to Mississippian arc sequences overlain by Mississippian to Permian carbonates and siliciclastic sediments(b) Shallow-water flow emplacement with increasing proportion of volcaniclastics resulted primarily from cooling^contraction granulationPicritic pillow basalts occur primarily near the transition between pillowed submarine flows and volcaniclastic units Pillowed and massiveflows occur locally within the volcaniclastic unit (c) Emergent stage of growth consists of overlapping subaerial flow fields similar to continen-tal flood basalts with pahoehoe flow structures Multi-tiered columnar jointing occurrs locally in response to ponding of flows (eg Bondre et al2004) Interflow carbonate deposits and plagioclase megacrystic flows developed locally during the waning stage of volcanism (d) Post-volcanicsubsidence and deposition of Late Carnian to Norian (c 212^223 Ma) platform carbonates on top of the plateau Units are labeled in the dia-grams Thicknesses of the underlying units are not to scale
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
33
within the volcaniclastic rocks indicate formation primar-ily via cooling^contraction granulation magma^water^steam interaction autobrecciation resedimentation andmass-wasting rather than pyroclastic eruption (Nixonet al 2008)A broad subaerial platform was constructed by emplace-
ment of subaerial lava flows during an emergent subaerialstage (Fig 16c) The subaerial unit predominantly com-prises well-layered sheet flows that formed in a similarway to continental flood basalts as hundred(s) of pahoehoeflow fields erupted on shallow slopes (Self et al 1997) how-ever volcaniclastic and pillowed flow units are locally pre-sent mostly in the upper parts of the stratigraphyLocalized ponding of flows led to the formation of multi-tiered columnar jointing in some subaerial flows (egBondre et al 2004) As volcanism waned local interflowcarbonate deposits developed and plagioclase megacrysticflows were erupted Subsidence of the plateau is recordedby the interflow lenses and deposits of platform carbonates(40^1500m thick) on top of the Karmutsen basalts thatcontain Late Carnian to Early Norian fossils (Fig 16d c216^227 Ma Nixon et al 2008)The evolution of the internal architecture of the
Wrangellia oceanic plateau was primarily related to theconditions of emplacement (ie water depth and eruptionenvironment) There does not appear to have been a sys-tematic coincident evolution or shift of lava compositionsonVancouver Island as has been recognized in most con-tinental flood basalts and parts of the Wrangellia oceanicplateau in Alaska and Yukon (Greene et al 2008 2009)In general with the exception of the picritic pillow lavasin the submarine stratigraphy and local variations betweenhigh-MgO and tholeiitic lavas there are no significantcompositional differences between lavas erupted duringthe different stages of growth of the plateau (ie submarineand subaerial flows) nor are substantial compositional dif-ferences generally found between massive submarine flowsin contact with pillowed flows tholeiitic basalt flows nearthe base of the Karmutsen Formation are compositionallysimilar to flows sampled near the top of the volcanicsequence Where it was possible to sample continuous sec-tions of volcanic stratigraphy greater than several hun-dreds of meters (eg Keogh^Maynard Lake area west ofthe Karmutsen Range Fig 2) there is no clear relationshipbetween stratigraphic position and compositionalvariation
CONCLUSIONSThe Karmutsen Formation covers extensive areas ofVancouver Island (20 000 km2) and forms part of a largeoceanic plateau erupted over a geologically short interval(c 225^231 Ma) The tripartite volcanic stratigraphyexposed on Vancouver Island is at least 6 km thick andcomprises a succession of submarine flows volcaniclastic
deposits and massive subaerial flowsThese volcanologicaldifferences are primarily related to the eruption environ-ment (deep-water shallow-water or subaerial) The rapidgrowth of the plateau prevented significant accumulationsof intervening sediment except in the uppermost stratigra-phy where isolated limestone lenses commonly associatedwith pillow basalt and hyaloclastite deposits preserve arecord of the final stages of subsidence of the plateauThe Wrangellia plateau on Vancouver Island was con-
structed dominantly of tholeiitic basalt with restrictedmajor- and trace-element and isotopic compositional var-iations Minor volumes of picritic pillow basalts eruptedlate in the submarine phase The Karmutsen basalts wereemplaced onto older Paleozoic arc volcanic sequences andlow NbLa for the picrites may indicate a contributionfrom the sub-arc lithospheric mantle Modeling of the pet-rogenesis of the Karmutsen picrites indicates melting ofanomalously hot mantle (15008C) and extensive degreesof partial melting (23^27) and is consistent with aplume initiation model Combined Sr^Nd^Hf^Pb isotopicsystematics indicate that the picritic and tholeiitic lavaswere derived from a depleted plume-type mantle sourcedistinct from the source of MORB that has compositionalsimilarities to the source of the lavas forming theCaribbean Plateau Picrites were generated by melting ofspinel lherzolite whereas the tholeiitic basalts formed asaggregate melts of both garnet and spinel lherzolite Thetholeiitic basalts underwent extensive low-pressure fractio-nation (52^3 kbar) and seismic studies indicate that someof theWrangellia crust beneathVancouver Island may cor-respond to the plutonic residues of the Karmutsen floodbasalts
ACKNOWLEDGEMENTSWe are grateful to Nick Arndt for helping us get this pro-ject started and Nick Massey for insights into VancouverIsland geology We greatly appreciate the assistance withPRIMELT modeling from Claude Herzberg We thankMikkel Schau for his insight and enthusiasm during field-work Jane Barling assisted with analyses by MC-ICP-MSThoughtful and thorough reviews of the manuscript byJohn Mahoney Fred Frey Wendy Bohrson DougalJerram and Teal Riley are greatly appreciated Fundingwas generously provided by the Rocks to Riches Programadministered by the BC amp Yukon Chamber of Mines in2004 by the BC Geological Survey and by NSERCDiscovery Grants to JSS and DW ARG was supportedby a University Graduate Fellowship at UBC
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
34
REFERENCESBarker F Brown A S Budahn J R amp Plafker G (1989) Back-arc
with frontal-arc component origin of Triassic Karmutsen basaltBritish Columbia Canada Chemical Geology 75 81^102
Blichert-Toft JWeis D Maerschalk C Agranier A amp Albare de F(2003) Hawaiian hot spot dynamics as inferred from the Hf and Pbisotope evolution of Mauna Kea volcano Geochemistry GeophysicsGeosystems 4(2) 1^27 doi1010292002GC000340
Bondre N R Duraiswami R A amp Dole G (2004) Morphologyand emplacement of flows from the Deccan Volcanic ProvinceIndia Bulletin of Volcanology 66 29^45
Carlisle D (1963) Pillow breccias and their aquagene tuffs QuadraIsland British Columbia Journal of Geology 71 48^71
Carlisle D (1972) Late Paleozoic to Mid-Triassic sedimentary-volcanic sequence on Northeastern Vancouver Island Geological
Society of Canada Paper Report of Activities 72-1 Part B 24^30Carlisle D amp SuzukiT (1974) Emergent basalt and submergent car-
bonate^clastic sequences including the Upper Triassic Dilleri andWelleri zones on Vancouver Island Canadian Journal of Earth
Sciences 11 254^279Chauvel C amp Blichert-Toft J (2001) A hafnium isotope and trace
element perspective on melting of the depleted mantle Earth and
Planetary Science Letters 190 137^151Chayes F (1954)The theory of thin-section analysis Journal of Geology
62 92^101Cho M Liou J G amp Maruyama S (1987) Transition from the zeo-
lite to prehnite^pumpellyite facies in the Karmutsen metabasitesVancouver Island British Columbia Journal of Petrology 27(2)467^494
Clowes R M Zelt C A Amor J R amp Ellis R M (1995)Lithospheric structure in the southern Canadian Cordillera froma network of seismic refraction lines Canadian Journal of Earth
Sciences 32(10) 1485^1513Coffin M F amp Eldholm O (1994) Large igneous provinces Crustal
structure dimensions and external consequences Reviews of
Geophysics 32(1) 1^36Cox K G (1980) A model for flood basalt volcanism Journal of
Petrology 21 629^650Eldholm O amp Coffin M F (2000) Large igneous provinces
and plate tectonics In Richards M A Gordon R G amp vander Hilst R D (eds) The History and Dynamics of Global
Plate Motions American Geophysical Union Geophysical Monograph 121309^326
Elliott T R Hawkesworth C J amp Grolaquo nvold K (1991) Dynamicmelting of the Iceland plume Nature 351 201^206
Escuder-Viruete J Pecurren rez-Estaucurren n A Contreras F Joubert MWeis D Ullrich T D amp Spadea P (2007) Plume mantle sourceheterogeneity through time Insights from the Duarte ComplexHispaniola northeastern Caribbean Journal of Geophysical Research112(B04203) doi1010292006JB004323
Farnetani C G Richards M A amp Ghiorso M S (1996)Petrological models of magma evolution and deep crustal structurebeneath hotspots and flood basalt provinces Earth and Planetary
Science Letters 143 81^94Fitton J G amp Godard M (2004) Origin and evolution of magmas
on the Ontong Java Plateau In Fitton J G Mahoney J JWallace P J amp Saunders A D (eds) Origin and Evolution of the
Ontong Java Plateau Geological Society London Special Publications 229151^178
Fitton J G Saunders A D Norry M J Hardarson B S ampTaylor R N (1997) Thermal and chemical structure of theIceland plume Earth and Planetary Science Letters 153 197^208
Fitton J G Saunders A D Kempton P D amp Hardarson B S(2003) Does depleted mantle form an intrinsic part of the Icelandplume Geochemistry Geophysics Geosystems 4(3) 1032 doi1010292002GC000424
Frey F A Huang S Blichert-Toft J Regelous M amp Boyet M(2005) Origin of depleted components in basalt related to theHawaiian hotspot Evidence from isotopic and incompatible ele-ment ratios Geochemistry Geophysics Geosystems 6(1) 23 doi1010292004GC000757
Galer S J G amp Abouchami W (1998) Practical application of leadtriple spiking for correction of instrumental mass discriminationMineralogical Magazine 62A 491^492
Ghiorso M S amp Sack R O (1995) Chemical mass transfer in mag-matic processes IV A revised and internally consistent thermody-namic model for the interpolation and extrapolation of liquid^solidequilibria in magmatic systems at elevated temperatures and pres-sures Contributions to Mineralogy and Petrology 119 197^212
Greene A R Scoates J S amp Weis D (2005)WrangelliaTerrane onVancouver Island Distribution of flood basalts with implicationsfor potential Ni^Cu^PGE mineralization In Grant B (ed)Geological Fieldwork 2004 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2005-1 209^220Greene A R Scoates J S Nixon G T amp Weis D (2006) Picritic
lavas and basal sills in the Karmutsen flood basalt provinceWrangellia northern Vancouver Island In Grant B (ed)Geological Fieldwork 2005 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2006-1 39^52Greene A R Scoates J S amp Weis D (2008) Wrangellia flood
basalts in Alaska A record of plume^lithosphere interaction in aLate Triassic accreted oceanic plateau Geochemistry Geophysics
Geosystems 9(Q12004) doi1010292008GC002092Greene A R Scoates J S Weis D amp Israel S (2009)
Geochemistry of triassic flood basalts from the Yukon (Canada)segment of the accreted Wrangellia oceanic plateau Lithosdoi101016jlithos200811010
Hauff F Hoernle K Tilton G Graham D W amp Kerr A C(2000) Large volume recycling of oceanic lithosphere over shorttime scales geochemical constraints from the Caribbean LargeIgneous Province Earth and Planetary Science Letters 174 247^263
Hauff F Hoernle K amp Schmidt A (2003) Sr^Nd^Pb compositionof Mesozoic Pacific oceanic crust (Site 1149 and 801 ODP Leg185)Implications for alteration of ocean crust and the input into theIzu^Bonin^Mariana subduction system Geochemistry Geophysics
Geosystems 4(8) doi1010292002GC000421Herzberg C (2004) Partial melting below the Ontong Java Plateau
In Fitton J G Mahoney J J Wallace P J amp Saunders A D(eds) Origin and Evolution of the Ontong Java Plateau Geological SocietyLondon Special Publications 229 179^183
Herzberg C (2006) Petrology and thermal structure of the Hawaiianplume from Mauna Kea volcano Nature 444(7119) 605
Herzberg C amp Asimow P D (2008) Petrology of some oceanicisland basalts PRIMELT2XLS software for primary magma cal-culation Geochemistry Geophysics Geosystems 9(Q09001) doi1010292008GC002057
Herzberg C amp OrsquoHara M J (2002) Plume-associated ultramaficmagmas of Phanerozoic age Journal of Petrology 43(10) 1857^1883
Herzberg C Asimow P D Arndt N Niu Y Lesher C MFitton J G Cheadle M J amp Saunders A D (2007)Temperatures in ambient mantle and plumes Constraints frombasalts picrites and komatiites Geochemistry Geophysics Geosystems8(Q02006) doi1010292006GC001390
Huang S amp Frey F A (2005) Trace element abundances of MaunaKea basalt from phase 2 of the Hawaii Scientific Drilling Project
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
35
Petrogenetic implications of correlations with major element con-tent and isotopic ratios Geochemistry Geophysics Geosystems 4(6)doi1010292002GC000322
Hussong D M Wipperman L K amp Kroenke L W (1979) Thecrustal structure of the Ontong Java and Manihiki OceanicPlateaus Journal of Geophysical Research 84(B11) 6003^6010
Jerram D A amp Widdowson M (2005) The anatomy of ContinentalFlood Basalt Provinces geological constraints on the processes andproducts of flood volcanism Lithos 79(3^4) 385^405
Jones D L Silberling N J amp Hillhouse J (1977)Wrangellia a dis-placed terrane in northwestern North America CanadianJournal ofEarth Sciences 14(11) 2565^2577
Kerr A C (2003) Oceanic Plateaus In Rudnick R (ed) The Crust
Treatise on GeochemistryVol 3 Oxford Elsevier Science pp 537^565Kerr A C amp Mahoney J J (2007) Oceanic plateaus Problematic
plumes potential paradigms Chemical Geology 241 332^353Kerr A CTarney J Marriner G F Klaver GT Saunders A D
amp Thirlwall M F (1996) The geochemistry and petrogenesis ofthe Late-Cretaceous picrites and basalts of Curacao NetherlandsAntilles a remnant of an oceanic plateau Contributions to
Mineralogy and Petrology 124(1) 29^43Kerr A C Marriner G F Tarney J Nivia A Saunders A D
Thirlwall M F amp Sinton A C (1997) Cretaceous basaltic ter-ranes in Western Colombia Elemental chronological and Sr-Ndisotopic constraints on petrogenesis Journal of Petrology 38(6)677^702
Kinzler R J (1997) Melting of mantle peridotite at pressuresapproaching the spinel to garnet transition Application to mid-ocean ridge basalt petrogenesis Journal of Geophysical Research
102(B1) 853^874Kinzler R J amp Grove T L (1992) Primary magmas of mid-ocean
ridge basalts 1 Experiments and methods Journal of Geophysical
Research 97(B5) 6885^6906Korenaga J (2005) Why did not the Ontong Java Plateau form sub-
aerially Earth and Planetary Science Letters 234 385^399Kuniyoshi S (1972) Petrology of the Karmutsen (Volcanic) Group
northeastern Vancouver Island British Columbia PhD disserta-tion University of California Los Angeles 242 pp
Lassiter J C DePaolo D J amp Mahoney J J (1995) Geochemistry ofthe Wrangellia flood basalt province Implications for the role ofcontinental and oceanic lithosphere in flood basalt genesis Journalof Petrology 36(4) 983^1009
LeBas M J (2000) IUGS reclassification of the high-Mg and picriticvolcanic rocks Journal of Petrology 41(10) 1467^1470
Longhi J (2002) Some phase equilibrium systematics of lherzolitemelting I Geochemistry Geophysics Geosystems 3(3) doi1010292001GC000204
MacDonald G A amp Katsura T (1964) Chemical composition ofHawaiian lavas Journal of Petrology 5 82^133
Mahoney J J (1987) An isotopic survey of Pacific oceanic plateausimplications for their nature and origin In Keating B HFryer P Batiza R amp Boehlert GW (eds) Seamounts Islands andAtolls American Geophysical Union Geophysical Monograph 43 207^220
Mahoney J LeRoex A P Peng Z Fisher R L amp Natland J H(1992) Southwestern limits of Indian Ocean Ridge mantle and theorigin of low 206Pb204Pb mid-ocean ridge basalt isotope systema-tics of the central Southwest Indian Ridge (178^508E) Journal ofGeophysical Research 97(B13) 19771^19790
Mahoney J J Sinton J M Kurz M D MacDougall J DSpencer K J amp Lugmair GW (1994) Isotope and trace elementcharacteristics of a super-fast spreading ridge East Pacific rise13^238S Earth and Planetary Science Letters 121 173^193
Massey N W D (1995a) Geology and mineral resources of theAlberni^Nanaimo Lakes sheet Vancouver Island 92F1W 92F2Eand part of 92F7E British Columbia Ministry of Energy Mines and
Petroleum Resources Paper 1992-2 132 ppMassey N W D (1995b) Geology and mineral resources of the
Cowichan Lake sheet Vancouver Island 92C16 British Columbia
Ministry of Energy Mines and Petroleum Resources Paper 1992-3 112 ppMassey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005a) Digital Geology Map of British Columbia Tile NM9 MidCoast BC British Columbia Ministry of Energy and Mines Geofile
2005-2Massey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005b) Digital Geology Map of British Columbia Tile NM10Southwest BC British Columbia Ministry of Energy and Mines Geofile
2005-3McDonough W F amp Sun S (1995) The composition of the Earth
Chemical Geology 120 223^253Monger JW H amp Journeay J M (1994) Basement geology and tec-
tonic evolution of theVancouver region In Monger JW H (ed)Geology and Geological Hazards of the Vancouver Region Southwestern
British Columbia Geological Survey of Canada Bulletin 481 3^25Muller J E (1977) Geology of Vancouver Island Geological Survey of
Canada Open File Map 463Muller J E (1980)The Paleozoic Sicker Group of Vancouver Island British
Columbia Geological Survey of Canada Paper 79-30 22 ppMuller J E Northcote K E amp Carlisle D (1974) Geology and
mineral deposits of Alert Bay^Cape Scott map area VancouverIsland British Columbia Geological Survey of Canada Paper 74-8 77
Neal C R Mahoney J J Kroenke L W Duncan R A ampPetterson M G (1997) The Ontong^Java Plateau InMahoney J J amp Coffin M F (eds) Large Igneous Provinces
Continental Oceanic and Planetary FloodVolcanism American Geophysical
Union Geophysical Monograph 100 183^216Niu Y Collerson K D Batiza R Wendt J I amp Regelous M
(1999) Origin of enriched-typed mid-ocean ridge basalt at ridgesfar from mantle plumes The East Pacific Rise at 11820rsquoN Journalof Geophysical Research 104 7067^7088
Nixon G T amp Orr A J (2007) Recent revisions to the EarlyMesozoic stratigraphy of Northern Vancouver Island (NTS 102I092L) and metallogenic implicatinos British Columbia InGrant B (ed) Geological Fieldwork 2006 British Columbia Ministry of
Energy Mines and Petroleum Resources Paper 2007-1 163^177Nixon GT Hammack J L KoyanagiV M Payie G J Haggart
JW Orchard M JTozerT Friedman R M Archibald D APalfy J amp Cordey F (2006a) Geology of the Quatsino^PortMcNeill area northernVancouver Island British Columbia Ministry
of Energy Mines and Petroleum Resources Geoscience Map 2006-2 scale150 000
Nixon G T Kelman M C Stevenson D Stokes L A ampJohnston K A (2006b) Preliminary geology of the Nimpkishmap area (NTS 092L07) northern Vancouver Island BritishColumbia In Grant B amp Newell J M (eds) Geological Fieldwork2005 British Columbia Ministry of Energy Mines and Petroleum Resources
Paper 2006-1 135^152Nixon G T Laroque J Pals A Styan J Greene A R amp
Scoates J S (2008) High-Mg lavas in the Karmutsen flood basaltsnorthernVancouver Island (NTS 092L) Stratigraphic setting andmetallogenic significance In Grant B (ed) Geological Fieldwork
2007 British Columbia Ministry of Energy Mines and Petroleum
Resources Paper 2008-1 175^190Nowell G M Kempton P D Noble S R Fitton J G
Saunders A Mahoney J J amp Taylor R N (1998) High precisionHf isotope measurements of MORB and OIB by thermal ionisation
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
36
mass spectrometry insights into the depleted mantle Chemical
Geology 149 211^233Parrish R R amp McNicoll V J (1992) U^Pb age determinations
from the southern Vancouver Island area British Columbia InRadiogenic Age and Isotopic Studies Report 5 Geological Survey of
Canada Paper 91-2 79^86Peate DW (1997) The Parana^Etendeka Province In Coffin M F
amp Mahoney J (eds) Large Igneous Provinces Continental Oceanic andPlanetary Flood Volcanism American Geophysical Union Geophysical
Monograph 100 217^245Perfit M R Fornari D J Ridley W I Kirk P D Casey J
Kastens K A Reynolds J R Edwards M Desonie DShuster R amp Paradis S (1996) Recent volcanism in the Siqueirostransform fault picritic basalts and implications for MORBmagma genesis Earth and Planetary Science Letters 141(1^4) 91^108
Pretorius W Weis D Williams G Hanano D Kieffer B ampScoates J S (2006) Complete trace elemental characterization ofgranitoid (USGSG-2GSP-2) reference materials by high resolutioninductively coupled plasma-mass spectrometry Geostandards and
Geoanalytical Research 30(1) 39^54Regelous M Niu Y Wendt J I Batiza R Greig A amp
Collerson K D (1999) Variations in the geochemistry of magma-tism on the East Pacific Rise at 10830rsquoN since 800 ka Earth and
Planetary Science Letters 168 45^63Recurren villon S Chauvel C Arndt N T Pik R Martineau F
Fourcade S amp Marty B (2002) Heterogeneity of the Caribbeanplateau mantle source Sr O and He isotopic compositions of oli-vine and clinopyroxene from Gorgona Island Earth and Planetary
Science Letters 205(1^2) 91Rhodes J M (1996) Geochemical stratigraphy of lava flows samples
by the Hawaii Scientific Drilling Project Journal of Geophysical
Research 101(B5) 11729^11746Rhodes J M amp Vollinger M J (2004) Composition of basaltic lavas
sampled by phase-2 of the Hawaii Scientific Drilling ProjectGeochemical stratigraphy and magma types Geochemistry
Geophysics Geosystems 5(3) doi1010292002GC000434Richards M A Jones D L Duncan R A amp DePaolo D J (1991)
A mantle plume initiation model for the Wrangellia flood basaltand other oceanic plateaus Science 254 263^267
Salters V J M amp Hart S R (1991) The mantle sources of oceanridges islands and arcs the Hf-isotope connection Earth and
Planetary Science Letters 104 364^380Salters V J M amp Stracke A (2004) Composition of the depleted
SaltersV J amp WhiteW (1998) Hf isotope constraints on mantle evo-lution Chemical Geology 145 447^460
SanoT amp Yamashita S (2004) Experimental petrology of basementlavas from Ocean Drilling Program Leg 192 implications for dif-ferentiation processes in Ontong Java Plateau magmas InFitton J G Mahoney J JWallace P J amp Saunders A D (eds)Origin and Evolution of the Ontong Java Plateau Geological Society
London Special Publications 229 185^218Saunders A D (2005) Large igneous provinces Origin and environ-
mental consequences Elements 1 259^263Self S ThordarsonT amp Keszthely L (1997) Emplacement of conti-
nental flood basalt lava flows In Mahoney J J amp Coffin M F(eds) Large Igneous Provinces Continental Oceanic and Planetary Flood
Volcanism American Geophysical Union Geophysical Monograph 100381^410
Shaw D M (2000) Continuous (dynamic) melting theory revisitedCanadian Mineralogist 38(5) 1041^1063
Sluggett C L (2003) Uranium^lead age and geochemical constraintson Paleozoic and Early Mesozoic magmatism in WrangelliaTerrane Saltspring Island British Columbia BSc thesisUniversity of British ColumbiaVancouver 84 pp
Storey M Mahoney J J amp Saunders A D (1997) Cretaceousbasalts in Madagascar and the transition between plume and con-tinental lithospheric mantle sources In Mahoney J J ampCoffin M F (eds) Large Igneous Provinces Continental Oceanic and
Planetary Flood Volcanism Aamerican Geophysical Union Geophysical
Monograph 100 95^122Surdam R C (1967) Low-grade metamorphism of the Karmutsen
Group PhD dissertation University of California Los Angeles288 pp
Sutherland-Brown A Yorath C J Anderson R G amp Dom K(1986) Geological maps of southern Vancouver IslandLITHOPROBE I 92C10 11 14 16 92F1 2 7 8 Geological Survey ofCanada Open File 1272
Tejada M L G Mahoney J J Castillo P R Ingle S P Sheth HC amp Weis D (2004) Pin-pricking the elephant evidence on theorigin of the Ontong Java Plateau from Pb^Sr^Hf^Nd isotopiccharacteristics of ODP Leg 192 basalts In Fitton J GMahoney J J Wallace P J amp Saunders A D (eds) Origin and
Evolution of the Ontong Java Plateau Geological Society London Special
Publications 229 133^150Thompson P M E Kempton P D amp Kerr A C (2008) Evaluation
of the effects of alteration and leaching on Sm^Nd and Lu^Hf sys-tematics in submarine mafic rocks Lithos 104(1^4) 164^176
Thompson P M E Kempton P D White R V Kerr A CTarney J Saunders A D Fitton J G amp McBirney A (2003)Hf-Nd isotope constraints on the origin of the CretaceousCaribbean plateau and its relationship to the Galapagos plumeEarth and Planetary Science Letters 217(1^2) 59^75
Vervoort J D Patchett P Blichert-Toft J amp Albare de F (1999)Relationships between Lu^Hf and Sm^Nd isotopic systems in theglobal sedimentary system Earth and Planetary Science Letters 16879^99
Walter M J (1998) Melting of garnet peridotite and the origin ofkomatiite and depleted lithosphere Journal of Petrology 39(1) 29^60
Weis D Kieffer B Maerschalk C Barling J de Jong JWilliams G A Hanano D Mattielli N Scoates J SGoolaerts A Friedman R A amp Mahoney J B (2006) High-pre-cision isotopic characterization of USGS reference materials byTIMS and MC-ICP-MS Geochemistry Geophysics Geosystems
7(Q08006) doi1010292006GC001283Weis D Kieffer B Hanano D Silva I N Barling J
Pretorius W Maerschalk C amp Mattielli N (2007) Hf isotopecompositions of US Geological Survey reference materialsGeochemistry Geophysics Geosystems 8(Q06006) doi1010292006GC001473
Wheeler J O amp McFeely P (1991) Tectonic assemblage map of theCanadian Cordillera and adjacent part of the United States ofAmerica Geological Survey of Canada Map 1712A
WhiteW M Albare de F amp Tecurren louk P (2000) High-precision analy-sis of Pb isotope ratios by multi-collector ICP-MS Chemical Geology167 257^270
Yorath C J Sutherland Brown A amp Massey N W D (1999)LITHOPROBE southern Vancouver Island British Columbia Geological
Survey of Canada Bulletin 498 145 ppZou H (1998) Trace element fractionation during modal and
non-model dynamic melting and open-system melting Amathematical treatment Geochimica et Cosmochimica Acta 62(11)1937^1945
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
37
Zou H amp Reid M R (2001) Quantitative modeling of trace elementfractionation during incongruent dynamic melting Geochimica et
Cosmochimica Acta 65(1) 153^162
APPENDIX SAMPLEPREPARAT ION ANDANALYT ICAL METHODSOnly the freshest rocks in the field were sampled and onlythe least altered samples were selected for chemical andisotopic analysis based on thorough petrographic inspec-tion Fifty-nine of the 129 collected samples were crushed(400 g) into small pieces 52mm in diameter in aRocklabs hydraulic piston crusher between WC plates tominimize contamination The coarse-crush was mixedand 100 g was powdered in a planetary mill using agatejars and balls cleaned with quartz sand between samples
ActLabs analytical methodsThe major- and trace-element compositions of the whole-rock powders were determined at Activation LaboratoriesLtd (Actlabs) in Ancaster Ontario Analytical techniquesand detection limits are also available from Actlabs (httpwwwactlabscommethsub_code4erehtm) The analyticalmethod for each of the elements analyzed is indicated inTable 2 For the major elements a 02 g sample was mixedwith a lithium metaborate^lithium tetraborate mixtureand fused in a graphite crucible The molten mixture waspoured into a 5 HNO3 solution and shaken until dis-solved (30min) The samples were analyzed for majoroxides and selected trace elements by inductively coupledplasma optical emission spectrometry (ICP-OES) on acombination simultaneous^sequential Thermo Jarrell-AshEnviro II system Internal calibration was achieved usinga variety of international reference materials (eg W-2BIR-1 DNC-1) and independent control samplesAdditional trace elements were analyzed by both instru-mental neutron activation analysis (INAA) and induc-tively couple plasma mass spectrometry (ICP-MS) ForINAA15^25 g of sample was weighed into small polyeth-ylene vials and irradiated with a control international ref-erence material CANMET WMS-1 and NiCr flux wires ata thermal neutron flux of 71012 neutronscm2 per s in theMcMaster Nuclear Reactor Following a 7 day waitingperiod the samples were measured on an Ortec high-purity Ge detector linked to a Canberra Series 95 multi-channel analyzer Activities for each element weredecay- and weight-corrected and compared with a detectorcalibration developed from multiple international certifiedreference materials For ICP-MS 025 g of sample wasdigested in HF followed by HNO3^HClO4 heated andtaken to dryness The samples were brought back into solu-tion with HCl Samples were analyzed using a PerkinElmer Optima 3000 ICP In-lab standards or certified ref-erence materials (eg W-2 BIR-1 DNC-1) were used for
quality control A total of 15 blind duplicates were ana-lyzed from the same sample powders to assessreproducibility
University of Massachusetts analyticalmethodsFourteen sample duplicate powders taken from the same100 g sample aliquot (high-MgO basalts and picrites)were also analyzed at the Ronald B Gilmore X-RayFluorescence (XRF) Laboratory at the University ofMassachusetts Major elements were measured on a fusedLa-bearing lithium borate glass disc using a SiemensMRS-400 spectrometer Trace element concentrations(Rb Sr Ba Ce Nb Zr Y Pb Zn Ga Ni Cr V) weremeasured on a separate powder pellet using a PhilipsPW2400 sequential spectrometer using a Rh tube LOIand ferrous iron measurements were made as describedby Rhodes amp Vollinger (2004) Precision and accuracy esti-mates for the data have been described by Rhodes (1996)and Rhodes amp Vollinger (2004) Results for each sampleare the average of two separate analyses and are shown inElectronic Appendix 2 for a comparison of analyses ofduplicate powders using XRFand ICP-MS
PCIGR analytical methodsA subset of 19 samples was selected for high-precisiontrace-element analysis and Sr Nd Pb and Hf isotopicanalysis at the Pacific Centre for Isotopic andGeochemical Research (PCIGR) at the University ofBritish Columbia (UBC) Samples were selected from the59 samples analyzed for whole-rock chemistry at ActLabsbased on major- and trace-element chemistry alteration(low LOI and petrographic alteration index) and samplelocation Samples were prepared for trace-element analysisat the PCIGR by the technique described by Pretorius et al(2006) on unleached rock powders Sample powders(100mg) were weighed in 7ml screw-top Savillex bea-kers and dissolved in 1ml 14N HNO3 and 5ml 48 HFon a hotplate for 48 h at 1308C with periodic ultrasonica-tion Samples were dried and redissolved in 6ml 6N HClon a hotplate for 24 h and then dried and redissolved in1ml concentrated HNO3 for 24 h before final dryingTrace element abundances were measured by high-resolution inductively coupled plasma-mass spectrometry(HR-ICP-MS) with aThermo Finnigan Element2 systemfollowing the procedures described by Pretorius et al(2006) within 24 h of redissolution HFSE and LILE weremeasured in medium-resolution mode at 2000 dilutionusing a PFATeflon spray chamber washed with aqua regiafor 3min between samples REE were measured in high-resolution mode and U and Pb in low-resolution modeat 2000 dilution using a glass spray chamber washedwith 2 HNO3 between samples Total procedural blanksand reference materials (BCR-2 BHVO-2) were analyzedwith the batch of samples Indium was used as an internal
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
38
standard in all samples and standard solutionsBackground and standard solutions were analyzed afterevery five samples to detect memory effects and massdrift Results for PCIGR trace-element analyses areshown in Electronic Appendix 3 for the full set of analyzedtrace elementsSample digestion for purification of Sr Nd Hf and Pb
for column chemistry began by weighing each samplepowder (400^500mg) prior to leaching All samples wereleached with 6N HCl and placed in an ultrasonic bath for15min Samples were rinsed twice with 18 M-cm H2Obetween each leaching step (15 total) until the supernatantliquid was clear [following the method of Mahoney (1987)]Samples were then dried on a hotplate for 24 h andweighed again Sample solutions were then prepared bydissolving 100^250mg of the leached powder dissolvedin 1ml 14N HNO3 and 10ml 48 HF on a hotplate for48 h at 1308C with periodic ultrasonication Samples weredried and redissolved in 6ml 6N HCl on a hotplate for24 h and then dried Pb was separated using anion exhangecolumns and the discard was used for Sr REE and Hfseparation Nd was separated from the other REE and Hfrequired two additional purification steps Detailed proce-dures for column chemistry for separating Sr Nd and Pbat the PCIGR have been described byWeis et al (2006) andHf purification has been described byWeis et al (2007) Srand Nd isotope ratios were measured by thermal ioniza-tion mass spectrometry (TIMS) on a Thermo FinniganTriton system in static mode with relay matrix rotation ona single Ta and double Re^Ta filament respectively Fourto five filaments per barrel of 21 were occupied by stan-dards (NBS 987 for Sr and LaJolla for Nd) for each barrelwhere samples were run Sample Sr and Nd isotopic com-positions were corrected for mass fractionation using86Sr88Srfrac14 01194 and 146Nd144Ndfrac14 07219 Each samplewas then normalized using the barrel average of the refer-ence material relative to the values of143Nd144Ndfrac14 0511858 and 87Sr86Srfrac14 0710248 (Weiset al 2006) During the course of the Vancouver Islandanalyses the La Jolla Nd standard gave an average valueof 05118566 (nfrac14 7) and NBS987 standard gave an aver-age of 07102408 (nfrac1411) (2 error is reported as times106) 147Sm144Nd errors are 15 or 0006 USGeological Survey (USGS) reference material BHVO-2was processed with the samples and yielded Sr and Nd iso-topic ratios of 07034607 and 05129786 respectivelyThese are in agreement with the published values of070347920 and 051298411 respectively (Weis et al2006)
Pb and Hf isotopic compositions were analyzed by mul-tiple collector-inductively coupled plasma-mass spectrome-try (MC-ICP-MS) on a static multi-collection on a NuPlasma (Nu Instruments) system The detailed analyticalprocedure for Pb isotopic analyses on the Nu Plasma atthe PCIGR has been described by Weis et al (2006) Theconfiguration for Pb analyses allows for collection of PbTl and Hg togetherTl and Hg are used to monitor instru-mental mass discrimination and isobaric overlap repec-tively All sample solutions were analyzed withapproximately the same PbTl ratio (4) as the referencematerial NIST SRM 981 To accomplish this a small ali-quot of each sample solution from the Pb columns was ana-lyzed on the Element2 to determine the precise amount ofPb available for analysis on the Nu Plasma The SRM 981standard was run after every two samples on the NuPlasma During the time samples were run analyses ofthe SRM 981 Pb reference material gave values of206Pb204Pbfrac1416940322 207Pb204Pbfrac1415495823and 208Pb204Pbfrac14 36713164 (nfrac14 61 2 error is reportedas times 104) these values are in excellent agreement withreported TIMS triple-spike values of Galer amp Abouchami(1998) Fractionation-corrected Pb isotopic ratios were fur-ther corrected by the sample-standard bracketing methodor the ln^ln correction method described by White et al(2000) and Blichert-Toft et al (2003) Leached powders ofUSGS reference material BHVO-2 yielded Pb isotopicratios of 206Pb204Pbfrac14186454 8 207Pb204Pbfrac14 1549105 and 208Pb204Pbfrac14 38222514 and BCR-2yielded 206Pb204Pbfrac141880466 207Pb204Pbfrac14 1562518 and 208Pb204Pbfrac14 3883496 (2 error isreported as times 104) These values are in agreement withthose for leached residues of BHVO-2 and BCR-2 fromWeis et al (2006)Hf isotopic compositions were analyzed following the
procedures detailed byWeis et al (2007) The configurationfor Hf analyses monitored Lu mass 175 andYb mass 172 toallow for interference correction to masses 174 and 176 Hfisotopic ratios were normalized internally for mass fractio-nation to a 179Hf177Hf ratio of 07325 using an exponentialcorrection Standards were run after every two samplesand sample results were normalized to the ratio of the in-run daily average and a 176Hf177Hf ratio for JMC-475 of0282160 During the course of analyses the Hf standardJMC-475 gave an average value 02821679 (nfrac14130)USGS reference materials BCR-2 and BHVO-2 were pro-cessed with the samples and yielded Hf isotopic ratios of02828675 and 02831005 respectively Publishedvalues for BCR-2 and BHVO-2 are 02828717 and0283104 8 respectively (Weis et al 2007)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
39
trace-element patterns to the Karmutsen Range except forthe two LREE-depleted coarse-grained mafic rocks andthe mineralized sill (Fig 7d)
Sr^Nd^Hf^Pb isotopic compositionsA total of 19 samples from the four main groups ofthe Karmutsen Formation were selected for radiogenicisotopic geochemistry on the basis of major- and trace-element variation stratigraphic position and geographicallocation The samples have overlapping age-correctedHf and Nd isotope ratios and distinct ranges of Sr isotopiccompositions (Fig 9) The tholeiitic basalts picriteshigh-MgO basalt and coarse-grained mafic rocks have
initial eHffrac14thorn87 to thorn126 and eNdfrac14thorn66 to thorn88 cor-rected for in situ radioactive decay since 230 Ma (Fig 9Tables 3 and 4) All the Karmutsen samples form a well-defined linear array in a Lu^Hf isochron diagram corre-sponding to an age of 24115 Ma (Fig 9) This is withinerror of the accepted age of the Karmutsen Formationand indicates that the Hf isotope systematics behaved as aclosed system since c 230 Ma The anomalous pillowedflow has similar initial eHf (thorn98) and slightly lower initialeNd (thorn62) compared with the other Karmutsen samplesThe picrites and high-MgO basalt have higher initial87Sr86Sr (070398^070518) than the tholeiitic basalts(initial 87Sr86Srfrac14 070306^070381) and the coarse-grained mafic rocks (initial 87Sr86Srfrac14 070265^070428)
05128
05129
05130
05131
05132
015 017 019 021 023 025
02829
02830
02831
02832
02833
02834
000 002 004 006 008 010
4
6
8
10
0702 0703 0704 0705 07066
8
10
12
14
5 6 7 8 9 10
12
0
1
2
3
4
5
6
7
0702 0703 0704 0705 0706
(d)
LOI (wt )
87Sr 86Sr
4723A2
Nd
143Nd144Nd
147Sm144Nd
87Sr 86Sr (230 Ma)
(230 Ma)
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
(a)
(c)
176Hf177Hf
Age=241 plusmn 15 Ma=+990 plusmn 064
176Lu177Hf
Hf (230 Ma)
Nd (230 Ma)
(e)
Hf (i)
(b)
Fig 9 Whole-rock Sr Nd and Hf isotopic compositions for the Karmutsen Formation (a) 143Nd144Nd vs 147Sm144Nd (b) 176Hf177Hf vs176Lu177Hf The slope of the best-fit line for all samples corresponds to an age of 24115 Ma (c) Initial eNd vs 87Sr86Sr Age-correction to230 Ma (d) LOI vs 87Sr86Sr (e) Initial eHf vs eNd Average 2 error bars are shown in a corner of each panel Complete chemical duplicatesshown inTables 3 and 4 (samples 4720A7 and 4722A4) are circled in each plot
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
overlap the ranges of the picrites and tholeiitic basalts(Fig 9 Table 3)The measured Pb isotopic compositions of the tholeiitic
basalts are more radiogenic than those of the picrites andthe most magnesian Keogh Lake picrites have the leastradiogenic Pb isotopic compositions The range ofinitial Pb isotope ratios for the picrites is206Pb204Pbfrac1417868^18580 207Pb204Pbfrac1415547^15562and 208Pb204Pbfrac14 37873^38257 and the range for thetholeiitic basalts is 206Pb204Pbfrac1418782^19098207Pb204Pbfrac1415570^15584 and 208Pb204Pbfrac14 37312^38587 (Fig 10 Table 5) The coarse-grained mafic rocksoverlap the range of initial Pb isotopic compositions forthe tholeiitic basalts with 206Pb204Pbfrac1418652^19155207Pb204Pbfrac1415568^15588 and 208Pb204Pbfrac14 38153^38735 One high-MgO basalt has the highest initial206Pb204Pb (19242) and 207Pb204Pb (15590) and thelowest initial 208Pb204Pb (37676) The Pb isotopic ratiosfor Karmutsen samples define broadly linearrelationships in Pb isotope plots The age-corrected Pb
isotopic compositions of several picrites have been affectedby U and Pb (andor Th) mobility during alteration(Fig 10)
ALTERAT IONThe Karmutsen basalts have retained most of their origi-nal igneous structures and textures however secondaryalteration and low-grade metamorphism have producedzeolitic- and prehnite^pumpellyite-bearing mineral assem-blages (Table 1 Cho et al 1987) Alteration and metamor-phism have primarily affected the distribution of the LILEand the Sr isotopic systematics The Keogh Lake picriteshave been affected more by alteration than the tholeiiticbasalts and have higher LOI (up to 55wt ) variableLILE and higher measured Sr Nd and Hf and lowermeasured Pb isotope ratios than the tholeiitic basalts Srand Pb isotope compositions for picrites and tholeiiticbasalts show a relationship with LOI (Figs 9 and 10)whereas Nd and Hf isotopic compositions do not correlate
Table 3 Sr and Nd isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Rb Sr 87Sr86Sr 2m87Rb86Sr 87Sr86Srt Sm Nd 143Nd144Nd 2m eNd
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses carried out at the PCIGR thecomplete trace element analyses are given in Electronic Appendix 3
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
21
with LOI (not shown) The relatively high apparent initialSr isotopic compositions for the high-MgO lavas (up to07052) probably resulted from post-magmatic loss of Rbor an increase in 87Sr86Sr through addition of seawater Sr(eg Hauff et al 2003) High-MgO lavas from theCaribbean plateau have similarly high 87Sr86Sr comparedwith basalts (Recurren villon et al 2002)The correction for in situdecay on initial Pb isotopic ratios has been affected bymobilization of U and Th (and Pb) in the whole-rockssince their formation A thorough acid leaching duringsample preparation was used that has been shown to effec-tively remove alteration phases (Weis et al 2006 NobreSilva et al in preparation) Whole-rock SmNd ratiosand age-correction of Nd isotopes may be affected bythe leaching procedure for submarine rocks older than50 Ma (Thompson et al 2008 Fig 9) there ishowever very little variation in Nd isotopic compositionsof the picrites or tholeiitic basalts and this system does notappear to be disturbed by alteration The HFSE abun-dances for tholeiitic and picritic basalts exhibit clear
linear correlations in binary diagrams (Fig 8) whereasplots of LILE vs HFSE are highly scattered (not shown)because of the mobility of some of the alkali and alkalineearth LILE (Cs Rb Ba K) and Sr for most samplesduring alterationThe degree of alteration in the Karmutsen samples does
not appear to be related to eruption environment or depthof burial Pillow rims are compositionally different frompillow cores (Surdam1967 Kuniyoshi1972) and aquagenetuffs are chemically different from isolated pillows withinpillow breccias however submarine and subaerial basaltsexhibit a similar degree of alterationThere is no clear cor-relation between the submarine and subaerial basalts andsome commonly used chemical alteration indices (eg BaRb vs K2OP2O5 Huang amp Frey 2005) There is also nodefinitive correlation between the petrographic alterationindex (Table 1) and chemical alteration indices although10 of 13 tholeiitic basalts with the highest K and LILEabundances have the highest petrographic alterationindex of three (seeTable 1)
Table 4 Hf isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Lu Hf 176Hf177Hf 2m eHf176Lu177Hf 176Hf177Hft eHf(t)
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
OLIV INE ACCUMULAT ION INH IGH-MgO AND PICR IT IC LAVASGeochemical trends and petrographic characteristics indi-cate that accumulation of olivine played an important rolein the formation of the high-MgO basalts and picrites fromthe Keogh Lake area on northern Vancouver Island TheKeogh Lake samples show a strong linear correlation inplots of Al2O3 TiO2 ScYb and Ni vs MgO and many ofthe picrites have abundant clusters of olivine pseudomorphs(Table1 Fig11)There is a strong linear correlationbetween
modal per cent olivine and whole-rock magnesium contentmagnesium contents range between 10 and 19wt MgOand the proportion of olivine phenocrysts in these samplesvaries from 0 to 42 vol (Fig 11)With a few exceptionsboth the size range and median size of the olivine pheno-crysts in most samples are comparable The clear correla-tion between the proportion of olivine phenocrysts andwhole-rock magnesium contents indicates that accumula-tion of olivine was directly responsible for the high MgOcontents (410wt ) in most of the picritic lavas
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
207Pb204Pb
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
208Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
(a) (b)
(c) (e)
4723A24723A2
4723A2
4723A2
4722A4
4723A4
4722A4
4723A4
4723A134723A3
4723A3
4723A13
1556
1558
1560
1562
1564
1566
1568
186 190 194 198 202 206 2101550
1552
1554
1556
1558
1560
178 180 182 184 186 188 190 192 194
380
384
388
392
396
186 190 194 198 202 206 210374
378
382
386
390
178 180 182 184 186 188 190 192 194
206Pb204Pb(d)
LOI (wt )
0
1
2
3
4
5
6
7
186 190 194 198 202 206 210
4723A2
Fig 10 Pb isotopic compositions of leached whole-rock samples measured by MC-ICP-MS for the Karmutsen Formation Error bars are smal-ler than symbols (a) Measured 207Pb204Pb vs 206Pb204Pb (b) Initial 207Pb204Pb vs 206Pb204Pb Age-correction to 230 Ma (c) Measured208Pb204Pb vs 206Pb204Pb (d) LOI vs 206Pb204Pb (e) Initial 208Pb204Pb vs 206Pb204Pb Complete chemical duplicates shown in Table 5(samples 4720A7 and 4722A4) are circled in each plot The dashed lines in (b) and (e) show the differences in age-corrections for two picrites(4723A4 4722A4) using the measured UTh and Pb concentrations for each sample and the age-corrections when concentrations are used fromthe two picrites (4723A3 4723A13) that appear to be least affected by alteration
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
23
DISCUSS IONThe compositional and spatial development of the eruptedlava sequences of the Wrangellia oceanic plateau onVancouver Island provide information about the evolutionof a mantle plume upwelling beneath oceanic lithospherein an analogous way to the interpretation of continentalflood basalts The Caribbean and Ontong Java plateausare the two primary examples of accreted parts of oceanicplateaus that have been studied to date and comparisonscan be made with the Wrangellia oceanic plateauHowever the Wrangellia oceanic plateau is a distinctiveoccurrence of an oceanic plateau because it formed on thecrust of a Devonian to Mississippian intra-oceanic islandarc overlain by Mississippian to Permian limestones andpelagic sediments The nature and thickness of oceaniclithospheric mantle are considered to play a fundamentalrole in the decompression and evolution of a mantleplume and thus the compositions of the erupted lavasequences (Greene et al 2008) The geochemical and
stratigraphic relationships observed in the KarmutsenFormation onVancouver Island and the focus of the subse-quent discussion sections provide constraints on (1) thetemperature and degree of melting required to generatethe primary picritic magmas (2) the composition of themantle source (3) the depth of melting and residual min-eralogy in the source region (4) the low-pressure evolutionof the Karmutsen tholeiitic basalts (5) the growth of theWrangellia oceanic plateau
Melting conditions and major-elementcomposition of the primary magmasThe primary magmas to most flood basalt provinces arebelieved to be picritic (eg Cox 1980) and they have beenfound in many continental and oceanic flood basaltsequences worldwide (eg Siberia Karoo Paranacurren ^Etendeka Caribbean Deccan Saunders 2005) Near-pri-mary picritic lavas with low total alkali abundances
Table 5 Pb isotopic compositions of Karmutsen basaltsVancouver Island BC
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
require high-degree partial melting of the mantle source(eg Herzberg amp OrsquoHara 2002) The Keogh Lake picritesfrom northernVancouver Island are the best candidates foridentifying the least modified partial melts of the mantleplume source despite having accumulated olivine pheno-crysts and can be used to estimate conditions of meltingand the composition of the primary magmas for theKarmutsen FormationPrimary magma compositions mantle potential tem-
peratures and source melt fractions for the picritic lavas ofthe Karmutsen Formation were calculated from primitivewhole-rock compositions using PRIMELT1XLS software(Herzberg et al 2007) and are presented in Fig 12 andTable 6 A detailed discussion of the computationalmethod has been given by Herzberg amp OrsquoHara (2002)and Herzberg et al (2007) key features of the approachare outlined belowPrimary magma composition is calculated for a primi-
tive lava by incremental addition of olivine and compari-son with primary melts of mantle peridotite Lavas thathad experienced plagioclase andor clinopyroxene fractio-nation are excluded from this analysis All calculated pri-mary magma compositions are assumed to be derived byaccumulated fractional melting of fertile peridotite KR-4003 Uncertainties in fertile peridotite composition donot propagate to significant variations in melt fractionand mantle potential temperature melting of depletedperidotite propagates to calculated melt fractions that aretoo high but with a negligible error in mantle potential
temperature PRIMELT1XLS calculates the olivine liqui-dus temperature at 1atm which should be similar to theactual eruption temperature of the primary magmaand is accurate to 308C at the 2 level of confidenceMantle potential temperature is model dependent witha precision that is similar to the accuracy of the olivineliquidus temperature (C Herzberg personal communica-tion 2008)With regard to the evidence for accumulated olivine in
the Keogh Lake picrites it should not matter whether themodeled lava composition is aphyric or laden with accu-mulated olivine phenocrysts PRIMELT1 computes theprimary magma composition by adding olivine to a lavathat has experienced olivine fractionation in this way itinverts the effects of fractional crystallization The onlyrequirement is that the olivine phenocrysts are not acci-dental or exotic to the lava compositions Support for thisprocedure can be seen in the calculated olivine liquid-line-of-descent shown by the curved arrays with 5 olivineaddition increments (Fig 12c) Fractional crystallization ofolivine from model primary magmas having 85^95wt FeO and 153^175wt MgO will yield arrays of deriva-tive liquids and randomly sorted olivines that in most butnot all cases connect the picrites and high-MgO basalts Anewer version of the PRIMELT software (Herzberg ampAsimow 2008) can perform olivine addition to and sub-traction from lavas with highly variably olivine phenocrystcontents and yields results that converge with those shownin Fig 12
Fig 11 Relationship between abundance of olivine phenocrysts and whole-rock MgO contents for the Keogh Lake picrites (a) Tracings ofolivine grains (gray) in six high-MgO samples made from high resolution (2400 dpi) scans of petrographic thin-sections Scale bars represent10mm sample numbers are indicated at the upper left (b) Modal olivine vs whole-rock MgO Continuous line is the best-fit line for 10 high-MgO samplesThe areal proportion of olivine was calculated from raster images of olivine grains using ImageJ image analysis software whichprovides an acceptable estimation of the modal abundance (eg Chayes 1954)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
25
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
Gray lines = final melting pressure
L + Ol + Opx
07
08
09
Pressure (GPa)
10
3
4
5
6
1
SOLIDUS
01
04
0506
L + Ol
2
03
02
PeridotiteKR-4003
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
Thick black lines = initial melting pressure
Dashed lines = melt fraction
Melt Fraction
2
3 4 5 6
L+Ol+Opx
SolidusSpinel
SolidusGarnet Peridotite
7
L+Ol
0 10 20 30 4040
45
50
55
60
MgO (wt)
SiO2
(wt)
PeridotiteKR-4003
Melt Fraction
0 10 20 305
10
15
CaO(wt)
MgO (wt)
3
4 5 6 7
2
46
2
Peridotite Partial Melts
Thick black line = initial melting pressureGray lines = final melting pressure
Peridotite
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
02
Pressure (GPa)
0302
04
Pressure (GPa)
Picrite
High-MgO basalt
Tholeiitic basalt
(c)
(d)
Karmutsen flood basaltsOlivine addition modelOlivine addition (5 increment)Primary magma
PicriteHigh-MgO basaltTholeiitic basalt
(a)
(b)
800 850 900
910
920
930
940
950
Karmutsen flood basalts
Olivine addition modelOlivine addition (5 increment)Primary magma
Gray lines = Fo contents of olivine
Fig 12 Estimated primary magma compositions for three Keogh Lake picrites and high-MgO basalts (samples 4723A4 4723A135616A7) using the forward and inverse modeling technique of Herzberg et al (2007) Karmutsen compositions and modeling results areoverlain on diagrams provided by C Herzberg (a) Whole-rock FeO vs MgO for Karmutsen samples from this study Total iron estimatedto be FeO is 090 Gray lines show olivine compositions that would precipitate from liquid of a given MgO^FeO composition Black lineswith crosses show results of olivine addition (inverse model) using PRIMELT1 (Herzberg et al 2007) Gray field highlights olivinecompositions with Fo contents of 91^92 predicted to precipitate (b) FeO vs MgO showing Karmutsen lava compositions and results ofa forward model for accumulated fractional melting of fertile peridotite KR-4003 (Kettle River Peridotite Walter 1998) (c) SiO2 vsMgO for Karmtusen lavas and model results (d) CaO vs MgO showing Karmtusen lava compositions and model results To brieflysummarize the technique [see Herzberg et al (2007) for complete description] potential parental magma compositions for the high-MgOlava series were selected (highest MgO and appropriate CaO) and using PRIMELT1 software olivine was incrementally added tothe selected compositions to show an array of potential primary magma compositions (inverse model) Then using PRIMELT1 theresults from the inverse model were compared with a range of accumulated fractional melts for fertile peridotite derived fromparameterization of the experimental results for KR-4003 from Walter (1998) forward model Herzberg amp OrsquoHara (2002) A melt fractionwas sought that was unique to both the inverse and forward models (Herzberg et al 2007) A unique solution was found when there was a com-mon melt fraction for both models in FeO^MgO and CaO^MgO^Al2O3^SiO2 (CMAS) projection space This modeling assumes thatolivine was the only phase crystallizing and ignores chromite precipitation and possible augite fractionation in the mantle (Herzbergamp OrsquoHara 2002) Results are best for a residue of spinel lherzolite (not pyroxenite) The presence of accumulated olivine in samples ofthe Keogh Lake picrites used as starting compositions does not significantly affect the results because we are modeling addition of olivine(see text for further description) The tholeiitic basalts cannot be used for modeling because they are all saturated in plagthorn cpxthornolThick black line for initial melting pressure at 4 GPa is not shown in (c) for clarity Gray area in (a) indicates parental magmas thatwill crystallize olivine with Fo 91^92 at the surface Dashed lines in (c) indicate melt fraction Solidi for spinel and garnet peridotite arelabelled in (c)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
The PRIMELT1 results indicate that if the Karmutsenpicritic lavas were derived from accumulated fractionalmelting of fertile peridotite the primary magmas wouldhave contained 15^17wt MgO and 10wt CaO(Fig 12 Table 6) formed from 23^27 partial meltingand would have crystallized olivine with a Fo content of 91 (Fig 12 Table 6) Calculated mantle temperatures forthe Karmutsen picrites (14908C) indicate melting ofanomalously hot mantle (100^2008C above ambient) com-pared with ambient mantle that produces mid-ocean ridgebasalt (MORB 1280^14008C Herzberg et al 2007)which initiated between 3 and 4 GPa Several picritesappear to be near-primary melts and required very littleaddition of olivine (eg sample 93G171 with measured158wt MgO) These temperature and melting esti-mates of the Karmutsen picrites are consistent withdecompression of hot mantle peridotite in an actively con-vecting plume head (ie plume initiation model) Threesamples (picrite 5614A1 and high-MgO basalts 5614A3and 5614A5) are lower in iron than the other Karmutsenlavas with FeO in the 65^80wt range (Fig 12c) andhave MgO and FeO contents that are similar to primitiveMORB from the Sequeiros fracture zone of the EastPacific Rise (EPR Perfit et al 1996) These samples
however differ from EPR MORB in having higher SiO2
and lower Al2O3 and they are restricted geographicallyto one area (Sara Lake Fig 2)We therefore have evidence for primary magmas with
MgO contents that range from a possible low of 110wt to as high as 174wt with inferred mantle potential tem-peratures from 1340 to 15208C This is similar to the rangedisplayed by lavas from the Ontong Java Plateau deter-mined by Herzberg (2004) who found that primarymagma compositions assuming a peridotite sourcewould contain 17wt MgO and 10wt CaO(Table 6) and that these magmas would have formed by27 melting at a mantle potential temperature of15008C (first melting occurring at 36 GPa) Fitton ampGodard (2004) estimated 30 partial melting for theOntong Java lavas based on the Zr contents of primarymagmas of Kroenke-type basalt calculated by incrementaladdition of equilibrium olivine However if a considerableamount of eclogite was involved in the formation of theOntong Java plateau magmas the excess temperatures esti-mated by Herzberg et al (2007) may not be required(Korenaga 2005) The variability in mantle potential tem-perature for the Keogh Lake picrites is also similar to thewide range of temperatures that have been calculated for
Table 6 Estimated primary magma compositions for Karmutsen basalts and other oceanic plateaus or islands
Sample 4723A4 4723A13 5616A7 93G171 Average OJP Mauna Keay Gorgonaz
Ontong Java primary magma composition for accumulated fraction melting (AFM) from Herzberg (2004)yMauna Kea primary magma composition is average of four samples of Herzberg (2006 table 1)zGorgona primary magma composition for AFM for 1F fertile source from Herzberg amp OrsquoHara (2002 table 4)Total iron estimated to be FeO is 090
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
27
lavas from single volcanoes in the Galapagos Hawaii andelsewhere by Herzberg amp Asimow (2008) Those workersinterpreted the range to reflect the transport of magmasfrom both the cool periphery and the hotter axis of aplume A thermally heterogeneous plume will yield pri-mary magmas with different MgO contents and the rangeof primary magma compositions and their inferred mantlepotential temperatures for Karmutsen lavas may reflectmelting from different parts of the plume
Source of the Karmutsen Formation lavasThe Sr^Nd^Hf^Pb isotopic compositions of theKarmutsen volcanic rocks on Vancouver Island indicatethat they formed from plume-type mantle containing adepleted component that is compositionally similar to butdistinct from other oceanic plateaus that formed in thePacific Ocean (eg Ontong Java and Caribbean Fig 13)Trace element and isotopic constraints can be used to testwhether the depleted component of the KarmutsenFormation was intrinsic to the mantle plume and distinctfrom the source of MORB or the result of mixing or invol-vement of ambient depleted MORB mantle with plume-type mantle The Karmutsen tholeiitic basalts have initialeNd and
87Sr86Sr ratios that fall within the range of basaltsfrom the Caribbean Plateau (eg Kerr et al 1996 1997Hauff et al 2000 Kerr 2003) and a range of Hawaiian iso-tope data (eg Frey et al 2005) and lie within and abovethe range for the Ontong Java Plateau (Tejada et al 2004Fig 13a) Karmutsen tholeiitic basalts with the highest ini-tial eNd and lowest initial 87Sr86Sr lie within and justbelow the field for age-corrected northern EPR MORB at230 Ma (eg Niu et al1999 Regelous et al1999) Initial Hfand Nd isotopic compositions for the Karmutsen samplesare noticeably offset to the low side of the ocean islandbasalt (OIB) array (Vervoort et al 1999) and lie belowand slightly overlap the low eHf end of fields for Hawaiiand the Caribbean Plateau (Fig 13c) the Karmutsen sam-ples mostly fall between and slightly overlap a field forage-corrected Pacific MORB at 230 Ma and Ontong Java(Mahoney et al 1992 1994 Nowell et al 1998 Chauvel ampBlichert-Toft 2001) Karmutsen lava Pb isotope ratios over-lap the broad range for the Caribbean Plateau and thelinear trends for the Karmutsen samples intersect thefields for EPR MORB Hawaii and Ontong Java in208Pb^206Pb space but do not intersect these fields in207Pb^206Pb space (Fig 13d and e) The more radiogenic207Pb204Pb for a given 206Pb204Pb of the Karmutsenbasalts requires high UPb ratios at an ancient time when235U was abundant similar to the Caribbean Plateau andenriched in comparison with EPR MORB Hawaii andOntong JavaThe low eHf^low eNd component of the Karmutsen
basalts that places them below the OIB mantle array andpartly within the field for age-corrected Pacific MORBcan form from low ratios of LuHf and high SmNd over
time (eg Salters amp White 1998) MORB will develop aHf^Nd isotopic signature over time that falls below themantle array Low LuHf can also lead to low 176Hf177Hffrom remelting of residues produced by melting in theabsence of garnet (eg Salters amp Hart 1991 Salters ampWhite 1998) The Hf^Nd isotopic compositions of theKarmutsen Formation indicate the possible involvementof depleted MORB mantle however similar to Iceland(Fitton et al 2003) Nb^Zr^Y variations provide a way todistinguish between a depleted component within theplume and a MORB-like component The Karmutsenbasalts and picrites fall within the Iceland array and donot overlap the MORB field in a logarithmic plot of NbYvs ZrY (Fig 13b) using the fields established by Fittonet al (1997) Similar to the depleted component within theCaribbean Plateau (Thompson et al 2003) the trend ofHf^Nd isotopic compositions towards Pacific MORB-likecompositions is not followed by MORB values in Nb^Zr^Y systematics and this suggests that the depleted compo-nent in the source of the Karmutsen basalts is distinctfrom the source of MORB and probably an intrinsic partof the plume The relatively radiogenic Pb isotopic ratiosand REE patterns distinct from MORB also support thisinterpretationTrace-element characteristics suggest the possible invol-
vement of sub-arc lithospheric mantle in the generation ofthe Karmutsen picrites Lassiter et al (1995) suggested thatmixing of a plume-type source with eNdthorn 6 to thorn 7 witharc material with low NbTh could reproduce the varia-tions observed in the Karmutsen basalts however theyconcluded that the absence of low NbLa ratios in theirsample of tholeiitic basalts restricts the amount of arc litho-sphere involved The study of Lassiter et al (1995) wasbased on major and trace element and Sr Nd and Pb iso-topic data for a suite of 29 samples from Buttle Lake in theStrathcona Provincial Park on central Vancouver Island(Fig 1) Whereas there are no clear HFSE depletions inthe Karmutsen tholeiitic basalts the low NbLa (and TaLa) of the Karmutsen picritic lavas in this study indicatethe possible involvement of Paleozoic arc lithosphere onVancouver Island (Fig 8)
REE modeling dynamic melting andsource mineralogyThe combined isotopic and trace-element variations of thepicritic and tholeiitic Karmutsen lavas indicate that differ-ences in trace elements may have originated from differentmelting histories For example the LuHf ratios of theKarmutsen picritic lavas (mean 008) are considerablyhigher than those in the tholeiitic basalts (mean 002)however the small range of overlapping initial eHf valuesfor the two lava suites suggests that these differences devel-oped during melting within the plume (Fig 9b) Below wemodel the effects of source mineralogy and depth of
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
28
melting on differences in trace-element concentrations inthe tholeiitic basalts and picritesDynamic melting models simulate progressive decom-
pression melting where some of the melt fraction isretained in the residue and only when the degree of partialmelting is greater than the critical mass porosity and the
source becomes permeable is excess melt extracted fromthe residue (Zou1998) For the Karmutsen lavas evolutionof trace element concentrations was simulated using theincongruent dynamic melting model developed by Zou ampReid (2001) Melting of the mantle at pressures greater than1 GPa involves incongruent melting (eg Longhi 2002)
OIB array
PacificMORB(230 Ma)
(0 Ma)
EPR(230 Ma)
-4
-2
0
2
4
8
10
12
14
16
18
20
22
-4 -2 0 2 4 6 8 10 12 14
minus2
0
2
4
6
8
10
12
14
0702 0703 0704 0705 0706
IndianMORB
87Sr 86Sr (initial)
Hf (initial)
(a) (c)
Nd (initial)
Karmutsen tholeiitic basalt
HawaiiOntong JavaCaribbean Plateau
Karmutsen high-MgO lava
high-MgO lavasKarmutsen
1540
1545
1550
1555
1560
1565
175 180 185 190 195 200375
380
385
390
395
175 180 185 190 195 200
(d) (e)
EPR
EPR
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb
Karmutsen Formation (initial)
HawaiiOntong JavaCaribbean Plateau
East Pacific Rise
Karmutsen Formation (measured)
Juan de FucaGorda
ε Nd
(initial)
Karmutsen Formation (different age-correction for 3 picrites)
(0 Ma)
NbY
001
01
1
1 10
ZrY
OIB
NMORB
(b)
Iceland array
6
Fig 13 Comparison of age-corrected (230 Ma) Sr^Nd^Hf^Pb isotopic compositions for Karmutsen flood basalts onVancouver Island withage-corrected OIB and MORB and Nb^Zr^Y variation (a) Initial eNd vs 87Sr86Sr (b) NbYand ZrY variation (after Fitton et al 1997)(c) Initial eHf vs eNd (d) Measured and initial 207Pb204Pb vs 206Pb204Pb (e) Measured and initial 208Pb204Pb vs 206Pb204Pb References fordata sources are listed in Electronic Appendix 4 Most of the compiled data were extracted from the GEOROC database (httpgeorocmpch-mainzgwdgdegeoroc) OIB array line in (c) is fromVervoort et al (1999) EPR East Pacific Rise Several high-MgO samples affected bysecondary alteration were not plotted for clarity in Pb isotope plots Estimation of fields for age-corrected Pacific MORB and EPR at 230 Mawere made using parent^daughter concentrations (in ppm) from Salters amp Stracke (2004) of Lufrac14 0063 Hffrac14 0202 Smfrac14 027 Ndfrac14 071Rbfrac14 0086 and Srfrac14 836 In the NbYvs ZrYplot the normal (N)-MORB and OIB fields not including Icelandic OIB are from data com-pilations of Fitton et al (1997 2003) The OIB field is data from 780 basalts (MgO45wt ) from major ocean islands given by Fitton et al(2003) The N-MORB field is based on data from the Reykjanes Ridge EPR and Southwest Indian Ridge from Fitton et al (1997) The twoparallel gray lines in (b) (Iceland array) are the limits of data for Icelandic rocks
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
29
with olivine being produced during the reactioncpxthornopxthorn spfrac14meltthornol for spinel peridotites The mod-eling used coefficients for incongruent melting reactionsbased on experiments on lherzolite melting (eg Kinzleramp Grove 1992 Longhi 2002) and was separated into (1)melting of spinel lherzolite (2) two combined intervals ofmelting for a garnet lherzolite (gt lherz) and spinel lher-zolite (sp lherz) source (3) melting of garnet lherzoliteThe model solutions are non-unique and a range indegree of melting partition coefficients melt reactioncoefficients and proportion of mixed source componentsis possible to achieve acceptable solutions (see Fig 14 cap-tion for description of modeling parameters anduncertainties)The LREE-depleted high-MgO lavas require melting of
a depleted spinel lherzolite source (Fig 14a) The modelingresults indicate a high degree of melting (22^25) similarto the results from PRIMELT1 based on major-elementvariations (discussed above) of a LREE-depleted sourceThe high-MgO lavas lack a residual garnet signature Theenriched tholeiitic basalts involved melting of both garnetand spinel lherzolite and represent aggregate melts pro-duced from continuous melting throughout the depth ofthe melting column (Fig 14b) Trace-element concentra-tions in the tholeiitic and picritic lavas are not consistentwith low-degree melting of garnet lherzolite alone(Fig 14c) The modeling results indicate that the aggregatemelts that formed the tholeiitic basalts would haveinvolved lesser proportions of enriched small-degreemelts (1^6 melting) generated at depth and greateramounts of depleted high-degree melts (12^25) gener-ated at lower pressure (a ratio of garnet lherzolite tospinel lherzolite melt of 14 provides the best fits seeFig 14b and description in figure caption) The totaldegree of melting for the tholeiitic basalts was high(23^27) similar to the degree of partial melting in thespinel lherzolite facies for the primary melts that eruptedas the Keogh Lake picrites of a source that was also prob-ably depleted in LREE
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
(c)
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
Melting of garnet lherzolite
High-MgO lavas
Melting of spinel lherzolite
source (07DM+03PM)
source (07DM+03PM)
6 melting
30 melting
(b)
(a)
30 melting
Sam
ple
Cho
ndrit
eS
ampl
eC
hond
rite 20 melting
1 melting
Melting of garnet lherzoliteand spinel lherzolite
x gt lherz + 4x sp lherz
5 meltingx gt lherz + 4x sp lherz
Sam
ple
Cho
ndrit
e
Tholeiitic lavas
Tholeiitic lavas
High-MgO lavas
source (07DM+03PM)
Fig 14 Trace-element modeling results for incongruent dynamicmantle melting for picritic and tholeiitic lavas from the KarmutsenFormation Three steps of modeling are shown (a) melting of spinellherzolite compared with high-MgO lavas (b) melting of garnet lher-zolite and spinel lherzolite compared with tholeiitic lavas (c) meltingof garnet lherzolite compared with tholeiitic and picritic lavasShaded fields in each panel for tholeiitic basalts and picrites arelabeled Patterns with symbols in each panel are modeling results(patterns are 5 melting increments in (a) and (b) and 1 incre-ments in (c) PM primitive mantle DM depleted MORB mantle gtlherz garnet lherzolite sp lherz spinel lherzolite In (b) the ratios ofper cent melting for garnet and spinel lherzolite are indicated (x gtlherzthorn 4x sp lherz means that for 5 melting 1 melt in garnetfacies and 4 melt in spinel facies where x is per cent melting in theinterval of modeling) The ratio of the respective melt proportions inthe aggregate melt (x gt lherzthorn 4x sp lherz) was kept constant andthis ratio of melt contribution provided the best fit to the data Arange of proportion of source components (07DMthorn 03PM to
09DMthorn 01PM) can reproduce the variation in the high-MgOlavas Melting modeling uses the formulation of Zou amp Reid (2001)an example calculation is shown in their Appendix PM fromMcDonough amp Sun (1995) and DM from Salters amp Stracke (2004)Melt reaction coefficients of spinel lherzolite from Kinzler amp Grove(1992) and garnet lherzolite fromWalter (1998) Partition coefficientsfrom Salters amp Stracke (2004) and Shaw (2000) were kept constantSource mineralogy for spinel lherzolite (018cpx027opx052ol003sp) from Kinzler (1997) and for garnet lherzolite(034cpx008opx053ol005gt) from Salters amp Stracke (2004) Arange of source compositions for melting of spinel lherzolite(07DMthorn 03PM to 03DMthorn 07PM) reproduce REE patternssimilar to those of the high-MgO lavas with best fits for 22ndash25melting and 07DMthorn 03PM source composition Concentrationsof tholeiitic basalts cannot be reproduced from an entirelyprimitive or depleted source
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
30
The uncertainty in the composition of the source in thespinel lherzolite melting model for the high-MgO lavasprecludes determining whether the source was moredepleted in incompatible trace elements than the source ofthe tholeiitic lavas (see Fig 14 caption for description) It ispossible that early high-pressure low-degree melting left aresidue within parts of the plume (possibly the hotter inte-rior portion) that was depleted in trace elements (egElliott et al 1991) further decompression and high-degreemelting of such depleted regions could then have generatedthe depleted high-MgO Karmutsen lavas Shallow high-degree melting would preferentially sample depletedregions whereas deeper low-degree melting may notsample more refractory depleted regions (eg CaribbeanEscuder-Viruete et al 2007) The picrites lie at the highend of the range of Nd isotopic compositions and havelow NbZr similar to depleted Icelandic basalts (Fittonet al 2003) therefore the picrites may represent the partsof the mantle plume that were the most depleted prior tothe initiation of melting As Fitton et al (2003) suggestedthese depleted melts may only rarely be erupted withoutcombining with more voluminous less depleted melts andthey may be detected only as undifferentiated small-volume flows like the Keogh Lake picrites within theKarmutsen Formation on Vancouver IslandDecompression melting within the mantle plume initiatedwithin the garnet stability field (35^25 GPa) at highmantle potential temperatures (Tp414508C) and pro-ceeded beneath oceanic arc lithosphere within the spinelstability field (25^09 GPa) where more extensivedegrees of melting could occur
Magmatic evolution of Karmutsentholeiitic basaltsPrimary magmas in large igneous provinces leave exten-sive coarse-grained cumulate residues within or below thecrust these mafic and ultramafic plutonic sequences repre-sent a significant proportion of the original magma thatpartially crystallized at depth (eg Farnetani et al 1996)For example seismic and petrological studies of OntongJava (eg Farnetani et al 1996) combined with the use ofMELTS (Ghiorso amp Sack 1995) indicate that the volcanicsequence may represent only 40 of the total magmareaching the Moho Neal et al (1997) estimated that30^45 fractional crystallization took place for theOntong Java magmas and produced cumulate thicknessesof 7^14 km corresponding to 11^22 km of flood basaltsdepending on the presence of pre-existing oceanic crustand the total crustal thickness (25^35 km) Interpretationsof seismic velocity measurements suggest the presence ofpyroxene and gabbroic cumulates 9^16 km thick beneaththe Ontong Java Plateau (eg Hussong et al 1979)Wrangellia is characterized by high crustal velocities in
seismic refraction lines which is indicative of mafic
plutonic rocks extending to depth beneath VancouverIsland (Clowes et al 1995)Wrangellia crust is 25^30 kmthick beneath Vancouver Island and is underlain by astrongly reflective zone of high velocity and density thathas been interpreted as a major shear zone where lowerWrangellia lithosphere was detached (Clowes et al 1995)The vast majority of the Karmutsen flows are evolved tho-leiitic lavas (eg low MgO high FeOMgO) indicating animportant role for partial crystallization of melts at lowpressure and a significant portion of the crust beneathVancouver Island has seismic properties that are consistentwith crystalline residues from partially crystallizedKarmutsen magmas To test the proportion of the primarymagma that fractionated within the crust MELTS(Ghiorso amp Sack 1995) was used to simulate fractionalcrystallization using several estimated primary magmacompositions from the PRIMELT1 modeling results Themajor-element composition of the primary magmas couldnot be estimated for the tholeiitic basalts because of exten-sive plagthorn cpxthornol fractionation and thus the MELTSmodeling of the picrites is used as a proxy for the evolutionof major elements in the volcanic sequence A pressure of 1kbar was used with variable water contents (anhydrousand 02wt H2O) calculated at the quartz^fayalite^magnetite (QFM) oxygen buffer Experimental resultsfrom Ontong Java indicate that most crystallizationoccurs in magma chambers shallower than 6 km deep(52 kbar Sano amp Yamashita 2004) however some crys-tallization may take place at greater depths (3^5 kbarFarnetani et al 1996)A selection of the MELTS results are shown in Fig 15
Olivine and spinel crystallize at high temperature until15^20wt of the liquid mass has fractionated and theresidual magma contains 9^10wt MgO (Fig 15f) At1235^12258C plagioclase begins crystallizing and between1235 and 11908C olivine ceases crystallizing and clinopyr-oxene saturates (Fig 15f) As expected the addition of clin-opyroxene and plagioclase to the crystallization sequencecauses substantial changes in the composition of the resid-ual liquid FeO and TiO2 increase and Al2O3 and CaOdecrease while the decrease in MgO lessens considerably(Fig 15) The near-vertical trends for the Karmutsen tho-leiitic basalts especially for CaO do not match the calcu-lated liquid-line-of-descent very well which misses manyof the high-CaO high-Al2O3 low-FeO basalts A positivecorrelation between Al2O3CaO and SrNd indicates thataccumulation of plagioclase played a major role in the evo-lution of the tholeiitic basalts (Fig 15g) Increasing thecrystallization pressure slightly and the water content ofthe melts does not systematically improve the match ofthe MELTS models to the observed dataThe MELTS results indicate that a significant propor-
tion of the crystallization takes place before plagioclase(15^20 crystallization) and clinopyroxene (35^45
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
31
crystallization) join the crystallization sequence Theseresults suggest that445 crystallization of the primarymagmas occurred (estimated from the residual liquid percent in Fig 15f) and if the original flood basalt stratigra-phy on Vancouver Island was 6 km thick at least anequivalent amount of complementary plutonic rocksshould have crystallized The Karmutsen basalts and theirplutonic residues thus represent a significant addition ofcrust (perhaps 412 km thickness) previously thickenedduring the formation of the underlying Paleozoic arc
Growth of the Wrangellia oceanic plateauon Vancouver IslandThe extrusive part of the Wrangellia oceanic plateau onVancouver Island was constructed in a three-layersequence of mostly tholeiitic basalts with a restrictedrange of composition (Fig 16) The plateau overliesDevonian to Mississippian arc volcanic sequences andmarine sediments of Mississippian to Permian ages thatare intruded by gabbroic rocks related to the flood basaltsThe volcanic stratigraphy of theWrangellia plateau formed
00
05
10
15
20
25
30
35
40
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
16
0 2 4 6 8 10 12 14 16 18 20
10
11
12
13
14
15
16
17
18
19
20
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
0 2 4 6 8 10 12 14 16 18 20
MgO (wt)
0 10 20 30 40 50
percent of total mass
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
) Mineral proportions
Sp (e)
Ol
CpxPlag
PicriteHigh-MgO basaltTholeiitic basalt
TiO2 (wt)
MgO (wt)
Al2O3 (wt) CaO (wt)
FeO (wt)
MgO(a) (b)
(c) (d)
MgO (wt)MgO (wt)
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
)
Residual liquid ()
1220
1190
Ol + Sp
Plag+Ol+Sp Plag+Cpx+Sp
02 wt H2O
no H2O
Residual liquid and crystallizing phases
(f)
0102030405060708090100
plagioclaseaccumulation
plagioclaseaccumulation
plagioclase
accumulatio
n
10
12
14
16
0 10 20 30 40
Al2O3CaO
SrNd
tholeiitic basalts
(g)
Fig 15 Forward fractional crystallization modeling results for major elements from MELTS (Ghiorso amp Sack 1995) compared with picriticand tholeiitic lavas from the Karmutsen Formation (a^d) Major-element compositions of residual liquids compared with Karmutsen lavacompositions For clarity the results are shown only for one starting composition (open circle) using the composition of the estimated primarymagma for the high-MgO lava series from the modeling technique of Herzberg et al (2007) for sample 4723A13 as shown in Fig 12 The esti-mated primary magmas for samples 4723A3 and 93G171 yield similar results A pressure of 1 kbar was used with no H2O and 02wt H2Ocalculated at QFM black line is anhydrous and grey line is 02wt H2OThe end-point for the anhydrous result is at 78 fractionation and11308C and the end-point for the result with 02wt H2O is 75 fractionation and11108C (e) Mineral proportions of crystallized phases fromone MELTrun (sample 4723A13) with no H2O (f) Per cent residual liquid vs temperature labelled with intervals of crystallizing phases (g)Al2O3CaO vs SrNd for the tholeiitic basalts
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
32
in a single eruptive phase (c 231^225 Ma) as an emergentbasalt sequence during a deeper-water shallow-water andsubaerial stage and subsided below sea level relativelysoon after its formation (Fig 16) The deeper-water stagecontains pillowed and massive submarine flows (3000mthick) that built up the submarine volcanic edifice from theseafloor The lowest pillow basalts were emplaced onunconsolidated and lithified fine-grained siliciclastic andcarbonaceous sediments containing Daonella beds(c 235^232 Ma) that are intruded by abundantKarmutsen mafic sills (Fig 16a) The different pillowedand massive submarine flows resulted primarily from dif-ferent effusive rates and local topography Some of themassive submarine flows commonly recognized by radiat-ing columnar jointing represent master tubes for the deliv-ery of lava to distal parts of submarine flow fields Thepillowed flows which commonly have uneven topography
in modern sequences may have served as levees to themassive submarine flows rather than allowing unob-structed spreading and inflation typical of the massive sub-aerial flows that form on shallow slopes in continentalflood basalt provinces (eg Self et al 1997) The rapid suc-cession of flows is evident by the near absence of sedimentbetween submarine flows Minor volumes of picritic pillowlavas occur in the upper part of the pillowed flow sequencein areas of the plateau exposed on northern VancouverIsland (Fig 16 Nixon et al 2008)The growth of the plateau in shallow water led to
emplacement of an increasing proportion of volcaniclastics(pillow breccias and hyaloclastites) that accumulated up to1500m thick in some areas (Fig 16b Nixon et al 2008)Horizons of pillowed flows occur locally within the volca-niclastic stratigraphy Sedimentary structures (eg gradedbedding fluidization structures) and petrographic textures
Late Devonian-Mississippian arc
Daonella beds
Middle Triassic
deeper water
Late Devonian-Mississippian arc
pillowed flows massive submarine
flows
pillow brecciahyaloclastiteshallow water
~4 km
~6 km
Late Triassic limestoneinterflow limestonelenses
Karmutsen Formation
mafic intrusions sediment-sill complex
Karmutsen Formation
(a) (b)
(c) (d)
subaerial flows
Late Devonian-Mississippian arc Late Devonian-Mississippian arc
picritic pillow basalts
Miss to Permian marine sediments
sills
Fig 16 Growth of the Wrangellia oceanic plateau onVancouver Island (a) Deep-water eruptive phase of pillowed and massive submarineflows with almost no intervening sediments onto MiddleTriassic shale containing Daonella beds (c 235^232 Ma) The basement of the plateauis composed of Devonian to Mississippian arc sequences overlain by Mississippian to Permian carbonates and siliciclastic sediments(b) Shallow-water flow emplacement with increasing proportion of volcaniclastics resulted primarily from cooling^contraction granulationPicritic pillow basalts occur primarily near the transition between pillowed submarine flows and volcaniclastic units Pillowed and massiveflows occur locally within the volcaniclastic unit (c) Emergent stage of growth consists of overlapping subaerial flow fields similar to continen-tal flood basalts with pahoehoe flow structures Multi-tiered columnar jointing occurrs locally in response to ponding of flows (eg Bondre et al2004) Interflow carbonate deposits and plagioclase megacrystic flows developed locally during the waning stage of volcanism (d) Post-volcanicsubsidence and deposition of Late Carnian to Norian (c 212^223 Ma) platform carbonates on top of the plateau Units are labeled in the dia-grams Thicknesses of the underlying units are not to scale
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
33
within the volcaniclastic rocks indicate formation primar-ily via cooling^contraction granulation magma^water^steam interaction autobrecciation resedimentation andmass-wasting rather than pyroclastic eruption (Nixonet al 2008)A broad subaerial platform was constructed by emplace-
ment of subaerial lava flows during an emergent subaerialstage (Fig 16c) The subaerial unit predominantly com-prises well-layered sheet flows that formed in a similarway to continental flood basalts as hundred(s) of pahoehoeflow fields erupted on shallow slopes (Self et al 1997) how-ever volcaniclastic and pillowed flow units are locally pre-sent mostly in the upper parts of the stratigraphyLocalized ponding of flows led to the formation of multi-tiered columnar jointing in some subaerial flows (egBondre et al 2004) As volcanism waned local interflowcarbonate deposits developed and plagioclase megacrysticflows were erupted Subsidence of the plateau is recordedby the interflow lenses and deposits of platform carbonates(40^1500m thick) on top of the Karmutsen basalts thatcontain Late Carnian to Early Norian fossils (Fig 16d c216^227 Ma Nixon et al 2008)The evolution of the internal architecture of the
Wrangellia oceanic plateau was primarily related to theconditions of emplacement (ie water depth and eruptionenvironment) There does not appear to have been a sys-tematic coincident evolution or shift of lava compositionsonVancouver Island as has been recognized in most con-tinental flood basalts and parts of the Wrangellia oceanicplateau in Alaska and Yukon (Greene et al 2008 2009)In general with the exception of the picritic pillow lavasin the submarine stratigraphy and local variations betweenhigh-MgO and tholeiitic lavas there are no significantcompositional differences between lavas erupted duringthe different stages of growth of the plateau (ie submarineand subaerial flows) nor are substantial compositional dif-ferences generally found between massive submarine flowsin contact with pillowed flows tholeiitic basalt flows nearthe base of the Karmutsen Formation are compositionallysimilar to flows sampled near the top of the volcanicsequence Where it was possible to sample continuous sec-tions of volcanic stratigraphy greater than several hun-dreds of meters (eg Keogh^Maynard Lake area west ofthe Karmutsen Range Fig 2) there is no clear relationshipbetween stratigraphic position and compositionalvariation
CONCLUSIONSThe Karmutsen Formation covers extensive areas ofVancouver Island (20 000 km2) and forms part of a largeoceanic plateau erupted over a geologically short interval(c 225^231 Ma) The tripartite volcanic stratigraphyexposed on Vancouver Island is at least 6 km thick andcomprises a succession of submarine flows volcaniclastic
deposits and massive subaerial flowsThese volcanologicaldifferences are primarily related to the eruption environ-ment (deep-water shallow-water or subaerial) The rapidgrowth of the plateau prevented significant accumulationsof intervening sediment except in the uppermost stratigra-phy where isolated limestone lenses commonly associatedwith pillow basalt and hyaloclastite deposits preserve arecord of the final stages of subsidence of the plateauThe Wrangellia plateau on Vancouver Island was con-
structed dominantly of tholeiitic basalt with restrictedmajor- and trace-element and isotopic compositional var-iations Minor volumes of picritic pillow basalts eruptedlate in the submarine phase The Karmutsen basalts wereemplaced onto older Paleozoic arc volcanic sequences andlow NbLa for the picrites may indicate a contributionfrom the sub-arc lithospheric mantle Modeling of the pet-rogenesis of the Karmutsen picrites indicates melting ofanomalously hot mantle (15008C) and extensive degreesof partial melting (23^27) and is consistent with aplume initiation model Combined Sr^Nd^Hf^Pb isotopicsystematics indicate that the picritic and tholeiitic lavaswere derived from a depleted plume-type mantle sourcedistinct from the source of MORB that has compositionalsimilarities to the source of the lavas forming theCaribbean Plateau Picrites were generated by melting ofspinel lherzolite whereas the tholeiitic basalts formed asaggregate melts of both garnet and spinel lherzolite Thetholeiitic basalts underwent extensive low-pressure fractio-nation (52^3 kbar) and seismic studies indicate that someof theWrangellia crust beneathVancouver Island may cor-respond to the plutonic residues of the Karmutsen floodbasalts
ACKNOWLEDGEMENTSWe are grateful to Nick Arndt for helping us get this pro-ject started and Nick Massey for insights into VancouverIsland geology We greatly appreciate the assistance withPRIMELT modeling from Claude Herzberg We thankMikkel Schau for his insight and enthusiasm during field-work Jane Barling assisted with analyses by MC-ICP-MSThoughtful and thorough reviews of the manuscript byJohn Mahoney Fred Frey Wendy Bohrson DougalJerram and Teal Riley are greatly appreciated Fundingwas generously provided by the Rocks to Riches Programadministered by the BC amp Yukon Chamber of Mines in2004 by the BC Geological Survey and by NSERCDiscovery Grants to JSS and DW ARG was supportedby a University Graduate Fellowship at UBC
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
34
REFERENCESBarker F Brown A S Budahn J R amp Plafker G (1989) Back-arc
with frontal-arc component origin of Triassic Karmutsen basaltBritish Columbia Canada Chemical Geology 75 81^102
Blichert-Toft JWeis D Maerschalk C Agranier A amp Albare de F(2003) Hawaiian hot spot dynamics as inferred from the Hf and Pbisotope evolution of Mauna Kea volcano Geochemistry GeophysicsGeosystems 4(2) 1^27 doi1010292002GC000340
Bondre N R Duraiswami R A amp Dole G (2004) Morphologyand emplacement of flows from the Deccan Volcanic ProvinceIndia Bulletin of Volcanology 66 29^45
Carlisle D (1963) Pillow breccias and their aquagene tuffs QuadraIsland British Columbia Journal of Geology 71 48^71
Carlisle D (1972) Late Paleozoic to Mid-Triassic sedimentary-volcanic sequence on Northeastern Vancouver Island Geological
Society of Canada Paper Report of Activities 72-1 Part B 24^30Carlisle D amp SuzukiT (1974) Emergent basalt and submergent car-
bonate^clastic sequences including the Upper Triassic Dilleri andWelleri zones on Vancouver Island Canadian Journal of Earth
Sciences 11 254^279Chauvel C amp Blichert-Toft J (2001) A hafnium isotope and trace
element perspective on melting of the depleted mantle Earth and
Planetary Science Letters 190 137^151Chayes F (1954)The theory of thin-section analysis Journal of Geology
62 92^101Cho M Liou J G amp Maruyama S (1987) Transition from the zeo-
lite to prehnite^pumpellyite facies in the Karmutsen metabasitesVancouver Island British Columbia Journal of Petrology 27(2)467^494
Clowes R M Zelt C A Amor J R amp Ellis R M (1995)Lithospheric structure in the southern Canadian Cordillera froma network of seismic refraction lines Canadian Journal of Earth
Sciences 32(10) 1485^1513Coffin M F amp Eldholm O (1994) Large igneous provinces Crustal
structure dimensions and external consequences Reviews of
Geophysics 32(1) 1^36Cox K G (1980) A model for flood basalt volcanism Journal of
Petrology 21 629^650Eldholm O amp Coffin M F (2000) Large igneous provinces
and plate tectonics In Richards M A Gordon R G amp vander Hilst R D (eds) The History and Dynamics of Global
Plate Motions American Geophysical Union Geophysical Monograph 121309^326
Elliott T R Hawkesworth C J amp Grolaquo nvold K (1991) Dynamicmelting of the Iceland plume Nature 351 201^206
Escuder-Viruete J Pecurren rez-Estaucurren n A Contreras F Joubert MWeis D Ullrich T D amp Spadea P (2007) Plume mantle sourceheterogeneity through time Insights from the Duarte ComplexHispaniola northeastern Caribbean Journal of Geophysical Research112(B04203) doi1010292006JB004323
Farnetani C G Richards M A amp Ghiorso M S (1996)Petrological models of magma evolution and deep crustal structurebeneath hotspots and flood basalt provinces Earth and Planetary
Science Letters 143 81^94Fitton J G amp Godard M (2004) Origin and evolution of magmas
on the Ontong Java Plateau In Fitton J G Mahoney J JWallace P J amp Saunders A D (eds) Origin and Evolution of the
Ontong Java Plateau Geological Society London Special Publications 229151^178
Fitton J G Saunders A D Norry M J Hardarson B S ampTaylor R N (1997) Thermal and chemical structure of theIceland plume Earth and Planetary Science Letters 153 197^208
Fitton J G Saunders A D Kempton P D amp Hardarson B S(2003) Does depleted mantle form an intrinsic part of the Icelandplume Geochemistry Geophysics Geosystems 4(3) 1032 doi1010292002GC000424
Frey F A Huang S Blichert-Toft J Regelous M amp Boyet M(2005) Origin of depleted components in basalt related to theHawaiian hotspot Evidence from isotopic and incompatible ele-ment ratios Geochemistry Geophysics Geosystems 6(1) 23 doi1010292004GC000757
Galer S J G amp Abouchami W (1998) Practical application of leadtriple spiking for correction of instrumental mass discriminationMineralogical Magazine 62A 491^492
Ghiorso M S amp Sack R O (1995) Chemical mass transfer in mag-matic processes IV A revised and internally consistent thermody-namic model for the interpolation and extrapolation of liquid^solidequilibria in magmatic systems at elevated temperatures and pres-sures Contributions to Mineralogy and Petrology 119 197^212
Greene A R Scoates J S amp Weis D (2005)WrangelliaTerrane onVancouver Island Distribution of flood basalts with implicationsfor potential Ni^Cu^PGE mineralization In Grant B (ed)Geological Fieldwork 2004 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2005-1 209^220Greene A R Scoates J S Nixon G T amp Weis D (2006) Picritic
lavas and basal sills in the Karmutsen flood basalt provinceWrangellia northern Vancouver Island In Grant B (ed)Geological Fieldwork 2005 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2006-1 39^52Greene A R Scoates J S amp Weis D (2008) Wrangellia flood
basalts in Alaska A record of plume^lithosphere interaction in aLate Triassic accreted oceanic plateau Geochemistry Geophysics
Geosystems 9(Q12004) doi1010292008GC002092Greene A R Scoates J S Weis D amp Israel S (2009)
Geochemistry of triassic flood basalts from the Yukon (Canada)segment of the accreted Wrangellia oceanic plateau Lithosdoi101016jlithos200811010
Hauff F Hoernle K Tilton G Graham D W amp Kerr A C(2000) Large volume recycling of oceanic lithosphere over shorttime scales geochemical constraints from the Caribbean LargeIgneous Province Earth and Planetary Science Letters 174 247^263
Hauff F Hoernle K amp Schmidt A (2003) Sr^Nd^Pb compositionof Mesozoic Pacific oceanic crust (Site 1149 and 801 ODP Leg185)Implications for alteration of ocean crust and the input into theIzu^Bonin^Mariana subduction system Geochemistry Geophysics
Geosystems 4(8) doi1010292002GC000421Herzberg C (2004) Partial melting below the Ontong Java Plateau
In Fitton J G Mahoney J J Wallace P J amp Saunders A D(eds) Origin and Evolution of the Ontong Java Plateau Geological SocietyLondon Special Publications 229 179^183
Herzberg C (2006) Petrology and thermal structure of the Hawaiianplume from Mauna Kea volcano Nature 444(7119) 605
Herzberg C amp Asimow P D (2008) Petrology of some oceanicisland basalts PRIMELT2XLS software for primary magma cal-culation Geochemistry Geophysics Geosystems 9(Q09001) doi1010292008GC002057
Herzberg C amp OrsquoHara M J (2002) Plume-associated ultramaficmagmas of Phanerozoic age Journal of Petrology 43(10) 1857^1883
Herzberg C Asimow P D Arndt N Niu Y Lesher C MFitton J G Cheadle M J amp Saunders A D (2007)Temperatures in ambient mantle and plumes Constraints frombasalts picrites and komatiites Geochemistry Geophysics Geosystems8(Q02006) doi1010292006GC001390
Huang S amp Frey F A (2005) Trace element abundances of MaunaKea basalt from phase 2 of the Hawaii Scientific Drilling Project
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
35
Petrogenetic implications of correlations with major element con-tent and isotopic ratios Geochemistry Geophysics Geosystems 4(6)doi1010292002GC000322
Hussong D M Wipperman L K amp Kroenke L W (1979) Thecrustal structure of the Ontong Java and Manihiki OceanicPlateaus Journal of Geophysical Research 84(B11) 6003^6010
Jerram D A amp Widdowson M (2005) The anatomy of ContinentalFlood Basalt Provinces geological constraints on the processes andproducts of flood volcanism Lithos 79(3^4) 385^405
Jones D L Silberling N J amp Hillhouse J (1977)Wrangellia a dis-placed terrane in northwestern North America CanadianJournal ofEarth Sciences 14(11) 2565^2577
Kerr A C (2003) Oceanic Plateaus In Rudnick R (ed) The Crust
Treatise on GeochemistryVol 3 Oxford Elsevier Science pp 537^565Kerr A C amp Mahoney J J (2007) Oceanic plateaus Problematic
plumes potential paradigms Chemical Geology 241 332^353Kerr A CTarney J Marriner G F Klaver GT Saunders A D
amp Thirlwall M F (1996) The geochemistry and petrogenesis ofthe Late-Cretaceous picrites and basalts of Curacao NetherlandsAntilles a remnant of an oceanic plateau Contributions to
Mineralogy and Petrology 124(1) 29^43Kerr A C Marriner G F Tarney J Nivia A Saunders A D
Thirlwall M F amp Sinton A C (1997) Cretaceous basaltic ter-ranes in Western Colombia Elemental chronological and Sr-Ndisotopic constraints on petrogenesis Journal of Petrology 38(6)677^702
Kinzler R J (1997) Melting of mantle peridotite at pressuresapproaching the spinel to garnet transition Application to mid-ocean ridge basalt petrogenesis Journal of Geophysical Research
102(B1) 853^874Kinzler R J amp Grove T L (1992) Primary magmas of mid-ocean
ridge basalts 1 Experiments and methods Journal of Geophysical
Research 97(B5) 6885^6906Korenaga J (2005) Why did not the Ontong Java Plateau form sub-
aerially Earth and Planetary Science Letters 234 385^399Kuniyoshi S (1972) Petrology of the Karmutsen (Volcanic) Group
northeastern Vancouver Island British Columbia PhD disserta-tion University of California Los Angeles 242 pp
Lassiter J C DePaolo D J amp Mahoney J J (1995) Geochemistry ofthe Wrangellia flood basalt province Implications for the role ofcontinental and oceanic lithosphere in flood basalt genesis Journalof Petrology 36(4) 983^1009
LeBas M J (2000) IUGS reclassification of the high-Mg and picriticvolcanic rocks Journal of Petrology 41(10) 1467^1470
Longhi J (2002) Some phase equilibrium systematics of lherzolitemelting I Geochemistry Geophysics Geosystems 3(3) doi1010292001GC000204
MacDonald G A amp Katsura T (1964) Chemical composition ofHawaiian lavas Journal of Petrology 5 82^133
Mahoney J J (1987) An isotopic survey of Pacific oceanic plateausimplications for their nature and origin In Keating B HFryer P Batiza R amp Boehlert GW (eds) Seamounts Islands andAtolls American Geophysical Union Geophysical Monograph 43 207^220
Mahoney J LeRoex A P Peng Z Fisher R L amp Natland J H(1992) Southwestern limits of Indian Ocean Ridge mantle and theorigin of low 206Pb204Pb mid-ocean ridge basalt isotope systema-tics of the central Southwest Indian Ridge (178^508E) Journal ofGeophysical Research 97(B13) 19771^19790
Mahoney J J Sinton J M Kurz M D MacDougall J DSpencer K J amp Lugmair GW (1994) Isotope and trace elementcharacteristics of a super-fast spreading ridge East Pacific rise13^238S Earth and Planetary Science Letters 121 173^193
Massey N W D (1995a) Geology and mineral resources of theAlberni^Nanaimo Lakes sheet Vancouver Island 92F1W 92F2Eand part of 92F7E British Columbia Ministry of Energy Mines and
Petroleum Resources Paper 1992-2 132 ppMassey N W D (1995b) Geology and mineral resources of the
Cowichan Lake sheet Vancouver Island 92C16 British Columbia
Ministry of Energy Mines and Petroleum Resources Paper 1992-3 112 ppMassey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005a) Digital Geology Map of British Columbia Tile NM9 MidCoast BC British Columbia Ministry of Energy and Mines Geofile
2005-2Massey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005b) Digital Geology Map of British Columbia Tile NM10Southwest BC British Columbia Ministry of Energy and Mines Geofile
2005-3McDonough W F amp Sun S (1995) The composition of the Earth
Chemical Geology 120 223^253Monger JW H amp Journeay J M (1994) Basement geology and tec-
tonic evolution of theVancouver region In Monger JW H (ed)Geology and Geological Hazards of the Vancouver Region Southwestern
British Columbia Geological Survey of Canada Bulletin 481 3^25Muller J E (1977) Geology of Vancouver Island Geological Survey of
Canada Open File Map 463Muller J E (1980)The Paleozoic Sicker Group of Vancouver Island British
Columbia Geological Survey of Canada Paper 79-30 22 ppMuller J E Northcote K E amp Carlisle D (1974) Geology and
mineral deposits of Alert Bay^Cape Scott map area VancouverIsland British Columbia Geological Survey of Canada Paper 74-8 77
Neal C R Mahoney J J Kroenke L W Duncan R A ampPetterson M G (1997) The Ontong^Java Plateau InMahoney J J amp Coffin M F (eds) Large Igneous Provinces
Continental Oceanic and Planetary FloodVolcanism American Geophysical
Union Geophysical Monograph 100 183^216Niu Y Collerson K D Batiza R Wendt J I amp Regelous M
(1999) Origin of enriched-typed mid-ocean ridge basalt at ridgesfar from mantle plumes The East Pacific Rise at 11820rsquoN Journalof Geophysical Research 104 7067^7088
Nixon G T amp Orr A J (2007) Recent revisions to the EarlyMesozoic stratigraphy of Northern Vancouver Island (NTS 102I092L) and metallogenic implicatinos British Columbia InGrant B (ed) Geological Fieldwork 2006 British Columbia Ministry of
Energy Mines and Petroleum Resources Paper 2007-1 163^177Nixon GT Hammack J L KoyanagiV M Payie G J Haggart
JW Orchard M JTozerT Friedman R M Archibald D APalfy J amp Cordey F (2006a) Geology of the Quatsino^PortMcNeill area northernVancouver Island British Columbia Ministry
of Energy Mines and Petroleum Resources Geoscience Map 2006-2 scale150 000
Nixon G T Kelman M C Stevenson D Stokes L A ampJohnston K A (2006b) Preliminary geology of the Nimpkishmap area (NTS 092L07) northern Vancouver Island BritishColumbia In Grant B amp Newell J M (eds) Geological Fieldwork2005 British Columbia Ministry of Energy Mines and Petroleum Resources
Paper 2006-1 135^152Nixon G T Laroque J Pals A Styan J Greene A R amp
Scoates J S (2008) High-Mg lavas in the Karmutsen flood basaltsnorthernVancouver Island (NTS 092L) Stratigraphic setting andmetallogenic significance In Grant B (ed) Geological Fieldwork
2007 British Columbia Ministry of Energy Mines and Petroleum
Resources Paper 2008-1 175^190Nowell G M Kempton P D Noble S R Fitton J G
Saunders A Mahoney J J amp Taylor R N (1998) High precisionHf isotope measurements of MORB and OIB by thermal ionisation
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
36
mass spectrometry insights into the depleted mantle Chemical
Geology 149 211^233Parrish R R amp McNicoll V J (1992) U^Pb age determinations
from the southern Vancouver Island area British Columbia InRadiogenic Age and Isotopic Studies Report 5 Geological Survey of
Canada Paper 91-2 79^86Peate DW (1997) The Parana^Etendeka Province In Coffin M F
amp Mahoney J (eds) Large Igneous Provinces Continental Oceanic andPlanetary Flood Volcanism American Geophysical Union Geophysical
Monograph 100 217^245Perfit M R Fornari D J Ridley W I Kirk P D Casey J
Kastens K A Reynolds J R Edwards M Desonie DShuster R amp Paradis S (1996) Recent volcanism in the Siqueirostransform fault picritic basalts and implications for MORBmagma genesis Earth and Planetary Science Letters 141(1^4) 91^108
Pretorius W Weis D Williams G Hanano D Kieffer B ampScoates J S (2006) Complete trace elemental characterization ofgranitoid (USGSG-2GSP-2) reference materials by high resolutioninductively coupled plasma-mass spectrometry Geostandards and
Geoanalytical Research 30(1) 39^54Regelous M Niu Y Wendt J I Batiza R Greig A amp
Collerson K D (1999) Variations in the geochemistry of magma-tism on the East Pacific Rise at 10830rsquoN since 800 ka Earth and
Planetary Science Letters 168 45^63Recurren villon S Chauvel C Arndt N T Pik R Martineau F
Fourcade S amp Marty B (2002) Heterogeneity of the Caribbeanplateau mantle source Sr O and He isotopic compositions of oli-vine and clinopyroxene from Gorgona Island Earth and Planetary
Science Letters 205(1^2) 91Rhodes J M (1996) Geochemical stratigraphy of lava flows samples
by the Hawaii Scientific Drilling Project Journal of Geophysical
Research 101(B5) 11729^11746Rhodes J M amp Vollinger M J (2004) Composition of basaltic lavas
sampled by phase-2 of the Hawaii Scientific Drilling ProjectGeochemical stratigraphy and magma types Geochemistry
Geophysics Geosystems 5(3) doi1010292002GC000434Richards M A Jones D L Duncan R A amp DePaolo D J (1991)
A mantle plume initiation model for the Wrangellia flood basaltand other oceanic plateaus Science 254 263^267
Salters V J M amp Hart S R (1991) The mantle sources of oceanridges islands and arcs the Hf-isotope connection Earth and
Planetary Science Letters 104 364^380Salters V J M amp Stracke A (2004) Composition of the depleted
SaltersV J amp WhiteW (1998) Hf isotope constraints on mantle evo-lution Chemical Geology 145 447^460
SanoT amp Yamashita S (2004) Experimental petrology of basementlavas from Ocean Drilling Program Leg 192 implications for dif-ferentiation processes in Ontong Java Plateau magmas InFitton J G Mahoney J JWallace P J amp Saunders A D (eds)Origin and Evolution of the Ontong Java Plateau Geological Society
London Special Publications 229 185^218Saunders A D (2005) Large igneous provinces Origin and environ-
mental consequences Elements 1 259^263Self S ThordarsonT amp Keszthely L (1997) Emplacement of conti-
nental flood basalt lava flows In Mahoney J J amp Coffin M F(eds) Large Igneous Provinces Continental Oceanic and Planetary Flood
Volcanism American Geophysical Union Geophysical Monograph 100381^410
Shaw D M (2000) Continuous (dynamic) melting theory revisitedCanadian Mineralogist 38(5) 1041^1063
Sluggett C L (2003) Uranium^lead age and geochemical constraintson Paleozoic and Early Mesozoic magmatism in WrangelliaTerrane Saltspring Island British Columbia BSc thesisUniversity of British ColumbiaVancouver 84 pp
Storey M Mahoney J J amp Saunders A D (1997) Cretaceousbasalts in Madagascar and the transition between plume and con-tinental lithospheric mantle sources In Mahoney J J ampCoffin M F (eds) Large Igneous Provinces Continental Oceanic and
Planetary Flood Volcanism Aamerican Geophysical Union Geophysical
Monograph 100 95^122Surdam R C (1967) Low-grade metamorphism of the Karmutsen
Group PhD dissertation University of California Los Angeles288 pp
Sutherland-Brown A Yorath C J Anderson R G amp Dom K(1986) Geological maps of southern Vancouver IslandLITHOPROBE I 92C10 11 14 16 92F1 2 7 8 Geological Survey ofCanada Open File 1272
Tejada M L G Mahoney J J Castillo P R Ingle S P Sheth HC amp Weis D (2004) Pin-pricking the elephant evidence on theorigin of the Ontong Java Plateau from Pb^Sr^Hf^Nd isotopiccharacteristics of ODP Leg 192 basalts In Fitton J GMahoney J J Wallace P J amp Saunders A D (eds) Origin and
Evolution of the Ontong Java Plateau Geological Society London Special
Publications 229 133^150Thompson P M E Kempton P D amp Kerr A C (2008) Evaluation
of the effects of alteration and leaching on Sm^Nd and Lu^Hf sys-tematics in submarine mafic rocks Lithos 104(1^4) 164^176
Thompson P M E Kempton P D White R V Kerr A CTarney J Saunders A D Fitton J G amp McBirney A (2003)Hf-Nd isotope constraints on the origin of the CretaceousCaribbean plateau and its relationship to the Galapagos plumeEarth and Planetary Science Letters 217(1^2) 59^75
Vervoort J D Patchett P Blichert-Toft J amp Albare de F (1999)Relationships between Lu^Hf and Sm^Nd isotopic systems in theglobal sedimentary system Earth and Planetary Science Letters 16879^99
Walter M J (1998) Melting of garnet peridotite and the origin ofkomatiite and depleted lithosphere Journal of Petrology 39(1) 29^60
Weis D Kieffer B Maerschalk C Barling J de Jong JWilliams G A Hanano D Mattielli N Scoates J SGoolaerts A Friedman R A amp Mahoney J B (2006) High-pre-cision isotopic characterization of USGS reference materials byTIMS and MC-ICP-MS Geochemistry Geophysics Geosystems
7(Q08006) doi1010292006GC001283Weis D Kieffer B Hanano D Silva I N Barling J
Pretorius W Maerschalk C amp Mattielli N (2007) Hf isotopecompositions of US Geological Survey reference materialsGeochemistry Geophysics Geosystems 8(Q06006) doi1010292006GC001473
Wheeler J O amp McFeely P (1991) Tectonic assemblage map of theCanadian Cordillera and adjacent part of the United States ofAmerica Geological Survey of Canada Map 1712A
WhiteW M Albare de F amp Tecurren louk P (2000) High-precision analy-sis of Pb isotope ratios by multi-collector ICP-MS Chemical Geology167 257^270
Yorath C J Sutherland Brown A amp Massey N W D (1999)LITHOPROBE southern Vancouver Island British Columbia Geological
Survey of Canada Bulletin 498 145 ppZou H (1998) Trace element fractionation during modal and
non-model dynamic melting and open-system melting Amathematical treatment Geochimica et Cosmochimica Acta 62(11)1937^1945
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
37
Zou H amp Reid M R (2001) Quantitative modeling of trace elementfractionation during incongruent dynamic melting Geochimica et
Cosmochimica Acta 65(1) 153^162
APPENDIX SAMPLEPREPARAT ION ANDANALYT ICAL METHODSOnly the freshest rocks in the field were sampled and onlythe least altered samples were selected for chemical andisotopic analysis based on thorough petrographic inspec-tion Fifty-nine of the 129 collected samples were crushed(400 g) into small pieces 52mm in diameter in aRocklabs hydraulic piston crusher between WC plates tominimize contamination The coarse-crush was mixedand 100 g was powdered in a planetary mill using agatejars and balls cleaned with quartz sand between samples
ActLabs analytical methodsThe major- and trace-element compositions of the whole-rock powders were determined at Activation LaboratoriesLtd (Actlabs) in Ancaster Ontario Analytical techniquesand detection limits are also available from Actlabs (httpwwwactlabscommethsub_code4erehtm) The analyticalmethod for each of the elements analyzed is indicated inTable 2 For the major elements a 02 g sample was mixedwith a lithium metaborate^lithium tetraborate mixtureand fused in a graphite crucible The molten mixture waspoured into a 5 HNO3 solution and shaken until dis-solved (30min) The samples were analyzed for majoroxides and selected trace elements by inductively coupledplasma optical emission spectrometry (ICP-OES) on acombination simultaneous^sequential Thermo Jarrell-AshEnviro II system Internal calibration was achieved usinga variety of international reference materials (eg W-2BIR-1 DNC-1) and independent control samplesAdditional trace elements were analyzed by both instru-mental neutron activation analysis (INAA) and induc-tively couple plasma mass spectrometry (ICP-MS) ForINAA15^25 g of sample was weighed into small polyeth-ylene vials and irradiated with a control international ref-erence material CANMET WMS-1 and NiCr flux wires ata thermal neutron flux of 71012 neutronscm2 per s in theMcMaster Nuclear Reactor Following a 7 day waitingperiod the samples were measured on an Ortec high-purity Ge detector linked to a Canberra Series 95 multi-channel analyzer Activities for each element weredecay- and weight-corrected and compared with a detectorcalibration developed from multiple international certifiedreference materials For ICP-MS 025 g of sample wasdigested in HF followed by HNO3^HClO4 heated andtaken to dryness The samples were brought back into solu-tion with HCl Samples were analyzed using a PerkinElmer Optima 3000 ICP In-lab standards or certified ref-erence materials (eg W-2 BIR-1 DNC-1) were used for
quality control A total of 15 blind duplicates were ana-lyzed from the same sample powders to assessreproducibility
University of Massachusetts analyticalmethodsFourteen sample duplicate powders taken from the same100 g sample aliquot (high-MgO basalts and picrites)were also analyzed at the Ronald B Gilmore X-RayFluorescence (XRF) Laboratory at the University ofMassachusetts Major elements were measured on a fusedLa-bearing lithium borate glass disc using a SiemensMRS-400 spectrometer Trace element concentrations(Rb Sr Ba Ce Nb Zr Y Pb Zn Ga Ni Cr V) weremeasured on a separate powder pellet using a PhilipsPW2400 sequential spectrometer using a Rh tube LOIand ferrous iron measurements were made as describedby Rhodes amp Vollinger (2004) Precision and accuracy esti-mates for the data have been described by Rhodes (1996)and Rhodes amp Vollinger (2004) Results for each sampleare the average of two separate analyses and are shown inElectronic Appendix 2 for a comparison of analyses ofduplicate powders using XRFand ICP-MS
PCIGR analytical methodsA subset of 19 samples was selected for high-precisiontrace-element analysis and Sr Nd Pb and Hf isotopicanalysis at the Pacific Centre for Isotopic andGeochemical Research (PCIGR) at the University ofBritish Columbia (UBC) Samples were selected from the59 samples analyzed for whole-rock chemistry at ActLabsbased on major- and trace-element chemistry alteration(low LOI and petrographic alteration index) and samplelocation Samples were prepared for trace-element analysisat the PCIGR by the technique described by Pretorius et al(2006) on unleached rock powders Sample powders(100mg) were weighed in 7ml screw-top Savillex bea-kers and dissolved in 1ml 14N HNO3 and 5ml 48 HFon a hotplate for 48 h at 1308C with periodic ultrasonica-tion Samples were dried and redissolved in 6ml 6N HClon a hotplate for 24 h and then dried and redissolved in1ml concentrated HNO3 for 24 h before final dryingTrace element abundances were measured by high-resolution inductively coupled plasma-mass spectrometry(HR-ICP-MS) with aThermo Finnigan Element2 systemfollowing the procedures described by Pretorius et al(2006) within 24 h of redissolution HFSE and LILE weremeasured in medium-resolution mode at 2000 dilutionusing a PFATeflon spray chamber washed with aqua regiafor 3min between samples REE were measured in high-resolution mode and U and Pb in low-resolution modeat 2000 dilution using a glass spray chamber washedwith 2 HNO3 between samples Total procedural blanksand reference materials (BCR-2 BHVO-2) were analyzedwith the batch of samples Indium was used as an internal
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
38
standard in all samples and standard solutionsBackground and standard solutions were analyzed afterevery five samples to detect memory effects and massdrift Results for PCIGR trace-element analyses areshown in Electronic Appendix 3 for the full set of analyzedtrace elementsSample digestion for purification of Sr Nd Hf and Pb
for column chemistry began by weighing each samplepowder (400^500mg) prior to leaching All samples wereleached with 6N HCl and placed in an ultrasonic bath for15min Samples were rinsed twice with 18 M-cm H2Obetween each leaching step (15 total) until the supernatantliquid was clear [following the method of Mahoney (1987)]Samples were then dried on a hotplate for 24 h andweighed again Sample solutions were then prepared bydissolving 100^250mg of the leached powder dissolvedin 1ml 14N HNO3 and 10ml 48 HF on a hotplate for48 h at 1308C with periodic ultrasonication Samples weredried and redissolved in 6ml 6N HCl on a hotplate for24 h and then dried Pb was separated using anion exhangecolumns and the discard was used for Sr REE and Hfseparation Nd was separated from the other REE and Hfrequired two additional purification steps Detailed proce-dures for column chemistry for separating Sr Nd and Pbat the PCIGR have been described byWeis et al (2006) andHf purification has been described byWeis et al (2007) Srand Nd isotope ratios were measured by thermal ioniza-tion mass spectrometry (TIMS) on a Thermo FinniganTriton system in static mode with relay matrix rotation ona single Ta and double Re^Ta filament respectively Fourto five filaments per barrel of 21 were occupied by stan-dards (NBS 987 for Sr and LaJolla for Nd) for each barrelwhere samples were run Sample Sr and Nd isotopic com-positions were corrected for mass fractionation using86Sr88Srfrac14 01194 and 146Nd144Ndfrac14 07219 Each samplewas then normalized using the barrel average of the refer-ence material relative to the values of143Nd144Ndfrac14 0511858 and 87Sr86Srfrac14 0710248 (Weiset al 2006) During the course of the Vancouver Islandanalyses the La Jolla Nd standard gave an average valueof 05118566 (nfrac14 7) and NBS987 standard gave an aver-age of 07102408 (nfrac1411) (2 error is reported as times106) 147Sm144Nd errors are 15 or 0006 USGeological Survey (USGS) reference material BHVO-2was processed with the samples and yielded Sr and Nd iso-topic ratios of 07034607 and 05129786 respectivelyThese are in agreement with the published values of070347920 and 051298411 respectively (Weis et al2006)
Pb and Hf isotopic compositions were analyzed by mul-tiple collector-inductively coupled plasma-mass spectrome-try (MC-ICP-MS) on a static multi-collection on a NuPlasma (Nu Instruments) system The detailed analyticalprocedure for Pb isotopic analyses on the Nu Plasma atthe PCIGR has been described by Weis et al (2006) Theconfiguration for Pb analyses allows for collection of PbTl and Hg togetherTl and Hg are used to monitor instru-mental mass discrimination and isobaric overlap repec-tively All sample solutions were analyzed withapproximately the same PbTl ratio (4) as the referencematerial NIST SRM 981 To accomplish this a small ali-quot of each sample solution from the Pb columns was ana-lyzed on the Element2 to determine the precise amount ofPb available for analysis on the Nu Plasma The SRM 981standard was run after every two samples on the NuPlasma During the time samples were run analyses ofthe SRM 981 Pb reference material gave values of206Pb204Pbfrac1416940322 207Pb204Pbfrac1415495823and 208Pb204Pbfrac14 36713164 (nfrac14 61 2 error is reportedas times 104) these values are in excellent agreement withreported TIMS triple-spike values of Galer amp Abouchami(1998) Fractionation-corrected Pb isotopic ratios were fur-ther corrected by the sample-standard bracketing methodor the ln^ln correction method described by White et al(2000) and Blichert-Toft et al (2003) Leached powders ofUSGS reference material BHVO-2 yielded Pb isotopicratios of 206Pb204Pbfrac14186454 8 207Pb204Pbfrac14 1549105 and 208Pb204Pbfrac14 38222514 and BCR-2yielded 206Pb204Pbfrac141880466 207Pb204Pbfrac14 1562518 and 208Pb204Pbfrac14 3883496 (2 error isreported as times 104) These values are in agreement withthose for leached residues of BHVO-2 and BCR-2 fromWeis et al (2006)Hf isotopic compositions were analyzed following the
procedures detailed byWeis et al (2007) The configurationfor Hf analyses monitored Lu mass 175 andYb mass 172 toallow for interference correction to masses 174 and 176 Hfisotopic ratios were normalized internally for mass fractio-nation to a 179Hf177Hf ratio of 07325 using an exponentialcorrection Standards were run after every two samplesand sample results were normalized to the ratio of the in-run daily average and a 176Hf177Hf ratio for JMC-475 of0282160 During the course of analyses the Hf standardJMC-475 gave an average value 02821679 (nfrac14130)USGS reference materials BCR-2 and BHVO-2 were pro-cessed with the samples and yielded Hf isotopic ratios of02828675 and 02831005 respectively Publishedvalues for BCR-2 and BHVO-2 are 02828717 and0283104 8 respectively (Weis et al 2007)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
39
overlap the ranges of the picrites and tholeiitic basalts(Fig 9 Table 3)The measured Pb isotopic compositions of the tholeiitic
basalts are more radiogenic than those of the picrites andthe most magnesian Keogh Lake picrites have the leastradiogenic Pb isotopic compositions The range ofinitial Pb isotope ratios for the picrites is206Pb204Pbfrac1417868^18580 207Pb204Pbfrac1415547^15562and 208Pb204Pbfrac14 37873^38257 and the range for thetholeiitic basalts is 206Pb204Pbfrac1418782^19098207Pb204Pbfrac1415570^15584 and 208Pb204Pbfrac14 37312^38587 (Fig 10 Table 5) The coarse-grained mafic rocksoverlap the range of initial Pb isotopic compositions forthe tholeiitic basalts with 206Pb204Pbfrac1418652^19155207Pb204Pbfrac1415568^15588 and 208Pb204Pbfrac14 38153^38735 One high-MgO basalt has the highest initial206Pb204Pb (19242) and 207Pb204Pb (15590) and thelowest initial 208Pb204Pb (37676) The Pb isotopic ratiosfor Karmutsen samples define broadly linearrelationships in Pb isotope plots The age-corrected Pb
isotopic compositions of several picrites have been affectedby U and Pb (andor Th) mobility during alteration(Fig 10)
ALTERAT IONThe Karmutsen basalts have retained most of their origi-nal igneous structures and textures however secondaryalteration and low-grade metamorphism have producedzeolitic- and prehnite^pumpellyite-bearing mineral assem-blages (Table 1 Cho et al 1987) Alteration and metamor-phism have primarily affected the distribution of the LILEand the Sr isotopic systematics The Keogh Lake picriteshave been affected more by alteration than the tholeiiticbasalts and have higher LOI (up to 55wt ) variableLILE and higher measured Sr Nd and Hf and lowermeasured Pb isotope ratios than the tholeiitic basalts Srand Pb isotope compositions for picrites and tholeiiticbasalts show a relationship with LOI (Figs 9 and 10)whereas Nd and Hf isotopic compositions do not correlate
Table 3 Sr and Nd isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Rb Sr 87Sr86Sr 2m87Rb86Sr 87Sr86Srt Sm Nd 143Nd144Nd 2m eNd
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses carried out at the PCIGR thecomplete trace element analyses are given in Electronic Appendix 3
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
21
with LOI (not shown) The relatively high apparent initialSr isotopic compositions for the high-MgO lavas (up to07052) probably resulted from post-magmatic loss of Rbor an increase in 87Sr86Sr through addition of seawater Sr(eg Hauff et al 2003) High-MgO lavas from theCaribbean plateau have similarly high 87Sr86Sr comparedwith basalts (Recurren villon et al 2002)The correction for in situdecay on initial Pb isotopic ratios has been affected bymobilization of U and Th (and Pb) in the whole-rockssince their formation A thorough acid leaching duringsample preparation was used that has been shown to effec-tively remove alteration phases (Weis et al 2006 NobreSilva et al in preparation) Whole-rock SmNd ratiosand age-correction of Nd isotopes may be affected bythe leaching procedure for submarine rocks older than50 Ma (Thompson et al 2008 Fig 9) there ishowever very little variation in Nd isotopic compositionsof the picrites or tholeiitic basalts and this system does notappear to be disturbed by alteration The HFSE abun-dances for tholeiitic and picritic basalts exhibit clear
linear correlations in binary diagrams (Fig 8) whereasplots of LILE vs HFSE are highly scattered (not shown)because of the mobility of some of the alkali and alkalineearth LILE (Cs Rb Ba K) and Sr for most samplesduring alterationThe degree of alteration in the Karmutsen samples does
not appear to be related to eruption environment or depthof burial Pillow rims are compositionally different frompillow cores (Surdam1967 Kuniyoshi1972) and aquagenetuffs are chemically different from isolated pillows withinpillow breccias however submarine and subaerial basaltsexhibit a similar degree of alterationThere is no clear cor-relation between the submarine and subaerial basalts andsome commonly used chemical alteration indices (eg BaRb vs K2OP2O5 Huang amp Frey 2005) There is also nodefinitive correlation between the petrographic alterationindex (Table 1) and chemical alteration indices although10 of 13 tholeiitic basalts with the highest K and LILEabundances have the highest petrographic alterationindex of three (seeTable 1)
Table 4 Hf isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Lu Hf 176Hf177Hf 2m eHf176Lu177Hf 176Hf177Hft eHf(t)
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
OLIV INE ACCUMULAT ION INH IGH-MgO AND PICR IT IC LAVASGeochemical trends and petrographic characteristics indi-cate that accumulation of olivine played an important rolein the formation of the high-MgO basalts and picrites fromthe Keogh Lake area on northern Vancouver Island TheKeogh Lake samples show a strong linear correlation inplots of Al2O3 TiO2 ScYb and Ni vs MgO and many ofthe picrites have abundant clusters of olivine pseudomorphs(Table1 Fig11)There is a strong linear correlationbetween
modal per cent olivine and whole-rock magnesium contentmagnesium contents range between 10 and 19wt MgOand the proportion of olivine phenocrysts in these samplesvaries from 0 to 42 vol (Fig 11)With a few exceptionsboth the size range and median size of the olivine pheno-crysts in most samples are comparable The clear correla-tion between the proportion of olivine phenocrysts andwhole-rock magnesium contents indicates that accumula-tion of olivine was directly responsible for the high MgOcontents (410wt ) in most of the picritic lavas
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
207Pb204Pb
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
208Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
(a) (b)
(c) (e)
4723A24723A2
4723A2
4723A2
4722A4
4723A4
4722A4
4723A4
4723A134723A3
4723A3
4723A13
1556
1558
1560
1562
1564
1566
1568
186 190 194 198 202 206 2101550
1552
1554
1556
1558
1560
178 180 182 184 186 188 190 192 194
380
384
388
392
396
186 190 194 198 202 206 210374
378
382
386
390
178 180 182 184 186 188 190 192 194
206Pb204Pb(d)
LOI (wt )
0
1
2
3
4
5
6
7
186 190 194 198 202 206 210
4723A2
Fig 10 Pb isotopic compositions of leached whole-rock samples measured by MC-ICP-MS for the Karmutsen Formation Error bars are smal-ler than symbols (a) Measured 207Pb204Pb vs 206Pb204Pb (b) Initial 207Pb204Pb vs 206Pb204Pb Age-correction to 230 Ma (c) Measured208Pb204Pb vs 206Pb204Pb (d) LOI vs 206Pb204Pb (e) Initial 208Pb204Pb vs 206Pb204Pb Complete chemical duplicates shown in Table 5(samples 4720A7 and 4722A4) are circled in each plot The dashed lines in (b) and (e) show the differences in age-corrections for two picrites(4723A4 4722A4) using the measured UTh and Pb concentrations for each sample and the age-corrections when concentrations are used fromthe two picrites (4723A3 4723A13) that appear to be least affected by alteration
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
23
DISCUSS IONThe compositional and spatial development of the eruptedlava sequences of the Wrangellia oceanic plateau onVancouver Island provide information about the evolutionof a mantle plume upwelling beneath oceanic lithospherein an analogous way to the interpretation of continentalflood basalts The Caribbean and Ontong Java plateausare the two primary examples of accreted parts of oceanicplateaus that have been studied to date and comparisonscan be made with the Wrangellia oceanic plateauHowever the Wrangellia oceanic plateau is a distinctiveoccurrence of an oceanic plateau because it formed on thecrust of a Devonian to Mississippian intra-oceanic islandarc overlain by Mississippian to Permian limestones andpelagic sediments The nature and thickness of oceaniclithospheric mantle are considered to play a fundamentalrole in the decompression and evolution of a mantleplume and thus the compositions of the erupted lavasequences (Greene et al 2008) The geochemical and
stratigraphic relationships observed in the KarmutsenFormation onVancouver Island and the focus of the subse-quent discussion sections provide constraints on (1) thetemperature and degree of melting required to generatethe primary picritic magmas (2) the composition of themantle source (3) the depth of melting and residual min-eralogy in the source region (4) the low-pressure evolutionof the Karmutsen tholeiitic basalts (5) the growth of theWrangellia oceanic plateau
Melting conditions and major-elementcomposition of the primary magmasThe primary magmas to most flood basalt provinces arebelieved to be picritic (eg Cox 1980) and they have beenfound in many continental and oceanic flood basaltsequences worldwide (eg Siberia Karoo Paranacurren ^Etendeka Caribbean Deccan Saunders 2005) Near-pri-mary picritic lavas with low total alkali abundances
Table 5 Pb isotopic compositions of Karmutsen basaltsVancouver Island BC
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
require high-degree partial melting of the mantle source(eg Herzberg amp OrsquoHara 2002) The Keogh Lake picritesfrom northernVancouver Island are the best candidates foridentifying the least modified partial melts of the mantleplume source despite having accumulated olivine pheno-crysts and can be used to estimate conditions of meltingand the composition of the primary magmas for theKarmutsen FormationPrimary magma compositions mantle potential tem-
peratures and source melt fractions for the picritic lavas ofthe Karmutsen Formation were calculated from primitivewhole-rock compositions using PRIMELT1XLS software(Herzberg et al 2007) and are presented in Fig 12 andTable 6 A detailed discussion of the computationalmethod has been given by Herzberg amp OrsquoHara (2002)and Herzberg et al (2007) key features of the approachare outlined belowPrimary magma composition is calculated for a primi-
tive lava by incremental addition of olivine and compari-son with primary melts of mantle peridotite Lavas thathad experienced plagioclase andor clinopyroxene fractio-nation are excluded from this analysis All calculated pri-mary magma compositions are assumed to be derived byaccumulated fractional melting of fertile peridotite KR-4003 Uncertainties in fertile peridotite composition donot propagate to significant variations in melt fractionand mantle potential temperature melting of depletedperidotite propagates to calculated melt fractions that aretoo high but with a negligible error in mantle potential
temperature PRIMELT1XLS calculates the olivine liqui-dus temperature at 1atm which should be similar to theactual eruption temperature of the primary magmaand is accurate to 308C at the 2 level of confidenceMantle potential temperature is model dependent witha precision that is similar to the accuracy of the olivineliquidus temperature (C Herzberg personal communica-tion 2008)With regard to the evidence for accumulated olivine in
the Keogh Lake picrites it should not matter whether themodeled lava composition is aphyric or laden with accu-mulated olivine phenocrysts PRIMELT1 computes theprimary magma composition by adding olivine to a lavathat has experienced olivine fractionation in this way itinverts the effects of fractional crystallization The onlyrequirement is that the olivine phenocrysts are not acci-dental or exotic to the lava compositions Support for thisprocedure can be seen in the calculated olivine liquid-line-of-descent shown by the curved arrays with 5 olivineaddition increments (Fig 12c) Fractional crystallization ofolivine from model primary magmas having 85^95wt FeO and 153^175wt MgO will yield arrays of deriva-tive liquids and randomly sorted olivines that in most butnot all cases connect the picrites and high-MgO basalts Anewer version of the PRIMELT software (Herzberg ampAsimow 2008) can perform olivine addition to and sub-traction from lavas with highly variably olivine phenocrystcontents and yields results that converge with those shownin Fig 12
Fig 11 Relationship between abundance of olivine phenocrysts and whole-rock MgO contents for the Keogh Lake picrites (a) Tracings ofolivine grains (gray) in six high-MgO samples made from high resolution (2400 dpi) scans of petrographic thin-sections Scale bars represent10mm sample numbers are indicated at the upper left (b) Modal olivine vs whole-rock MgO Continuous line is the best-fit line for 10 high-MgO samplesThe areal proportion of olivine was calculated from raster images of olivine grains using ImageJ image analysis software whichprovides an acceptable estimation of the modal abundance (eg Chayes 1954)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
25
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
Gray lines = final melting pressure
L + Ol + Opx
07
08
09
Pressure (GPa)
10
3
4
5
6
1
SOLIDUS
01
04
0506
L + Ol
2
03
02
PeridotiteKR-4003
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
Thick black lines = initial melting pressure
Dashed lines = melt fraction
Melt Fraction
2
3 4 5 6
L+Ol+Opx
SolidusSpinel
SolidusGarnet Peridotite
7
L+Ol
0 10 20 30 4040
45
50
55
60
MgO (wt)
SiO2
(wt)
PeridotiteKR-4003
Melt Fraction
0 10 20 305
10
15
CaO(wt)
MgO (wt)
3
4 5 6 7
2
46
2
Peridotite Partial Melts
Thick black line = initial melting pressureGray lines = final melting pressure
Peridotite
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
02
Pressure (GPa)
0302
04
Pressure (GPa)
Picrite
High-MgO basalt
Tholeiitic basalt
(c)
(d)
Karmutsen flood basaltsOlivine addition modelOlivine addition (5 increment)Primary magma
PicriteHigh-MgO basaltTholeiitic basalt
(a)
(b)
800 850 900
910
920
930
940
950
Karmutsen flood basalts
Olivine addition modelOlivine addition (5 increment)Primary magma
Gray lines = Fo contents of olivine
Fig 12 Estimated primary magma compositions for three Keogh Lake picrites and high-MgO basalts (samples 4723A4 4723A135616A7) using the forward and inverse modeling technique of Herzberg et al (2007) Karmutsen compositions and modeling results areoverlain on diagrams provided by C Herzberg (a) Whole-rock FeO vs MgO for Karmutsen samples from this study Total iron estimatedto be FeO is 090 Gray lines show olivine compositions that would precipitate from liquid of a given MgO^FeO composition Black lineswith crosses show results of olivine addition (inverse model) using PRIMELT1 (Herzberg et al 2007) Gray field highlights olivinecompositions with Fo contents of 91^92 predicted to precipitate (b) FeO vs MgO showing Karmutsen lava compositions and results ofa forward model for accumulated fractional melting of fertile peridotite KR-4003 (Kettle River Peridotite Walter 1998) (c) SiO2 vsMgO for Karmtusen lavas and model results (d) CaO vs MgO showing Karmtusen lava compositions and model results To brieflysummarize the technique [see Herzberg et al (2007) for complete description] potential parental magma compositions for the high-MgOlava series were selected (highest MgO and appropriate CaO) and using PRIMELT1 software olivine was incrementally added tothe selected compositions to show an array of potential primary magma compositions (inverse model) Then using PRIMELT1 theresults from the inverse model were compared with a range of accumulated fractional melts for fertile peridotite derived fromparameterization of the experimental results for KR-4003 from Walter (1998) forward model Herzberg amp OrsquoHara (2002) A melt fractionwas sought that was unique to both the inverse and forward models (Herzberg et al 2007) A unique solution was found when there was a com-mon melt fraction for both models in FeO^MgO and CaO^MgO^Al2O3^SiO2 (CMAS) projection space This modeling assumes thatolivine was the only phase crystallizing and ignores chromite precipitation and possible augite fractionation in the mantle (Herzbergamp OrsquoHara 2002) Results are best for a residue of spinel lherzolite (not pyroxenite) The presence of accumulated olivine in samples ofthe Keogh Lake picrites used as starting compositions does not significantly affect the results because we are modeling addition of olivine(see text for further description) The tholeiitic basalts cannot be used for modeling because they are all saturated in plagthorn cpxthornolThick black line for initial melting pressure at 4 GPa is not shown in (c) for clarity Gray area in (a) indicates parental magmas thatwill crystallize olivine with Fo 91^92 at the surface Dashed lines in (c) indicate melt fraction Solidi for spinel and garnet peridotite arelabelled in (c)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
The PRIMELT1 results indicate that if the Karmutsenpicritic lavas were derived from accumulated fractionalmelting of fertile peridotite the primary magmas wouldhave contained 15^17wt MgO and 10wt CaO(Fig 12 Table 6) formed from 23^27 partial meltingand would have crystallized olivine with a Fo content of 91 (Fig 12 Table 6) Calculated mantle temperatures forthe Karmutsen picrites (14908C) indicate melting ofanomalously hot mantle (100^2008C above ambient) com-pared with ambient mantle that produces mid-ocean ridgebasalt (MORB 1280^14008C Herzberg et al 2007)which initiated between 3 and 4 GPa Several picritesappear to be near-primary melts and required very littleaddition of olivine (eg sample 93G171 with measured158wt MgO) These temperature and melting esti-mates of the Karmutsen picrites are consistent withdecompression of hot mantle peridotite in an actively con-vecting plume head (ie plume initiation model) Threesamples (picrite 5614A1 and high-MgO basalts 5614A3and 5614A5) are lower in iron than the other Karmutsenlavas with FeO in the 65^80wt range (Fig 12c) andhave MgO and FeO contents that are similar to primitiveMORB from the Sequeiros fracture zone of the EastPacific Rise (EPR Perfit et al 1996) These samples
however differ from EPR MORB in having higher SiO2
and lower Al2O3 and they are restricted geographicallyto one area (Sara Lake Fig 2)We therefore have evidence for primary magmas with
MgO contents that range from a possible low of 110wt to as high as 174wt with inferred mantle potential tem-peratures from 1340 to 15208C This is similar to the rangedisplayed by lavas from the Ontong Java Plateau deter-mined by Herzberg (2004) who found that primarymagma compositions assuming a peridotite sourcewould contain 17wt MgO and 10wt CaO(Table 6) and that these magmas would have formed by27 melting at a mantle potential temperature of15008C (first melting occurring at 36 GPa) Fitton ampGodard (2004) estimated 30 partial melting for theOntong Java lavas based on the Zr contents of primarymagmas of Kroenke-type basalt calculated by incrementaladdition of equilibrium olivine However if a considerableamount of eclogite was involved in the formation of theOntong Java plateau magmas the excess temperatures esti-mated by Herzberg et al (2007) may not be required(Korenaga 2005) The variability in mantle potential tem-perature for the Keogh Lake picrites is also similar to thewide range of temperatures that have been calculated for
Table 6 Estimated primary magma compositions for Karmutsen basalts and other oceanic plateaus or islands
Sample 4723A4 4723A13 5616A7 93G171 Average OJP Mauna Keay Gorgonaz
Ontong Java primary magma composition for accumulated fraction melting (AFM) from Herzberg (2004)yMauna Kea primary magma composition is average of four samples of Herzberg (2006 table 1)zGorgona primary magma composition for AFM for 1F fertile source from Herzberg amp OrsquoHara (2002 table 4)Total iron estimated to be FeO is 090
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
27
lavas from single volcanoes in the Galapagos Hawaii andelsewhere by Herzberg amp Asimow (2008) Those workersinterpreted the range to reflect the transport of magmasfrom both the cool periphery and the hotter axis of aplume A thermally heterogeneous plume will yield pri-mary magmas with different MgO contents and the rangeof primary magma compositions and their inferred mantlepotential temperatures for Karmutsen lavas may reflectmelting from different parts of the plume
Source of the Karmutsen Formation lavasThe Sr^Nd^Hf^Pb isotopic compositions of theKarmutsen volcanic rocks on Vancouver Island indicatethat they formed from plume-type mantle containing adepleted component that is compositionally similar to butdistinct from other oceanic plateaus that formed in thePacific Ocean (eg Ontong Java and Caribbean Fig 13)Trace element and isotopic constraints can be used to testwhether the depleted component of the KarmutsenFormation was intrinsic to the mantle plume and distinctfrom the source of MORB or the result of mixing or invol-vement of ambient depleted MORB mantle with plume-type mantle The Karmutsen tholeiitic basalts have initialeNd and
87Sr86Sr ratios that fall within the range of basaltsfrom the Caribbean Plateau (eg Kerr et al 1996 1997Hauff et al 2000 Kerr 2003) and a range of Hawaiian iso-tope data (eg Frey et al 2005) and lie within and abovethe range for the Ontong Java Plateau (Tejada et al 2004Fig 13a) Karmutsen tholeiitic basalts with the highest ini-tial eNd and lowest initial 87Sr86Sr lie within and justbelow the field for age-corrected northern EPR MORB at230 Ma (eg Niu et al1999 Regelous et al1999) Initial Hfand Nd isotopic compositions for the Karmutsen samplesare noticeably offset to the low side of the ocean islandbasalt (OIB) array (Vervoort et al 1999) and lie belowand slightly overlap the low eHf end of fields for Hawaiiand the Caribbean Plateau (Fig 13c) the Karmutsen sam-ples mostly fall between and slightly overlap a field forage-corrected Pacific MORB at 230 Ma and Ontong Java(Mahoney et al 1992 1994 Nowell et al 1998 Chauvel ampBlichert-Toft 2001) Karmutsen lava Pb isotope ratios over-lap the broad range for the Caribbean Plateau and thelinear trends for the Karmutsen samples intersect thefields for EPR MORB Hawaii and Ontong Java in208Pb^206Pb space but do not intersect these fields in207Pb^206Pb space (Fig 13d and e) The more radiogenic207Pb204Pb for a given 206Pb204Pb of the Karmutsenbasalts requires high UPb ratios at an ancient time when235U was abundant similar to the Caribbean Plateau andenriched in comparison with EPR MORB Hawaii andOntong JavaThe low eHf^low eNd component of the Karmutsen
basalts that places them below the OIB mantle array andpartly within the field for age-corrected Pacific MORBcan form from low ratios of LuHf and high SmNd over
time (eg Salters amp White 1998) MORB will develop aHf^Nd isotopic signature over time that falls below themantle array Low LuHf can also lead to low 176Hf177Hffrom remelting of residues produced by melting in theabsence of garnet (eg Salters amp Hart 1991 Salters ampWhite 1998) The Hf^Nd isotopic compositions of theKarmutsen Formation indicate the possible involvementof depleted MORB mantle however similar to Iceland(Fitton et al 2003) Nb^Zr^Y variations provide a way todistinguish between a depleted component within theplume and a MORB-like component The Karmutsenbasalts and picrites fall within the Iceland array and donot overlap the MORB field in a logarithmic plot of NbYvs ZrY (Fig 13b) using the fields established by Fittonet al (1997) Similar to the depleted component within theCaribbean Plateau (Thompson et al 2003) the trend ofHf^Nd isotopic compositions towards Pacific MORB-likecompositions is not followed by MORB values in Nb^Zr^Y systematics and this suggests that the depleted compo-nent in the source of the Karmutsen basalts is distinctfrom the source of MORB and probably an intrinsic partof the plume The relatively radiogenic Pb isotopic ratiosand REE patterns distinct from MORB also support thisinterpretationTrace-element characteristics suggest the possible invol-
vement of sub-arc lithospheric mantle in the generation ofthe Karmutsen picrites Lassiter et al (1995) suggested thatmixing of a plume-type source with eNdthorn 6 to thorn 7 witharc material with low NbTh could reproduce the varia-tions observed in the Karmutsen basalts however theyconcluded that the absence of low NbLa ratios in theirsample of tholeiitic basalts restricts the amount of arc litho-sphere involved The study of Lassiter et al (1995) wasbased on major and trace element and Sr Nd and Pb iso-topic data for a suite of 29 samples from Buttle Lake in theStrathcona Provincial Park on central Vancouver Island(Fig 1) Whereas there are no clear HFSE depletions inthe Karmutsen tholeiitic basalts the low NbLa (and TaLa) of the Karmutsen picritic lavas in this study indicatethe possible involvement of Paleozoic arc lithosphere onVancouver Island (Fig 8)
REE modeling dynamic melting andsource mineralogyThe combined isotopic and trace-element variations of thepicritic and tholeiitic Karmutsen lavas indicate that differ-ences in trace elements may have originated from differentmelting histories For example the LuHf ratios of theKarmutsen picritic lavas (mean 008) are considerablyhigher than those in the tholeiitic basalts (mean 002)however the small range of overlapping initial eHf valuesfor the two lava suites suggests that these differences devel-oped during melting within the plume (Fig 9b) Below wemodel the effects of source mineralogy and depth of
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
28
melting on differences in trace-element concentrations inthe tholeiitic basalts and picritesDynamic melting models simulate progressive decom-
pression melting where some of the melt fraction isretained in the residue and only when the degree of partialmelting is greater than the critical mass porosity and the
source becomes permeable is excess melt extracted fromthe residue (Zou1998) For the Karmutsen lavas evolutionof trace element concentrations was simulated using theincongruent dynamic melting model developed by Zou ampReid (2001) Melting of the mantle at pressures greater than1 GPa involves incongruent melting (eg Longhi 2002)
OIB array
PacificMORB(230 Ma)
(0 Ma)
EPR(230 Ma)
-4
-2
0
2
4
8
10
12
14
16
18
20
22
-4 -2 0 2 4 6 8 10 12 14
minus2
0
2
4
6
8
10
12
14
0702 0703 0704 0705 0706
IndianMORB
87Sr 86Sr (initial)
Hf (initial)
(a) (c)
Nd (initial)
Karmutsen tholeiitic basalt
HawaiiOntong JavaCaribbean Plateau
Karmutsen high-MgO lava
high-MgO lavasKarmutsen
1540
1545
1550
1555
1560
1565
175 180 185 190 195 200375
380
385
390
395
175 180 185 190 195 200
(d) (e)
EPR
EPR
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb
Karmutsen Formation (initial)
HawaiiOntong JavaCaribbean Plateau
East Pacific Rise
Karmutsen Formation (measured)
Juan de FucaGorda
ε Nd
(initial)
Karmutsen Formation (different age-correction for 3 picrites)
(0 Ma)
NbY
001
01
1
1 10
ZrY
OIB
NMORB
(b)
Iceland array
6
Fig 13 Comparison of age-corrected (230 Ma) Sr^Nd^Hf^Pb isotopic compositions for Karmutsen flood basalts onVancouver Island withage-corrected OIB and MORB and Nb^Zr^Y variation (a) Initial eNd vs 87Sr86Sr (b) NbYand ZrY variation (after Fitton et al 1997)(c) Initial eHf vs eNd (d) Measured and initial 207Pb204Pb vs 206Pb204Pb (e) Measured and initial 208Pb204Pb vs 206Pb204Pb References fordata sources are listed in Electronic Appendix 4 Most of the compiled data were extracted from the GEOROC database (httpgeorocmpch-mainzgwdgdegeoroc) OIB array line in (c) is fromVervoort et al (1999) EPR East Pacific Rise Several high-MgO samples affected bysecondary alteration were not plotted for clarity in Pb isotope plots Estimation of fields for age-corrected Pacific MORB and EPR at 230 Mawere made using parent^daughter concentrations (in ppm) from Salters amp Stracke (2004) of Lufrac14 0063 Hffrac14 0202 Smfrac14 027 Ndfrac14 071Rbfrac14 0086 and Srfrac14 836 In the NbYvs ZrYplot the normal (N)-MORB and OIB fields not including Icelandic OIB are from data com-pilations of Fitton et al (1997 2003) The OIB field is data from 780 basalts (MgO45wt ) from major ocean islands given by Fitton et al(2003) The N-MORB field is based on data from the Reykjanes Ridge EPR and Southwest Indian Ridge from Fitton et al (1997) The twoparallel gray lines in (b) (Iceland array) are the limits of data for Icelandic rocks
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
29
with olivine being produced during the reactioncpxthornopxthorn spfrac14meltthornol for spinel peridotites The mod-eling used coefficients for incongruent melting reactionsbased on experiments on lherzolite melting (eg Kinzleramp Grove 1992 Longhi 2002) and was separated into (1)melting of spinel lherzolite (2) two combined intervals ofmelting for a garnet lherzolite (gt lherz) and spinel lher-zolite (sp lherz) source (3) melting of garnet lherzoliteThe model solutions are non-unique and a range indegree of melting partition coefficients melt reactioncoefficients and proportion of mixed source componentsis possible to achieve acceptable solutions (see Fig 14 cap-tion for description of modeling parameters anduncertainties)The LREE-depleted high-MgO lavas require melting of
a depleted spinel lherzolite source (Fig 14a) The modelingresults indicate a high degree of melting (22^25) similarto the results from PRIMELT1 based on major-elementvariations (discussed above) of a LREE-depleted sourceThe high-MgO lavas lack a residual garnet signature Theenriched tholeiitic basalts involved melting of both garnetand spinel lherzolite and represent aggregate melts pro-duced from continuous melting throughout the depth ofthe melting column (Fig 14b) Trace-element concentra-tions in the tholeiitic and picritic lavas are not consistentwith low-degree melting of garnet lherzolite alone(Fig 14c) The modeling results indicate that the aggregatemelts that formed the tholeiitic basalts would haveinvolved lesser proportions of enriched small-degreemelts (1^6 melting) generated at depth and greateramounts of depleted high-degree melts (12^25) gener-ated at lower pressure (a ratio of garnet lherzolite tospinel lherzolite melt of 14 provides the best fits seeFig 14b and description in figure caption) The totaldegree of melting for the tholeiitic basalts was high(23^27) similar to the degree of partial melting in thespinel lherzolite facies for the primary melts that eruptedas the Keogh Lake picrites of a source that was also prob-ably depleted in LREE
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
(c)
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
Melting of garnet lherzolite
High-MgO lavas
Melting of spinel lherzolite
source (07DM+03PM)
source (07DM+03PM)
6 melting
30 melting
(b)
(a)
30 melting
Sam
ple
Cho
ndrit
eS
ampl
eC
hond
rite 20 melting
1 melting
Melting of garnet lherzoliteand spinel lherzolite
x gt lherz + 4x sp lherz
5 meltingx gt lherz + 4x sp lherz
Sam
ple
Cho
ndrit
e
Tholeiitic lavas
Tholeiitic lavas
High-MgO lavas
source (07DM+03PM)
Fig 14 Trace-element modeling results for incongruent dynamicmantle melting for picritic and tholeiitic lavas from the KarmutsenFormation Three steps of modeling are shown (a) melting of spinellherzolite compared with high-MgO lavas (b) melting of garnet lher-zolite and spinel lherzolite compared with tholeiitic lavas (c) meltingof garnet lherzolite compared with tholeiitic and picritic lavasShaded fields in each panel for tholeiitic basalts and picrites arelabeled Patterns with symbols in each panel are modeling results(patterns are 5 melting increments in (a) and (b) and 1 incre-ments in (c) PM primitive mantle DM depleted MORB mantle gtlherz garnet lherzolite sp lherz spinel lherzolite In (b) the ratios ofper cent melting for garnet and spinel lherzolite are indicated (x gtlherzthorn 4x sp lherz means that for 5 melting 1 melt in garnetfacies and 4 melt in spinel facies where x is per cent melting in theinterval of modeling) The ratio of the respective melt proportions inthe aggregate melt (x gt lherzthorn 4x sp lherz) was kept constant andthis ratio of melt contribution provided the best fit to the data Arange of proportion of source components (07DMthorn 03PM to
09DMthorn 01PM) can reproduce the variation in the high-MgOlavas Melting modeling uses the formulation of Zou amp Reid (2001)an example calculation is shown in their Appendix PM fromMcDonough amp Sun (1995) and DM from Salters amp Stracke (2004)Melt reaction coefficients of spinel lherzolite from Kinzler amp Grove(1992) and garnet lherzolite fromWalter (1998) Partition coefficientsfrom Salters amp Stracke (2004) and Shaw (2000) were kept constantSource mineralogy for spinel lherzolite (018cpx027opx052ol003sp) from Kinzler (1997) and for garnet lherzolite(034cpx008opx053ol005gt) from Salters amp Stracke (2004) Arange of source compositions for melting of spinel lherzolite(07DMthorn 03PM to 03DMthorn 07PM) reproduce REE patternssimilar to those of the high-MgO lavas with best fits for 22ndash25melting and 07DMthorn 03PM source composition Concentrationsof tholeiitic basalts cannot be reproduced from an entirelyprimitive or depleted source
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
30
The uncertainty in the composition of the source in thespinel lherzolite melting model for the high-MgO lavasprecludes determining whether the source was moredepleted in incompatible trace elements than the source ofthe tholeiitic lavas (see Fig 14 caption for description) It ispossible that early high-pressure low-degree melting left aresidue within parts of the plume (possibly the hotter inte-rior portion) that was depleted in trace elements (egElliott et al 1991) further decompression and high-degreemelting of such depleted regions could then have generatedthe depleted high-MgO Karmutsen lavas Shallow high-degree melting would preferentially sample depletedregions whereas deeper low-degree melting may notsample more refractory depleted regions (eg CaribbeanEscuder-Viruete et al 2007) The picrites lie at the highend of the range of Nd isotopic compositions and havelow NbZr similar to depleted Icelandic basalts (Fittonet al 2003) therefore the picrites may represent the partsof the mantle plume that were the most depleted prior tothe initiation of melting As Fitton et al (2003) suggestedthese depleted melts may only rarely be erupted withoutcombining with more voluminous less depleted melts andthey may be detected only as undifferentiated small-volume flows like the Keogh Lake picrites within theKarmutsen Formation on Vancouver IslandDecompression melting within the mantle plume initiatedwithin the garnet stability field (35^25 GPa) at highmantle potential temperatures (Tp414508C) and pro-ceeded beneath oceanic arc lithosphere within the spinelstability field (25^09 GPa) where more extensivedegrees of melting could occur
Magmatic evolution of Karmutsentholeiitic basaltsPrimary magmas in large igneous provinces leave exten-sive coarse-grained cumulate residues within or below thecrust these mafic and ultramafic plutonic sequences repre-sent a significant proportion of the original magma thatpartially crystallized at depth (eg Farnetani et al 1996)For example seismic and petrological studies of OntongJava (eg Farnetani et al 1996) combined with the use ofMELTS (Ghiorso amp Sack 1995) indicate that the volcanicsequence may represent only 40 of the total magmareaching the Moho Neal et al (1997) estimated that30^45 fractional crystallization took place for theOntong Java magmas and produced cumulate thicknessesof 7^14 km corresponding to 11^22 km of flood basaltsdepending on the presence of pre-existing oceanic crustand the total crustal thickness (25^35 km) Interpretationsof seismic velocity measurements suggest the presence ofpyroxene and gabbroic cumulates 9^16 km thick beneaththe Ontong Java Plateau (eg Hussong et al 1979)Wrangellia is characterized by high crustal velocities in
seismic refraction lines which is indicative of mafic
plutonic rocks extending to depth beneath VancouverIsland (Clowes et al 1995)Wrangellia crust is 25^30 kmthick beneath Vancouver Island and is underlain by astrongly reflective zone of high velocity and density thathas been interpreted as a major shear zone where lowerWrangellia lithosphere was detached (Clowes et al 1995)The vast majority of the Karmutsen flows are evolved tho-leiitic lavas (eg low MgO high FeOMgO) indicating animportant role for partial crystallization of melts at lowpressure and a significant portion of the crust beneathVancouver Island has seismic properties that are consistentwith crystalline residues from partially crystallizedKarmutsen magmas To test the proportion of the primarymagma that fractionated within the crust MELTS(Ghiorso amp Sack 1995) was used to simulate fractionalcrystallization using several estimated primary magmacompositions from the PRIMELT1 modeling results Themajor-element composition of the primary magmas couldnot be estimated for the tholeiitic basalts because of exten-sive plagthorn cpxthornol fractionation and thus the MELTSmodeling of the picrites is used as a proxy for the evolutionof major elements in the volcanic sequence A pressure of 1kbar was used with variable water contents (anhydrousand 02wt H2O) calculated at the quartz^fayalite^magnetite (QFM) oxygen buffer Experimental resultsfrom Ontong Java indicate that most crystallizationoccurs in magma chambers shallower than 6 km deep(52 kbar Sano amp Yamashita 2004) however some crys-tallization may take place at greater depths (3^5 kbarFarnetani et al 1996)A selection of the MELTS results are shown in Fig 15
Olivine and spinel crystallize at high temperature until15^20wt of the liquid mass has fractionated and theresidual magma contains 9^10wt MgO (Fig 15f) At1235^12258C plagioclase begins crystallizing and between1235 and 11908C olivine ceases crystallizing and clinopyr-oxene saturates (Fig 15f) As expected the addition of clin-opyroxene and plagioclase to the crystallization sequencecauses substantial changes in the composition of the resid-ual liquid FeO and TiO2 increase and Al2O3 and CaOdecrease while the decrease in MgO lessens considerably(Fig 15) The near-vertical trends for the Karmutsen tho-leiitic basalts especially for CaO do not match the calcu-lated liquid-line-of-descent very well which misses manyof the high-CaO high-Al2O3 low-FeO basalts A positivecorrelation between Al2O3CaO and SrNd indicates thataccumulation of plagioclase played a major role in the evo-lution of the tholeiitic basalts (Fig 15g) Increasing thecrystallization pressure slightly and the water content ofthe melts does not systematically improve the match ofthe MELTS models to the observed dataThe MELTS results indicate that a significant propor-
tion of the crystallization takes place before plagioclase(15^20 crystallization) and clinopyroxene (35^45
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
31
crystallization) join the crystallization sequence Theseresults suggest that445 crystallization of the primarymagmas occurred (estimated from the residual liquid percent in Fig 15f) and if the original flood basalt stratigra-phy on Vancouver Island was 6 km thick at least anequivalent amount of complementary plutonic rocksshould have crystallized The Karmutsen basalts and theirplutonic residues thus represent a significant addition ofcrust (perhaps 412 km thickness) previously thickenedduring the formation of the underlying Paleozoic arc
Growth of the Wrangellia oceanic plateauon Vancouver IslandThe extrusive part of the Wrangellia oceanic plateau onVancouver Island was constructed in a three-layersequence of mostly tholeiitic basalts with a restrictedrange of composition (Fig 16) The plateau overliesDevonian to Mississippian arc volcanic sequences andmarine sediments of Mississippian to Permian ages thatare intruded by gabbroic rocks related to the flood basaltsThe volcanic stratigraphy of theWrangellia plateau formed
00
05
10
15
20
25
30
35
40
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
16
0 2 4 6 8 10 12 14 16 18 20
10
11
12
13
14
15
16
17
18
19
20
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
0 2 4 6 8 10 12 14 16 18 20
MgO (wt)
0 10 20 30 40 50
percent of total mass
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
) Mineral proportions
Sp (e)
Ol
CpxPlag
PicriteHigh-MgO basaltTholeiitic basalt
TiO2 (wt)
MgO (wt)
Al2O3 (wt) CaO (wt)
FeO (wt)
MgO(a) (b)
(c) (d)
MgO (wt)MgO (wt)
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
)
Residual liquid ()
1220
1190
Ol + Sp
Plag+Ol+Sp Plag+Cpx+Sp
02 wt H2O
no H2O
Residual liquid and crystallizing phases
(f)
0102030405060708090100
plagioclaseaccumulation
plagioclaseaccumulation
plagioclase
accumulatio
n
10
12
14
16
0 10 20 30 40
Al2O3CaO
SrNd
tholeiitic basalts
(g)
Fig 15 Forward fractional crystallization modeling results for major elements from MELTS (Ghiorso amp Sack 1995) compared with picriticand tholeiitic lavas from the Karmutsen Formation (a^d) Major-element compositions of residual liquids compared with Karmutsen lavacompositions For clarity the results are shown only for one starting composition (open circle) using the composition of the estimated primarymagma for the high-MgO lava series from the modeling technique of Herzberg et al (2007) for sample 4723A13 as shown in Fig 12 The esti-mated primary magmas for samples 4723A3 and 93G171 yield similar results A pressure of 1 kbar was used with no H2O and 02wt H2Ocalculated at QFM black line is anhydrous and grey line is 02wt H2OThe end-point for the anhydrous result is at 78 fractionation and11308C and the end-point for the result with 02wt H2O is 75 fractionation and11108C (e) Mineral proportions of crystallized phases fromone MELTrun (sample 4723A13) with no H2O (f) Per cent residual liquid vs temperature labelled with intervals of crystallizing phases (g)Al2O3CaO vs SrNd for the tholeiitic basalts
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
32
in a single eruptive phase (c 231^225 Ma) as an emergentbasalt sequence during a deeper-water shallow-water andsubaerial stage and subsided below sea level relativelysoon after its formation (Fig 16) The deeper-water stagecontains pillowed and massive submarine flows (3000mthick) that built up the submarine volcanic edifice from theseafloor The lowest pillow basalts were emplaced onunconsolidated and lithified fine-grained siliciclastic andcarbonaceous sediments containing Daonella beds(c 235^232 Ma) that are intruded by abundantKarmutsen mafic sills (Fig 16a) The different pillowedand massive submarine flows resulted primarily from dif-ferent effusive rates and local topography Some of themassive submarine flows commonly recognized by radiat-ing columnar jointing represent master tubes for the deliv-ery of lava to distal parts of submarine flow fields Thepillowed flows which commonly have uneven topography
in modern sequences may have served as levees to themassive submarine flows rather than allowing unob-structed spreading and inflation typical of the massive sub-aerial flows that form on shallow slopes in continentalflood basalt provinces (eg Self et al 1997) The rapid suc-cession of flows is evident by the near absence of sedimentbetween submarine flows Minor volumes of picritic pillowlavas occur in the upper part of the pillowed flow sequencein areas of the plateau exposed on northern VancouverIsland (Fig 16 Nixon et al 2008)The growth of the plateau in shallow water led to
emplacement of an increasing proportion of volcaniclastics(pillow breccias and hyaloclastites) that accumulated up to1500m thick in some areas (Fig 16b Nixon et al 2008)Horizons of pillowed flows occur locally within the volca-niclastic stratigraphy Sedimentary structures (eg gradedbedding fluidization structures) and petrographic textures
Late Devonian-Mississippian arc
Daonella beds
Middle Triassic
deeper water
Late Devonian-Mississippian arc
pillowed flows massive submarine
flows
pillow brecciahyaloclastiteshallow water
~4 km
~6 km
Late Triassic limestoneinterflow limestonelenses
Karmutsen Formation
mafic intrusions sediment-sill complex
Karmutsen Formation
(a) (b)
(c) (d)
subaerial flows
Late Devonian-Mississippian arc Late Devonian-Mississippian arc
picritic pillow basalts
Miss to Permian marine sediments
sills
Fig 16 Growth of the Wrangellia oceanic plateau onVancouver Island (a) Deep-water eruptive phase of pillowed and massive submarineflows with almost no intervening sediments onto MiddleTriassic shale containing Daonella beds (c 235^232 Ma) The basement of the plateauis composed of Devonian to Mississippian arc sequences overlain by Mississippian to Permian carbonates and siliciclastic sediments(b) Shallow-water flow emplacement with increasing proportion of volcaniclastics resulted primarily from cooling^contraction granulationPicritic pillow basalts occur primarily near the transition between pillowed submarine flows and volcaniclastic units Pillowed and massiveflows occur locally within the volcaniclastic unit (c) Emergent stage of growth consists of overlapping subaerial flow fields similar to continen-tal flood basalts with pahoehoe flow structures Multi-tiered columnar jointing occurrs locally in response to ponding of flows (eg Bondre et al2004) Interflow carbonate deposits and plagioclase megacrystic flows developed locally during the waning stage of volcanism (d) Post-volcanicsubsidence and deposition of Late Carnian to Norian (c 212^223 Ma) platform carbonates on top of the plateau Units are labeled in the dia-grams Thicknesses of the underlying units are not to scale
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
33
within the volcaniclastic rocks indicate formation primar-ily via cooling^contraction granulation magma^water^steam interaction autobrecciation resedimentation andmass-wasting rather than pyroclastic eruption (Nixonet al 2008)A broad subaerial platform was constructed by emplace-
ment of subaerial lava flows during an emergent subaerialstage (Fig 16c) The subaerial unit predominantly com-prises well-layered sheet flows that formed in a similarway to continental flood basalts as hundred(s) of pahoehoeflow fields erupted on shallow slopes (Self et al 1997) how-ever volcaniclastic and pillowed flow units are locally pre-sent mostly in the upper parts of the stratigraphyLocalized ponding of flows led to the formation of multi-tiered columnar jointing in some subaerial flows (egBondre et al 2004) As volcanism waned local interflowcarbonate deposits developed and plagioclase megacrysticflows were erupted Subsidence of the plateau is recordedby the interflow lenses and deposits of platform carbonates(40^1500m thick) on top of the Karmutsen basalts thatcontain Late Carnian to Early Norian fossils (Fig 16d c216^227 Ma Nixon et al 2008)The evolution of the internal architecture of the
Wrangellia oceanic plateau was primarily related to theconditions of emplacement (ie water depth and eruptionenvironment) There does not appear to have been a sys-tematic coincident evolution or shift of lava compositionsonVancouver Island as has been recognized in most con-tinental flood basalts and parts of the Wrangellia oceanicplateau in Alaska and Yukon (Greene et al 2008 2009)In general with the exception of the picritic pillow lavasin the submarine stratigraphy and local variations betweenhigh-MgO and tholeiitic lavas there are no significantcompositional differences between lavas erupted duringthe different stages of growth of the plateau (ie submarineand subaerial flows) nor are substantial compositional dif-ferences generally found between massive submarine flowsin contact with pillowed flows tholeiitic basalt flows nearthe base of the Karmutsen Formation are compositionallysimilar to flows sampled near the top of the volcanicsequence Where it was possible to sample continuous sec-tions of volcanic stratigraphy greater than several hun-dreds of meters (eg Keogh^Maynard Lake area west ofthe Karmutsen Range Fig 2) there is no clear relationshipbetween stratigraphic position and compositionalvariation
CONCLUSIONSThe Karmutsen Formation covers extensive areas ofVancouver Island (20 000 km2) and forms part of a largeoceanic plateau erupted over a geologically short interval(c 225^231 Ma) The tripartite volcanic stratigraphyexposed on Vancouver Island is at least 6 km thick andcomprises a succession of submarine flows volcaniclastic
deposits and massive subaerial flowsThese volcanologicaldifferences are primarily related to the eruption environ-ment (deep-water shallow-water or subaerial) The rapidgrowth of the plateau prevented significant accumulationsof intervening sediment except in the uppermost stratigra-phy where isolated limestone lenses commonly associatedwith pillow basalt and hyaloclastite deposits preserve arecord of the final stages of subsidence of the plateauThe Wrangellia plateau on Vancouver Island was con-
structed dominantly of tholeiitic basalt with restrictedmajor- and trace-element and isotopic compositional var-iations Minor volumes of picritic pillow basalts eruptedlate in the submarine phase The Karmutsen basalts wereemplaced onto older Paleozoic arc volcanic sequences andlow NbLa for the picrites may indicate a contributionfrom the sub-arc lithospheric mantle Modeling of the pet-rogenesis of the Karmutsen picrites indicates melting ofanomalously hot mantle (15008C) and extensive degreesof partial melting (23^27) and is consistent with aplume initiation model Combined Sr^Nd^Hf^Pb isotopicsystematics indicate that the picritic and tholeiitic lavaswere derived from a depleted plume-type mantle sourcedistinct from the source of MORB that has compositionalsimilarities to the source of the lavas forming theCaribbean Plateau Picrites were generated by melting ofspinel lherzolite whereas the tholeiitic basalts formed asaggregate melts of both garnet and spinel lherzolite Thetholeiitic basalts underwent extensive low-pressure fractio-nation (52^3 kbar) and seismic studies indicate that someof theWrangellia crust beneathVancouver Island may cor-respond to the plutonic residues of the Karmutsen floodbasalts
ACKNOWLEDGEMENTSWe are grateful to Nick Arndt for helping us get this pro-ject started and Nick Massey for insights into VancouverIsland geology We greatly appreciate the assistance withPRIMELT modeling from Claude Herzberg We thankMikkel Schau for his insight and enthusiasm during field-work Jane Barling assisted with analyses by MC-ICP-MSThoughtful and thorough reviews of the manuscript byJohn Mahoney Fred Frey Wendy Bohrson DougalJerram and Teal Riley are greatly appreciated Fundingwas generously provided by the Rocks to Riches Programadministered by the BC amp Yukon Chamber of Mines in2004 by the BC Geological Survey and by NSERCDiscovery Grants to JSS and DW ARG was supportedby a University Graduate Fellowship at UBC
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
34
REFERENCESBarker F Brown A S Budahn J R amp Plafker G (1989) Back-arc
with frontal-arc component origin of Triassic Karmutsen basaltBritish Columbia Canada Chemical Geology 75 81^102
Blichert-Toft JWeis D Maerschalk C Agranier A amp Albare de F(2003) Hawaiian hot spot dynamics as inferred from the Hf and Pbisotope evolution of Mauna Kea volcano Geochemistry GeophysicsGeosystems 4(2) 1^27 doi1010292002GC000340
Bondre N R Duraiswami R A amp Dole G (2004) Morphologyand emplacement of flows from the Deccan Volcanic ProvinceIndia Bulletin of Volcanology 66 29^45
Carlisle D (1963) Pillow breccias and their aquagene tuffs QuadraIsland British Columbia Journal of Geology 71 48^71
Carlisle D (1972) Late Paleozoic to Mid-Triassic sedimentary-volcanic sequence on Northeastern Vancouver Island Geological
Society of Canada Paper Report of Activities 72-1 Part B 24^30Carlisle D amp SuzukiT (1974) Emergent basalt and submergent car-
bonate^clastic sequences including the Upper Triassic Dilleri andWelleri zones on Vancouver Island Canadian Journal of Earth
Sciences 11 254^279Chauvel C amp Blichert-Toft J (2001) A hafnium isotope and trace
element perspective on melting of the depleted mantle Earth and
Planetary Science Letters 190 137^151Chayes F (1954)The theory of thin-section analysis Journal of Geology
62 92^101Cho M Liou J G amp Maruyama S (1987) Transition from the zeo-
lite to prehnite^pumpellyite facies in the Karmutsen metabasitesVancouver Island British Columbia Journal of Petrology 27(2)467^494
Clowes R M Zelt C A Amor J R amp Ellis R M (1995)Lithospheric structure in the southern Canadian Cordillera froma network of seismic refraction lines Canadian Journal of Earth
Sciences 32(10) 1485^1513Coffin M F amp Eldholm O (1994) Large igneous provinces Crustal
structure dimensions and external consequences Reviews of
Geophysics 32(1) 1^36Cox K G (1980) A model for flood basalt volcanism Journal of
Petrology 21 629^650Eldholm O amp Coffin M F (2000) Large igneous provinces
and plate tectonics In Richards M A Gordon R G amp vander Hilst R D (eds) The History and Dynamics of Global
Plate Motions American Geophysical Union Geophysical Monograph 121309^326
Elliott T R Hawkesworth C J amp Grolaquo nvold K (1991) Dynamicmelting of the Iceland plume Nature 351 201^206
Escuder-Viruete J Pecurren rez-Estaucurren n A Contreras F Joubert MWeis D Ullrich T D amp Spadea P (2007) Plume mantle sourceheterogeneity through time Insights from the Duarte ComplexHispaniola northeastern Caribbean Journal of Geophysical Research112(B04203) doi1010292006JB004323
Farnetani C G Richards M A amp Ghiorso M S (1996)Petrological models of magma evolution and deep crustal structurebeneath hotspots and flood basalt provinces Earth and Planetary
Science Letters 143 81^94Fitton J G amp Godard M (2004) Origin and evolution of magmas
on the Ontong Java Plateau In Fitton J G Mahoney J JWallace P J amp Saunders A D (eds) Origin and Evolution of the
Ontong Java Plateau Geological Society London Special Publications 229151^178
Fitton J G Saunders A D Norry M J Hardarson B S ampTaylor R N (1997) Thermal and chemical structure of theIceland plume Earth and Planetary Science Letters 153 197^208
Fitton J G Saunders A D Kempton P D amp Hardarson B S(2003) Does depleted mantle form an intrinsic part of the Icelandplume Geochemistry Geophysics Geosystems 4(3) 1032 doi1010292002GC000424
Frey F A Huang S Blichert-Toft J Regelous M amp Boyet M(2005) Origin of depleted components in basalt related to theHawaiian hotspot Evidence from isotopic and incompatible ele-ment ratios Geochemistry Geophysics Geosystems 6(1) 23 doi1010292004GC000757
Galer S J G amp Abouchami W (1998) Practical application of leadtriple spiking for correction of instrumental mass discriminationMineralogical Magazine 62A 491^492
Ghiorso M S amp Sack R O (1995) Chemical mass transfer in mag-matic processes IV A revised and internally consistent thermody-namic model for the interpolation and extrapolation of liquid^solidequilibria in magmatic systems at elevated temperatures and pres-sures Contributions to Mineralogy and Petrology 119 197^212
Greene A R Scoates J S amp Weis D (2005)WrangelliaTerrane onVancouver Island Distribution of flood basalts with implicationsfor potential Ni^Cu^PGE mineralization In Grant B (ed)Geological Fieldwork 2004 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2005-1 209^220Greene A R Scoates J S Nixon G T amp Weis D (2006) Picritic
lavas and basal sills in the Karmutsen flood basalt provinceWrangellia northern Vancouver Island In Grant B (ed)Geological Fieldwork 2005 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2006-1 39^52Greene A R Scoates J S amp Weis D (2008) Wrangellia flood
basalts in Alaska A record of plume^lithosphere interaction in aLate Triassic accreted oceanic plateau Geochemistry Geophysics
Geosystems 9(Q12004) doi1010292008GC002092Greene A R Scoates J S Weis D amp Israel S (2009)
Geochemistry of triassic flood basalts from the Yukon (Canada)segment of the accreted Wrangellia oceanic plateau Lithosdoi101016jlithos200811010
Hauff F Hoernle K Tilton G Graham D W amp Kerr A C(2000) Large volume recycling of oceanic lithosphere over shorttime scales geochemical constraints from the Caribbean LargeIgneous Province Earth and Planetary Science Letters 174 247^263
Hauff F Hoernle K amp Schmidt A (2003) Sr^Nd^Pb compositionof Mesozoic Pacific oceanic crust (Site 1149 and 801 ODP Leg185)Implications for alteration of ocean crust and the input into theIzu^Bonin^Mariana subduction system Geochemistry Geophysics
Geosystems 4(8) doi1010292002GC000421Herzberg C (2004) Partial melting below the Ontong Java Plateau
In Fitton J G Mahoney J J Wallace P J amp Saunders A D(eds) Origin and Evolution of the Ontong Java Plateau Geological SocietyLondon Special Publications 229 179^183
Herzberg C (2006) Petrology and thermal structure of the Hawaiianplume from Mauna Kea volcano Nature 444(7119) 605
Herzberg C amp Asimow P D (2008) Petrology of some oceanicisland basalts PRIMELT2XLS software for primary magma cal-culation Geochemistry Geophysics Geosystems 9(Q09001) doi1010292008GC002057
Herzberg C amp OrsquoHara M J (2002) Plume-associated ultramaficmagmas of Phanerozoic age Journal of Petrology 43(10) 1857^1883
Herzberg C Asimow P D Arndt N Niu Y Lesher C MFitton J G Cheadle M J amp Saunders A D (2007)Temperatures in ambient mantle and plumes Constraints frombasalts picrites and komatiites Geochemistry Geophysics Geosystems8(Q02006) doi1010292006GC001390
Huang S amp Frey F A (2005) Trace element abundances of MaunaKea basalt from phase 2 of the Hawaii Scientific Drilling Project
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
35
Petrogenetic implications of correlations with major element con-tent and isotopic ratios Geochemistry Geophysics Geosystems 4(6)doi1010292002GC000322
Hussong D M Wipperman L K amp Kroenke L W (1979) Thecrustal structure of the Ontong Java and Manihiki OceanicPlateaus Journal of Geophysical Research 84(B11) 6003^6010
Jerram D A amp Widdowson M (2005) The anatomy of ContinentalFlood Basalt Provinces geological constraints on the processes andproducts of flood volcanism Lithos 79(3^4) 385^405
Jones D L Silberling N J amp Hillhouse J (1977)Wrangellia a dis-placed terrane in northwestern North America CanadianJournal ofEarth Sciences 14(11) 2565^2577
Kerr A C (2003) Oceanic Plateaus In Rudnick R (ed) The Crust
Treatise on GeochemistryVol 3 Oxford Elsevier Science pp 537^565Kerr A C amp Mahoney J J (2007) Oceanic plateaus Problematic
plumes potential paradigms Chemical Geology 241 332^353Kerr A CTarney J Marriner G F Klaver GT Saunders A D
amp Thirlwall M F (1996) The geochemistry and petrogenesis ofthe Late-Cretaceous picrites and basalts of Curacao NetherlandsAntilles a remnant of an oceanic plateau Contributions to
Mineralogy and Petrology 124(1) 29^43Kerr A C Marriner G F Tarney J Nivia A Saunders A D
Thirlwall M F amp Sinton A C (1997) Cretaceous basaltic ter-ranes in Western Colombia Elemental chronological and Sr-Ndisotopic constraints on petrogenesis Journal of Petrology 38(6)677^702
Kinzler R J (1997) Melting of mantle peridotite at pressuresapproaching the spinel to garnet transition Application to mid-ocean ridge basalt petrogenesis Journal of Geophysical Research
102(B1) 853^874Kinzler R J amp Grove T L (1992) Primary magmas of mid-ocean
ridge basalts 1 Experiments and methods Journal of Geophysical
Research 97(B5) 6885^6906Korenaga J (2005) Why did not the Ontong Java Plateau form sub-
aerially Earth and Planetary Science Letters 234 385^399Kuniyoshi S (1972) Petrology of the Karmutsen (Volcanic) Group
northeastern Vancouver Island British Columbia PhD disserta-tion University of California Los Angeles 242 pp
Lassiter J C DePaolo D J amp Mahoney J J (1995) Geochemistry ofthe Wrangellia flood basalt province Implications for the role ofcontinental and oceanic lithosphere in flood basalt genesis Journalof Petrology 36(4) 983^1009
LeBas M J (2000) IUGS reclassification of the high-Mg and picriticvolcanic rocks Journal of Petrology 41(10) 1467^1470
Longhi J (2002) Some phase equilibrium systematics of lherzolitemelting I Geochemistry Geophysics Geosystems 3(3) doi1010292001GC000204
MacDonald G A amp Katsura T (1964) Chemical composition ofHawaiian lavas Journal of Petrology 5 82^133
Mahoney J J (1987) An isotopic survey of Pacific oceanic plateausimplications for their nature and origin In Keating B HFryer P Batiza R amp Boehlert GW (eds) Seamounts Islands andAtolls American Geophysical Union Geophysical Monograph 43 207^220
Mahoney J LeRoex A P Peng Z Fisher R L amp Natland J H(1992) Southwestern limits of Indian Ocean Ridge mantle and theorigin of low 206Pb204Pb mid-ocean ridge basalt isotope systema-tics of the central Southwest Indian Ridge (178^508E) Journal ofGeophysical Research 97(B13) 19771^19790
Mahoney J J Sinton J M Kurz M D MacDougall J DSpencer K J amp Lugmair GW (1994) Isotope and trace elementcharacteristics of a super-fast spreading ridge East Pacific rise13^238S Earth and Planetary Science Letters 121 173^193
Massey N W D (1995a) Geology and mineral resources of theAlberni^Nanaimo Lakes sheet Vancouver Island 92F1W 92F2Eand part of 92F7E British Columbia Ministry of Energy Mines and
Petroleum Resources Paper 1992-2 132 ppMassey N W D (1995b) Geology and mineral resources of the
Cowichan Lake sheet Vancouver Island 92C16 British Columbia
Ministry of Energy Mines and Petroleum Resources Paper 1992-3 112 ppMassey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005a) Digital Geology Map of British Columbia Tile NM9 MidCoast BC British Columbia Ministry of Energy and Mines Geofile
2005-2Massey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005b) Digital Geology Map of British Columbia Tile NM10Southwest BC British Columbia Ministry of Energy and Mines Geofile
2005-3McDonough W F amp Sun S (1995) The composition of the Earth
Chemical Geology 120 223^253Monger JW H amp Journeay J M (1994) Basement geology and tec-
tonic evolution of theVancouver region In Monger JW H (ed)Geology and Geological Hazards of the Vancouver Region Southwestern
British Columbia Geological Survey of Canada Bulletin 481 3^25Muller J E (1977) Geology of Vancouver Island Geological Survey of
Canada Open File Map 463Muller J E (1980)The Paleozoic Sicker Group of Vancouver Island British
Columbia Geological Survey of Canada Paper 79-30 22 ppMuller J E Northcote K E amp Carlisle D (1974) Geology and
mineral deposits of Alert Bay^Cape Scott map area VancouverIsland British Columbia Geological Survey of Canada Paper 74-8 77
Neal C R Mahoney J J Kroenke L W Duncan R A ampPetterson M G (1997) The Ontong^Java Plateau InMahoney J J amp Coffin M F (eds) Large Igneous Provinces
Continental Oceanic and Planetary FloodVolcanism American Geophysical
Union Geophysical Monograph 100 183^216Niu Y Collerson K D Batiza R Wendt J I amp Regelous M
(1999) Origin of enriched-typed mid-ocean ridge basalt at ridgesfar from mantle plumes The East Pacific Rise at 11820rsquoN Journalof Geophysical Research 104 7067^7088
Nixon G T amp Orr A J (2007) Recent revisions to the EarlyMesozoic stratigraphy of Northern Vancouver Island (NTS 102I092L) and metallogenic implicatinos British Columbia InGrant B (ed) Geological Fieldwork 2006 British Columbia Ministry of
Energy Mines and Petroleum Resources Paper 2007-1 163^177Nixon GT Hammack J L KoyanagiV M Payie G J Haggart
JW Orchard M JTozerT Friedman R M Archibald D APalfy J amp Cordey F (2006a) Geology of the Quatsino^PortMcNeill area northernVancouver Island British Columbia Ministry
of Energy Mines and Petroleum Resources Geoscience Map 2006-2 scale150 000
Nixon G T Kelman M C Stevenson D Stokes L A ampJohnston K A (2006b) Preliminary geology of the Nimpkishmap area (NTS 092L07) northern Vancouver Island BritishColumbia In Grant B amp Newell J M (eds) Geological Fieldwork2005 British Columbia Ministry of Energy Mines and Petroleum Resources
Paper 2006-1 135^152Nixon G T Laroque J Pals A Styan J Greene A R amp
Scoates J S (2008) High-Mg lavas in the Karmutsen flood basaltsnorthernVancouver Island (NTS 092L) Stratigraphic setting andmetallogenic significance In Grant B (ed) Geological Fieldwork
2007 British Columbia Ministry of Energy Mines and Petroleum
Resources Paper 2008-1 175^190Nowell G M Kempton P D Noble S R Fitton J G
Saunders A Mahoney J J amp Taylor R N (1998) High precisionHf isotope measurements of MORB and OIB by thermal ionisation
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
36
mass spectrometry insights into the depleted mantle Chemical
Geology 149 211^233Parrish R R amp McNicoll V J (1992) U^Pb age determinations
from the southern Vancouver Island area British Columbia InRadiogenic Age and Isotopic Studies Report 5 Geological Survey of
Canada Paper 91-2 79^86Peate DW (1997) The Parana^Etendeka Province In Coffin M F
amp Mahoney J (eds) Large Igneous Provinces Continental Oceanic andPlanetary Flood Volcanism American Geophysical Union Geophysical
Monograph 100 217^245Perfit M R Fornari D J Ridley W I Kirk P D Casey J
Kastens K A Reynolds J R Edwards M Desonie DShuster R amp Paradis S (1996) Recent volcanism in the Siqueirostransform fault picritic basalts and implications for MORBmagma genesis Earth and Planetary Science Letters 141(1^4) 91^108
Pretorius W Weis D Williams G Hanano D Kieffer B ampScoates J S (2006) Complete trace elemental characterization ofgranitoid (USGSG-2GSP-2) reference materials by high resolutioninductively coupled plasma-mass spectrometry Geostandards and
Geoanalytical Research 30(1) 39^54Regelous M Niu Y Wendt J I Batiza R Greig A amp
Collerson K D (1999) Variations in the geochemistry of magma-tism on the East Pacific Rise at 10830rsquoN since 800 ka Earth and
Planetary Science Letters 168 45^63Recurren villon S Chauvel C Arndt N T Pik R Martineau F
Fourcade S amp Marty B (2002) Heterogeneity of the Caribbeanplateau mantle source Sr O and He isotopic compositions of oli-vine and clinopyroxene from Gorgona Island Earth and Planetary
Science Letters 205(1^2) 91Rhodes J M (1996) Geochemical stratigraphy of lava flows samples
by the Hawaii Scientific Drilling Project Journal of Geophysical
Research 101(B5) 11729^11746Rhodes J M amp Vollinger M J (2004) Composition of basaltic lavas
sampled by phase-2 of the Hawaii Scientific Drilling ProjectGeochemical stratigraphy and magma types Geochemistry
Geophysics Geosystems 5(3) doi1010292002GC000434Richards M A Jones D L Duncan R A amp DePaolo D J (1991)
A mantle plume initiation model for the Wrangellia flood basaltand other oceanic plateaus Science 254 263^267
Salters V J M amp Hart S R (1991) The mantle sources of oceanridges islands and arcs the Hf-isotope connection Earth and
Planetary Science Letters 104 364^380Salters V J M amp Stracke A (2004) Composition of the depleted
SaltersV J amp WhiteW (1998) Hf isotope constraints on mantle evo-lution Chemical Geology 145 447^460
SanoT amp Yamashita S (2004) Experimental petrology of basementlavas from Ocean Drilling Program Leg 192 implications for dif-ferentiation processes in Ontong Java Plateau magmas InFitton J G Mahoney J JWallace P J amp Saunders A D (eds)Origin and Evolution of the Ontong Java Plateau Geological Society
London Special Publications 229 185^218Saunders A D (2005) Large igneous provinces Origin and environ-
mental consequences Elements 1 259^263Self S ThordarsonT amp Keszthely L (1997) Emplacement of conti-
nental flood basalt lava flows In Mahoney J J amp Coffin M F(eds) Large Igneous Provinces Continental Oceanic and Planetary Flood
Volcanism American Geophysical Union Geophysical Monograph 100381^410
Shaw D M (2000) Continuous (dynamic) melting theory revisitedCanadian Mineralogist 38(5) 1041^1063
Sluggett C L (2003) Uranium^lead age and geochemical constraintson Paleozoic and Early Mesozoic magmatism in WrangelliaTerrane Saltspring Island British Columbia BSc thesisUniversity of British ColumbiaVancouver 84 pp
Storey M Mahoney J J amp Saunders A D (1997) Cretaceousbasalts in Madagascar and the transition between plume and con-tinental lithospheric mantle sources In Mahoney J J ampCoffin M F (eds) Large Igneous Provinces Continental Oceanic and
Planetary Flood Volcanism Aamerican Geophysical Union Geophysical
Monograph 100 95^122Surdam R C (1967) Low-grade metamorphism of the Karmutsen
Group PhD dissertation University of California Los Angeles288 pp
Sutherland-Brown A Yorath C J Anderson R G amp Dom K(1986) Geological maps of southern Vancouver IslandLITHOPROBE I 92C10 11 14 16 92F1 2 7 8 Geological Survey ofCanada Open File 1272
Tejada M L G Mahoney J J Castillo P R Ingle S P Sheth HC amp Weis D (2004) Pin-pricking the elephant evidence on theorigin of the Ontong Java Plateau from Pb^Sr^Hf^Nd isotopiccharacteristics of ODP Leg 192 basalts In Fitton J GMahoney J J Wallace P J amp Saunders A D (eds) Origin and
Evolution of the Ontong Java Plateau Geological Society London Special
Publications 229 133^150Thompson P M E Kempton P D amp Kerr A C (2008) Evaluation
of the effects of alteration and leaching on Sm^Nd and Lu^Hf sys-tematics in submarine mafic rocks Lithos 104(1^4) 164^176
Thompson P M E Kempton P D White R V Kerr A CTarney J Saunders A D Fitton J G amp McBirney A (2003)Hf-Nd isotope constraints on the origin of the CretaceousCaribbean plateau and its relationship to the Galapagos plumeEarth and Planetary Science Letters 217(1^2) 59^75
Vervoort J D Patchett P Blichert-Toft J amp Albare de F (1999)Relationships between Lu^Hf and Sm^Nd isotopic systems in theglobal sedimentary system Earth and Planetary Science Letters 16879^99
Walter M J (1998) Melting of garnet peridotite and the origin ofkomatiite and depleted lithosphere Journal of Petrology 39(1) 29^60
Weis D Kieffer B Maerschalk C Barling J de Jong JWilliams G A Hanano D Mattielli N Scoates J SGoolaerts A Friedman R A amp Mahoney J B (2006) High-pre-cision isotopic characterization of USGS reference materials byTIMS and MC-ICP-MS Geochemistry Geophysics Geosystems
7(Q08006) doi1010292006GC001283Weis D Kieffer B Hanano D Silva I N Barling J
Pretorius W Maerschalk C amp Mattielli N (2007) Hf isotopecompositions of US Geological Survey reference materialsGeochemistry Geophysics Geosystems 8(Q06006) doi1010292006GC001473
Wheeler J O amp McFeely P (1991) Tectonic assemblage map of theCanadian Cordillera and adjacent part of the United States ofAmerica Geological Survey of Canada Map 1712A
WhiteW M Albare de F amp Tecurren louk P (2000) High-precision analy-sis of Pb isotope ratios by multi-collector ICP-MS Chemical Geology167 257^270
Yorath C J Sutherland Brown A amp Massey N W D (1999)LITHOPROBE southern Vancouver Island British Columbia Geological
Survey of Canada Bulletin 498 145 ppZou H (1998) Trace element fractionation during modal and
non-model dynamic melting and open-system melting Amathematical treatment Geochimica et Cosmochimica Acta 62(11)1937^1945
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
37
Zou H amp Reid M R (2001) Quantitative modeling of trace elementfractionation during incongruent dynamic melting Geochimica et
Cosmochimica Acta 65(1) 153^162
APPENDIX SAMPLEPREPARAT ION ANDANALYT ICAL METHODSOnly the freshest rocks in the field were sampled and onlythe least altered samples were selected for chemical andisotopic analysis based on thorough petrographic inspec-tion Fifty-nine of the 129 collected samples were crushed(400 g) into small pieces 52mm in diameter in aRocklabs hydraulic piston crusher between WC plates tominimize contamination The coarse-crush was mixedand 100 g was powdered in a planetary mill using agatejars and balls cleaned with quartz sand between samples
ActLabs analytical methodsThe major- and trace-element compositions of the whole-rock powders were determined at Activation LaboratoriesLtd (Actlabs) in Ancaster Ontario Analytical techniquesand detection limits are also available from Actlabs (httpwwwactlabscommethsub_code4erehtm) The analyticalmethod for each of the elements analyzed is indicated inTable 2 For the major elements a 02 g sample was mixedwith a lithium metaborate^lithium tetraborate mixtureand fused in a graphite crucible The molten mixture waspoured into a 5 HNO3 solution and shaken until dis-solved (30min) The samples were analyzed for majoroxides and selected trace elements by inductively coupledplasma optical emission spectrometry (ICP-OES) on acombination simultaneous^sequential Thermo Jarrell-AshEnviro II system Internal calibration was achieved usinga variety of international reference materials (eg W-2BIR-1 DNC-1) and independent control samplesAdditional trace elements were analyzed by both instru-mental neutron activation analysis (INAA) and induc-tively couple plasma mass spectrometry (ICP-MS) ForINAA15^25 g of sample was weighed into small polyeth-ylene vials and irradiated with a control international ref-erence material CANMET WMS-1 and NiCr flux wires ata thermal neutron flux of 71012 neutronscm2 per s in theMcMaster Nuclear Reactor Following a 7 day waitingperiod the samples were measured on an Ortec high-purity Ge detector linked to a Canberra Series 95 multi-channel analyzer Activities for each element weredecay- and weight-corrected and compared with a detectorcalibration developed from multiple international certifiedreference materials For ICP-MS 025 g of sample wasdigested in HF followed by HNO3^HClO4 heated andtaken to dryness The samples were brought back into solu-tion with HCl Samples were analyzed using a PerkinElmer Optima 3000 ICP In-lab standards or certified ref-erence materials (eg W-2 BIR-1 DNC-1) were used for
quality control A total of 15 blind duplicates were ana-lyzed from the same sample powders to assessreproducibility
University of Massachusetts analyticalmethodsFourteen sample duplicate powders taken from the same100 g sample aliquot (high-MgO basalts and picrites)were also analyzed at the Ronald B Gilmore X-RayFluorescence (XRF) Laboratory at the University ofMassachusetts Major elements were measured on a fusedLa-bearing lithium borate glass disc using a SiemensMRS-400 spectrometer Trace element concentrations(Rb Sr Ba Ce Nb Zr Y Pb Zn Ga Ni Cr V) weremeasured on a separate powder pellet using a PhilipsPW2400 sequential spectrometer using a Rh tube LOIand ferrous iron measurements were made as describedby Rhodes amp Vollinger (2004) Precision and accuracy esti-mates for the data have been described by Rhodes (1996)and Rhodes amp Vollinger (2004) Results for each sampleare the average of two separate analyses and are shown inElectronic Appendix 2 for a comparison of analyses ofduplicate powders using XRFand ICP-MS
PCIGR analytical methodsA subset of 19 samples was selected for high-precisiontrace-element analysis and Sr Nd Pb and Hf isotopicanalysis at the Pacific Centre for Isotopic andGeochemical Research (PCIGR) at the University ofBritish Columbia (UBC) Samples were selected from the59 samples analyzed for whole-rock chemistry at ActLabsbased on major- and trace-element chemistry alteration(low LOI and petrographic alteration index) and samplelocation Samples were prepared for trace-element analysisat the PCIGR by the technique described by Pretorius et al(2006) on unleached rock powders Sample powders(100mg) were weighed in 7ml screw-top Savillex bea-kers and dissolved in 1ml 14N HNO3 and 5ml 48 HFon a hotplate for 48 h at 1308C with periodic ultrasonica-tion Samples were dried and redissolved in 6ml 6N HClon a hotplate for 24 h and then dried and redissolved in1ml concentrated HNO3 for 24 h before final dryingTrace element abundances were measured by high-resolution inductively coupled plasma-mass spectrometry(HR-ICP-MS) with aThermo Finnigan Element2 systemfollowing the procedures described by Pretorius et al(2006) within 24 h of redissolution HFSE and LILE weremeasured in medium-resolution mode at 2000 dilutionusing a PFATeflon spray chamber washed with aqua regiafor 3min between samples REE were measured in high-resolution mode and U and Pb in low-resolution modeat 2000 dilution using a glass spray chamber washedwith 2 HNO3 between samples Total procedural blanksand reference materials (BCR-2 BHVO-2) were analyzedwith the batch of samples Indium was used as an internal
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
38
standard in all samples and standard solutionsBackground and standard solutions were analyzed afterevery five samples to detect memory effects and massdrift Results for PCIGR trace-element analyses areshown in Electronic Appendix 3 for the full set of analyzedtrace elementsSample digestion for purification of Sr Nd Hf and Pb
for column chemistry began by weighing each samplepowder (400^500mg) prior to leaching All samples wereleached with 6N HCl and placed in an ultrasonic bath for15min Samples were rinsed twice with 18 M-cm H2Obetween each leaching step (15 total) until the supernatantliquid was clear [following the method of Mahoney (1987)]Samples were then dried on a hotplate for 24 h andweighed again Sample solutions were then prepared bydissolving 100^250mg of the leached powder dissolvedin 1ml 14N HNO3 and 10ml 48 HF on a hotplate for48 h at 1308C with periodic ultrasonication Samples weredried and redissolved in 6ml 6N HCl on a hotplate for24 h and then dried Pb was separated using anion exhangecolumns and the discard was used for Sr REE and Hfseparation Nd was separated from the other REE and Hfrequired two additional purification steps Detailed proce-dures for column chemistry for separating Sr Nd and Pbat the PCIGR have been described byWeis et al (2006) andHf purification has been described byWeis et al (2007) Srand Nd isotope ratios were measured by thermal ioniza-tion mass spectrometry (TIMS) on a Thermo FinniganTriton system in static mode with relay matrix rotation ona single Ta and double Re^Ta filament respectively Fourto five filaments per barrel of 21 were occupied by stan-dards (NBS 987 for Sr and LaJolla for Nd) for each barrelwhere samples were run Sample Sr and Nd isotopic com-positions were corrected for mass fractionation using86Sr88Srfrac14 01194 and 146Nd144Ndfrac14 07219 Each samplewas then normalized using the barrel average of the refer-ence material relative to the values of143Nd144Ndfrac14 0511858 and 87Sr86Srfrac14 0710248 (Weiset al 2006) During the course of the Vancouver Islandanalyses the La Jolla Nd standard gave an average valueof 05118566 (nfrac14 7) and NBS987 standard gave an aver-age of 07102408 (nfrac1411) (2 error is reported as times106) 147Sm144Nd errors are 15 or 0006 USGeological Survey (USGS) reference material BHVO-2was processed with the samples and yielded Sr and Nd iso-topic ratios of 07034607 and 05129786 respectivelyThese are in agreement with the published values of070347920 and 051298411 respectively (Weis et al2006)
Pb and Hf isotopic compositions were analyzed by mul-tiple collector-inductively coupled plasma-mass spectrome-try (MC-ICP-MS) on a static multi-collection on a NuPlasma (Nu Instruments) system The detailed analyticalprocedure for Pb isotopic analyses on the Nu Plasma atthe PCIGR has been described by Weis et al (2006) Theconfiguration for Pb analyses allows for collection of PbTl and Hg togetherTl and Hg are used to monitor instru-mental mass discrimination and isobaric overlap repec-tively All sample solutions were analyzed withapproximately the same PbTl ratio (4) as the referencematerial NIST SRM 981 To accomplish this a small ali-quot of each sample solution from the Pb columns was ana-lyzed on the Element2 to determine the precise amount ofPb available for analysis on the Nu Plasma The SRM 981standard was run after every two samples on the NuPlasma During the time samples were run analyses ofthe SRM 981 Pb reference material gave values of206Pb204Pbfrac1416940322 207Pb204Pbfrac1415495823and 208Pb204Pbfrac14 36713164 (nfrac14 61 2 error is reportedas times 104) these values are in excellent agreement withreported TIMS triple-spike values of Galer amp Abouchami(1998) Fractionation-corrected Pb isotopic ratios were fur-ther corrected by the sample-standard bracketing methodor the ln^ln correction method described by White et al(2000) and Blichert-Toft et al (2003) Leached powders ofUSGS reference material BHVO-2 yielded Pb isotopicratios of 206Pb204Pbfrac14186454 8 207Pb204Pbfrac14 1549105 and 208Pb204Pbfrac14 38222514 and BCR-2yielded 206Pb204Pbfrac141880466 207Pb204Pbfrac14 1562518 and 208Pb204Pbfrac14 3883496 (2 error isreported as times 104) These values are in agreement withthose for leached residues of BHVO-2 and BCR-2 fromWeis et al (2006)Hf isotopic compositions were analyzed following the
procedures detailed byWeis et al (2007) The configurationfor Hf analyses monitored Lu mass 175 andYb mass 172 toallow for interference correction to masses 174 and 176 Hfisotopic ratios were normalized internally for mass fractio-nation to a 179Hf177Hf ratio of 07325 using an exponentialcorrection Standards were run after every two samplesand sample results were normalized to the ratio of the in-run daily average and a 176Hf177Hf ratio for JMC-475 of0282160 During the course of analyses the Hf standardJMC-475 gave an average value 02821679 (nfrac14130)USGS reference materials BCR-2 and BHVO-2 were pro-cessed with the samples and yielded Hf isotopic ratios of02828675 and 02831005 respectively Publishedvalues for BCR-2 and BHVO-2 are 02828717 and0283104 8 respectively (Weis et al 2007)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
39
with LOI (not shown) The relatively high apparent initialSr isotopic compositions for the high-MgO lavas (up to07052) probably resulted from post-magmatic loss of Rbor an increase in 87Sr86Sr through addition of seawater Sr(eg Hauff et al 2003) High-MgO lavas from theCaribbean plateau have similarly high 87Sr86Sr comparedwith basalts (Recurren villon et al 2002)The correction for in situdecay on initial Pb isotopic ratios has been affected bymobilization of U and Th (and Pb) in the whole-rockssince their formation A thorough acid leaching duringsample preparation was used that has been shown to effec-tively remove alteration phases (Weis et al 2006 NobreSilva et al in preparation) Whole-rock SmNd ratiosand age-correction of Nd isotopes may be affected bythe leaching procedure for submarine rocks older than50 Ma (Thompson et al 2008 Fig 9) there ishowever very little variation in Nd isotopic compositionsof the picrites or tholeiitic basalts and this system does notappear to be disturbed by alteration The HFSE abun-dances for tholeiitic and picritic basalts exhibit clear
linear correlations in binary diagrams (Fig 8) whereasplots of LILE vs HFSE are highly scattered (not shown)because of the mobility of some of the alkali and alkalineearth LILE (Cs Rb Ba K) and Sr for most samplesduring alterationThe degree of alteration in the Karmutsen samples does
not appear to be related to eruption environment or depthof burial Pillow rims are compositionally different frompillow cores (Surdam1967 Kuniyoshi1972) and aquagenetuffs are chemically different from isolated pillows withinpillow breccias however submarine and subaerial basaltsexhibit a similar degree of alterationThere is no clear cor-relation between the submarine and subaerial basalts andsome commonly used chemical alteration indices (eg BaRb vs K2OP2O5 Huang amp Frey 2005) There is also nodefinitive correlation between the petrographic alterationindex (Table 1) and chemical alteration indices although10 of 13 tholeiitic basalts with the highest K and LILEabundances have the highest petrographic alterationindex of three (seeTable 1)
Table 4 Hf isotopic compositions of Karmutsen basaltsVancouver Island BC
Sample Group Area Lu Hf 176Hf177Hf 2m eHf176Lu177Hf 176Hf177Hft eHf(t)
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
OLIV INE ACCUMULAT ION INH IGH-MgO AND PICR IT IC LAVASGeochemical trends and petrographic characteristics indi-cate that accumulation of olivine played an important rolein the formation of the high-MgO basalts and picrites fromthe Keogh Lake area on northern Vancouver Island TheKeogh Lake samples show a strong linear correlation inplots of Al2O3 TiO2 ScYb and Ni vs MgO and many ofthe picrites have abundant clusters of olivine pseudomorphs(Table1 Fig11)There is a strong linear correlationbetween
modal per cent olivine and whole-rock magnesium contentmagnesium contents range between 10 and 19wt MgOand the proportion of olivine phenocrysts in these samplesvaries from 0 to 42 vol (Fig 11)With a few exceptionsboth the size range and median size of the olivine pheno-crysts in most samples are comparable The clear correla-tion between the proportion of olivine phenocrysts andwhole-rock magnesium contents indicates that accumula-tion of olivine was directly responsible for the high MgOcontents (410wt ) in most of the picritic lavas
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
207Pb204Pb
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
208Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
(a) (b)
(c) (e)
4723A24723A2
4723A2
4723A2
4722A4
4723A4
4722A4
4723A4
4723A134723A3
4723A3
4723A13
1556
1558
1560
1562
1564
1566
1568
186 190 194 198 202 206 2101550
1552
1554
1556
1558
1560
178 180 182 184 186 188 190 192 194
380
384
388
392
396
186 190 194 198 202 206 210374
378
382
386
390
178 180 182 184 186 188 190 192 194
206Pb204Pb(d)
LOI (wt )
0
1
2
3
4
5
6
7
186 190 194 198 202 206 210
4723A2
Fig 10 Pb isotopic compositions of leached whole-rock samples measured by MC-ICP-MS for the Karmutsen Formation Error bars are smal-ler than symbols (a) Measured 207Pb204Pb vs 206Pb204Pb (b) Initial 207Pb204Pb vs 206Pb204Pb Age-correction to 230 Ma (c) Measured208Pb204Pb vs 206Pb204Pb (d) LOI vs 206Pb204Pb (e) Initial 208Pb204Pb vs 206Pb204Pb Complete chemical duplicates shown in Table 5(samples 4720A7 and 4722A4) are circled in each plot The dashed lines in (b) and (e) show the differences in age-corrections for two picrites(4723A4 4722A4) using the measured UTh and Pb concentrations for each sample and the age-corrections when concentrations are used fromthe two picrites (4723A3 4723A13) that appear to be least affected by alteration
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
23
DISCUSS IONThe compositional and spatial development of the eruptedlava sequences of the Wrangellia oceanic plateau onVancouver Island provide information about the evolutionof a mantle plume upwelling beneath oceanic lithospherein an analogous way to the interpretation of continentalflood basalts The Caribbean and Ontong Java plateausare the two primary examples of accreted parts of oceanicplateaus that have been studied to date and comparisonscan be made with the Wrangellia oceanic plateauHowever the Wrangellia oceanic plateau is a distinctiveoccurrence of an oceanic plateau because it formed on thecrust of a Devonian to Mississippian intra-oceanic islandarc overlain by Mississippian to Permian limestones andpelagic sediments The nature and thickness of oceaniclithospheric mantle are considered to play a fundamentalrole in the decompression and evolution of a mantleplume and thus the compositions of the erupted lavasequences (Greene et al 2008) The geochemical and
stratigraphic relationships observed in the KarmutsenFormation onVancouver Island and the focus of the subse-quent discussion sections provide constraints on (1) thetemperature and degree of melting required to generatethe primary picritic magmas (2) the composition of themantle source (3) the depth of melting and residual min-eralogy in the source region (4) the low-pressure evolutionof the Karmutsen tholeiitic basalts (5) the growth of theWrangellia oceanic plateau
Melting conditions and major-elementcomposition of the primary magmasThe primary magmas to most flood basalt provinces arebelieved to be picritic (eg Cox 1980) and they have beenfound in many continental and oceanic flood basaltsequences worldwide (eg Siberia Karoo Paranacurren ^Etendeka Caribbean Deccan Saunders 2005) Near-pri-mary picritic lavas with low total alkali abundances
Table 5 Pb isotopic compositions of Karmutsen basaltsVancouver Island BC
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
require high-degree partial melting of the mantle source(eg Herzberg amp OrsquoHara 2002) The Keogh Lake picritesfrom northernVancouver Island are the best candidates foridentifying the least modified partial melts of the mantleplume source despite having accumulated olivine pheno-crysts and can be used to estimate conditions of meltingand the composition of the primary magmas for theKarmutsen FormationPrimary magma compositions mantle potential tem-
peratures and source melt fractions for the picritic lavas ofthe Karmutsen Formation were calculated from primitivewhole-rock compositions using PRIMELT1XLS software(Herzberg et al 2007) and are presented in Fig 12 andTable 6 A detailed discussion of the computationalmethod has been given by Herzberg amp OrsquoHara (2002)and Herzberg et al (2007) key features of the approachare outlined belowPrimary magma composition is calculated for a primi-
tive lava by incremental addition of olivine and compari-son with primary melts of mantle peridotite Lavas thathad experienced plagioclase andor clinopyroxene fractio-nation are excluded from this analysis All calculated pri-mary magma compositions are assumed to be derived byaccumulated fractional melting of fertile peridotite KR-4003 Uncertainties in fertile peridotite composition donot propagate to significant variations in melt fractionand mantle potential temperature melting of depletedperidotite propagates to calculated melt fractions that aretoo high but with a negligible error in mantle potential
temperature PRIMELT1XLS calculates the olivine liqui-dus temperature at 1atm which should be similar to theactual eruption temperature of the primary magmaand is accurate to 308C at the 2 level of confidenceMantle potential temperature is model dependent witha precision that is similar to the accuracy of the olivineliquidus temperature (C Herzberg personal communica-tion 2008)With regard to the evidence for accumulated olivine in
the Keogh Lake picrites it should not matter whether themodeled lava composition is aphyric or laden with accu-mulated olivine phenocrysts PRIMELT1 computes theprimary magma composition by adding olivine to a lavathat has experienced olivine fractionation in this way itinverts the effects of fractional crystallization The onlyrequirement is that the olivine phenocrysts are not acci-dental or exotic to the lava compositions Support for thisprocedure can be seen in the calculated olivine liquid-line-of-descent shown by the curved arrays with 5 olivineaddition increments (Fig 12c) Fractional crystallization ofolivine from model primary magmas having 85^95wt FeO and 153^175wt MgO will yield arrays of deriva-tive liquids and randomly sorted olivines that in most butnot all cases connect the picrites and high-MgO basalts Anewer version of the PRIMELT software (Herzberg ampAsimow 2008) can perform olivine addition to and sub-traction from lavas with highly variably olivine phenocrystcontents and yields results that converge with those shownin Fig 12
Fig 11 Relationship between abundance of olivine phenocrysts and whole-rock MgO contents for the Keogh Lake picrites (a) Tracings ofolivine grains (gray) in six high-MgO samples made from high resolution (2400 dpi) scans of petrographic thin-sections Scale bars represent10mm sample numbers are indicated at the upper left (b) Modal olivine vs whole-rock MgO Continuous line is the best-fit line for 10 high-MgO samplesThe areal proportion of olivine was calculated from raster images of olivine grains using ImageJ image analysis software whichprovides an acceptable estimation of the modal abundance (eg Chayes 1954)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
25
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
Gray lines = final melting pressure
L + Ol + Opx
07
08
09
Pressure (GPa)
10
3
4
5
6
1
SOLIDUS
01
04
0506
L + Ol
2
03
02
PeridotiteKR-4003
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
Thick black lines = initial melting pressure
Dashed lines = melt fraction
Melt Fraction
2
3 4 5 6
L+Ol+Opx
SolidusSpinel
SolidusGarnet Peridotite
7
L+Ol
0 10 20 30 4040
45
50
55
60
MgO (wt)
SiO2
(wt)
PeridotiteKR-4003
Melt Fraction
0 10 20 305
10
15
CaO(wt)
MgO (wt)
3
4 5 6 7
2
46
2
Peridotite Partial Melts
Thick black line = initial melting pressureGray lines = final melting pressure
Peridotite
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
02
Pressure (GPa)
0302
04
Pressure (GPa)
Picrite
High-MgO basalt
Tholeiitic basalt
(c)
(d)
Karmutsen flood basaltsOlivine addition modelOlivine addition (5 increment)Primary magma
PicriteHigh-MgO basaltTholeiitic basalt
(a)
(b)
800 850 900
910
920
930
940
950
Karmutsen flood basalts
Olivine addition modelOlivine addition (5 increment)Primary magma
Gray lines = Fo contents of olivine
Fig 12 Estimated primary magma compositions for three Keogh Lake picrites and high-MgO basalts (samples 4723A4 4723A135616A7) using the forward and inverse modeling technique of Herzberg et al (2007) Karmutsen compositions and modeling results areoverlain on diagrams provided by C Herzberg (a) Whole-rock FeO vs MgO for Karmutsen samples from this study Total iron estimatedto be FeO is 090 Gray lines show olivine compositions that would precipitate from liquid of a given MgO^FeO composition Black lineswith crosses show results of olivine addition (inverse model) using PRIMELT1 (Herzberg et al 2007) Gray field highlights olivinecompositions with Fo contents of 91^92 predicted to precipitate (b) FeO vs MgO showing Karmutsen lava compositions and results ofa forward model for accumulated fractional melting of fertile peridotite KR-4003 (Kettle River Peridotite Walter 1998) (c) SiO2 vsMgO for Karmtusen lavas and model results (d) CaO vs MgO showing Karmtusen lava compositions and model results To brieflysummarize the technique [see Herzberg et al (2007) for complete description] potential parental magma compositions for the high-MgOlava series were selected (highest MgO and appropriate CaO) and using PRIMELT1 software olivine was incrementally added tothe selected compositions to show an array of potential primary magma compositions (inverse model) Then using PRIMELT1 theresults from the inverse model were compared with a range of accumulated fractional melts for fertile peridotite derived fromparameterization of the experimental results for KR-4003 from Walter (1998) forward model Herzberg amp OrsquoHara (2002) A melt fractionwas sought that was unique to both the inverse and forward models (Herzberg et al 2007) A unique solution was found when there was a com-mon melt fraction for both models in FeO^MgO and CaO^MgO^Al2O3^SiO2 (CMAS) projection space This modeling assumes thatolivine was the only phase crystallizing and ignores chromite precipitation and possible augite fractionation in the mantle (Herzbergamp OrsquoHara 2002) Results are best for a residue of spinel lherzolite (not pyroxenite) The presence of accumulated olivine in samples ofthe Keogh Lake picrites used as starting compositions does not significantly affect the results because we are modeling addition of olivine(see text for further description) The tholeiitic basalts cannot be used for modeling because they are all saturated in plagthorn cpxthornolThick black line for initial melting pressure at 4 GPa is not shown in (c) for clarity Gray area in (a) indicates parental magmas thatwill crystallize olivine with Fo 91^92 at the surface Dashed lines in (c) indicate melt fraction Solidi for spinel and garnet peridotite arelabelled in (c)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
The PRIMELT1 results indicate that if the Karmutsenpicritic lavas were derived from accumulated fractionalmelting of fertile peridotite the primary magmas wouldhave contained 15^17wt MgO and 10wt CaO(Fig 12 Table 6) formed from 23^27 partial meltingand would have crystallized olivine with a Fo content of 91 (Fig 12 Table 6) Calculated mantle temperatures forthe Karmutsen picrites (14908C) indicate melting ofanomalously hot mantle (100^2008C above ambient) com-pared with ambient mantle that produces mid-ocean ridgebasalt (MORB 1280^14008C Herzberg et al 2007)which initiated between 3 and 4 GPa Several picritesappear to be near-primary melts and required very littleaddition of olivine (eg sample 93G171 with measured158wt MgO) These temperature and melting esti-mates of the Karmutsen picrites are consistent withdecompression of hot mantle peridotite in an actively con-vecting plume head (ie plume initiation model) Threesamples (picrite 5614A1 and high-MgO basalts 5614A3and 5614A5) are lower in iron than the other Karmutsenlavas with FeO in the 65^80wt range (Fig 12c) andhave MgO and FeO contents that are similar to primitiveMORB from the Sequeiros fracture zone of the EastPacific Rise (EPR Perfit et al 1996) These samples
however differ from EPR MORB in having higher SiO2
and lower Al2O3 and they are restricted geographicallyto one area (Sara Lake Fig 2)We therefore have evidence for primary magmas with
MgO contents that range from a possible low of 110wt to as high as 174wt with inferred mantle potential tem-peratures from 1340 to 15208C This is similar to the rangedisplayed by lavas from the Ontong Java Plateau deter-mined by Herzberg (2004) who found that primarymagma compositions assuming a peridotite sourcewould contain 17wt MgO and 10wt CaO(Table 6) and that these magmas would have formed by27 melting at a mantle potential temperature of15008C (first melting occurring at 36 GPa) Fitton ampGodard (2004) estimated 30 partial melting for theOntong Java lavas based on the Zr contents of primarymagmas of Kroenke-type basalt calculated by incrementaladdition of equilibrium olivine However if a considerableamount of eclogite was involved in the formation of theOntong Java plateau magmas the excess temperatures esti-mated by Herzberg et al (2007) may not be required(Korenaga 2005) The variability in mantle potential tem-perature for the Keogh Lake picrites is also similar to thewide range of temperatures that have been calculated for
Table 6 Estimated primary magma compositions for Karmutsen basalts and other oceanic plateaus or islands
Sample 4723A4 4723A13 5616A7 93G171 Average OJP Mauna Keay Gorgonaz
Ontong Java primary magma composition for accumulated fraction melting (AFM) from Herzberg (2004)yMauna Kea primary magma composition is average of four samples of Herzberg (2006 table 1)zGorgona primary magma composition for AFM for 1F fertile source from Herzberg amp OrsquoHara (2002 table 4)Total iron estimated to be FeO is 090
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
27
lavas from single volcanoes in the Galapagos Hawaii andelsewhere by Herzberg amp Asimow (2008) Those workersinterpreted the range to reflect the transport of magmasfrom both the cool periphery and the hotter axis of aplume A thermally heterogeneous plume will yield pri-mary magmas with different MgO contents and the rangeof primary magma compositions and their inferred mantlepotential temperatures for Karmutsen lavas may reflectmelting from different parts of the plume
Source of the Karmutsen Formation lavasThe Sr^Nd^Hf^Pb isotopic compositions of theKarmutsen volcanic rocks on Vancouver Island indicatethat they formed from plume-type mantle containing adepleted component that is compositionally similar to butdistinct from other oceanic plateaus that formed in thePacific Ocean (eg Ontong Java and Caribbean Fig 13)Trace element and isotopic constraints can be used to testwhether the depleted component of the KarmutsenFormation was intrinsic to the mantle plume and distinctfrom the source of MORB or the result of mixing or invol-vement of ambient depleted MORB mantle with plume-type mantle The Karmutsen tholeiitic basalts have initialeNd and
87Sr86Sr ratios that fall within the range of basaltsfrom the Caribbean Plateau (eg Kerr et al 1996 1997Hauff et al 2000 Kerr 2003) and a range of Hawaiian iso-tope data (eg Frey et al 2005) and lie within and abovethe range for the Ontong Java Plateau (Tejada et al 2004Fig 13a) Karmutsen tholeiitic basalts with the highest ini-tial eNd and lowest initial 87Sr86Sr lie within and justbelow the field for age-corrected northern EPR MORB at230 Ma (eg Niu et al1999 Regelous et al1999) Initial Hfand Nd isotopic compositions for the Karmutsen samplesare noticeably offset to the low side of the ocean islandbasalt (OIB) array (Vervoort et al 1999) and lie belowand slightly overlap the low eHf end of fields for Hawaiiand the Caribbean Plateau (Fig 13c) the Karmutsen sam-ples mostly fall between and slightly overlap a field forage-corrected Pacific MORB at 230 Ma and Ontong Java(Mahoney et al 1992 1994 Nowell et al 1998 Chauvel ampBlichert-Toft 2001) Karmutsen lava Pb isotope ratios over-lap the broad range for the Caribbean Plateau and thelinear trends for the Karmutsen samples intersect thefields for EPR MORB Hawaii and Ontong Java in208Pb^206Pb space but do not intersect these fields in207Pb^206Pb space (Fig 13d and e) The more radiogenic207Pb204Pb for a given 206Pb204Pb of the Karmutsenbasalts requires high UPb ratios at an ancient time when235U was abundant similar to the Caribbean Plateau andenriched in comparison with EPR MORB Hawaii andOntong JavaThe low eHf^low eNd component of the Karmutsen
basalts that places them below the OIB mantle array andpartly within the field for age-corrected Pacific MORBcan form from low ratios of LuHf and high SmNd over
time (eg Salters amp White 1998) MORB will develop aHf^Nd isotopic signature over time that falls below themantle array Low LuHf can also lead to low 176Hf177Hffrom remelting of residues produced by melting in theabsence of garnet (eg Salters amp Hart 1991 Salters ampWhite 1998) The Hf^Nd isotopic compositions of theKarmutsen Formation indicate the possible involvementof depleted MORB mantle however similar to Iceland(Fitton et al 2003) Nb^Zr^Y variations provide a way todistinguish between a depleted component within theplume and a MORB-like component The Karmutsenbasalts and picrites fall within the Iceland array and donot overlap the MORB field in a logarithmic plot of NbYvs ZrY (Fig 13b) using the fields established by Fittonet al (1997) Similar to the depleted component within theCaribbean Plateau (Thompson et al 2003) the trend ofHf^Nd isotopic compositions towards Pacific MORB-likecompositions is not followed by MORB values in Nb^Zr^Y systematics and this suggests that the depleted compo-nent in the source of the Karmutsen basalts is distinctfrom the source of MORB and probably an intrinsic partof the plume The relatively radiogenic Pb isotopic ratiosand REE patterns distinct from MORB also support thisinterpretationTrace-element characteristics suggest the possible invol-
vement of sub-arc lithospheric mantle in the generation ofthe Karmutsen picrites Lassiter et al (1995) suggested thatmixing of a plume-type source with eNdthorn 6 to thorn 7 witharc material with low NbTh could reproduce the varia-tions observed in the Karmutsen basalts however theyconcluded that the absence of low NbLa ratios in theirsample of tholeiitic basalts restricts the amount of arc litho-sphere involved The study of Lassiter et al (1995) wasbased on major and trace element and Sr Nd and Pb iso-topic data for a suite of 29 samples from Buttle Lake in theStrathcona Provincial Park on central Vancouver Island(Fig 1) Whereas there are no clear HFSE depletions inthe Karmutsen tholeiitic basalts the low NbLa (and TaLa) of the Karmutsen picritic lavas in this study indicatethe possible involvement of Paleozoic arc lithosphere onVancouver Island (Fig 8)
REE modeling dynamic melting andsource mineralogyThe combined isotopic and trace-element variations of thepicritic and tholeiitic Karmutsen lavas indicate that differ-ences in trace elements may have originated from differentmelting histories For example the LuHf ratios of theKarmutsen picritic lavas (mean 008) are considerablyhigher than those in the tholeiitic basalts (mean 002)however the small range of overlapping initial eHf valuesfor the two lava suites suggests that these differences devel-oped during melting within the plume (Fig 9b) Below wemodel the effects of source mineralogy and depth of
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
28
melting on differences in trace-element concentrations inthe tholeiitic basalts and picritesDynamic melting models simulate progressive decom-
pression melting where some of the melt fraction isretained in the residue and only when the degree of partialmelting is greater than the critical mass porosity and the
source becomes permeable is excess melt extracted fromthe residue (Zou1998) For the Karmutsen lavas evolutionof trace element concentrations was simulated using theincongruent dynamic melting model developed by Zou ampReid (2001) Melting of the mantle at pressures greater than1 GPa involves incongruent melting (eg Longhi 2002)
OIB array
PacificMORB(230 Ma)
(0 Ma)
EPR(230 Ma)
-4
-2
0
2
4
8
10
12
14
16
18
20
22
-4 -2 0 2 4 6 8 10 12 14
minus2
0
2
4
6
8
10
12
14
0702 0703 0704 0705 0706
IndianMORB
87Sr 86Sr (initial)
Hf (initial)
(a) (c)
Nd (initial)
Karmutsen tholeiitic basalt
HawaiiOntong JavaCaribbean Plateau
Karmutsen high-MgO lava
high-MgO lavasKarmutsen
1540
1545
1550
1555
1560
1565
175 180 185 190 195 200375
380
385
390
395
175 180 185 190 195 200
(d) (e)
EPR
EPR
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb
Karmutsen Formation (initial)
HawaiiOntong JavaCaribbean Plateau
East Pacific Rise
Karmutsen Formation (measured)
Juan de FucaGorda
ε Nd
(initial)
Karmutsen Formation (different age-correction for 3 picrites)
(0 Ma)
NbY
001
01
1
1 10
ZrY
OIB
NMORB
(b)
Iceland array
6
Fig 13 Comparison of age-corrected (230 Ma) Sr^Nd^Hf^Pb isotopic compositions for Karmutsen flood basalts onVancouver Island withage-corrected OIB and MORB and Nb^Zr^Y variation (a) Initial eNd vs 87Sr86Sr (b) NbYand ZrY variation (after Fitton et al 1997)(c) Initial eHf vs eNd (d) Measured and initial 207Pb204Pb vs 206Pb204Pb (e) Measured and initial 208Pb204Pb vs 206Pb204Pb References fordata sources are listed in Electronic Appendix 4 Most of the compiled data were extracted from the GEOROC database (httpgeorocmpch-mainzgwdgdegeoroc) OIB array line in (c) is fromVervoort et al (1999) EPR East Pacific Rise Several high-MgO samples affected bysecondary alteration were not plotted for clarity in Pb isotope plots Estimation of fields for age-corrected Pacific MORB and EPR at 230 Mawere made using parent^daughter concentrations (in ppm) from Salters amp Stracke (2004) of Lufrac14 0063 Hffrac14 0202 Smfrac14 027 Ndfrac14 071Rbfrac14 0086 and Srfrac14 836 In the NbYvs ZrYplot the normal (N)-MORB and OIB fields not including Icelandic OIB are from data com-pilations of Fitton et al (1997 2003) The OIB field is data from 780 basalts (MgO45wt ) from major ocean islands given by Fitton et al(2003) The N-MORB field is based on data from the Reykjanes Ridge EPR and Southwest Indian Ridge from Fitton et al (1997) The twoparallel gray lines in (b) (Iceland array) are the limits of data for Icelandic rocks
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
29
with olivine being produced during the reactioncpxthornopxthorn spfrac14meltthornol for spinel peridotites The mod-eling used coefficients for incongruent melting reactionsbased on experiments on lherzolite melting (eg Kinzleramp Grove 1992 Longhi 2002) and was separated into (1)melting of spinel lherzolite (2) two combined intervals ofmelting for a garnet lherzolite (gt lherz) and spinel lher-zolite (sp lherz) source (3) melting of garnet lherzoliteThe model solutions are non-unique and a range indegree of melting partition coefficients melt reactioncoefficients and proportion of mixed source componentsis possible to achieve acceptable solutions (see Fig 14 cap-tion for description of modeling parameters anduncertainties)The LREE-depleted high-MgO lavas require melting of
a depleted spinel lherzolite source (Fig 14a) The modelingresults indicate a high degree of melting (22^25) similarto the results from PRIMELT1 based on major-elementvariations (discussed above) of a LREE-depleted sourceThe high-MgO lavas lack a residual garnet signature Theenriched tholeiitic basalts involved melting of both garnetand spinel lherzolite and represent aggregate melts pro-duced from continuous melting throughout the depth ofthe melting column (Fig 14b) Trace-element concentra-tions in the tholeiitic and picritic lavas are not consistentwith low-degree melting of garnet lherzolite alone(Fig 14c) The modeling results indicate that the aggregatemelts that formed the tholeiitic basalts would haveinvolved lesser proportions of enriched small-degreemelts (1^6 melting) generated at depth and greateramounts of depleted high-degree melts (12^25) gener-ated at lower pressure (a ratio of garnet lherzolite tospinel lherzolite melt of 14 provides the best fits seeFig 14b and description in figure caption) The totaldegree of melting for the tholeiitic basalts was high(23^27) similar to the degree of partial melting in thespinel lherzolite facies for the primary melts that eruptedas the Keogh Lake picrites of a source that was also prob-ably depleted in LREE
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
(c)
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
Melting of garnet lherzolite
High-MgO lavas
Melting of spinel lherzolite
source (07DM+03PM)
source (07DM+03PM)
6 melting
30 melting
(b)
(a)
30 melting
Sam
ple
Cho
ndrit
eS
ampl
eC
hond
rite 20 melting
1 melting
Melting of garnet lherzoliteand spinel lherzolite
x gt lherz + 4x sp lherz
5 meltingx gt lherz + 4x sp lherz
Sam
ple
Cho
ndrit
e
Tholeiitic lavas
Tholeiitic lavas
High-MgO lavas
source (07DM+03PM)
Fig 14 Trace-element modeling results for incongruent dynamicmantle melting for picritic and tholeiitic lavas from the KarmutsenFormation Three steps of modeling are shown (a) melting of spinellherzolite compared with high-MgO lavas (b) melting of garnet lher-zolite and spinel lherzolite compared with tholeiitic lavas (c) meltingof garnet lherzolite compared with tholeiitic and picritic lavasShaded fields in each panel for tholeiitic basalts and picrites arelabeled Patterns with symbols in each panel are modeling results(patterns are 5 melting increments in (a) and (b) and 1 incre-ments in (c) PM primitive mantle DM depleted MORB mantle gtlherz garnet lherzolite sp lherz spinel lherzolite In (b) the ratios ofper cent melting for garnet and spinel lherzolite are indicated (x gtlherzthorn 4x sp lherz means that for 5 melting 1 melt in garnetfacies and 4 melt in spinel facies where x is per cent melting in theinterval of modeling) The ratio of the respective melt proportions inthe aggregate melt (x gt lherzthorn 4x sp lherz) was kept constant andthis ratio of melt contribution provided the best fit to the data Arange of proportion of source components (07DMthorn 03PM to
09DMthorn 01PM) can reproduce the variation in the high-MgOlavas Melting modeling uses the formulation of Zou amp Reid (2001)an example calculation is shown in their Appendix PM fromMcDonough amp Sun (1995) and DM from Salters amp Stracke (2004)Melt reaction coefficients of spinel lherzolite from Kinzler amp Grove(1992) and garnet lherzolite fromWalter (1998) Partition coefficientsfrom Salters amp Stracke (2004) and Shaw (2000) were kept constantSource mineralogy for spinel lherzolite (018cpx027opx052ol003sp) from Kinzler (1997) and for garnet lherzolite(034cpx008opx053ol005gt) from Salters amp Stracke (2004) Arange of source compositions for melting of spinel lherzolite(07DMthorn 03PM to 03DMthorn 07PM) reproduce REE patternssimilar to those of the high-MgO lavas with best fits for 22ndash25melting and 07DMthorn 03PM source composition Concentrationsof tholeiitic basalts cannot be reproduced from an entirelyprimitive or depleted source
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
30
The uncertainty in the composition of the source in thespinel lherzolite melting model for the high-MgO lavasprecludes determining whether the source was moredepleted in incompatible trace elements than the source ofthe tholeiitic lavas (see Fig 14 caption for description) It ispossible that early high-pressure low-degree melting left aresidue within parts of the plume (possibly the hotter inte-rior portion) that was depleted in trace elements (egElliott et al 1991) further decompression and high-degreemelting of such depleted regions could then have generatedthe depleted high-MgO Karmutsen lavas Shallow high-degree melting would preferentially sample depletedregions whereas deeper low-degree melting may notsample more refractory depleted regions (eg CaribbeanEscuder-Viruete et al 2007) The picrites lie at the highend of the range of Nd isotopic compositions and havelow NbZr similar to depleted Icelandic basalts (Fittonet al 2003) therefore the picrites may represent the partsof the mantle plume that were the most depleted prior tothe initiation of melting As Fitton et al (2003) suggestedthese depleted melts may only rarely be erupted withoutcombining with more voluminous less depleted melts andthey may be detected only as undifferentiated small-volume flows like the Keogh Lake picrites within theKarmutsen Formation on Vancouver IslandDecompression melting within the mantle plume initiatedwithin the garnet stability field (35^25 GPa) at highmantle potential temperatures (Tp414508C) and pro-ceeded beneath oceanic arc lithosphere within the spinelstability field (25^09 GPa) where more extensivedegrees of melting could occur
Magmatic evolution of Karmutsentholeiitic basaltsPrimary magmas in large igneous provinces leave exten-sive coarse-grained cumulate residues within or below thecrust these mafic and ultramafic plutonic sequences repre-sent a significant proportion of the original magma thatpartially crystallized at depth (eg Farnetani et al 1996)For example seismic and petrological studies of OntongJava (eg Farnetani et al 1996) combined with the use ofMELTS (Ghiorso amp Sack 1995) indicate that the volcanicsequence may represent only 40 of the total magmareaching the Moho Neal et al (1997) estimated that30^45 fractional crystallization took place for theOntong Java magmas and produced cumulate thicknessesof 7^14 km corresponding to 11^22 km of flood basaltsdepending on the presence of pre-existing oceanic crustand the total crustal thickness (25^35 km) Interpretationsof seismic velocity measurements suggest the presence ofpyroxene and gabbroic cumulates 9^16 km thick beneaththe Ontong Java Plateau (eg Hussong et al 1979)Wrangellia is characterized by high crustal velocities in
seismic refraction lines which is indicative of mafic
plutonic rocks extending to depth beneath VancouverIsland (Clowes et al 1995)Wrangellia crust is 25^30 kmthick beneath Vancouver Island and is underlain by astrongly reflective zone of high velocity and density thathas been interpreted as a major shear zone where lowerWrangellia lithosphere was detached (Clowes et al 1995)The vast majority of the Karmutsen flows are evolved tho-leiitic lavas (eg low MgO high FeOMgO) indicating animportant role for partial crystallization of melts at lowpressure and a significant portion of the crust beneathVancouver Island has seismic properties that are consistentwith crystalline residues from partially crystallizedKarmutsen magmas To test the proportion of the primarymagma that fractionated within the crust MELTS(Ghiorso amp Sack 1995) was used to simulate fractionalcrystallization using several estimated primary magmacompositions from the PRIMELT1 modeling results Themajor-element composition of the primary magmas couldnot be estimated for the tholeiitic basalts because of exten-sive plagthorn cpxthornol fractionation and thus the MELTSmodeling of the picrites is used as a proxy for the evolutionof major elements in the volcanic sequence A pressure of 1kbar was used with variable water contents (anhydrousand 02wt H2O) calculated at the quartz^fayalite^magnetite (QFM) oxygen buffer Experimental resultsfrom Ontong Java indicate that most crystallizationoccurs in magma chambers shallower than 6 km deep(52 kbar Sano amp Yamashita 2004) however some crys-tallization may take place at greater depths (3^5 kbarFarnetani et al 1996)A selection of the MELTS results are shown in Fig 15
Olivine and spinel crystallize at high temperature until15^20wt of the liquid mass has fractionated and theresidual magma contains 9^10wt MgO (Fig 15f) At1235^12258C plagioclase begins crystallizing and between1235 and 11908C olivine ceases crystallizing and clinopyr-oxene saturates (Fig 15f) As expected the addition of clin-opyroxene and plagioclase to the crystallization sequencecauses substantial changes in the composition of the resid-ual liquid FeO and TiO2 increase and Al2O3 and CaOdecrease while the decrease in MgO lessens considerably(Fig 15) The near-vertical trends for the Karmutsen tho-leiitic basalts especially for CaO do not match the calcu-lated liquid-line-of-descent very well which misses manyof the high-CaO high-Al2O3 low-FeO basalts A positivecorrelation between Al2O3CaO and SrNd indicates thataccumulation of plagioclase played a major role in the evo-lution of the tholeiitic basalts (Fig 15g) Increasing thecrystallization pressure slightly and the water content ofthe melts does not systematically improve the match ofthe MELTS models to the observed dataThe MELTS results indicate that a significant propor-
tion of the crystallization takes place before plagioclase(15^20 crystallization) and clinopyroxene (35^45
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
31
crystallization) join the crystallization sequence Theseresults suggest that445 crystallization of the primarymagmas occurred (estimated from the residual liquid percent in Fig 15f) and if the original flood basalt stratigra-phy on Vancouver Island was 6 km thick at least anequivalent amount of complementary plutonic rocksshould have crystallized The Karmutsen basalts and theirplutonic residues thus represent a significant addition ofcrust (perhaps 412 km thickness) previously thickenedduring the formation of the underlying Paleozoic arc
Growth of the Wrangellia oceanic plateauon Vancouver IslandThe extrusive part of the Wrangellia oceanic plateau onVancouver Island was constructed in a three-layersequence of mostly tholeiitic basalts with a restrictedrange of composition (Fig 16) The plateau overliesDevonian to Mississippian arc volcanic sequences andmarine sediments of Mississippian to Permian ages thatare intruded by gabbroic rocks related to the flood basaltsThe volcanic stratigraphy of theWrangellia plateau formed
00
05
10
15
20
25
30
35
40
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
16
0 2 4 6 8 10 12 14 16 18 20
10
11
12
13
14
15
16
17
18
19
20
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
0 2 4 6 8 10 12 14 16 18 20
MgO (wt)
0 10 20 30 40 50
percent of total mass
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
) Mineral proportions
Sp (e)
Ol
CpxPlag
PicriteHigh-MgO basaltTholeiitic basalt
TiO2 (wt)
MgO (wt)
Al2O3 (wt) CaO (wt)
FeO (wt)
MgO(a) (b)
(c) (d)
MgO (wt)MgO (wt)
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
)
Residual liquid ()
1220
1190
Ol + Sp
Plag+Ol+Sp Plag+Cpx+Sp
02 wt H2O
no H2O
Residual liquid and crystallizing phases
(f)
0102030405060708090100
plagioclaseaccumulation
plagioclaseaccumulation
plagioclase
accumulatio
n
10
12
14
16
0 10 20 30 40
Al2O3CaO
SrNd
tholeiitic basalts
(g)
Fig 15 Forward fractional crystallization modeling results for major elements from MELTS (Ghiorso amp Sack 1995) compared with picriticand tholeiitic lavas from the Karmutsen Formation (a^d) Major-element compositions of residual liquids compared with Karmutsen lavacompositions For clarity the results are shown only for one starting composition (open circle) using the composition of the estimated primarymagma for the high-MgO lava series from the modeling technique of Herzberg et al (2007) for sample 4723A13 as shown in Fig 12 The esti-mated primary magmas for samples 4723A3 and 93G171 yield similar results A pressure of 1 kbar was used with no H2O and 02wt H2Ocalculated at QFM black line is anhydrous and grey line is 02wt H2OThe end-point for the anhydrous result is at 78 fractionation and11308C and the end-point for the result with 02wt H2O is 75 fractionation and11108C (e) Mineral proportions of crystallized phases fromone MELTrun (sample 4723A13) with no H2O (f) Per cent residual liquid vs temperature labelled with intervals of crystallizing phases (g)Al2O3CaO vs SrNd for the tholeiitic basalts
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
32
in a single eruptive phase (c 231^225 Ma) as an emergentbasalt sequence during a deeper-water shallow-water andsubaerial stage and subsided below sea level relativelysoon after its formation (Fig 16) The deeper-water stagecontains pillowed and massive submarine flows (3000mthick) that built up the submarine volcanic edifice from theseafloor The lowest pillow basalts were emplaced onunconsolidated and lithified fine-grained siliciclastic andcarbonaceous sediments containing Daonella beds(c 235^232 Ma) that are intruded by abundantKarmutsen mafic sills (Fig 16a) The different pillowedand massive submarine flows resulted primarily from dif-ferent effusive rates and local topography Some of themassive submarine flows commonly recognized by radiat-ing columnar jointing represent master tubes for the deliv-ery of lava to distal parts of submarine flow fields Thepillowed flows which commonly have uneven topography
in modern sequences may have served as levees to themassive submarine flows rather than allowing unob-structed spreading and inflation typical of the massive sub-aerial flows that form on shallow slopes in continentalflood basalt provinces (eg Self et al 1997) The rapid suc-cession of flows is evident by the near absence of sedimentbetween submarine flows Minor volumes of picritic pillowlavas occur in the upper part of the pillowed flow sequencein areas of the plateau exposed on northern VancouverIsland (Fig 16 Nixon et al 2008)The growth of the plateau in shallow water led to
emplacement of an increasing proportion of volcaniclastics(pillow breccias and hyaloclastites) that accumulated up to1500m thick in some areas (Fig 16b Nixon et al 2008)Horizons of pillowed flows occur locally within the volca-niclastic stratigraphy Sedimentary structures (eg gradedbedding fluidization structures) and petrographic textures
Late Devonian-Mississippian arc
Daonella beds
Middle Triassic
deeper water
Late Devonian-Mississippian arc
pillowed flows massive submarine
flows
pillow brecciahyaloclastiteshallow water
~4 km
~6 km
Late Triassic limestoneinterflow limestonelenses
Karmutsen Formation
mafic intrusions sediment-sill complex
Karmutsen Formation
(a) (b)
(c) (d)
subaerial flows
Late Devonian-Mississippian arc Late Devonian-Mississippian arc
picritic pillow basalts
Miss to Permian marine sediments
sills
Fig 16 Growth of the Wrangellia oceanic plateau onVancouver Island (a) Deep-water eruptive phase of pillowed and massive submarineflows with almost no intervening sediments onto MiddleTriassic shale containing Daonella beds (c 235^232 Ma) The basement of the plateauis composed of Devonian to Mississippian arc sequences overlain by Mississippian to Permian carbonates and siliciclastic sediments(b) Shallow-water flow emplacement with increasing proportion of volcaniclastics resulted primarily from cooling^contraction granulationPicritic pillow basalts occur primarily near the transition between pillowed submarine flows and volcaniclastic units Pillowed and massiveflows occur locally within the volcaniclastic unit (c) Emergent stage of growth consists of overlapping subaerial flow fields similar to continen-tal flood basalts with pahoehoe flow structures Multi-tiered columnar jointing occurrs locally in response to ponding of flows (eg Bondre et al2004) Interflow carbonate deposits and plagioclase megacrystic flows developed locally during the waning stage of volcanism (d) Post-volcanicsubsidence and deposition of Late Carnian to Norian (c 212^223 Ma) platform carbonates on top of the plateau Units are labeled in the dia-grams Thicknesses of the underlying units are not to scale
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
33
within the volcaniclastic rocks indicate formation primar-ily via cooling^contraction granulation magma^water^steam interaction autobrecciation resedimentation andmass-wasting rather than pyroclastic eruption (Nixonet al 2008)A broad subaerial platform was constructed by emplace-
ment of subaerial lava flows during an emergent subaerialstage (Fig 16c) The subaerial unit predominantly com-prises well-layered sheet flows that formed in a similarway to continental flood basalts as hundred(s) of pahoehoeflow fields erupted on shallow slopes (Self et al 1997) how-ever volcaniclastic and pillowed flow units are locally pre-sent mostly in the upper parts of the stratigraphyLocalized ponding of flows led to the formation of multi-tiered columnar jointing in some subaerial flows (egBondre et al 2004) As volcanism waned local interflowcarbonate deposits developed and plagioclase megacrysticflows were erupted Subsidence of the plateau is recordedby the interflow lenses and deposits of platform carbonates(40^1500m thick) on top of the Karmutsen basalts thatcontain Late Carnian to Early Norian fossils (Fig 16d c216^227 Ma Nixon et al 2008)The evolution of the internal architecture of the
Wrangellia oceanic plateau was primarily related to theconditions of emplacement (ie water depth and eruptionenvironment) There does not appear to have been a sys-tematic coincident evolution or shift of lava compositionsonVancouver Island as has been recognized in most con-tinental flood basalts and parts of the Wrangellia oceanicplateau in Alaska and Yukon (Greene et al 2008 2009)In general with the exception of the picritic pillow lavasin the submarine stratigraphy and local variations betweenhigh-MgO and tholeiitic lavas there are no significantcompositional differences between lavas erupted duringthe different stages of growth of the plateau (ie submarineand subaerial flows) nor are substantial compositional dif-ferences generally found between massive submarine flowsin contact with pillowed flows tholeiitic basalt flows nearthe base of the Karmutsen Formation are compositionallysimilar to flows sampled near the top of the volcanicsequence Where it was possible to sample continuous sec-tions of volcanic stratigraphy greater than several hun-dreds of meters (eg Keogh^Maynard Lake area west ofthe Karmutsen Range Fig 2) there is no clear relationshipbetween stratigraphic position and compositionalvariation
CONCLUSIONSThe Karmutsen Formation covers extensive areas ofVancouver Island (20 000 km2) and forms part of a largeoceanic plateau erupted over a geologically short interval(c 225^231 Ma) The tripartite volcanic stratigraphyexposed on Vancouver Island is at least 6 km thick andcomprises a succession of submarine flows volcaniclastic
deposits and massive subaerial flowsThese volcanologicaldifferences are primarily related to the eruption environ-ment (deep-water shallow-water or subaerial) The rapidgrowth of the plateau prevented significant accumulationsof intervening sediment except in the uppermost stratigra-phy where isolated limestone lenses commonly associatedwith pillow basalt and hyaloclastite deposits preserve arecord of the final stages of subsidence of the plateauThe Wrangellia plateau on Vancouver Island was con-
structed dominantly of tholeiitic basalt with restrictedmajor- and trace-element and isotopic compositional var-iations Minor volumes of picritic pillow basalts eruptedlate in the submarine phase The Karmutsen basalts wereemplaced onto older Paleozoic arc volcanic sequences andlow NbLa for the picrites may indicate a contributionfrom the sub-arc lithospheric mantle Modeling of the pet-rogenesis of the Karmutsen picrites indicates melting ofanomalously hot mantle (15008C) and extensive degreesof partial melting (23^27) and is consistent with aplume initiation model Combined Sr^Nd^Hf^Pb isotopicsystematics indicate that the picritic and tholeiitic lavaswere derived from a depleted plume-type mantle sourcedistinct from the source of MORB that has compositionalsimilarities to the source of the lavas forming theCaribbean Plateau Picrites were generated by melting ofspinel lherzolite whereas the tholeiitic basalts formed asaggregate melts of both garnet and spinel lherzolite Thetholeiitic basalts underwent extensive low-pressure fractio-nation (52^3 kbar) and seismic studies indicate that someof theWrangellia crust beneathVancouver Island may cor-respond to the plutonic residues of the Karmutsen floodbasalts
ACKNOWLEDGEMENTSWe are grateful to Nick Arndt for helping us get this pro-ject started and Nick Massey for insights into VancouverIsland geology We greatly appreciate the assistance withPRIMELT modeling from Claude Herzberg We thankMikkel Schau for his insight and enthusiasm during field-work Jane Barling assisted with analyses by MC-ICP-MSThoughtful and thorough reviews of the manuscript byJohn Mahoney Fred Frey Wendy Bohrson DougalJerram and Teal Riley are greatly appreciated Fundingwas generously provided by the Rocks to Riches Programadministered by the BC amp Yukon Chamber of Mines in2004 by the BC Geological Survey and by NSERCDiscovery Grants to JSS and DW ARG was supportedby a University Graduate Fellowship at UBC
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
34
REFERENCESBarker F Brown A S Budahn J R amp Plafker G (1989) Back-arc
with frontal-arc component origin of Triassic Karmutsen basaltBritish Columbia Canada Chemical Geology 75 81^102
Blichert-Toft JWeis D Maerschalk C Agranier A amp Albare de F(2003) Hawaiian hot spot dynamics as inferred from the Hf and Pbisotope evolution of Mauna Kea volcano Geochemistry GeophysicsGeosystems 4(2) 1^27 doi1010292002GC000340
Bondre N R Duraiswami R A amp Dole G (2004) Morphologyand emplacement of flows from the Deccan Volcanic ProvinceIndia Bulletin of Volcanology 66 29^45
Carlisle D (1963) Pillow breccias and their aquagene tuffs QuadraIsland British Columbia Journal of Geology 71 48^71
Carlisle D (1972) Late Paleozoic to Mid-Triassic sedimentary-volcanic sequence on Northeastern Vancouver Island Geological
Society of Canada Paper Report of Activities 72-1 Part B 24^30Carlisle D amp SuzukiT (1974) Emergent basalt and submergent car-
bonate^clastic sequences including the Upper Triassic Dilleri andWelleri zones on Vancouver Island Canadian Journal of Earth
Sciences 11 254^279Chauvel C amp Blichert-Toft J (2001) A hafnium isotope and trace
element perspective on melting of the depleted mantle Earth and
Planetary Science Letters 190 137^151Chayes F (1954)The theory of thin-section analysis Journal of Geology
62 92^101Cho M Liou J G amp Maruyama S (1987) Transition from the zeo-
lite to prehnite^pumpellyite facies in the Karmutsen metabasitesVancouver Island British Columbia Journal of Petrology 27(2)467^494
Clowes R M Zelt C A Amor J R amp Ellis R M (1995)Lithospheric structure in the southern Canadian Cordillera froma network of seismic refraction lines Canadian Journal of Earth
Sciences 32(10) 1485^1513Coffin M F amp Eldholm O (1994) Large igneous provinces Crustal
structure dimensions and external consequences Reviews of
Geophysics 32(1) 1^36Cox K G (1980) A model for flood basalt volcanism Journal of
Petrology 21 629^650Eldholm O amp Coffin M F (2000) Large igneous provinces
and plate tectonics In Richards M A Gordon R G amp vander Hilst R D (eds) The History and Dynamics of Global
Plate Motions American Geophysical Union Geophysical Monograph 121309^326
Elliott T R Hawkesworth C J amp Grolaquo nvold K (1991) Dynamicmelting of the Iceland plume Nature 351 201^206
Escuder-Viruete J Pecurren rez-Estaucurren n A Contreras F Joubert MWeis D Ullrich T D amp Spadea P (2007) Plume mantle sourceheterogeneity through time Insights from the Duarte ComplexHispaniola northeastern Caribbean Journal of Geophysical Research112(B04203) doi1010292006JB004323
Farnetani C G Richards M A amp Ghiorso M S (1996)Petrological models of magma evolution and deep crustal structurebeneath hotspots and flood basalt provinces Earth and Planetary
Science Letters 143 81^94Fitton J G amp Godard M (2004) Origin and evolution of magmas
on the Ontong Java Plateau In Fitton J G Mahoney J JWallace P J amp Saunders A D (eds) Origin and Evolution of the
Ontong Java Plateau Geological Society London Special Publications 229151^178
Fitton J G Saunders A D Norry M J Hardarson B S ampTaylor R N (1997) Thermal and chemical structure of theIceland plume Earth and Planetary Science Letters 153 197^208
Fitton J G Saunders A D Kempton P D amp Hardarson B S(2003) Does depleted mantle form an intrinsic part of the Icelandplume Geochemistry Geophysics Geosystems 4(3) 1032 doi1010292002GC000424
Frey F A Huang S Blichert-Toft J Regelous M amp Boyet M(2005) Origin of depleted components in basalt related to theHawaiian hotspot Evidence from isotopic and incompatible ele-ment ratios Geochemistry Geophysics Geosystems 6(1) 23 doi1010292004GC000757
Galer S J G amp Abouchami W (1998) Practical application of leadtriple spiking for correction of instrumental mass discriminationMineralogical Magazine 62A 491^492
Ghiorso M S amp Sack R O (1995) Chemical mass transfer in mag-matic processes IV A revised and internally consistent thermody-namic model for the interpolation and extrapolation of liquid^solidequilibria in magmatic systems at elevated temperatures and pres-sures Contributions to Mineralogy and Petrology 119 197^212
Greene A R Scoates J S amp Weis D (2005)WrangelliaTerrane onVancouver Island Distribution of flood basalts with implicationsfor potential Ni^Cu^PGE mineralization In Grant B (ed)Geological Fieldwork 2004 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2005-1 209^220Greene A R Scoates J S Nixon G T amp Weis D (2006) Picritic
lavas and basal sills in the Karmutsen flood basalt provinceWrangellia northern Vancouver Island In Grant B (ed)Geological Fieldwork 2005 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2006-1 39^52Greene A R Scoates J S amp Weis D (2008) Wrangellia flood
basalts in Alaska A record of plume^lithosphere interaction in aLate Triassic accreted oceanic plateau Geochemistry Geophysics
Geosystems 9(Q12004) doi1010292008GC002092Greene A R Scoates J S Weis D amp Israel S (2009)
Geochemistry of triassic flood basalts from the Yukon (Canada)segment of the accreted Wrangellia oceanic plateau Lithosdoi101016jlithos200811010
Hauff F Hoernle K Tilton G Graham D W amp Kerr A C(2000) Large volume recycling of oceanic lithosphere over shorttime scales geochemical constraints from the Caribbean LargeIgneous Province Earth and Planetary Science Letters 174 247^263
Hauff F Hoernle K amp Schmidt A (2003) Sr^Nd^Pb compositionof Mesozoic Pacific oceanic crust (Site 1149 and 801 ODP Leg185)Implications for alteration of ocean crust and the input into theIzu^Bonin^Mariana subduction system Geochemistry Geophysics
Geosystems 4(8) doi1010292002GC000421Herzberg C (2004) Partial melting below the Ontong Java Plateau
In Fitton J G Mahoney J J Wallace P J amp Saunders A D(eds) Origin and Evolution of the Ontong Java Plateau Geological SocietyLondon Special Publications 229 179^183
Herzberg C (2006) Petrology and thermal structure of the Hawaiianplume from Mauna Kea volcano Nature 444(7119) 605
Herzberg C amp Asimow P D (2008) Petrology of some oceanicisland basalts PRIMELT2XLS software for primary magma cal-culation Geochemistry Geophysics Geosystems 9(Q09001) doi1010292008GC002057
Herzberg C amp OrsquoHara M J (2002) Plume-associated ultramaficmagmas of Phanerozoic age Journal of Petrology 43(10) 1857^1883
Herzberg C Asimow P D Arndt N Niu Y Lesher C MFitton J G Cheadle M J amp Saunders A D (2007)Temperatures in ambient mantle and plumes Constraints frombasalts picrites and komatiites Geochemistry Geophysics Geosystems8(Q02006) doi1010292006GC001390
Huang S amp Frey F A (2005) Trace element abundances of MaunaKea basalt from phase 2 of the Hawaii Scientific Drilling Project
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
35
Petrogenetic implications of correlations with major element con-tent and isotopic ratios Geochemistry Geophysics Geosystems 4(6)doi1010292002GC000322
Hussong D M Wipperman L K amp Kroenke L W (1979) Thecrustal structure of the Ontong Java and Manihiki OceanicPlateaus Journal of Geophysical Research 84(B11) 6003^6010
Jerram D A amp Widdowson M (2005) The anatomy of ContinentalFlood Basalt Provinces geological constraints on the processes andproducts of flood volcanism Lithos 79(3^4) 385^405
Jones D L Silberling N J amp Hillhouse J (1977)Wrangellia a dis-placed terrane in northwestern North America CanadianJournal ofEarth Sciences 14(11) 2565^2577
Kerr A C (2003) Oceanic Plateaus In Rudnick R (ed) The Crust
Treatise on GeochemistryVol 3 Oxford Elsevier Science pp 537^565Kerr A C amp Mahoney J J (2007) Oceanic plateaus Problematic
plumes potential paradigms Chemical Geology 241 332^353Kerr A CTarney J Marriner G F Klaver GT Saunders A D
amp Thirlwall M F (1996) The geochemistry and petrogenesis ofthe Late-Cretaceous picrites and basalts of Curacao NetherlandsAntilles a remnant of an oceanic plateau Contributions to
Mineralogy and Petrology 124(1) 29^43Kerr A C Marriner G F Tarney J Nivia A Saunders A D
Thirlwall M F amp Sinton A C (1997) Cretaceous basaltic ter-ranes in Western Colombia Elemental chronological and Sr-Ndisotopic constraints on petrogenesis Journal of Petrology 38(6)677^702
Kinzler R J (1997) Melting of mantle peridotite at pressuresapproaching the spinel to garnet transition Application to mid-ocean ridge basalt petrogenesis Journal of Geophysical Research
102(B1) 853^874Kinzler R J amp Grove T L (1992) Primary magmas of mid-ocean
ridge basalts 1 Experiments and methods Journal of Geophysical
Research 97(B5) 6885^6906Korenaga J (2005) Why did not the Ontong Java Plateau form sub-
aerially Earth and Planetary Science Letters 234 385^399Kuniyoshi S (1972) Petrology of the Karmutsen (Volcanic) Group
northeastern Vancouver Island British Columbia PhD disserta-tion University of California Los Angeles 242 pp
Lassiter J C DePaolo D J amp Mahoney J J (1995) Geochemistry ofthe Wrangellia flood basalt province Implications for the role ofcontinental and oceanic lithosphere in flood basalt genesis Journalof Petrology 36(4) 983^1009
LeBas M J (2000) IUGS reclassification of the high-Mg and picriticvolcanic rocks Journal of Petrology 41(10) 1467^1470
Longhi J (2002) Some phase equilibrium systematics of lherzolitemelting I Geochemistry Geophysics Geosystems 3(3) doi1010292001GC000204
MacDonald G A amp Katsura T (1964) Chemical composition ofHawaiian lavas Journal of Petrology 5 82^133
Mahoney J J (1987) An isotopic survey of Pacific oceanic plateausimplications for their nature and origin In Keating B HFryer P Batiza R amp Boehlert GW (eds) Seamounts Islands andAtolls American Geophysical Union Geophysical Monograph 43 207^220
Mahoney J LeRoex A P Peng Z Fisher R L amp Natland J H(1992) Southwestern limits of Indian Ocean Ridge mantle and theorigin of low 206Pb204Pb mid-ocean ridge basalt isotope systema-tics of the central Southwest Indian Ridge (178^508E) Journal ofGeophysical Research 97(B13) 19771^19790
Mahoney J J Sinton J M Kurz M D MacDougall J DSpencer K J amp Lugmair GW (1994) Isotope and trace elementcharacteristics of a super-fast spreading ridge East Pacific rise13^238S Earth and Planetary Science Letters 121 173^193
Massey N W D (1995a) Geology and mineral resources of theAlberni^Nanaimo Lakes sheet Vancouver Island 92F1W 92F2Eand part of 92F7E British Columbia Ministry of Energy Mines and
Petroleum Resources Paper 1992-2 132 ppMassey N W D (1995b) Geology and mineral resources of the
Cowichan Lake sheet Vancouver Island 92C16 British Columbia
Ministry of Energy Mines and Petroleum Resources Paper 1992-3 112 ppMassey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005a) Digital Geology Map of British Columbia Tile NM9 MidCoast BC British Columbia Ministry of Energy and Mines Geofile
2005-2Massey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005b) Digital Geology Map of British Columbia Tile NM10Southwest BC British Columbia Ministry of Energy and Mines Geofile
2005-3McDonough W F amp Sun S (1995) The composition of the Earth
Chemical Geology 120 223^253Monger JW H amp Journeay J M (1994) Basement geology and tec-
tonic evolution of theVancouver region In Monger JW H (ed)Geology and Geological Hazards of the Vancouver Region Southwestern
British Columbia Geological Survey of Canada Bulletin 481 3^25Muller J E (1977) Geology of Vancouver Island Geological Survey of
Canada Open File Map 463Muller J E (1980)The Paleozoic Sicker Group of Vancouver Island British
Columbia Geological Survey of Canada Paper 79-30 22 ppMuller J E Northcote K E amp Carlisle D (1974) Geology and
mineral deposits of Alert Bay^Cape Scott map area VancouverIsland British Columbia Geological Survey of Canada Paper 74-8 77
Neal C R Mahoney J J Kroenke L W Duncan R A ampPetterson M G (1997) The Ontong^Java Plateau InMahoney J J amp Coffin M F (eds) Large Igneous Provinces
Continental Oceanic and Planetary FloodVolcanism American Geophysical
Union Geophysical Monograph 100 183^216Niu Y Collerson K D Batiza R Wendt J I amp Regelous M
(1999) Origin of enriched-typed mid-ocean ridge basalt at ridgesfar from mantle plumes The East Pacific Rise at 11820rsquoN Journalof Geophysical Research 104 7067^7088
Nixon G T amp Orr A J (2007) Recent revisions to the EarlyMesozoic stratigraphy of Northern Vancouver Island (NTS 102I092L) and metallogenic implicatinos British Columbia InGrant B (ed) Geological Fieldwork 2006 British Columbia Ministry of
Energy Mines and Petroleum Resources Paper 2007-1 163^177Nixon GT Hammack J L KoyanagiV M Payie G J Haggart
JW Orchard M JTozerT Friedman R M Archibald D APalfy J amp Cordey F (2006a) Geology of the Quatsino^PortMcNeill area northernVancouver Island British Columbia Ministry
of Energy Mines and Petroleum Resources Geoscience Map 2006-2 scale150 000
Nixon G T Kelman M C Stevenson D Stokes L A ampJohnston K A (2006b) Preliminary geology of the Nimpkishmap area (NTS 092L07) northern Vancouver Island BritishColumbia In Grant B amp Newell J M (eds) Geological Fieldwork2005 British Columbia Ministry of Energy Mines and Petroleum Resources
Paper 2006-1 135^152Nixon G T Laroque J Pals A Styan J Greene A R amp
Scoates J S (2008) High-Mg lavas in the Karmutsen flood basaltsnorthernVancouver Island (NTS 092L) Stratigraphic setting andmetallogenic significance In Grant B (ed) Geological Fieldwork
2007 British Columbia Ministry of Energy Mines and Petroleum
Resources Paper 2008-1 175^190Nowell G M Kempton P D Noble S R Fitton J G
Saunders A Mahoney J J amp Taylor R N (1998) High precisionHf isotope measurements of MORB and OIB by thermal ionisation
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
36
mass spectrometry insights into the depleted mantle Chemical
Geology 149 211^233Parrish R R amp McNicoll V J (1992) U^Pb age determinations
from the southern Vancouver Island area British Columbia InRadiogenic Age and Isotopic Studies Report 5 Geological Survey of
Canada Paper 91-2 79^86Peate DW (1997) The Parana^Etendeka Province In Coffin M F
amp Mahoney J (eds) Large Igneous Provinces Continental Oceanic andPlanetary Flood Volcanism American Geophysical Union Geophysical
Monograph 100 217^245Perfit M R Fornari D J Ridley W I Kirk P D Casey J
Kastens K A Reynolds J R Edwards M Desonie DShuster R amp Paradis S (1996) Recent volcanism in the Siqueirostransform fault picritic basalts and implications for MORBmagma genesis Earth and Planetary Science Letters 141(1^4) 91^108
Pretorius W Weis D Williams G Hanano D Kieffer B ampScoates J S (2006) Complete trace elemental characterization ofgranitoid (USGSG-2GSP-2) reference materials by high resolutioninductively coupled plasma-mass spectrometry Geostandards and
Geoanalytical Research 30(1) 39^54Regelous M Niu Y Wendt J I Batiza R Greig A amp
Collerson K D (1999) Variations in the geochemistry of magma-tism on the East Pacific Rise at 10830rsquoN since 800 ka Earth and
Planetary Science Letters 168 45^63Recurren villon S Chauvel C Arndt N T Pik R Martineau F
Fourcade S amp Marty B (2002) Heterogeneity of the Caribbeanplateau mantle source Sr O and He isotopic compositions of oli-vine and clinopyroxene from Gorgona Island Earth and Planetary
Science Letters 205(1^2) 91Rhodes J M (1996) Geochemical stratigraphy of lava flows samples
by the Hawaii Scientific Drilling Project Journal of Geophysical
Research 101(B5) 11729^11746Rhodes J M amp Vollinger M J (2004) Composition of basaltic lavas
sampled by phase-2 of the Hawaii Scientific Drilling ProjectGeochemical stratigraphy and magma types Geochemistry
Geophysics Geosystems 5(3) doi1010292002GC000434Richards M A Jones D L Duncan R A amp DePaolo D J (1991)
A mantle plume initiation model for the Wrangellia flood basaltand other oceanic plateaus Science 254 263^267
Salters V J M amp Hart S R (1991) The mantle sources of oceanridges islands and arcs the Hf-isotope connection Earth and
Planetary Science Letters 104 364^380Salters V J M amp Stracke A (2004) Composition of the depleted
SaltersV J amp WhiteW (1998) Hf isotope constraints on mantle evo-lution Chemical Geology 145 447^460
SanoT amp Yamashita S (2004) Experimental petrology of basementlavas from Ocean Drilling Program Leg 192 implications for dif-ferentiation processes in Ontong Java Plateau magmas InFitton J G Mahoney J JWallace P J amp Saunders A D (eds)Origin and Evolution of the Ontong Java Plateau Geological Society
London Special Publications 229 185^218Saunders A D (2005) Large igneous provinces Origin and environ-
mental consequences Elements 1 259^263Self S ThordarsonT amp Keszthely L (1997) Emplacement of conti-
nental flood basalt lava flows In Mahoney J J amp Coffin M F(eds) Large Igneous Provinces Continental Oceanic and Planetary Flood
Volcanism American Geophysical Union Geophysical Monograph 100381^410
Shaw D M (2000) Continuous (dynamic) melting theory revisitedCanadian Mineralogist 38(5) 1041^1063
Sluggett C L (2003) Uranium^lead age and geochemical constraintson Paleozoic and Early Mesozoic magmatism in WrangelliaTerrane Saltspring Island British Columbia BSc thesisUniversity of British ColumbiaVancouver 84 pp
Storey M Mahoney J J amp Saunders A D (1997) Cretaceousbasalts in Madagascar and the transition between plume and con-tinental lithospheric mantle sources In Mahoney J J ampCoffin M F (eds) Large Igneous Provinces Continental Oceanic and
Planetary Flood Volcanism Aamerican Geophysical Union Geophysical
Monograph 100 95^122Surdam R C (1967) Low-grade metamorphism of the Karmutsen
Group PhD dissertation University of California Los Angeles288 pp
Sutherland-Brown A Yorath C J Anderson R G amp Dom K(1986) Geological maps of southern Vancouver IslandLITHOPROBE I 92C10 11 14 16 92F1 2 7 8 Geological Survey ofCanada Open File 1272
Tejada M L G Mahoney J J Castillo P R Ingle S P Sheth HC amp Weis D (2004) Pin-pricking the elephant evidence on theorigin of the Ontong Java Plateau from Pb^Sr^Hf^Nd isotopiccharacteristics of ODP Leg 192 basalts In Fitton J GMahoney J J Wallace P J amp Saunders A D (eds) Origin and
Evolution of the Ontong Java Plateau Geological Society London Special
Publications 229 133^150Thompson P M E Kempton P D amp Kerr A C (2008) Evaluation
of the effects of alteration and leaching on Sm^Nd and Lu^Hf sys-tematics in submarine mafic rocks Lithos 104(1^4) 164^176
Thompson P M E Kempton P D White R V Kerr A CTarney J Saunders A D Fitton J G amp McBirney A (2003)Hf-Nd isotope constraints on the origin of the CretaceousCaribbean plateau and its relationship to the Galapagos plumeEarth and Planetary Science Letters 217(1^2) 59^75
Vervoort J D Patchett P Blichert-Toft J amp Albare de F (1999)Relationships between Lu^Hf and Sm^Nd isotopic systems in theglobal sedimentary system Earth and Planetary Science Letters 16879^99
Walter M J (1998) Melting of garnet peridotite and the origin ofkomatiite and depleted lithosphere Journal of Petrology 39(1) 29^60
Weis D Kieffer B Maerschalk C Barling J de Jong JWilliams G A Hanano D Mattielli N Scoates J SGoolaerts A Friedman R A amp Mahoney J B (2006) High-pre-cision isotopic characterization of USGS reference materials byTIMS and MC-ICP-MS Geochemistry Geophysics Geosystems
7(Q08006) doi1010292006GC001283Weis D Kieffer B Hanano D Silva I N Barling J
Pretorius W Maerschalk C amp Mattielli N (2007) Hf isotopecompositions of US Geological Survey reference materialsGeochemistry Geophysics Geosystems 8(Q06006) doi1010292006GC001473
Wheeler J O amp McFeely P (1991) Tectonic assemblage map of theCanadian Cordillera and adjacent part of the United States ofAmerica Geological Survey of Canada Map 1712A
WhiteW M Albare de F amp Tecurren louk P (2000) High-precision analy-sis of Pb isotope ratios by multi-collector ICP-MS Chemical Geology167 257^270
Yorath C J Sutherland Brown A amp Massey N W D (1999)LITHOPROBE southern Vancouver Island British Columbia Geological
Survey of Canada Bulletin 498 145 ppZou H (1998) Trace element fractionation during modal and
non-model dynamic melting and open-system melting Amathematical treatment Geochimica et Cosmochimica Acta 62(11)1937^1945
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
37
Zou H amp Reid M R (2001) Quantitative modeling of trace elementfractionation during incongruent dynamic melting Geochimica et
Cosmochimica Acta 65(1) 153^162
APPENDIX SAMPLEPREPARAT ION ANDANALYT ICAL METHODSOnly the freshest rocks in the field were sampled and onlythe least altered samples were selected for chemical andisotopic analysis based on thorough petrographic inspec-tion Fifty-nine of the 129 collected samples were crushed(400 g) into small pieces 52mm in diameter in aRocklabs hydraulic piston crusher between WC plates tominimize contamination The coarse-crush was mixedand 100 g was powdered in a planetary mill using agatejars and balls cleaned with quartz sand between samples
ActLabs analytical methodsThe major- and trace-element compositions of the whole-rock powders were determined at Activation LaboratoriesLtd (Actlabs) in Ancaster Ontario Analytical techniquesand detection limits are also available from Actlabs (httpwwwactlabscommethsub_code4erehtm) The analyticalmethod for each of the elements analyzed is indicated inTable 2 For the major elements a 02 g sample was mixedwith a lithium metaborate^lithium tetraborate mixtureand fused in a graphite crucible The molten mixture waspoured into a 5 HNO3 solution and shaken until dis-solved (30min) The samples were analyzed for majoroxides and selected trace elements by inductively coupledplasma optical emission spectrometry (ICP-OES) on acombination simultaneous^sequential Thermo Jarrell-AshEnviro II system Internal calibration was achieved usinga variety of international reference materials (eg W-2BIR-1 DNC-1) and independent control samplesAdditional trace elements were analyzed by both instru-mental neutron activation analysis (INAA) and induc-tively couple plasma mass spectrometry (ICP-MS) ForINAA15^25 g of sample was weighed into small polyeth-ylene vials and irradiated with a control international ref-erence material CANMET WMS-1 and NiCr flux wires ata thermal neutron flux of 71012 neutronscm2 per s in theMcMaster Nuclear Reactor Following a 7 day waitingperiod the samples were measured on an Ortec high-purity Ge detector linked to a Canberra Series 95 multi-channel analyzer Activities for each element weredecay- and weight-corrected and compared with a detectorcalibration developed from multiple international certifiedreference materials For ICP-MS 025 g of sample wasdigested in HF followed by HNO3^HClO4 heated andtaken to dryness The samples were brought back into solu-tion with HCl Samples were analyzed using a PerkinElmer Optima 3000 ICP In-lab standards or certified ref-erence materials (eg W-2 BIR-1 DNC-1) were used for
quality control A total of 15 blind duplicates were ana-lyzed from the same sample powders to assessreproducibility
University of Massachusetts analyticalmethodsFourteen sample duplicate powders taken from the same100 g sample aliquot (high-MgO basalts and picrites)were also analyzed at the Ronald B Gilmore X-RayFluorescence (XRF) Laboratory at the University ofMassachusetts Major elements were measured on a fusedLa-bearing lithium borate glass disc using a SiemensMRS-400 spectrometer Trace element concentrations(Rb Sr Ba Ce Nb Zr Y Pb Zn Ga Ni Cr V) weremeasured on a separate powder pellet using a PhilipsPW2400 sequential spectrometer using a Rh tube LOIand ferrous iron measurements were made as describedby Rhodes amp Vollinger (2004) Precision and accuracy esti-mates for the data have been described by Rhodes (1996)and Rhodes amp Vollinger (2004) Results for each sampleare the average of two separate analyses and are shown inElectronic Appendix 2 for a comparison of analyses ofduplicate powders using XRFand ICP-MS
PCIGR analytical methodsA subset of 19 samples was selected for high-precisiontrace-element analysis and Sr Nd Pb and Hf isotopicanalysis at the Pacific Centre for Isotopic andGeochemical Research (PCIGR) at the University ofBritish Columbia (UBC) Samples were selected from the59 samples analyzed for whole-rock chemistry at ActLabsbased on major- and trace-element chemistry alteration(low LOI and petrographic alteration index) and samplelocation Samples were prepared for trace-element analysisat the PCIGR by the technique described by Pretorius et al(2006) on unleached rock powders Sample powders(100mg) were weighed in 7ml screw-top Savillex bea-kers and dissolved in 1ml 14N HNO3 and 5ml 48 HFon a hotplate for 48 h at 1308C with periodic ultrasonica-tion Samples were dried and redissolved in 6ml 6N HClon a hotplate for 24 h and then dried and redissolved in1ml concentrated HNO3 for 24 h before final dryingTrace element abundances were measured by high-resolution inductively coupled plasma-mass spectrometry(HR-ICP-MS) with aThermo Finnigan Element2 systemfollowing the procedures described by Pretorius et al(2006) within 24 h of redissolution HFSE and LILE weremeasured in medium-resolution mode at 2000 dilutionusing a PFATeflon spray chamber washed with aqua regiafor 3min between samples REE were measured in high-resolution mode and U and Pb in low-resolution modeat 2000 dilution using a glass spray chamber washedwith 2 HNO3 between samples Total procedural blanksand reference materials (BCR-2 BHVO-2) were analyzedwith the batch of samples Indium was used as an internal
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
38
standard in all samples and standard solutionsBackground and standard solutions were analyzed afterevery five samples to detect memory effects and massdrift Results for PCIGR trace-element analyses areshown in Electronic Appendix 3 for the full set of analyzedtrace elementsSample digestion for purification of Sr Nd Hf and Pb
for column chemistry began by weighing each samplepowder (400^500mg) prior to leaching All samples wereleached with 6N HCl and placed in an ultrasonic bath for15min Samples were rinsed twice with 18 M-cm H2Obetween each leaching step (15 total) until the supernatantliquid was clear [following the method of Mahoney (1987)]Samples were then dried on a hotplate for 24 h andweighed again Sample solutions were then prepared bydissolving 100^250mg of the leached powder dissolvedin 1ml 14N HNO3 and 10ml 48 HF on a hotplate for48 h at 1308C with periodic ultrasonication Samples weredried and redissolved in 6ml 6N HCl on a hotplate for24 h and then dried Pb was separated using anion exhangecolumns and the discard was used for Sr REE and Hfseparation Nd was separated from the other REE and Hfrequired two additional purification steps Detailed proce-dures for column chemistry for separating Sr Nd and Pbat the PCIGR have been described byWeis et al (2006) andHf purification has been described byWeis et al (2007) Srand Nd isotope ratios were measured by thermal ioniza-tion mass spectrometry (TIMS) on a Thermo FinniganTriton system in static mode with relay matrix rotation ona single Ta and double Re^Ta filament respectively Fourto five filaments per barrel of 21 were occupied by stan-dards (NBS 987 for Sr and LaJolla for Nd) for each barrelwhere samples were run Sample Sr and Nd isotopic com-positions were corrected for mass fractionation using86Sr88Srfrac14 01194 and 146Nd144Ndfrac14 07219 Each samplewas then normalized using the barrel average of the refer-ence material relative to the values of143Nd144Ndfrac14 0511858 and 87Sr86Srfrac14 0710248 (Weiset al 2006) During the course of the Vancouver Islandanalyses the La Jolla Nd standard gave an average valueof 05118566 (nfrac14 7) and NBS987 standard gave an aver-age of 07102408 (nfrac1411) (2 error is reported as times106) 147Sm144Nd errors are 15 or 0006 USGeological Survey (USGS) reference material BHVO-2was processed with the samples and yielded Sr and Nd iso-topic ratios of 07034607 and 05129786 respectivelyThese are in agreement with the published values of070347920 and 051298411 respectively (Weis et al2006)
Pb and Hf isotopic compositions were analyzed by mul-tiple collector-inductively coupled plasma-mass spectrome-try (MC-ICP-MS) on a static multi-collection on a NuPlasma (Nu Instruments) system The detailed analyticalprocedure for Pb isotopic analyses on the Nu Plasma atthe PCIGR has been described by Weis et al (2006) Theconfiguration for Pb analyses allows for collection of PbTl and Hg togetherTl and Hg are used to monitor instru-mental mass discrimination and isobaric overlap repec-tively All sample solutions were analyzed withapproximately the same PbTl ratio (4) as the referencematerial NIST SRM 981 To accomplish this a small ali-quot of each sample solution from the Pb columns was ana-lyzed on the Element2 to determine the precise amount ofPb available for analysis on the Nu Plasma The SRM 981standard was run after every two samples on the NuPlasma During the time samples were run analyses ofthe SRM 981 Pb reference material gave values of206Pb204Pbfrac1416940322 207Pb204Pbfrac1415495823and 208Pb204Pbfrac14 36713164 (nfrac14 61 2 error is reportedas times 104) these values are in excellent agreement withreported TIMS triple-spike values of Galer amp Abouchami(1998) Fractionation-corrected Pb isotopic ratios were fur-ther corrected by the sample-standard bracketing methodor the ln^ln correction method described by White et al(2000) and Blichert-Toft et al (2003) Leached powders ofUSGS reference material BHVO-2 yielded Pb isotopicratios of 206Pb204Pbfrac14186454 8 207Pb204Pbfrac14 1549105 and 208Pb204Pbfrac14 38222514 and BCR-2yielded 206Pb204Pbfrac141880466 207Pb204Pbfrac14 1562518 and 208Pb204Pbfrac14 3883496 (2 error isreported as times 104) These values are in agreement withthose for leached residues of BHVO-2 and BCR-2 fromWeis et al (2006)Hf isotopic compositions were analyzed following the
procedures detailed byWeis et al (2007) The configurationfor Hf analyses monitored Lu mass 175 andYb mass 172 toallow for interference correction to masses 174 and 176 Hfisotopic ratios were normalized internally for mass fractio-nation to a 179Hf177Hf ratio of 07325 using an exponentialcorrection Standards were run after every two samplesand sample results were normalized to the ratio of the in-run daily average and a 176Hf177Hf ratio for JMC-475 of0282160 During the course of analyses the Hf standardJMC-475 gave an average value 02821679 (nfrac14130)USGS reference materials BCR-2 and BHVO-2 were pro-cessed with the samples and yielded Hf isotopic ratios of02828675 and 02831005 respectively Publishedvalues for BCR-2 and BHVO-2 are 02828717 and0283104 8 respectively (Weis et al 2007)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
39
OLIV INE ACCUMULAT ION INH IGH-MgO AND PICR IT IC LAVASGeochemical trends and petrographic characteristics indi-cate that accumulation of olivine played an important rolein the formation of the high-MgO basalts and picrites fromthe Keogh Lake area on northern Vancouver Island TheKeogh Lake samples show a strong linear correlation inplots of Al2O3 TiO2 ScYb and Ni vs MgO and many ofthe picrites have abundant clusters of olivine pseudomorphs(Table1 Fig11)There is a strong linear correlationbetween
modal per cent olivine and whole-rock magnesium contentmagnesium contents range between 10 and 19wt MgOand the proportion of olivine phenocrysts in these samplesvaries from 0 to 42 vol (Fig 11)With a few exceptionsboth the size range and median size of the olivine pheno-crysts in most samples are comparable The clear correla-tion between the proportion of olivine phenocrysts andwhole-rock magnesium contents indicates that accumula-tion of olivine was directly responsible for the high MgOcontents (410wt ) in most of the picritic lavas
PicriteHigh-MgO basalt
Tholeiitic basaltCoarse-grained
Pillowed flow
207Pb204Pb
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
208Pb204Pb (230 Ma)
206Pb204Pb (230 Ma)
(a) (b)
(c) (e)
4723A24723A2
4723A2
4723A2
4722A4
4723A4
4722A4
4723A4
4723A134723A3
4723A3
4723A13
1556
1558
1560
1562
1564
1566
1568
186 190 194 198 202 206 2101550
1552
1554
1556
1558
1560
178 180 182 184 186 188 190 192 194
380
384
388
392
396
186 190 194 198 202 206 210374
378
382
386
390
178 180 182 184 186 188 190 192 194
206Pb204Pb(d)
LOI (wt )
0
1
2
3
4
5
6
7
186 190 194 198 202 206 210
4723A2
Fig 10 Pb isotopic compositions of leached whole-rock samples measured by MC-ICP-MS for the Karmutsen Formation Error bars are smal-ler than symbols (a) Measured 207Pb204Pb vs 206Pb204Pb (b) Initial 207Pb204Pb vs 206Pb204Pb Age-correction to 230 Ma (c) Measured208Pb204Pb vs 206Pb204Pb (d) LOI vs 206Pb204Pb (e) Initial 208Pb204Pb vs 206Pb204Pb Complete chemical duplicates shown in Table 5(samples 4720A7 and 4722A4) are circled in each plot The dashed lines in (b) and (e) show the differences in age-corrections for two picrites(4723A4 4722A4) using the measured UTh and Pb concentrations for each sample and the age-corrections when concentrations are used fromthe two picrites (4723A3 4723A13) that appear to be least affected by alteration
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
23
DISCUSS IONThe compositional and spatial development of the eruptedlava sequences of the Wrangellia oceanic plateau onVancouver Island provide information about the evolutionof a mantle plume upwelling beneath oceanic lithospherein an analogous way to the interpretation of continentalflood basalts The Caribbean and Ontong Java plateausare the two primary examples of accreted parts of oceanicplateaus that have been studied to date and comparisonscan be made with the Wrangellia oceanic plateauHowever the Wrangellia oceanic plateau is a distinctiveoccurrence of an oceanic plateau because it formed on thecrust of a Devonian to Mississippian intra-oceanic islandarc overlain by Mississippian to Permian limestones andpelagic sediments The nature and thickness of oceaniclithospheric mantle are considered to play a fundamentalrole in the decompression and evolution of a mantleplume and thus the compositions of the erupted lavasequences (Greene et al 2008) The geochemical and
stratigraphic relationships observed in the KarmutsenFormation onVancouver Island and the focus of the subse-quent discussion sections provide constraints on (1) thetemperature and degree of melting required to generatethe primary picritic magmas (2) the composition of themantle source (3) the depth of melting and residual min-eralogy in the source region (4) the low-pressure evolutionof the Karmutsen tholeiitic basalts (5) the growth of theWrangellia oceanic plateau
Melting conditions and major-elementcomposition of the primary magmasThe primary magmas to most flood basalt provinces arebelieved to be picritic (eg Cox 1980) and they have beenfound in many continental and oceanic flood basaltsequences worldwide (eg Siberia Karoo Paranacurren ^Etendeka Caribbean Deccan Saunders 2005) Near-pri-mary picritic lavas with low total alkali abundances
Table 5 Pb isotopic compositions of Karmutsen basaltsVancouver Island BC
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
require high-degree partial melting of the mantle source(eg Herzberg amp OrsquoHara 2002) The Keogh Lake picritesfrom northernVancouver Island are the best candidates foridentifying the least modified partial melts of the mantleplume source despite having accumulated olivine pheno-crysts and can be used to estimate conditions of meltingand the composition of the primary magmas for theKarmutsen FormationPrimary magma compositions mantle potential tem-
peratures and source melt fractions for the picritic lavas ofthe Karmutsen Formation were calculated from primitivewhole-rock compositions using PRIMELT1XLS software(Herzberg et al 2007) and are presented in Fig 12 andTable 6 A detailed discussion of the computationalmethod has been given by Herzberg amp OrsquoHara (2002)and Herzberg et al (2007) key features of the approachare outlined belowPrimary magma composition is calculated for a primi-
tive lava by incremental addition of olivine and compari-son with primary melts of mantle peridotite Lavas thathad experienced plagioclase andor clinopyroxene fractio-nation are excluded from this analysis All calculated pri-mary magma compositions are assumed to be derived byaccumulated fractional melting of fertile peridotite KR-4003 Uncertainties in fertile peridotite composition donot propagate to significant variations in melt fractionand mantle potential temperature melting of depletedperidotite propagates to calculated melt fractions that aretoo high but with a negligible error in mantle potential
temperature PRIMELT1XLS calculates the olivine liqui-dus temperature at 1atm which should be similar to theactual eruption temperature of the primary magmaand is accurate to 308C at the 2 level of confidenceMantle potential temperature is model dependent witha precision that is similar to the accuracy of the olivineliquidus temperature (C Herzberg personal communica-tion 2008)With regard to the evidence for accumulated olivine in
the Keogh Lake picrites it should not matter whether themodeled lava composition is aphyric or laden with accu-mulated olivine phenocrysts PRIMELT1 computes theprimary magma composition by adding olivine to a lavathat has experienced olivine fractionation in this way itinverts the effects of fractional crystallization The onlyrequirement is that the olivine phenocrysts are not acci-dental or exotic to the lava compositions Support for thisprocedure can be seen in the calculated olivine liquid-line-of-descent shown by the curved arrays with 5 olivineaddition increments (Fig 12c) Fractional crystallization ofolivine from model primary magmas having 85^95wt FeO and 153^175wt MgO will yield arrays of deriva-tive liquids and randomly sorted olivines that in most butnot all cases connect the picrites and high-MgO basalts Anewer version of the PRIMELT software (Herzberg ampAsimow 2008) can perform olivine addition to and sub-traction from lavas with highly variably olivine phenocrystcontents and yields results that converge with those shownin Fig 12
Fig 11 Relationship between abundance of olivine phenocrysts and whole-rock MgO contents for the Keogh Lake picrites (a) Tracings ofolivine grains (gray) in six high-MgO samples made from high resolution (2400 dpi) scans of petrographic thin-sections Scale bars represent10mm sample numbers are indicated at the upper left (b) Modal olivine vs whole-rock MgO Continuous line is the best-fit line for 10 high-MgO samplesThe areal proportion of olivine was calculated from raster images of olivine grains using ImageJ image analysis software whichprovides an acceptable estimation of the modal abundance (eg Chayes 1954)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
25
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
Gray lines = final melting pressure
L + Ol + Opx
07
08
09
Pressure (GPa)
10
3
4
5
6
1
SOLIDUS
01
04
0506
L + Ol
2
03
02
PeridotiteKR-4003
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
Thick black lines = initial melting pressure
Dashed lines = melt fraction
Melt Fraction
2
3 4 5 6
L+Ol+Opx
SolidusSpinel
SolidusGarnet Peridotite
7
L+Ol
0 10 20 30 4040
45
50
55
60
MgO (wt)
SiO2
(wt)
PeridotiteKR-4003
Melt Fraction
0 10 20 305
10
15
CaO(wt)
MgO (wt)
3
4 5 6 7
2
46
2
Peridotite Partial Melts
Thick black line = initial melting pressureGray lines = final melting pressure
Peridotite
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
02
Pressure (GPa)
0302
04
Pressure (GPa)
Picrite
High-MgO basalt
Tholeiitic basalt
(c)
(d)
Karmutsen flood basaltsOlivine addition modelOlivine addition (5 increment)Primary magma
PicriteHigh-MgO basaltTholeiitic basalt
(a)
(b)
800 850 900
910
920
930
940
950
Karmutsen flood basalts
Olivine addition modelOlivine addition (5 increment)Primary magma
Gray lines = Fo contents of olivine
Fig 12 Estimated primary magma compositions for three Keogh Lake picrites and high-MgO basalts (samples 4723A4 4723A135616A7) using the forward and inverse modeling technique of Herzberg et al (2007) Karmutsen compositions and modeling results areoverlain on diagrams provided by C Herzberg (a) Whole-rock FeO vs MgO for Karmutsen samples from this study Total iron estimatedto be FeO is 090 Gray lines show olivine compositions that would precipitate from liquid of a given MgO^FeO composition Black lineswith crosses show results of olivine addition (inverse model) using PRIMELT1 (Herzberg et al 2007) Gray field highlights olivinecompositions with Fo contents of 91^92 predicted to precipitate (b) FeO vs MgO showing Karmutsen lava compositions and results ofa forward model for accumulated fractional melting of fertile peridotite KR-4003 (Kettle River Peridotite Walter 1998) (c) SiO2 vsMgO for Karmtusen lavas and model results (d) CaO vs MgO showing Karmtusen lava compositions and model results To brieflysummarize the technique [see Herzberg et al (2007) for complete description] potential parental magma compositions for the high-MgOlava series were selected (highest MgO and appropriate CaO) and using PRIMELT1 software olivine was incrementally added tothe selected compositions to show an array of potential primary magma compositions (inverse model) Then using PRIMELT1 theresults from the inverse model were compared with a range of accumulated fractional melts for fertile peridotite derived fromparameterization of the experimental results for KR-4003 from Walter (1998) forward model Herzberg amp OrsquoHara (2002) A melt fractionwas sought that was unique to both the inverse and forward models (Herzberg et al 2007) A unique solution was found when there was a com-mon melt fraction for both models in FeO^MgO and CaO^MgO^Al2O3^SiO2 (CMAS) projection space This modeling assumes thatolivine was the only phase crystallizing and ignores chromite precipitation and possible augite fractionation in the mantle (Herzbergamp OrsquoHara 2002) Results are best for a residue of spinel lherzolite (not pyroxenite) The presence of accumulated olivine in samples ofthe Keogh Lake picrites used as starting compositions does not significantly affect the results because we are modeling addition of olivine(see text for further description) The tholeiitic basalts cannot be used for modeling because they are all saturated in plagthorn cpxthornolThick black line for initial melting pressure at 4 GPa is not shown in (c) for clarity Gray area in (a) indicates parental magmas thatwill crystallize olivine with Fo 91^92 at the surface Dashed lines in (c) indicate melt fraction Solidi for spinel and garnet peridotite arelabelled in (c)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
The PRIMELT1 results indicate that if the Karmutsenpicritic lavas were derived from accumulated fractionalmelting of fertile peridotite the primary magmas wouldhave contained 15^17wt MgO and 10wt CaO(Fig 12 Table 6) formed from 23^27 partial meltingand would have crystallized olivine with a Fo content of 91 (Fig 12 Table 6) Calculated mantle temperatures forthe Karmutsen picrites (14908C) indicate melting ofanomalously hot mantle (100^2008C above ambient) com-pared with ambient mantle that produces mid-ocean ridgebasalt (MORB 1280^14008C Herzberg et al 2007)which initiated between 3 and 4 GPa Several picritesappear to be near-primary melts and required very littleaddition of olivine (eg sample 93G171 with measured158wt MgO) These temperature and melting esti-mates of the Karmutsen picrites are consistent withdecompression of hot mantle peridotite in an actively con-vecting plume head (ie plume initiation model) Threesamples (picrite 5614A1 and high-MgO basalts 5614A3and 5614A5) are lower in iron than the other Karmutsenlavas with FeO in the 65^80wt range (Fig 12c) andhave MgO and FeO contents that are similar to primitiveMORB from the Sequeiros fracture zone of the EastPacific Rise (EPR Perfit et al 1996) These samples
however differ from EPR MORB in having higher SiO2
and lower Al2O3 and they are restricted geographicallyto one area (Sara Lake Fig 2)We therefore have evidence for primary magmas with
MgO contents that range from a possible low of 110wt to as high as 174wt with inferred mantle potential tem-peratures from 1340 to 15208C This is similar to the rangedisplayed by lavas from the Ontong Java Plateau deter-mined by Herzberg (2004) who found that primarymagma compositions assuming a peridotite sourcewould contain 17wt MgO and 10wt CaO(Table 6) and that these magmas would have formed by27 melting at a mantle potential temperature of15008C (first melting occurring at 36 GPa) Fitton ampGodard (2004) estimated 30 partial melting for theOntong Java lavas based on the Zr contents of primarymagmas of Kroenke-type basalt calculated by incrementaladdition of equilibrium olivine However if a considerableamount of eclogite was involved in the formation of theOntong Java plateau magmas the excess temperatures esti-mated by Herzberg et al (2007) may not be required(Korenaga 2005) The variability in mantle potential tem-perature for the Keogh Lake picrites is also similar to thewide range of temperatures that have been calculated for
Table 6 Estimated primary magma compositions for Karmutsen basalts and other oceanic plateaus or islands
Sample 4723A4 4723A13 5616A7 93G171 Average OJP Mauna Keay Gorgonaz
Ontong Java primary magma composition for accumulated fraction melting (AFM) from Herzberg (2004)yMauna Kea primary magma composition is average of four samples of Herzberg (2006 table 1)zGorgona primary magma composition for AFM for 1F fertile source from Herzberg amp OrsquoHara (2002 table 4)Total iron estimated to be FeO is 090
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
27
lavas from single volcanoes in the Galapagos Hawaii andelsewhere by Herzberg amp Asimow (2008) Those workersinterpreted the range to reflect the transport of magmasfrom both the cool periphery and the hotter axis of aplume A thermally heterogeneous plume will yield pri-mary magmas with different MgO contents and the rangeof primary magma compositions and their inferred mantlepotential temperatures for Karmutsen lavas may reflectmelting from different parts of the plume
Source of the Karmutsen Formation lavasThe Sr^Nd^Hf^Pb isotopic compositions of theKarmutsen volcanic rocks on Vancouver Island indicatethat they formed from plume-type mantle containing adepleted component that is compositionally similar to butdistinct from other oceanic plateaus that formed in thePacific Ocean (eg Ontong Java and Caribbean Fig 13)Trace element and isotopic constraints can be used to testwhether the depleted component of the KarmutsenFormation was intrinsic to the mantle plume and distinctfrom the source of MORB or the result of mixing or invol-vement of ambient depleted MORB mantle with plume-type mantle The Karmutsen tholeiitic basalts have initialeNd and
87Sr86Sr ratios that fall within the range of basaltsfrom the Caribbean Plateau (eg Kerr et al 1996 1997Hauff et al 2000 Kerr 2003) and a range of Hawaiian iso-tope data (eg Frey et al 2005) and lie within and abovethe range for the Ontong Java Plateau (Tejada et al 2004Fig 13a) Karmutsen tholeiitic basalts with the highest ini-tial eNd and lowest initial 87Sr86Sr lie within and justbelow the field for age-corrected northern EPR MORB at230 Ma (eg Niu et al1999 Regelous et al1999) Initial Hfand Nd isotopic compositions for the Karmutsen samplesare noticeably offset to the low side of the ocean islandbasalt (OIB) array (Vervoort et al 1999) and lie belowand slightly overlap the low eHf end of fields for Hawaiiand the Caribbean Plateau (Fig 13c) the Karmutsen sam-ples mostly fall between and slightly overlap a field forage-corrected Pacific MORB at 230 Ma and Ontong Java(Mahoney et al 1992 1994 Nowell et al 1998 Chauvel ampBlichert-Toft 2001) Karmutsen lava Pb isotope ratios over-lap the broad range for the Caribbean Plateau and thelinear trends for the Karmutsen samples intersect thefields for EPR MORB Hawaii and Ontong Java in208Pb^206Pb space but do not intersect these fields in207Pb^206Pb space (Fig 13d and e) The more radiogenic207Pb204Pb for a given 206Pb204Pb of the Karmutsenbasalts requires high UPb ratios at an ancient time when235U was abundant similar to the Caribbean Plateau andenriched in comparison with EPR MORB Hawaii andOntong JavaThe low eHf^low eNd component of the Karmutsen
basalts that places them below the OIB mantle array andpartly within the field for age-corrected Pacific MORBcan form from low ratios of LuHf and high SmNd over
time (eg Salters amp White 1998) MORB will develop aHf^Nd isotopic signature over time that falls below themantle array Low LuHf can also lead to low 176Hf177Hffrom remelting of residues produced by melting in theabsence of garnet (eg Salters amp Hart 1991 Salters ampWhite 1998) The Hf^Nd isotopic compositions of theKarmutsen Formation indicate the possible involvementof depleted MORB mantle however similar to Iceland(Fitton et al 2003) Nb^Zr^Y variations provide a way todistinguish between a depleted component within theplume and a MORB-like component The Karmutsenbasalts and picrites fall within the Iceland array and donot overlap the MORB field in a logarithmic plot of NbYvs ZrY (Fig 13b) using the fields established by Fittonet al (1997) Similar to the depleted component within theCaribbean Plateau (Thompson et al 2003) the trend ofHf^Nd isotopic compositions towards Pacific MORB-likecompositions is not followed by MORB values in Nb^Zr^Y systematics and this suggests that the depleted compo-nent in the source of the Karmutsen basalts is distinctfrom the source of MORB and probably an intrinsic partof the plume The relatively radiogenic Pb isotopic ratiosand REE patterns distinct from MORB also support thisinterpretationTrace-element characteristics suggest the possible invol-
vement of sub-arc lithospheric mantle in the generation ofthe Karmutsen picrites Lassiter et al (1995) suggested thatmixing of a plume-type source with eNdthorn 6 to thorn 7 witharc material with low NbTh could reproduce the varia-tions observed in the Karmutsen basalts however theyconcluded that the absence of low NbLa ratios in theirsample of tholeiitic basalts restricts the amount of arc litho-sphere involved The study of Lassiter et al (1995) wasbased on major and trace element and Sr Nd and Pb iso-topic data for a suite of 29 samples from Buttle Lake in theStrathcona Provincial Park on central Vancouver Island(Fig 1) Whereas there are no clear HFSE depletions inthe Karmutsen tholeiitic basalts the low NbLa (and TaLa) of the Karmutsen picritic lavas in this study indicatethe possible involvement of Paleozoic arc lithosphere onVancouver Island (Fig 8)
REE modeling dynamic melting andsource mineralogyThe combined isotopic and trace-element variations of thepicritic and tholeiitic Karmutsen lavas indicate that differ-ences in trace elements may have originated from differentmelting histories For example the LuHf ratios of theKarmutsen picritic lavas (mean 008) are considerablyhigher than those in the tholeiitic basalts (mean 002)however the small range of overlapping initial eHf valuesfor the two lava suites suggests that these differences devel-oped during melting within the plume (Fig 9b) Below wemodel the effects of source mineralogy and depth of
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
28
melting on differences in trace-element concentrations inthe tholeiitic basalts and picritesDynamic melting models simulate progressive decom-
pression melting where some of the melt fraction isretained in the residue and only when the degree of partialmelting is greater than the critical mass porosity and the
source becomes permeable is excess melt extracted fromthe residue (Zou1998) For the Karmutsen lavas evolutionof trace element concentrations was simulated using theincongruent dynamic melting model developed by Zou ampReid (2001) Melting of the mantle at pressures greater than1 GPa involves incongruent melting (eg Longhi 2002)
OIB array
PacificMORB(230 Ma)
(0 Ma)
EPR(230 Ma)
-4
-2
0
2
4
8
10
12
14
16
18
20
22
-4 -2 0 2 4 6 8 10 12 14
minus2
0
2
4
6
8
10
12
14
0702 0703 0704 0705 0706
IndianMORB
87Sr 86Sr (initial)
Hf (initial)
(a) (c)
Nd (initial)
Karmutsen tholeiitic basalt
HawaiiOntong JavaCaribbean Plateau
Karmutsen high-MgO lava
high-MgO lavasKarmutsen
1540
1545
1550
1555
1560
1565
175 180 185 190 195 200375
380
385
390
395
175 180 185 190 195 200
(d) (e)
EPR
EPR
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb
Karmutsen Formation (initial)
HawaiiOntong JavaCaribbean Plateau
East Pacific Rise
Karmutsen Formation (measured)
Juan de FucaGorda
ε Nd
(initial)
Karmutsen Formation (different age-correction for 3 picrites)
(0 Ma)
NbY
001
01
1
1 10
ZrY
OIB
NMORB
(b)
Iceland array
6
Fig 13 Comparison of age-corrected (230 Ma) Sr^Nd^Hf^Pb isotopic compositions for Karmutsen flood basalts onVancouver Island withage-corrected OIB and MORB and Nb^Zr^Y variation (a) Initial eNd vs 87Sr86Sr (b) NbYand ZrY variation (after Fitton et al 1997)(c) Initial eHf vs eNd (d) Measured and initial 207Pb204Pb vs 206Pb204Pb (e) Measured and initial 208Pb204Pb vs 206Pb204Pb References fordata sources are listed in Electronic Appendix 4 Most of the compiled data were extracted from the GEOROC database (httpgeorocmpch-mainzgwdgdegeoroc) OIB array line in (c) is fromVervoort et al (1999) EPR East Pacific Rise Several high-MgO samples affected bysecondary alteration were not plotted for clarity in Pb isotope plots Estimation of fields for age-corrected Pacific MORB and EPR at 230 Mawere made using parent^daughter concentrations (in ppm) from Salters amp Stracke (2004) of Lufrac14 0063 Hffrac14 0202 Smfrac14 027 Ndfrac14 071Rbfrac14 0086 and Srfrac14 836 In the NbYvs ZrYplot the normal (N)-MORB and OIB fields not including Icelandic OIB are from data com-pilations of Fitton et al (1997 2003) The OIB field is data from 780 basalts (MgO45wt ) from major ocean islands given by Fitton et al(2003) The N-MORB field is based on data from the Reykjanes Ridge EPR and Southwest Indian Ridge from Fitton et al (1997) The twoparallel gray lines in (b) (Iceland array) are the limits of data for Icelandic rocks
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
29
with olivine being produced during the reactioncpxthornopxthorn spfrac14meltthornol for spinel peridotites The mod-eling used coefficients for incongruent melting reactionsbased on experiments on lherzolite melting (eg Kinzleramp Grove 1992 Longhi 2002) and was separated into (1)melting of spinel lherzolite (2) two combined intervals ofmelting for a garnet lherzolite (gt lherz) and spinel lher-zolite (sp lherz) source (3) melting of garnet lherzoliteThe model solutions are non-unique and a range indegree of melting partition coefficients melt reactioncoefficients and proportion of mixed source componentsis possible to achieve acceptable solutions (see Fig 14 cap-tion for description of modeling parameters anduncertainties)The LREE-depleted high-MgO lavas require melting of
a depleted spinel lherzolite source (Fig 14a) The modelingresults indicate a high degree of melting (22^25) similarto the results from PRIMELT1 based on major-elementvariations (discussed above) of a LREE-depleted sourceThe high-MgO lavas lack a residual garnet signature Theenriched tholeiitic basalts involved melting of both garnetand spinel lherzolite and represent aggregate melts pro-duced from continuous melting throughout the depth ofthe melting column (Fig 14b) Trace-element concentra-tions in the tholeiitic and picritic lavas are not consistentwith low-degree melting of garnet lherzolite alone(Fig 14c) The modeling results indicate that the aggregatemelts that formed the tholeiitic basalts would haveinvolved lesser proportions of enriched small-degreemelts (1^6 melting) generated at depth and greateramounts of depleted high-degree melts (12^25) gener-ated at lower pressure (a ratio of garnet lherzolite tospinel lherzolite melt of 14 provides the best fits seeFig 14b and description in figure caption) The totaldegree of melting for the tholeiitic basalts was high(23^27) similar to the degree of partial melting in thespinel lherzolite facies for the primary melts that eruptedas the Keogh Lake picrites of a source that was also prob-ably depleted in LREE
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
(c)
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
Melting of garnet lherzolite
High-MgO lavas
Melting of spinel lherzolite
source (07DM+03PM)
source (07DM+03PM)
6 melting
30 melting
(b)
(a)
30 melting
Sam
ple
Cho
ndrit
eS
ampl
eC
hond
rite 20 melting
1 melting
Melting of garnet lherzoliteand spinel lherzolite
x gt lherz + 4x sp lherz
5 meltingx gt lherz + 4x sp lherz
Sam
ple
Cho
ndrit
e
Tholeiitic lavas
Tholeiitic lavas
High-MgO lavas
source (07DM+03PM)
Fig 14 Trace-element modeling results for incongruent dynamicmantle melting for picritic and tholeiitic lavas from the KarmutsenFormation Three steps of modeling are shown (a) melting of spinellherzolite compared with high-MgO lavas (b) melting of garnet lher-zolite and spinel lherzolite compared with tholeiitic lavas (c) meltingof garnet lherzolite compared with tholeiitic and picritic lavasShaded fields in each panel for tholeiitic basalts and picrites arelabeled Patterns with symbols in each panel are modeling results(patterns are 5 melting increments in (a) and (b) and 1 incre-ments in (c) PM primitive mantle DM depleted MORB mantle gtlherz garnet lherzolite sp lherz spinel lherzolite In (b) the ratios ofper cent melting for garnet and spinel lherzolite are indicated (x gtlherzthorn 4x sp lherz means that for 5 melting 1 melt in garnetfacies and 4 melt in spinel facies where x is per cent melting in theinterval of modeling) The ratio of the respective melt proportions inthe aggregate melt (x gt lherzthorn 4x sp lherz) was kept constant andthis ratio of melt contribution provided the best fit to the data Arange of proportion of source components (07DMthorn 03PM to
09DMthorn 01PM) can reproduce the variation in the high-MgOlavas Melting modeling uses the formulation of Zou amp Reid (2001)an example calculation is shown in their Appendix PM fromMcDonough amp Sun (1995) and DM from Salters amp Stracke (2004)Melt reaction coefficients of spinel lherzolite from Kinzler amp Grove(1992) and garnet lherzolite fromWalter (1998) Partition coefficientsfrom Salters amp Stracke (2004) and Shaw (2000) were kept constantSource mineralogy for spinel lherzolite (018cpx027opx052ol003sp) from Kinzler (1997) and for garnet lherzolite(034cpx008opx053ol005gt) from Salters amp Stracke (2004) Arange of source compositions for melting of spinel lherzolite(07DMthorn 03PM to 03DMthorn 07PM) reproduce REE patternssimilar to those of the high-MgO lavas with best fits for 22ndash25melting and 07DMthorn 03PM source composition Concentrationsof tholeiitic basalts cannot be reproduced from an entirelyprimitive or depleted source
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
30
The uncertainty in the composition of the source in thespinel lherzolite melting model for the high-MgO lavasprecludes determining whether the source was moredepleted in incompatible trace elements than the source ofthe tholeiitic lavas (see Fig 14 caption for description) It ispossible that early high-pressure low-degree melting left aresidue within parts of the plume (possibly the hotter inte-rior portion) that was depleted in trace elements (egElliott et al 1991) further decompression and high-degreemelting of such depleted regions could then have generatedthe depleted high-MgO Karmutsen lavas Shallow high-degree melting would preferentially sample depletedregions whereas deeper low-degree melting may notsample more refractory depleted regions (eg CaribbeanEscuder-Viruete et al 2007) The picrites lie at the highend of the range of Nd isotopic compositions and havelow NbZr similar to depleted Icelandic basalts (Fittonet al 2003) therefore the picrites may represent the partsof the mantle plume that were the most depleted prior tothe initiation of melting As Fitton et al (2003) suggestedthese depleted melts may only rarely be erupted withoutcombining with more voluminous less depleted melts andthey may be detected only as undifferentiated small-volume flows like the Keogh Lake picrites within theKarmutsen Formation on Vancouver IslandDecompression melting within the mantle plume initiatedwithin the garnet stability field (35^25 GPa) at highmantle potential temperatures (Tp414508C) and pro-ceeded beneath oceanic arc lithosphere within the spinelstability field (25^09 GPa) where more extensivedegrees of melting could occur
Magmatic evolution of Karmutsentholeiitic basaltsPrimary magmas in large igneous provinces leave exten-sive coarse-grained cumulate residues within or below thecrust these mafic and ultramafic plutonic sequences repre-sent a significant proportion of the original magma thatpartially crystallized at depth (eg Farnetani et al 1996)For example seismic and petrological studies of OntongJava (eg Farnetani et al 1996) combined with the use ofMELTS (Ghiorso amp Sack 1995) indicate that the volcanicsequence may represent only 40 of the total magmareaching the Moho Neal et al (1997) estimated that30^45 fractional crystallization took place for theOntong Java magmas and produced cumulate thicknessesof 7^14 km corresponding to 11^22 km of flood basaltsdepending on the presence of pre-existing oceanic crustand the total crustal thickness (25^35 km) Interpretationsof seismic velocity measurements suggest the presence ofpyroxene and gabbroic cumulates 9^16 km thick beneaththe Ontong Java Plateau (eg Hussong et al 1979)Wrangellia is characterized by high crustal velocities in
seismic refraction lines which is indicative of mafic
plutonic rocks extending to depth beneath VancouverIsland (Clowes et al 1995)Wrangellia crust is 25^30 kmthick beneath Vancouver Island and is underlain by astrongly reflective zone of high velocity and density thathas been interpreted as a major shear zone where lowerWrangellia lithosphere was detached (Clowes et al 1995)The vast majority of the Karmutsen flows are evolved tho-leiitic lavas (eg low MgO high FeOMgO) indicating animportant role for partial crystallization of melts at lowpressure and a significant portion of the crust beneathVancouver Island has seismic properties that are consistentwith crystalline residues from partially crystallizedKarmutsen magmas To test the proportion of the primarymagma that fractionated within the crust MELTS(Ghiorso amp Sack 1995) was used to simulate fractionalcrystallization using several estimated primary magmacompositions from the PRIMELT1 modeling results Themajor-element composition of the primary magmas couldnot be estimated for the tholeiitic basalts because of exten-sive plagthorn cpxthornol fractionation and thus the MELTSmodeling of the picrites is used as a proxy for the evolutionof major elements in the volcanic sequence A pressure of 1kbar was used with variable water contents (anhydrousand 02wt H2O) calculated at the quartz^fayalite^magnetite (QFM) oxygen buffer Experimental resultsfrom Ontong Java indicate that most crystallizationoccurs in magma chambers shallower than 6 km deep(52 kbar Sano amp Yamashita 2004) however some crys-tallization may take place at greater depths (3^5 kbarFarnetani et al 1996)A selection of the MELTS results are shown in Fig 15
Olivine and spinel crystallize at high temperature until15^20wt of the liquid mass has fractionated and theresidual magma contains 9^10wt MgO (Fig 15f) At1235^12258C plagioclase begins crystallizing and between1235 and 11908C olivine ceases crystallizing and clinopyr-oxene saturates (Fig 15f) As expected the addition of clin-opyroxene and plagioclase to the crystallization sequencecauses substantial changes in the composition of the resid-ual liquid FeO and TiO2 increase and Al2O3 and CaOdecrease while the decrease in MgO lessens considerably(Fig 15) The near-vertical trends for the Karmutsen tho-leiitic basalts especially for CaO do not match the calcu-lated liquid-line-of-descent very well which misses manyof the high-CaO high-Al2O3 low-FeO basalts A positivecorrelation between Al2O3CaO and SrNd indicates thataccumulation of plagioclase played a major role in the evo-lution of the tholeiitic basalts (Fig 15g) Increasing thecrystallization pressure slightly and the water content ofthe melts does not systematically improve the match ofthe MELTS models to the observed dataThe MELTS results indicate that a significant propor-
tion of the crystallization takes place before plagioclase(15^20 crystallization) and clinopyroxene (35^45
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
31
crystallization) join the crystallization sequence Theseresults suggest that445 crystallization of the primarymagmas occurred (estimated from the residual liquid percent in Fig 15f) and if the original flood basalt stratigra-phy on Vancouver Island was 6 km thick at least anequivalent amount of complementary plutonic rocksshould have crystallized The Karmutsen basalts and theirplutonic residues thus represent a significant addition ofcrust (perhaps 412 km thickness) previously thickenedduring the formation of the underlying Paleozoic arc
Growth of the Wrangellia oceanic plateauon Vancouver IslandThe extrusive part of the Wrangellia oceanic plateau onVancouver Island was constructed in a three-layersequence of mostly tholeiitic basalts with a restrictedrange of composition (Fig 16) The plateau overliesDevonian to Mississippian arc volcanic sequences andmarine sediments of Mississippian to Permian ages thatare intruded by gabbroic rocks related to the flood basaltsThe volcanic stratigraphy of theWrangellia plateau formed
00
05
10
15
20
25
30
35
40
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
16
0 2 4 6 8 10 12 14 16 18 20
10
11
12
13
14
15
16
17
18
19
20
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
0 2 4 6 8 10 12 14 16 18 20
MgO (wt)
0 10 20 30 40 50
percent of total mass
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
) Mineral proportions
Sp (e)
Ol
CpxPlag
PicriteHigh-MgO basaltTholeiitic basalt
TiO2 (wt)
MgO (wt)
Al2O3 (wt) CaO (wt)
FeO (wt)
MgO(a) (b)
(c) (d)
MgO (wt)MgO (wt)
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
)
Residual liquid ()
1220
1190
Ol + Sp
Plag+Ol+Sp Plag+Cpx+Sp
02 wt H2O
no H2O
Residual liquid and crystallizing phases
(f)
0102030405060708090100
plagioclaseaccumulation
plagioclaseaccumulation
plagioclase
accumulatio
n
10
12
14
16
0 10 20 30 40
Al2O3CaO
SrNd
tholeiitic basalts
(g)
Fig 15 Forward fractional crystallization modeling results for major elements from MELTS (Ghiorso amp Sack 1995) compared with picriticand tholeiitic lavas from the Karmutsen Formation (a^d) Major-element compositions of residual liquids compared with Karmutsen lavacompositions For clarity the results are shown only for one starting composition (open circle) using the composition of the estimated primarymagma for the high-MgO lava series from the modeling technique of Herzberg et al (2007) for sample 4723A13 as shown in Fig 12 The esti-mated primary magmas for samples 4723A3 and 93G171 yield similar results A pressure of 1 kbar was used with no H2O and 02wt H2Ocalculated at QFM black line is anhydrous and grey line is 02wt H2OThe end-point for the anhydrous result is at 78 fractionation and11308C and the end-point for the result with 02wt H2O is 75 fractionation and11108C (e) Mineral proportions of crystallized phases fromone MELTrun (sample 4723A13) with no H2O (f) Per cent residual liquid vs temperature labelled with intervals of crystallizing phases (g)Al2O3CaO vs SrNd for the tholeiitic basalts
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
32
in a single eruptive phase (c 231^225 Ma) as an emergentbasalt sequence during a deeper-water shallow-water andsubaerial stage and subsided below sea level relativelysoon after its formation (Fig 16) The deeper-water stagecontains pillowed and massive submarine flows (3000mthick) that built up the submarine volcanic edifice from theseafloor The lowest pillow basalts were emplaced onunconsolidated and lithified fine-grained siliciclastic andcarbonaceous sediments containing Daonella beds(c 235^232 Ma) that are intruded by abundantKarmutsen mafic sills (Fig 16a) The different pillowedand massive submarine flows resulted primarily from dif-ferent effusive rates and local topography Some of themassive submarine flows commonly recognized by radiat-ing columnar jointing represent master tubes for the deliv-ery of lava to distal parts of submarine flow fields Thepillowed flows which commonly have uneven topography
in modern sequences may have served as levees to themassive submarine flows rather than allowing unob-structed spreading and inflation typical of the massive sub-aerial flows that form on shallow slopes in continentalflood basalt provinces (eg Self et al 1997) The rapid suc-cession of flows is evident by the near absence of sedimentbetween submarine flows Minor volumes of picritic pillowlavas occur in the upper part of the pillowed flow sequencein areas of the plateau exposed on northern VancouverIsland (Fig 16 Nixon et al 2008)The growth of the plateau in shallow water led to
emplacement of an increasing proportion of volcaniclastics(pillow breccias and hyaloclastites) that accumulated up to1500m thick in some areas (Fig 16b Nixon et al 2008)Horizons of pillowed flows occur locally within the volca-niclastic stratigraphy Sedimentary structures (eg gradedbedding fluidization structures) and petrographic textures
Late Devonian-Mississippian arc
Daonella beds
Middle Triassic
deeper water
Late Devonian-Mississippian arc
pillowed flows massive submarine
flows
pillow brecciahyaloclastiteshallow water
~4 km
~6 km
Late Triassic limestoneinterflow limestonelenses
Karmutsen Formation
mafic intrusions sediment-sill complex
Karmutsen Formation
(a) (b)
(c) (d)
subaerial flows
Late Devonian-Mississippian arc Late Devonian-Mississippian arc
picritic pillow basalts
Miss to Permian marine sediments
sills
Fig 16 Growth of the Wrangellia oceanic plateau onVancouver Island (a) Deep-water eruptive phase of pillowed and massive submarineflows with almost no intervening sediments onto MiddleTriassic shale containing Daonella beds (c 235^232 Ma) The basement of the plateauis composed of Devonian to Mississippian arc sequences overlain by Mississippian to Permian carbonates and siliciclastic sediments(b) Shallow-water flow emplacement with increasing proportion of volcaniclastics resulted primarily from cooling^contraction granulationPicritic pillow basalts occur primarily near the transition between pillowed submarine flows and volcaniclastic units Pillowed and massiveflows occur locally within the volcaniclastic unit (c) Emergent stage of growth consists of overlapping subaerial flow fields similar to continen-tal flood basalts with pahoehoe flow structures Multi-tiered columnar jointing occurrs locally in response to ponding of flows (eg Bondre et al2004) Interflow carbonate deposits and plagioclase megacrystic flows developed locally during the waning stage of volcanism (d) Post-volcanicsubsidence and deposition of Late Carnian to Norian (c 212^223 Ma) platform carbonates on top of the plateau Units are labeled in the dia-grams Thicknesses of the underlying units are not to scale
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
33
within the volcaniclastic rocks indicate formation primar-ily via cooling^contraction granulation magma^water^steam interaction autobrecciation resedimentation andmass-wasting rather than pyroclastic eruption (Nixonet al 2008)A broad subaerial platform was constructed by emplace-
ment of subaerial lava flows during an emergent subaerialstage (Fig 16c) The subaerial unit predominantly com-prises well-layered sheet flows that formed in a similarway to continental flood basalts as hundred(s) of pahoehoeflow fields erupted on shallow slopes (Self et al 1997) how-ever volcaniclastic and pillowed flow units are locally pre-sent mostly in the upper parts of the stratigraphyLocalized ponding of flows led to the formation of multi-tiered columnar jointing in some subaerial flows (egBondre et al 2004) As volcanism waned local interflowcarbonate deposits developed and plagioclase megacrysticflows were erupted Subsidence of the plateau is recordedby the interflow lenses and deposits of platform carbonates(40^1500m thick) on top of the Karmutsen basalts thatcontain Late Carnian to Early Norian fossils (Fig 16d c216^227 Ma Nixon et al 2008)The evolution of the internal architecture of the
Wrangellia oceanic plateau was primarily related to theconditions of emplacement (ie water depth and eruptionenvironment) There does not appear to have been a sys-tematic coincident evolution or shift of lava compositionsonVancouver Island as has been recognized in most con-tinental flood basalts and parts of the Wrangellia oceanicplateau in Alaska and Yukon (Greene et al 2008 2009)In general with the exception of the picritic pillow lavasin the submarine stratigraphy and local variations betweenhigh-MgO and tholeiitic lavas there are no significantcompositional differences between lavas erupted duringthe different stages of growth of the plateau (ie submarineand subaerial flows) nor are substantial compositional dif-ferences generally found between massive submarine flowsin contact with pillowed flows tholeiitic basalt flows nearthe base of the Karmutsen Formation are compositionallysimilar to flows sampled near the top of the volcanicsequence Where it was possible to sample continuous sec-tions of volcanic stratigraphy greater than several hun-dreds of meters (eg Keogh^Maynard Lake area west ofthe Karmutsen Range Fig 2) there is no clear relationshipbetween stratigraphic position and compositionalvariation
CONCLUSIONSThe Karmutsen Formation covers extensive areas ofVancouver Island (20 000 km2) and forms part of a largeoceanic plateau erupted over a geologically short interval(c 225^231 Ma) The tripartite volcanic stratigraphyexposed on Vancouver Island is at least 6 km thick andcomprises a succession of submarine flows volcaniclastic
deposits and massive subaerial flowsThese volcanologicaldifferences are primarily related to the eruption environ-ment (deep-water shallow-water or subaerial) The rapidgrowth of the plateau prevented significant accumulationsof intervening sediment except in the uppermost stratigra-phy where isolated limestone lenses commonly associatedwith pillow basalt and hyaloclastite deposits preserve arecord of the final stages of subsidence of the plateauThe Wrangellia plateau on Vancouver Island was con-
structed dominantly of tholeiitic basalt with restrictedmajor- and trace-element and isotopic compositional var-iations Minor volumes of picritic pillow basalts eruptedlate in the submarine phase The Karmutsen basalts wereemplaced onto older Paleozoic arc volcanic sequences andlow NbLa for the picrites may indicate a contributionfrom the sub-arc lithospheric mantle Modeling of the pet-rogenesis of the Karmutsen picrites indicates melting ofanomalously hot mantle (15008C) and extensive degreesof partial melting (23^27) and is consistent with aplume initiation model Combined Sr^Nd^Hf^Pb isotopicsystematics indicate that the picritic and tholeiitic lavaswere derived from a depleted plume-type mantle sourcedistinct from the source of MORB that has compositionalsimilarities to the source of the lavas forming theCaribbean Plateau Picrites were generated by melting ofspinel lherzolite whereas the tholeiitic basalts formed asaggregate melts of both garnet and spinel lherzolite Thetholeiitic basalts underwent extensive low-pressure fractio-nation (52^3 kbar) and seismic studies indicate that someof theWrangellia crust beneathVancouver Island may cor-respond to the plutonic residues of the Karmutsen floodbasalts
ACKNOWLEDGEMENTSWe are grateful to Nick Arndt for helping us get this pro-ject started and Nick Massey for insights into VancouverIsland geology We greatly appreciate the assistance withPRIMELT modeling from Claude Herzberg We thankMikkel Schau for his insight and enthusiasm during field-work Jane Barling assisted with analyses by MC-ICP-MSThoughtful and thorough reviews of the manuscript byJohn Mahoney Fred Frey Wendy Bohrson DougalJerram and Teal Riley are greatly appreciated Fundingwas generously provided by the Rocks to Riches Programadministered by the BC amp Yukon Chamber of Mines in2004 by the BC Geological Survey and by NSERCDiscovery Grants to JSS and DW ARG was supportedby a University Graduate Fellowship at UBC
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
34
REFERENCESBarker F Brown A S Budahn J R amp Plafker G (1989) Back-arc
with frontal-arc component origin of Triassic Karmutsen basaltBritish Columbia Canada Chemical Geology 75 81^102
Blichert-Toft JWeis D Maerschalk C Agranier A amp Albare de F(2003) Hawaiian hot spot dynamics as inferred from the Hf and Pbisotope evolution of Mauna Kea volcano Geochemistry GeophysicsGeosystems 4(2) 1^27 doi1010292002GC000340
Bondre N R Duraiswami R A amp Dole G (2004) Morphologyand emplacement of flows from the Deccan Volcanic ProvinceIndia Bulletin of Volcanology 66 29^45
Carlisle D (1963) Pillow breccias and their aquagene tuffs QuadraIsland British Columbia Journal of Geology 71 48^71
Carlisle D (1972) Late Paleozoic to Mid-Triassic sedimentary-volcanic sequence on Northeastern Vancouver Island Geological
Society of Canada Paper Report of Activities 72-1 Part B 24^30Carlisle D amp SuzukiT (1974) Emergent basalt and submergent car-
bonate^clastic sequences including the Upper Triassic Dilleri andWelleri zones on Vancouver Island Canadian Journal of Earth
Sciences 11 254^279Chauvel C amp Blichert-Toft J (2001) A hafnium isotope and trace
element perspective on melting of the depleted mantle Earth and
Planetary Science Letters 190 137^151Chayes F (1954)The theory of thin-section analysis Journal of Geology
62 92^101Cho M Liou J G amp Maruyama S (1987) Transition from the zeo-
lite to prehnite^pumpellyite facies in the Karmutsen metabasitesVancouver Island British Columbia Journal of Petrology 27(2)467^494
Clowes R M Zelt C A Amor J R amp Ellis R M (1995)Lithospheric structure in the southern Canadian Cordillera froma network of seismic refraction lines Canadian Journal of Earth
Sciences 32(10) 1485^1513Coffin M F amp Eldholm O (1994) Large igneous provinces Crustal
structure dimensions and external consequences Reviews of
Geophysics 32(1) 1^36Cox K G (1980) A model for flood basalt volcanism Journal of
Petrology 21 629^650Eldholm O amp Coffin M F (2000) Large igneous provinces
and plate tectonics In Richards M A Gordon R G amp vander Hilst R D (eds) The History and Dynamics of Global
Plate Motions American Geophysical Union Geophysical Monograph 121309^326
Elliott T R Hawkesworth C J amp Grolaquo nvold K (1991) Dynamicmelting of the Iceland plume Nature 351 201^206
Escuder-Viruete J Pecurren rez-Estaucurren n A Contreras F Joubert MWeis D Ullrich T D amp Spadea P (2007) Plume mantle sourceheterogeneity through time Insights from the Duarte ComplexHispaniola northeastern Caribbean Journal of Geophysical Research112(B04203) doi1010292006JB004323
Farnetani C G Richards M A amp Ghiorso M S (1996)Petrological models of magma evolution and deep crustal structurebeneath hotspots and flood basalt provinces Earth and Planetary
Science Letters 143 81^94Fitton J G amp Godard M (2004) Origin and evolution of magmas
on the Ontong Java Plateau In Fitton J G Mahoney J JWallace P J amp Saunders A D (eds) Origin and Evolution of the
Ontong Java Plateau Geological Society London Special Publications 229151^178
Fitton J G Saunders A D Norry M J Hardarson B S ampTaylor R N (1997) Thermal and chemical structure of theIceland plume Earth and Planetary Science Letters 153 197^208
Fitton J G Saunders A D Kempton P D amp Hardarson B S(2003) Does depleted mantle form an intrinsic part of the Icelandplume Geochemistry Geophysics Geosystems 4(3) 1032 doi1010292002GC000424
Frey F A Huang S Blichert-Toft J Regelous M amp Boyet M(2005) Origin of depleted components in basalt related to theHawaiian hotspot Evidence from isotopic and incompatible ele-ment ratios Geochemistry Geophysics Geosystems 6(1) 23 doi1010292004GC000757
Galer S J G amp Abouchami W (1998) Practical application of leadtriple spiking for correction of instrumental mass discriminationMineralogical Magazine 62A 491^492
Ghiorso M S amp Sack R O (1995) Chemical mass transfer in mag-matic processes IV A revised and internally consistent thermody-namic model for the interpolation and extrapolation of liquid^solidequilibria in magmatic systems at elevated temperatures and pres-sures Contributions to Mineralogy and Petrology 119 197^212
Greene A R Scoates J S amp Weis D (2005)WrangelliaTerrane onVancouver Island Distribution of flood basalts with implicationsfor potential Ni^Cu^PGE mineralization In Grant B (ed)Geological Fieldwork 2004 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2005-1 209^220Greene A R Scoates J S Nixon G T amp Weis D (2006) Picritic
lavas and basal sills in the Karmutsen flood basalt provinceWrangellia northern Vancouver Island In Grant B (ed)Geological Fieldwork 2005 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2006-1 39^52Greene A R Scoates J S amp Weis D (2008) Wrangellia flood
basalts in Alaska A record of plume^lithosphere interaction in aLate Triassic accreted oceanic plateau Geochemistry Geophysics
Geosystems 9(Q12004) doi1010292008GC002092Greene A R Scoates J S Weis D amp Israel S (2009)
Geochemistry of triassic flood basalts from the Yukon (Canada)segment of the accreted Wrangellia oceanic plateau Lithosdoi101016jlithos200811010
Hauff F Hoernle K Tilton G Graham D W amp Kerr A C(2000) Large volume recycling of oceanic lithosphere over shorttime scales geochemical constraints from the Caribbean LargeIgneous Province Earth and Planetary Science Letters 174 247^263
Hauff F Hoernle K amp Schmidt A (2003) Sr^Nd^Pb compositionof Mesozoic Pacific oceanic crust (Site 1149 and 801 ODP Leg185)Implications for alteration of ocean crust and the input into theIzu^Bonin^Mariana subduction system Geochemistry Geophysics
Geosystems 4(8) doi1010292002GC000421Herzberg C (2004) Partial melting below the Ontong Java Plateau
In Fitton J G Mahoney J J Wallace P J amp Saunders A D(eds) Origin and Evolution of the Ontong Java Plateau Geological SocietyLondon Special Publications 229 179^183
Herzberg C (2006) Petrology and thermal structure of the Hawaiianplume from Mauna Kea volcano Nature 444(7119) 605
Herzberg C amp Asimow P D (2008) Petrology of some oceanicisland basalts PRIMELT2XLS software for primary magma cal-culation Geochemistry Geophysics Geosystems 9(Q09001) doi1010292008GC002057
Herzberg C amp OrsquoHara M J (2002) Plume-associated ultramaficmagmas of Phanerozoic age Journal of Petrology 43(10) 1857^1883
Herzberg C Asimow P D Arndt N Niu Y Lesher C MFitton J G Cheadle M J amp Saunders A D (2007)Temperatures in ambient mantle and plumes Constraints frombasalts picrites and komatiites Geochemistry Geophysics Geosystems8(Q02006) doi1010292006GC001390
Huang S amp Frey F A (2005) Trace element abundances of MaunaKea basalt from phase 2 of the Hawaii Scientific Drilling Project
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
35
Petrogenetic implications of correlations with major element con-tent and isotopic ratios Geochemistry Geophysics Geosystems 4(6)doi1010292002GC000322
Hussong D M Wipperman L K amp Kroenke L W (1979) Thecrustal structure of the Ontong Java and Manihiki OceanicPlateaus Journal of Geophysical Research 84(B11) 6003^6010
Jerram D A amp Widdowson M (2005) The anatomy of ContinentalFlood Basalt Provinces geological constraints on the processes andproducts of flood volcanism Lithos 79(3^4) 385^405
Jones D L Silberling N J amp Hillhouse J (1977)Wrangellia a dis-placed terrane in northwestern North America CanadianJournal ofEarth Sciences 14(11) 2565^2577
Kerr A C (2003) Oceanic Plateaus In Rudnick R (ed) The Crust
Treatise on GeochemistryVol 3 Oxford Elsevier Science pp 537^565Kerr A C amp Mahoney J J (2007) Oceanic plateaus Problematic
plumes potential paradigms Chemical Geology 241 332^353Kerr A CTarney J Marriner G F Klaver GT Saunders A D
amp Thirlwall M F (1996) The geochemistry and petrogenesis ofthe Late-Cretaceous picrites and basalts of Curacao NetherlandsAntilles a remnant of an oceanic plateau Contributions to
Mineralogy and Petrology 124(1) 29^43Kerr A C Marriner G F Tarney J Nivia A Saunders A D
Thirlwall M F amp Sinton A C (1997) Cretaceous basaltic ter-ranes in Western Colombia Elemental chronological and Sr-Ndisotopic constraints on petrogenesis Journal of Petrology 38(6)677^702
Kinzler R J (1997) Melting of mantle peridotite at pressuresapproaching the spinel to garnet transition Application to mid-ocean ridge basalt petrogenesis Journal of Geophysical Research
102(B1) 853^874Kinzler R J amp Grove T L (1992) Primary magmas of mid-ocean
ridge basalts 1 Experiments and methods Journal of Geophysical
Research 97(B5) 6885^6906Korenaga J (2005) Why did not the Ontong Java Plateau form sub-
aerially Earth and Planetary Science Letters 234 385^399Kuniyoshi S (1972) Petrology of the Karmutsen (Volcanic) Group
northeastern Vancouver Island British Columbia PhD disserta-tion University of California Los Angeles 242 pp
Lassiter J C DePaolo D J amp Mahoney J J (1995) Geochemistry ofthe Wrangellia flood basalt province Implications for the role ofcontinental and oceanic lithosphere in flood basalt genesis Journalof Petrology 36(4) 983^1009
LeBas M J (2000) IUGS reclassification of the high-Mg and picriticvolcanic rocks Journal of Petrology 41(10) 1467^1470
Longhi J (2002) Some phase equilibrium systematics of lherzolitemelting I Geochemistry Geophysics Geosystems 3(3) doi1010292001GC000204
MacDonald G A amp Katsura T (1964) Chemical composition ofHawaiian lavas Journal of Petrology 5 82^133
Mahoney J J (1987) An isotopic survey of Pacific oceanic plateausimplications for their nature and origin In Keating B HFryer P Batiza R amp Boehlert GW (eds) Seamounts Islands andAtolls American Geophysical Union Geophysical Monograph 43 207^220
Mahoney J LeRoex A P Peng Z Fisher R L amp Natland J H(1992) Southwestern limits of Indian Ocean Ridge mantle and theorigin of low 206Pb204Pb mid-ocean ridge basalt isotope systema-tics of the central Southwest Indian Ridge (178^508E) Journal ofGeophysical Research 97(B13) 19771^19790
Mahoney J J Sinton J M Kurz M D MacDougall J DSpencer K J amp Lugmair GW (1994) Isotope and trace elementcharacteristics of a super-fast spreading ridge East Pacific rise13^238S Earth and Planetary Science Letters 121 173^193
Massey N W D (1995a) Geology and mineral resources of theAlberni^Nanaimo Lakes sheet Vancouver Island 92F1W 92F2Eand part of 92F7E British Columbia Ministry of Energy Mines and
Petroleum Resources Paper 1992-2 132 ppMassey N W D (1995b) Geology and mineral resources of the
Cowichan Lake sheet Vancouver Island 92C16 British Columbia
Ministry of Energy Mines and Petroleum Resources Paper 1992-3 112 ppMassey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005a) Digital Geology Map of British Columbia Tile NM9 MidCoast BC British Columbia Ministry of Energy and Mines Geofile
2005-2Massey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005b) Digital Geology Map of British Columbia Tile NM10Southwest BC British Columbia Ministry of Energy and Mines Geofile
2005-3McDonough W F amp Sun S (1995) The composition of the Earth
Chemical Geology 120 223^253Monger JW H amp Journeay J M (1994) Basement geology and tec-
tonic evolution of theVancouver region In Monger JW H (ed)Geology and Geological Hazards of the Vancouver Region Southwestern
British Columbia Geological Survey of Canada Bulletin 481 3^25Muller J E (1977) Geology of Vancouver Island Geological Survey of
Canada Open File Map 463Muller J E (1980)The Paleozoic Sicker Group of Vancouver Island British
Columbia Geological Survey of Canada Paper 79-30 22 ppMuller J E Northcote K E amp Carlisle D (1974) Geology and
mineral deposits of Alert Bay^Cape Scott map area VancouverIsland British Columbia Geological Survey of Canada Paper 74-8 77
Neal C R Mahoney J J Kroenke L W Duncan R A ampPetterson M G (1997) The Ontong^Java Plateau InMahoney J J amp Coffin M F (eds) Large Igneous Provinces
Continental Oceanic and Planetary FloodVolcanism American Geophysical
Union Geophysical Monograph 100 183^216Niu Y Collerson K D Batiza R Wendt J I amp Regelous M
(1999) Origin of enriched-typed mid-ocean ridge basalt at ridgesfar from mantle plumes The East Pacific Rise at 11820rsquoN Journalof Geophysical Research 104 7067^7088
Nixon G T amp Orr A J (2007) Recent revisions to the EarlyMesozoic stratigraphy of Northern Vancouver Island (NTS 102I092L) and metallogenic implicatinos British Columbia InGrant B (ed) Geological Fieldwork 2006 British Columbia Ministry of
Energy Mines and Petroleum Resources Paper 2007-1 163^177Nixon GT Hammack J L KoyanagiV M Payie G J Haggart
JW Orchard M JTozerT Friedman R M Archibald D APalfy J amp Cordey F (2006a) Geology of the Quatsino^PortMcNeill area northernVancouver Island British Columbia Ministry
of Energy Mines and Petroleum Resources Geoscience Map 2006-2 scale150 000
Nixon G T Kelman M C Stevenson D Stokes L A ampJohnston K A (2006b) Preliminary geology of the Nimpkishmap area (NTS 092L07) northern Vancouver Island BritishColumbia In Grant B amp Newell J M (eds) Geological Fieldwork2005 British Columbia Ministry of Energy Mines and Petroleum Resources
Paper 2006-1 135^152Nixon G T Laroque J Pals A Styan J Greene A R amp
Scoates J S (2008) High-Mg lavas in the Karmutsen flood basaltsnorthernVancouver Island (NTS 092L) Stratigraphic setting andmetallogenic significance In Grant B (ed) Geological Fieldwork
2007 British Columbia Ministry of Energy Mines and Petroleum
Resources Paper 2008-1 175^190Nowell G M Kempton P D Noble S R Fitton J G
Saunders A Mahoney J J amp Taylor R N (1998) High precisionHf isotope measurements of MORB and OIB by thermal ionisation
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
36
mass spectrometry insights into the depleted mantle Chemical
Geology 149 211^233Parrish R R amp McNicoll V J (1992) U^Pb age determinations
from the southern Vancouver Island area British Columbia InRadiogenic Age and Isotopic Studies Report 5 Geological Survey of
Canada Paper 91-2 79^86Peate DW (1997) The Parana^Etendeka Province In Coffin M F
amp Mahoney J (eds) Large Igneous Provinces Continental Oceanic andPlanetary Flood Volcanism American Geophysical Union Geophysical
Monograph 100 217^245Perfit M R Fornari D J Ridley W I Kirk P D Casey J
Kastens K A Reynolds J R Edwards M Desonie DShuster R amp Paradis S (1996) Recent volcanism in the Siqueirostransform fault picritic basalts and implications for MORBmagma genesis Earth and Planetary Science Letters 141(1^4) 91^108
Pretorius W Weis D Williams G Hanano D Kieffer B ampScoates J S (2006) Complete trace elemental characterization ofgranitoid (USGSG-2GSP-2) reference materials by high resolutioninductively coupled plasma-mass spectrometry Geostandards and
Geoanalytical Research 30(1) 39^54Regelous M Niu Y Wendt J I Batiza R Greig A amp
Collerson K D (1999) Variations in the geochemistry of magma-tism on the East Pacific Rise at 10830rsquoN since 800 ka Earth and
Planetary Science Letters 168 45^63Recurren villon S Chauvel C Arndt N T Pik R Martineau F
Fourcade S amp Marty B (2002) Heterogeneity of the Caribbeanplateau mantle source Sr O and He isotopic compositions of oli-vine and clinopyroxene from Gorgona Island Earth and Planetary
Science Letters 205(1^2) 91Rhodes J M (1996) Geochemical stratigraphy of lava flows samples
by the Hawaii Scientific Drilling Project Journal of Geophysical
Research 101(B5) 11729^11746Rhodes J M amp Vollinger M J (2004) Composition of basaltic lavas
sampled by phase-2 of the Hawaii Scientific Drilling ProjectGeochemical stratigraphy and magma types Geochemistry
Geophysics Geosystems 5(3) doi1010292002GC000434Richards M A Jones D L Duncan R A amp DePaolo D J (1991)
A mantle plume initiation model for the Wrangellia flood basaltand other oceanic plateaus Science 254 263^267
Salters V J M amp Hart S R (1991) The mantle sources of oceanridges islands and arcs the Hf-isotope connection Earth and
Planetary Science Letters 104 364^380Salters V J M amp Stracke A (2004) Composition of the depleted
SaltersV J amp WhiteW (1998) Hf isotope constraints on mantle evo-lution Chemical Geology 145 447^460
SanoT amp Yamashita S (2004) Experimental petrology of basementlavas from Ocean Drilling Program Leg 192 implications for dif-ferentiation processes in Ontong Java Plateau magmas InFitton J G Mahoney J JWallace P J amp Saunders A D (eds)Origin and Evolution of the Ontong Java Plateau Geological Society
London Special Publications 229 185^218Saunders A D (2005) Large igneous provinces Origin and environ-
mental consequences Elements 1 259^263Self S ThordarsonT amp Keszthely L (1997) Emplacement of conti-
nental flood basalt lava flows In Mahoney J J amp Coffin M F(eds) Large Igneous Provinces Continental Oceanic and Planetary Flood
Volcanism American Geophysical Union Geophysical Monograph 100381^410
Shaw D M (2000) Continuous (dynamic) melting theory revisitedCanadian Mineralogist 38(5) 1041^1063
Sluggett C L (2003) Uranium^lead age and geochemical constraintson Paleozoic and Early Mesozoic magmatism in WrangelliaTerrane Saltspring Island British Columbia BSc thesisUniversity of British ColumbiaVancouver 84 pp
Storey M Mahoney J J amp Saunders A D (1997) Cretaceousbasalts in Madagascar and the transition between plume and con-tinental lithospheric mantle sources In Mahoney J J ampCoffin M F (eds) Large Igneous Provinces Continental Oceanic and
Planetary Flood Volcanism Aamerican Geophysical Union Geophysical
Monograph 100 95^122Surdam R C (1967) Low-grade metamorphism of the Karmutsen
Group PhD dissertation University of California Los Angeles288 pp
Sutherland-Brown A Yorath C J Anderson R G amp Dom K(1986) Geological maps of southern Vancouver IslandLITHOPROBE I 92C10 11 14 16 92F1 2 7 8 Geological Survey ofCanada Open File 1272
Tejada M L G Mahoney J J Castillo P R Ingle S P Sheth HC amp Weis D (2004) Pin-pricking the elephant evidence on theorigin of the Ontong Java Plateau from Pb^Sr^Hf^Nd isotopiccharacteristics of ODP Leg 192 basalts In Fitton J GMahoney J J Wallace P J amp Saunders A D (eds) Origin and
Evolution of the Ontong Java Plateau Geological Society London Special
Publications 229 133^150Thompson P M E Kempton P D amp Kerr A C (2008) Evaluation
of the effects of alteration and leaching on Sm^Nd and Lu^Hf sys-tematics in submarine mafic rocks Lithos 104(1^4) 164^176
Thompson P M E Kempton P D White R V Kerr A CTarney J Saunders A D Fitton J G amp McBirney A (2003)Hf-Nd isotope constraints on the origin of the CretaceousCaribbean plateau and its relationship to the Galapagos plumeEarth and Planetary Science Letters 217(1^2) 59^75
Vervoort J D Patchett P Blichert-Toft J amp Albare de F (1999)Relationships between Lu^Hf and Sm^Nd isotopic systems in theglobal sedimentary system Earth and Planetary Science Letters 16879^99
Walter M J (1998) Melting of garnet peridotite and the origin ofkomatiite and depleted lithosphere Journal of Petrology 39(1) 29^60
Weis D Kieffer B Maerschalk C Barling J de Jong JWilliams G A Hanano D Mattielli N Scoates J SGoolaerts A Friedman R A amp Mahoney J B (2006) High-pre-cision isotopic characterization of USGS reference materials byTIMS and MC-ICP-MS Geochemistry Geophysics Geosystems
7(Q08006) doi1010292006GC001283Weis D Kieffer B Hanano D Silva I N Barling J
Pretorius W Maerschalk C amp Mattielli N (2007) Hf isotopecompositions of US Geological Survey reference materialsGeochemistry Geophysics Geosystems 8(Q06006) doi1010292006GC001473
Wheeler J O amp McFeely P (1991) Tectonic assemblage map of theCanadian Cordillera and adjacent part of the United States ofAmerica Geological Survey of Canada Map 1712A
WhiteW M Albare de F amp Tecurren louk P (2000) High-precision analy-sis of Pb isotope ratios by multi-collector ICP-MS Chemical Geology167 257^270
Yorath C J Sutherland Brown A amp Massey N W D (1999)LITHOPROBE southern Vancouver Island British Columbia Geological
Survey of Canada Bulletin 498 145 ppZou H (1998) Trace element fractionation during modal and
non-model dynamic melting and open-system melting Amathematical treatment Geochimica et Cosmochimica Acta 62(11)1937^1945
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
37
Zou H amp Reid M R (2001) Quantitative modeling of trace elementfractionation during incongruent dynamic melting Geochimica et
Cosmochimica Acta 65(1) 153^162
APPENDIX SAMPLEPREPARAT ION ANDANALYT ICAL METHODSOnly the freshest rocks in the field were sampled and onlythe least altered samples were selected for chemical andisotopic analysis based on thorough petrographic inspec-tion Fifty-nine of the 129 collected samples were crushed(400 g) into small pieces 52mm in diameter in aRocklabs hydraulic piston crusher between WC plates tominimize contamination The coarse-crush was mixedand 100 g was powdered in a planetary mill using agatejars and balls cleaned with quartz sand between samples
ActLabs analytical methodsThe major- and trace-element compositions of the whole-rock powders were determined at Activation LaboratoriesLtd (Actlabs) in Ancaster Ontario Analytical techniquesand detection limits are also available from Actlabs (httpwwwactlabscommethsub_code4erehtm) The analyticalmethod for each of the elements analyzed is indicated inTable 2 For the major elements a 02 g sample was mixedwith a lithium metaborate^lithium tetraborate mixtureand fused in a graphite crucible The molten mixture waspoured into a 5 HNO3 solution and shaken until dis-solved (30min) The samples were analyzed for majoroxides and selected trace elements by inductively coupledplasma optical emission spectrometry (ICP-OES) on acombination simultaneous^sequential Thermo Jarrell-AshEnviro II system Internal calibration was achieved usinga variety of international reference materials (eg W-2BIR-1 DNC-1) and independent control samplesAdditional trace elements were analyzed by both instru-mental neutron activation analysis (INAA) and induc-tively couple plasma mass spectrometry (ICP-MS) ForINAA15^25 g of sample was weighed into small polyeth-ylene vials and irradiated with a control international ref-erence material CANMET WMS-1 and NiCr flux wires ata thermal neutron flux of 71012 neutronscm2 per s in theMcMaster Nuclear Reactor Following a 7 day waitingperiod the samples were measured on an Ortec high-purity Ge detector linked to a Canberra Series 95 multi-channel analyzer Activities for each element weredecay- and weight-corrected and compared with a detectorcalibration developed from multiple international certifiedreference materials For ICP-MS 025 g of sample wasdigested in HF followed by HNO3^HClO4 heated andtaken to dryness The samples were brought back into solu-tion with HCl Samples were analyzed using a PerkinElmer Optima 3000 ICP In-lab standards or certified ref-erence materials (eg W-2 BIR-1 DNC-1) were used for
quality control A total of 15 blind duplicates were ana-lyzed from the same sample powders to assessreproducibility
University of Massachusetts analyticalmethodsFourteen sample duplicate powders taken from the same100 g sample aliquot (high-MgO basalts and picrites)were also analyzed at the Ronald B Gilmore X-RayFluorescence (XRF) Laboratory at the University ofMassachusetts Major elements were measured on a fusedLa-bearing lithium borate glass disc using a SiemensMRS-400 spectrometer Trace element concentrations(Rb Sr Ba Ce Nb Zr Y Pb Zn Ga Ni Cr V) weremeasured on a separate powder pellet using a PhilipsPW2400 sequential spectrometer using a Rh tube LOIand ferrous iron measurements were made as describedby Rhodes amp Vollinger (2004) Precision and accuracy esti-mates for the data have been described by Rhodes (1996)and Rhodes amp Vollinger (2004) Results for each sampleare the average of two separate analyses and are shown inElectronic Appendix 2 for a comparison of analyses ofduplicate powders using XRFand ICP-MS
PCIGR analytical methodsA subset of 19 samples was selected for high-precisiontrace-element analysis and Sr Nd Pb and Hf isotopicanalysis at the Pacific Centre for Isotopic andGeochemical Research (PCIGR) at the University ofBritish Columbia (UBC) Samples were selected from the59 samples analyzed for whole-rock chemistry at ActLabsbased on major- and trace-element chemistry alteration(low LOI and petrographic alteration index) and samplelocation Samples were prepared for trace-element analysisat the PCIGR by the technique described by Pretorius et al(2006) on unleached rock powders Sample powders(100mg) were weighed in 7ml screw-top Savillex bea-kers and dissolved in 1ml 14N HNO3 and 5ml 48 HFon a hotplate for 48 h at 1308C with periodic ultrasonica-tion Samples were dried and redissolved in 6ml 6N HClon a hotplate for 24 h and then dried and redissolved in1ml concentrated HNO3 for 24 h before final dryingTrace element abundances were measured by high-resolution inductively coupled plasma-mass spectrometry(HR-ICP-MS) with aThermo Finnigan Element2 systemfollowing the procedures described by Pretorius et al(2006) within 24 h of redissolution HFSE and LILE weremeasured in medium-resolution mode at 2000 dilutionusing a PFATeflon spray chamber washed with aqua regiafor 3min between samples REE were measured in high-resolution mode and U and Pb in low-resolution modeat 2000 dilution using a glass spray chamber washedwith 2 HNO3 between samples Total procedural blanksand reference materials (BCR-2 BHVO-2) were analyzedwith the batch of samples Indium was used as an internal
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
38
standard in all samples and standard solutionsBackground and standard solutions were analyzed afterevery five samples to detect memory effects and massdrift Results for PCIGR trace-element analyses areshown in Electronic Appendix 3 for the full set of analyzedtrace elementsSample digestion for purification of Sr Nd Hf and Pb
for column chemistry began by weighing each samplepowder (400^500mg) prior to leaching All samples wereleached with 6N HCl and placed in an ultrasonic bath for15min Samples were rinsed twice with 18 M-cm H2Obetween each leaching step (15 total) until the supernatantliquid was clear [following the method of Mahoney (1987)]Samples were then dried on a hotplate for 24 h andweighed again Sample solutions were then prepared bydissolving 100^250mg of the leached powder dissolvedin 1ml 14N HNO3 and 10ml 48 HF on a hotplate for48 h at 1308C with periodic ultrasonication Samples weredried and redissolved in 6ml 6N HCl on a hotplate for24 h and then dried Pb was separated using anion exhangecolumns and the discard was used for Sr REE and Hfseparation Nd was separated from the other REE and Hfrequired two additional purification steps Detailed proce-dures for column chemistry for separating Sr Nd and Pbat the PCIGR have been described byWeis et al (2006) andHf purification has been described byWeis et al (2007) Srand Nd isotope ratios were measured by thermal ioniza-tion mass spectrometry (TIMS) on a Thermo FinniganTriton system in static mode with relay matrix rotation ona single Ta and double Re^Ta filament respectively Fourto five filaments per barrel of 21 were occupied by stan-dards (NBS 987 for Sr and LaJolla for Nd) for each barrelwhere samples were run Sample Sr and Nd isotopic com-positions were corrected for mass fractionation using86Sr88Srfrac14 01194 and 146Nd144Ndfrac14 07219 Each samplewas then normalized using the barrel average of the refer-ence material relative to the values of143Nd144Ndfrac14 0511858 and 87Sr86Srfrac14 0710248 (Weiset al 2006) During the course of the Vancouver Islandanalyses the La Jolla Nd standard gave an average valueof 05118566 (nfrac14 7) and NBS987 standard gave an aver-age of 07102408 (nfrac1411) (2 error is reported as times106) 147Sm144Nd errors are 15 or 0006 USGeological Survey (USGS) reference material BHVO-2was processed with the samples and yielded Sr and Nd iso-topic ratios of 07034607 and 05129786 respectivelyThese are in agreement with the published values of070347920 and 051298411 respectively (Weis et al2006)
Pb and Hf isotopic compositions were analyzed by mul-tiple collector-inductively coupled plasma-mass spectrome-try (MC-ICP-MS) on a static multi-collection on a NuPlasma (Nu Instruments) system The detailed analyticalprocedure for Pb isotopic analyses on the Nu Plasma atthe PCIGR has been described by Weis et al (2006) Theconfiguration for Pb analyses allows for collection of PbTl and Hg togetherTl and Hg are used to monitor instru-mental mass discrimination and isobaric overlap repec-tively All sample solutions were analyzed withapproximately the same PbTl ratio (4) as the referencematerial NIST SRM 981 To accomplish this a small ali-quot of each sample solution from the Pb columns was ana-lyzed on the Element2 to determine the precise amount ofPb available for analysis on the Nu Plasma The SRM 981standard was run after every two samples on the NuPlasma During the time samples were run analyses ofthe SRM 981 Pb reference material gave values of206Pb204Pbfrac1416940322 207Pb204Pbfrac1415495823and 208Pb204Pbfrac14 36713164 (nfrac14 61 2 error is reportedas times 104) these values are in excellent agreement withreported TIMS triple-spike values of Galer amp Abouchami(1998) Fractionation-corrected Pb isotopic ratios were fur-ther corrected by the sample-standard bracketing methodor the ln^ln correction method described by White et al(2000) and Blichert-Toft et al (2003) Leached powders ofUSGS reference material BHVO-2 yielded Pb isotopicratios of 206Pb204Pbfrac14186454 8 207Pb204Pbfrac14 1549105 and 208Pb204Pbfrac14 38222514 and BCR-2yielded 206Pb204Pbfrac141880466 207Pb204Pbfrac14 1562518 and 208Pb204Pbfrac14 3883496 (2 error isreported as times 104) These values are in agreement withthose for leached residues of BHVO-2 and BCR-2 fromWeis et al (2006)Hf isotopic compositions were analyzed following the
procedures detailed byWeis et al (2007) The configurationfor Hf analyses monitored Lu mass 175 andYb mass 172 toallow for interference correction to masses 174 and 176 Hfisotopic ratios were normalized internally for mass fractio-nation to a 179Hf177Hf ratio of 07325 using an exponentialcorrection Standards were run after every two samplesand sample results were normalized to the ratio of the in-run daily average and a 176Hf177Hf ratio for JMC-475 of0282160 During the course of analyses the Hf standardJMC-475 gave an average value 02821679 (nfrac14130)USGS reference materials BCR-2 and BHVO-2 were pro-cessed with the samples and yielded Hf isotopic ratios of02828675 and 02831005 respectively Publishedvalues for BCR-2 and BHVO-2 are 02828717 and0283104 8 respectively (Weis et al 2007)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
39
DISCUSS IONThe compositional and spatial development of the eruptedlava sequences of the Wrangellia oceanic plateau onVancouver Island provide information about the evolutionof a mantle plume upwelling beneath oceanic lithospherein an analogous way to the interpretation of continentalflood basalts The Caribbean and Ontong Java plateausare the two primary examples of accreted parts of oceanicplateaus that have been studied to date and comparisonscan be made with the Wrangellia oceanic plateauHowever the Wrangellia oceanic plateau is a distinctiveoccurrence of an oceanic plateau because it formed on thecrust of a Devonian to Mississippian intra-oceanic islandarc overlain by Mississippian to Permian limestones andpelagic sediments The nature and thickness of oceaniclithospheric mantle are considered to play a fundamentalrole in the decompression and evolution of a mantleplume and thus the compositions of the erupted lavasequences (Greene et al 2008) The geochemical and
stratigraphic relationships observed in the KarmutsenFormation onVancouver Island and the focus of the subse-quent discussion sections provide constraints on (1) thetemperature and degree of melting required to generatethe primary picritic magmas (2) the composition of themantle source (3) the depth of melting and residual min-eralogy in the source region (4) the low-pressure evolutionof the Karmutsen tholeiitic basalts (5) the growth of theWrangellia oceanic plateau
Melting conditions and major-elementcomposition of the primary magmasThe primary magmas to most flood basalt provinces arebelieved to be picritic (eg Cox 1980) and they have beenfound in many continental and oceanic flood basaltsequences worldwide (eg Siberia Karoo Paranacurren ^Etendeka Caribbean Deccan Saunders 2005) Near-pri-mary picritic lavas with low total alkali abundances
Table 5 Pb isotopic compositions of Karmutsen basaltsVancouver Island BC
THOL tholeiitic basalt CG coarse-grained mafic rock PIC picrite HI-MG high-MgO basalt OUTLIER anomalouspillowed flow in plots KR Karmutsen Range SL Schoen Lake MA Mount Arrowsmith (dup) indicates completechemistry duplicate from the sample sample powder All isotopic and elemental analyses were carried out at the PCIGRthe complete trace element analyses are given in Electronic Appendix 3
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
require high-degree partial melting of the mantle source(eg Herzberg amp OrsquoHara 2002) The Keogh Lake picritesfrom northernVancouver Island are the best candidates foridentifying the least modified partial melts of the mantleplume source despite having accumulated olivine pheno-crysts and can be used to estimate conditions of meltingand the composition of the primary magmas for theKarmutsen FormationPrimary magma compositions mantle potential tem-
peratures and source melt fractions for the picritic lavas ofthe Karmutsen Formation were calculated from primitivewhole-rock compositions using PRIMELT1XLS software(Herzberg et al 2007) and are presented in Fig 12 andTable 6 A detailed discussion of the computationalmethod has been given by Herzberg amp OrsquoHara (2002)and Herzberg et al (2007) key features of the approachare outlined belowPrimary magma composition is calculated for a primi-
tive lava by incremental addition of olivine and compari-son with primary melts of mantle peridotite Lavas thathad experienced plagioclase andor clinopyroxene fractio-nation are excluded from this analysis All calculated pri-mary magma compositions are assumed to be derived byaccumulated fractional melting of fertile peridotite KR-4003 Uncertainties in fertile peridotite composition donot propagate to significant variations in melt fractionand mantle potential temperature melting of depletedperidotite propagates to calculated melt fractions that aretoo high but with a negligible error in mantle potential
temperature PRIMELT1XLS calculates the olivine liqui-dus temperature at 1atm which should be similar to theactual eruption temperature of the primary magmaand is accurate to 308C at the 2 level of confidenceMantle potential temperature is model dependent witha precision that is similar to the accuracy of the olivineliquidus temperature (C Herzberg personal communica-tion 2008)With regard to the evidence for accumulated olivine in
the Keogh Lake picrites it should not matter whether themodeled lava composition is aphyric or laden with accu-mulated olivine phenocrysts PRIMELT1 computes theprimary magma composition by adding olivine to a lavathat has experienced olivine fractionation in this way itinverts the effects of fractional crystallization The onlyrequirement is that the olivine phenocrysts are not acci-dental or exotic to the lava compositions Support for thisprocedure can be seen in the calculated olivine liquid-line-of-descent shown by the curved arrays with 5 olivineaddition increments (Fig 12c) Fractional crystallization ofolivine from model primary magmas having 85^95wt FeO and 153^175wt MgO will yield arrays of deriva-tive liquids and randomly sorted olivines that in most butnot all cases connect the picrites and high-MgO basalts Anewer version of the PRIMELT software (Herzberg ampAsimow 2008) can perform olivine addition to and sub-traction from lavas with highly variably olivine phenocrystcontents and yields results that converge with those shownin Fig 12
Fig 11 Relationship between abundance of olivine phenocrysts and whole-rock MgO contents for the Keogh Lake picrites (a) Tracings ofolivine grains (gray) in six high-MgO samples made from high resolution (2400 dpi) scans of petrographic thin-sections Scale bars represent10mm sample numbers are indicated at the upper left (b) Modal olivine vs whole-rock MgO Continuous line is the best-fit line for 10 high-MgO samplesThe areal proportion of olivine was calculated from raster images of olivine grains using ImageJ image analysis software whichprovides an acceptable estimation of the modal abundance (eg Chayes 1954)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
25
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
Gray lines = final melting pressure
L + Ol + Opx
07
08
09
Pressure (GPa)
10
3
4
5
6
1
SOLIDUS
01
04
0506
L + Ol
2
03
02
PeridotiteKR-4003
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
Thick black lines = initial melting pressure
Dashed lines = melt fraction
Melt Fraction
2
3 4 5 6
L+Ol+Opx
SolidusSpinel
SolidusGarnet Peridotite
7
L+Ol
0 10 20 30 4040
45
50
55
60
MgO (wt)
SiO2
(wt)
PeridotiteKR-4003
Melt Fraction
0 10 20 305
10
15
CaO(wt)
MgO (wt)
3
4 5 6 7
2
46
2
Peridotite Partial Melts
Thick black line = initial melting pressureGray lines = final melting pressure
Peridotite
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
02
Pressure (GPa)
0302
04
Pressure (GPa)
Picrite
High-MgO basalt
Tholeiitic basalt
(c)
(d)
Karmutsen flood basaltsOlivine addition modelOlivine addition (5 increment)Primary magma
PicriteHigh-MgO basaltTholeiitic basalt
(a)
(b)
800 850 900
910
920
930
940
950
Karmutsen flood basalts
Olivine addition modelOlivine addition (5 increment)Primary magma
Gray lines = Fo contents of olivine
Fig 12 Estimated primary magma compositions for three Keogh Lake picrites and high-MgO basalts (samples 4723A4 4723A135616A7) using the forward and inverse modeling technique of Herzberg et al (2007) Karmutsen compositions and modeling results areoverlain on diagrams provided by C Herzberg (a) Whole-rock FeO vs MgO for Karmutsen samples from this study Total iron estimatedto be FeO is 090 Gray lines show olivine compositions that would precipitate from liquid of a given MgO^FeO composition Black lineswith crosses show results of olivine addition (inverse model) using PRIMELT1 (Herzberg et al 2007) Gray field highlights olivinecompositions with Fo contents of 91^92 predicted to precipitate (b) FeO vs MgO showing Karmutsen lava compositions and results ofa forward model for accumulated fractional melting of fertile peridotite KR-4003 (Kettle River Peridotite Walter 1998) (c) SiO2 vsMgO for Karmtusen lavas and model results (d) CaO vs MgO showing Karmtusen lava compositions and model results To brieflysummarize the technique [see Herzberg et al (2007) for complete description] potential parental magma compositions for the high-MgOlava series were selected (highest MgO and appropriate CaO) and using PRIMELT1 software olivine was incrementally added tothe selected compositions to show an array of potential primary magma compositions (inverse model) Then using PRIMELT1 theresults from the inverse model were compared with a range of accumulated fractional melts for fertile peridotite derived fromparameterization of the experimental results for KR-4003 from Walter (1998) forward model Herzberg amp OrsquoHara (2002) A melt fractionwas sought that was unique to both the inverse and forward models (Herzberg et al 2007) A unique solution was found when there was a com-mon melt fraction for both models in FeO^MgO and CaO^MgO^Al2O3^SiO2 (CMAS) projection space This modeling assumes thatolivine was the only phase crystallizing and ignores chromite precipitation and possible augite fractionation in the mantle (Herzbergamp OrsquoHara 2002) Results are best for a residue of spinel lherzolite (not pyroxenite) The presence of accumulated olivine in samples ofthe Keogh Lake picrites used as starting compositions does not significantly affect the results because we are modeling addition of olivine(see text for further description) The tholeiitic basalts cannot be used for modeling because they are all saturated in plagthorn cpxthornolThick black line for initial melting pressure at 4 GPa is not shown in (c) for clarity Gray area in (a) indicates parental magmas thatwill crystallize olivine with Fo 91^92 at the surface Dashed lines in (c) indicate melt fraction Solidi for spinel and garnet peridotite arelabelled in (c)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
The PRIMELT1 results indicate that if the Karmutsenpicritic lavas were derived from accumulated fractionalmelting of fertile peridotite the primary magmas wouldhave contained 15^17wt MgO and 10wt CaO(Fig 12 Table 6) formed from 23^27 partial meltingand would have crystallized olivine with a Fo content of 91 (Fig 12 Table 6) Calculated mantle temperatures forthe Karmutsen picrites (14908C) indicate melting ofanomalously hot mantle (100^2008C above ambient) com-pared with ambient mantle that produces mid-ocean ridgebasalt (MORB 1280^14008C Herzberg et al 2007)which initiated between 3 and 4 GPa Several picritesappear to be near-primary melts and required very littleaddition of olivine (eg sample 93G171 with measured158wt MgO) These temperature and melting esti-mates of the Karmutsen picrites are consistent withdecompression of hot mantle peridotite in an actively con-vecting plume head (ie plume initiation model) Threesamples (picrite 5614A1 and high-MgO basalts 5614A3and 5614A5) are lower in iron than the other Karmutsenlavas with FeO in the 65^80wt range (Fig 12c) andhave MgO and FeO contents that are similar to primitiveMORB from the Sequeiros fracture zone of the EastPacific Rise (EPR Perfit et al 1996) These samples
however differ from EPR MORB in having higher SiO2
and lower Al2O3 and they are restricted geographicallyto one area (Sara Lake Fig 2)We therefore have evidence for primary magmas with
MgO contents that range from a possible low of 110wt to as high as 174wt with inferred mantle potential tem-peratures from 1340 to 15208C This is similar to the rangedisplayed by lavas from the Ontong Java Plateau deter-mined by Herzberg (2004) who found that primarymagma compositions assuming a peridotite sourcewould contain 17wt MgO and 10wt CaO(Table 6) and that these magmas would have formed by27 melting at a mantle potential temperature of15008C (first melting occurring at 36 GPa) Fitton ampGodard (2004) estimated 30 partial melting for theOntong Java lavas based on the Zr contents of primarymagmas of Kroenke-type basalt calculated by incrementaladdition of equilibrium olivine However if a considerableamount of eclogite was involved in the formation of theOntong Java plateau magmas the excess temperatures esti-mated by Herzberg et al (2007) may not be required(Korenaga 2005) The variability in mantle potential tem-perature for the Keogh Lake picrites is also similar to thewide range of temperatures that have been calculated for
Table 6 Estimated primary magma compositions for Karmutsen basalts and other oceanic plateaus or islands
Sample 4723A4 4723A13 5616A7 93G171 Average OJP Mauna Keay Gorgonaz
Ontong Java primary magma composition for accumulated fraction melting (AFM) from Herzberg (2004)yMauna Kea primary magma composition is average of four samples of Herzberg (2006 table 1)zGorgona primary magma composition for AFM for 1F fertile source from Herzberg amp OrsquoHara (2002 table 4)Total iron estimated to be FeO is 090
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
27
lavas from single volcanoes in the Galapagos Hawaii andelsewhere by Herzberg amp Asimow (2008) Those workersinterpreted the range to reflect the transport of magmasfrom both the cool periphery and the hotter axis of aplume A thermally heterogeneous plume will yield pri-mary magmas with different MgO contents and the rangeof primary magma compositions and their inferred mantlepotential temperatures for Karmutsen lavas may reflectmelting from different parts of the plume
Source of the Karmutsen Formation lavasThe Sr^Nd^Hf^Pb isotopic compositions of theKarmutsen volcanic rocks on Vancouver Island indicatethat they formed from plume-type mantle containing adepleted component that is compositionally similar to butdistinct from other oceanic plateaus that formed in thePacific Ocean (eg Ontong Java and Caribbean Fig 13)Trace element and isotopic constraints can be used to testwhether the depleted component of the KarmutsenFormation was intrinsic to the mantle plume and distinctfrom the source of MORB or the result of mixing or invol-vement of ambient depleted MORB mantle with plume-type mantle The Karmutsen tholeiitic basalts have initialeNd and
87Sr86Sr ratios that fall within the range of basaltsfrom the Caribbean Plateau (eg Kerr et al 1996 1997Hauff et al 2000 Kerr 2003) and a range of Hawaiian iso-tope data (eg Frey et al 2005) and lie within and abovethe range for the Ontong Java Plateau (Tejada et al 2004Fig 13a) Karmutsen tholeiitic basalts with the highest ini-tial eNd and lowest initial 87Sr86Sr lie within and justbelow the field for age-corrected northern EPR MORB at230 Ma (eg Niu et al1999 Regelous et al1999) Initial Hfand Nd isotopic compositions for the Karmutsen samplesare noticeably offset to the low side of the ocean islandbasalt (OIB) array (Vervoort et al 1999) and lie belowand slightly overlap the low eHf end of fields for Hawaiiand the Caribbean Plateau (Fig 13c) the Karmutsen sam-ples mostly fall between and slightly overlap a field forage-corrected Pacific MORB at 230 Ma and Ontong Java(Mahoney et al 1992 1994 Nowell et al 1998 Chauvel ampBlichert-Toft 2001) Karmutsen lava Pb isotope ratios over-lap the broad range for the Caribbean Plateau and thelinear trends for the Karmutsen samples intersect thefields for EPR MORB Hawaii and Ontong Java in208Pb^206Pb space but do not intersect these fields in207Pb^206Pb space (Fig 13d and e) The more radiogenic207Pb204Pb for a given 206Pb204Pb of the Karmutsenbasalts requires high UPb ratios at an ancient time when235U was abundant similar to the Caribbean Plateau andenriched in comparison with EPR MORB Hawaii andOntong JavaThe low eHf^low eNd component of the Karmutsen
basalts that places them below the OIB mantle array andpartly within the field for age-corrected Pacific MORBcan form from low ratios of LuHf and high SmNd over
time (eg Salters amp White 1998) MORB will develop aHf^Nd isotopic signature over time that falls below themantle array Low LuHf can also lead to low 176Hf177Hffrom remelting of residues produced by melting in theabsence of garnet (eg Salters amp Hart 1991 Salters ampWhite 1998) The Hf^Nd isotopic compositions of theKarmutsen Formation indicate the possible involvementof depleted MORB mantle however similar to Iceland(Fitton et al 2003) Nb^Zr^Y variations provide a way todistinguish between a depleted component within theplume and a MORB-like component The Karmutsenbasalts and picrites fall within the Iceland array and donot overlap the MORB field in a logarithmic plot of NbYvs ZrY (Fig 13b) using the fields established by Fittonet al (1997) Similar to the depleted component within theCaribbean Plateau (Thompson et al 2003) the trend ofHf^Nd isotopic compositions towards Pacific MORB-likecompositions is not followed by MORB values in Nb^Zr^Y systematics and this suggests that the depleted compo-nent in the source of the Karmutsen basalts is distinctfrom the source of MORB and probably an intrinsic partof the plume The relatively radiogenic Pb isotopic ratiosand REE patterns distinct from MORB also support thisinterpretationTrace-element characteristics suggest the possible invol-
vement of sub-arc lithospheric mantle in the generation ofthe Karmutsen picrites Lassiter et al (1995) suggested thatmixing of a plume-type source with eNdthorn 6 to thorn 7 witharc material with low NbTh could reproduce the varia-tions observed in the Karmutsen basalts however theyconcluded that the absence of low NbLa ratios in theirsample of tholeiitic basalts restricts the amount of arc litho-sphere involved The study of Lassiter et al (1995) wasbased on major and trace element and Sr Nd and Pb iso-topic data for a suite of 29 samples from Buttle Lake in theStrathcona Provincial Park on central Vancouver Island(Fig 1) Whereas there are no clear HFSE depletions inthe Karmutsen tholeiitic basalts the low NbLa (and TaLa) of the Karmutsen picritic lavas in this study indicatethe possible involvement of Paleozoic arc lithosphere onVancouver Island (Fig 8)
REE modeling dynamic melting andsource mineralogyThe combined isotopic and trace-element variations of thepicritic and tholeiitic Karmutsen lavas indicate that differ-ences in trace elements may have originated from differentmelting histories For example the LuHf ratios of theKarmutsen picritic lavas (mean 008) are considerablyhigher than those in the tholeiitic basalts (mean 002)however the small range of overlapping initial eHf valuesfor the two lava suites suggests that these differences devel-oped during melting within the plume (Fig 9b) Below wemodel the effects of source mineralogy and depth of
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
28
melting on differences in trace-element concentrations inthe tholeiitic basalts and picritesDynamic melting models simulate progressive decom-
pression melting where some of the melt fraction isretained in the residue and only when the degree of partialmelting is greater than the critical mass porosity and the
source becomes permeable is excess melt extracted fromthe residue (Zou1998) For the Karmutsen lavas evolutionof trace element concentrations was simulated using theincongruent dynamic melting model developed by Zou ampReid (2001) Melting of the mantle at pressures greater than1 GPa involves incongruent melting (eg Longhi 2002)
OIB array
PacificMORB(230 Ma)
(0 Ma)
EPR(230 Ma)
-4
-2
0
2
4
8
10
12
14
16
18
20
22
-4 -2 0 2 4 6 8 10 12 14
minus2
0
2
4
6
8
10
12
14
0702 0703 0704 0705 0706
IndianMORB
87Sr 86Sr (initial)
Hf (initial)
(a) (c)
Nd (initial)
Karmutsen tholeiitic basalt
HawaiiOntong JavaCaribbean Plateau
Karmutsen high-MgO lava
high-MgO lavasKarmutsen
1540
1545
1550
1555
1560
1565
175 180 185 190 195 200375
380
385
390
395
175 180 185 190 195 200
(d) (e)
EPR
EPR
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb
Karmutsen Formation (initial)
HawaiiOntong JavaCaribbean Plateau
East Pacific Rise
Karmutsen Formation (measured)
Juan de FucaGorda
ε Nd
(initial)
Karmutsen Formation (different age-correction for 3 picrites)
(0 Ma)
NbY
001
01
1
1 10
ZrY
OIB
NMORB
(b)
Iceland array
6
Fig 13 Comparison of age-corrected (230 Ma) Sr^Nd^Hf^Pb isotopic compositions for Karmutsen flood basalts onVancouver Island withage-corrected OIB and MORB and Nb^Zr^Y variation (a) Initial eNd vs 87Sr86Sr (b) NbYand ZrY variation (after Fitton et al 1997)(c) Initial eHf vs eNd (d) Measured and initial 207Pb204Pb vs 206Pb204Pb (e) Measured and initial 208Pb204Pb vs 206Pb204Pb References fordata sources are listed in Electronic Appendix 4 Most of the compiled data were extracted from the GEOROC database (httpgeorocmpch-mainzgwdgdegeoroc) OIB array line in (c) is fromVervoort et al (1999) EPR East Pacific Rise Several high-MgO samples affected bysecondary alteration were not plotted for clarity in Pb isotope plots Estimation of fields for age-corrected Pacific MORB and EPR at 230 Mawere made using parent^daughter concentrations (in ppm) from Salters amp Stracke (2004) of Lufrac14 0063 Hffrac14 0202 Smfrac14 027 Ndfrac14 071Rbfrac14 0086 and Srfrac14 836 In the NbYvs ZrYplot the normal (N)-MORB and OIB fields not including Icelandic OIB are from data com-pilations of Fitton et al (1997 2003) The OIB field is data from 780 basalts (MgO45wt ) from major ocean islands given by Fitton et al(2003) The N-MORB field is based on data from the Reykjanes Ridge EPR and Southwest Indian Ridge from Fitton et al (1997) The twoparallel gray lines in (b) (Iceland array) are the limits of data for Icelandic rocks
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
29
with olivine being produced during the reactioncpxthornopxthorn spfrac14meltthornol for spinel peridotites The mod-eling used coefficients for incongruent melting reactionsbased on experiments on lherzolite melting (eg Kinzleramp Grove 1992 Longhi 2002) and was separated into (1)melting of spinel lherzolite (2) two combined intervals ofmelting for a garnet lherzolite (gt lherz) and spinel lher-zolite (sp lherz) source (3) melting of garnet lherzoliteThe model solutions are non-unique and a range indegree of melting partition coefficients melt reactioncoefficients and proportion of mixed source componentsis possible to achieve acceptable solutions (see Fig 14 cap-tion for description of modeling parameters anduncertainties)The LREE-depleted high-MgO lavas require melting of
a depleted spinel lherzolite source (Fig 14a) The modelingresults indicate a high degree of melting (22^25) similarto the results from PRIMELT1 based on major-elementvariations (discussed above) of a LREE-depleted sourceThe high-MgO lavas lack a residual garnet signature Theenriched tholeiitic basalts involved melting of both garnetand spinel lherzolite and represent aggregate melts pro-duced from continuous melting throughout the depth ofthe melting column (Fig 14b) Trace-element concentra-tions in the tholeiitic and picritic lavas are not consistentwith low-degree melting of garnet lherzolite alone(Fig 14c) The modeling results indicate that the aggregatemelts that formed the tholeiitic basalts would haveinvolved lesser proportions of enriched small-degreemelts (1^6 melting) generated at depth and greateramounts of depleted high-degree melts (12^25) gener-ated at lower pressure (a ratio of garnet lherzolite tospinel lherzolite melt of 14 provides the best fits seeFig 14b and description in figure caption) The totaldegree of melting for the tholeiitic basalts was high(23^27) similar to the degree of partial melting in thespinel lherzolite facies for the primary melts that eruptedas the Keogh Lake picrites of a source that was also prob-ably depleted in LREE
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
(c)
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
Melting of garnet lherzolite
High-MgO lavas
Melting of spinel lherzolite
source (07DM+03PM)
source (07DM+03PM)
6 melting
30 melting
(b)
(a)
30 melting
Sam
ple
Cho
ndrit
eS
ampl
eC
hond
rite 20 melting
1 melting
Melting of garnet lherzoliteand spinel lherzolite
x gt lherz + 4x sp lherz
5 meltingx gt lherz + 4x sp lherz
Sam
ple
Cho
ndrit
e
Tholeiitic lavas
Tholeiitic lavas
High-MgO lavas
source (07DM+03PM)
Fig 14 Trace-element modeling results for incongruent dynamicmantle melting for picritic and tholeiitic lavas from the KarmutsenFormation Three steps of modeling are shown (a) melting of spinellherzolite compared with high-MgO lavas (b) melting of garnet lher-zolite and spinel lherzolite compared with tholeiitic lavas (c) meltingof garnet lherzolite compared with tholeiitic and picritic lavasShaded fields in each panel for tholeiitic basalts and picrites arelabeled Patterns with symbols in each panel are modeling results(patterns are 5 melting increments in (a) and (b) and 1 incre-ments in (c) PM primitive mantle DM depleted MORB mantle gtlherz garnet lherzolite sp lherz spinel lherzolite In (b) the ratios ofper cent melting for garnet and spinel lherzolite are indicated (x gtlherzthorn 4x sp lherz means that for 5 melting 1 melt in garnetfacies and 4 melt in spinel facies where x is per cent melting in theinterval of modeling) The ratio of the respective melt proportions inthe aggregate melt (x gt lherzthorn 4x sp lherz) was kept constant andthis ratio of melt contribution provided the best fit to the data Arange of proportion of source components (07DMthorn 03PM to
09DMthorn 01PM) can reproduce the variation in the high-MgOlavas Melting modeling uses the formulation of Zou amp Reid (2001)an example calculation is shown in their Appendix PM fromMcDonough amp Sun (1995) and DM from Salters amp Stracke (2004)Melt reaction coefficients of spinel lherzolite from Kinzler amp Grove(1992) and garnet lherzolite fromWalter (1998) Partition coefficientsfrom Salters amp Stracke (2004) and Shaw (2000) were kept constantSource mineralogy for spinel lherzolite (018cpx027opx052ol003sp) from Kinzler (1997) and for garnet lherzolite(034cpx008opx053ol005gt) from Salters amp Stracke (2004) Arange of source compositions for melting of spinel lherzolite(07DMthorn 03PM to 03DMthorn 07PM) reproduce REE patternssimilar to those of the high-MgO lavas with best fits for 22ndash25melting and 07DMthorn 03PM source composition Concentrationsof tholeiitic basalts cannot be reproduced from an entirelyprimitive or depleted source
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
30
The uncertainty in the composition of the source in thespinel lherzolite melting model for the high-MgO lavasprecludes determining whether the source was moredepleted in incompatible trace elements than the source ofthe tholeiitic lavas (see Fig 14 caption for description) It ispossible that early high-pressure low-degree melting left aresidue within parts of the plume (possibly the hotter inte-rior portion) that was depleted in trace elements (egElliott et al 1991) further decompression and high-degreemelting of such depleted regions could then have generatedthe depleted high-MgO Karmutsen lavas Shallow high-degree melting would preferentially sample depletedregions whereas deeper low-degree melting may notsample more refractory depleted regions (eg CaribbeanEscuder-Viruete et al 2007) The picrites lie at the highend of the range of Nd isotopic compositions and havelow NbZr similar to depleted Icelandic basalts (Fittonet al 2003) therefore the picrites may represent the partsof the mantle plume that were the most depleted prior tothe initiation of melting As Fitton et al (2003) suggestedthese depleted melts may only rarely be erupted withoutcombining with more voluminous less depleted melts andthey may be detected only as undifferentiated small-volume flows like the Keogh Lake picrites within theKarmutsen Formation on Vancouver IslandDecompression melting within the mantle plume initiatedwithin the garnet stability field (35^25 GPa) at highmantle potential temperatures (Tp414508C) and pro-ceeded beneath oceanic arc lithosphere within the spinelstability field (25^09 GPa) where more extensivedegrees of melting could occur
Magmatic evolution of Karmutsentholeiitic basaltsPrimary magmas in large igneous provinces leave exten-sive coarse-grained cumulate residues within or below thecrust these mafic and ultramafic plutonic sequences repre-sent a significant proportion of the original magma thatpartially crystallized at depth (eg Farnetani et al 1996)For example seismic and petrological studies of OntongJava (eg Farnetani et al 1996) combined with the use ofMELTS (Ghiorso amp Sack 1995) indicate that the volcanicsequence may represent only 40 of the total magmareaching the Moho Neal et al (1997) estimated that30^45 fractional crystallization took place for theOntong Java magmas and produced cumulate thicknessesof 7^14 km corresponding to 11^22 km of flood basaltsdepending on the presence of pre-existing oceanic crustand the total crustal thickness (25^35 km) Interpretationsof seismic velocity measurements suggest the presence ofpyroxene and gabbroic cumulates 9^16 km thick beneaththe Ontong Java Plateau (eg Hussong et al 1979)Wrangellia is characterized by high crustal velocities in
seismic refraction lines which is indicative of mafic
plutonic rocks extending to depth beneath VancouverIsland (Clowes et al 1995)Wrangellia crust is 25^30 kmthick beneath Vancouver Island and is underlain by astrongly reflective zone of high velocity and density thathas been interpreted as a major shear zone where lowerWrangellia lithosphere was detached (Clowes et al 1995)The vast majority of the Karmutsen flows are evolved tho-leiitic lavas (eg low MgO high FeOMgO) indicating animportant role for partial crystallization of melts at lowpressure and a significant portion of the crust beneathVancouver Island has seismic properties that are consistentwith crystalline residues from partially crystallizedKarmutsen magmas To test the proportion of the primarymagma that fractionated within the crust MELTS(Ghiorso amp Sack 1995) was used to simulate fractionalcrystallization using several estimated primary magmacompositions from the PRIMELT1 modeling results Themajor-element composition of the primary magmas couldnot be estimated for the tholeiitic basalts because of exten-sive plagthorn cpxthornol fractionation and thus the MELTSmodeling of the picrites is used as a proxy for the evolutionof major elements in the volcanic sequence A pressure of 1kbar was used with variable water contents (anhydrousand 02wt H2O) calculated at the quartz^fayalite^magnetite (QFM) oxygen buffer Experimental resultsfrom Ontong Java indicate that most crystallizationoccurs in magma chambers shallower than 6 km deep(52 kbar Sano amp Yamashita 2004) however some crys-tallization may take place at greater depths (3^5 kbarFarnetani et al 1996)A selection of the MELTS results are shown in Fig 15
Olivine and spinel crystallize at high temperature until15^20wt of the liquid mass has fractionated and theresidual magma contains 9^10wt MgO (Fig 15f) At1235^12258C plagioclase begins crystallizing and between1235 and 11908C olivine ceases crystallizing and clinopyr-oxene saturates (Fig 15f) As expected the addition of clin-opyroxene and plagioclase to the crystallization sequencecauses substantial changes in the composition of the resid-ual liquid FeO and TiO2 increase and Al2O3 and CaOdecrease while the decrease in MgO lessens considerably(Fig 15) The near-vertical trends for the Karmutsen tho-leiitic basalts especially for CaO do not match the calcu-lated liquid-line-of-descent very well which misses manyof the high-CaO high-Al2O3 low-FeO basalts A positivecorrelation between Al2O3CaO and SrNd indicates thataccumulation of plagioclase played a major role in the evo-lution of the tholeiitic basalts (Fig 15g) Increasing thecrystallization pressure slightly and the water content ofthe melts does not systematically improve the match ofthe MELTS models to the observed dataThe MELTS results indicate that a significant propor-
tion of the crystallization takes place before plagioclase(15^20 crystallization) and clinopyroxene (35^45
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
31
crystallization) join the crystallization sequence Theseresults suggest that445 crystallization of the primarymagmas occurred (estimated from the residual liquid percent in Fig 15f) and if the original flood basalt stratigra-phy on Vancouver Island was 6 km thick at least anequivalent amount of complementary plutonic rocksshould have crystallized The Karmutsen basalts and theirplutonic residues thus represent a significant addition ofcrust (perhaps 412 km thickness) previously thickenedduring the formation of the underlying Paleozoic arc
Growth of the Wrangellia oceanic plateauon Vancouver IslandThe extrusive part of the Wrangellia oceanic plateau onVancouver Island was constructed in a three-layersequence of mostly tholeiitic basalts with a restrictedrange of composition (Fig 16) The plateau overliesDevonian to Mississippian arc volcanic sequences andmarine sediments of Mississippian to Permian ages thatare intruded by gabbroic rocks related to the flood basaltsThe volcanic stratigraphy of theWrangellia plateau formed
00
05
10
15
20
25
30
35
40
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
16
0 2 4 6 8 10 12 14 16 18 20
10
11
12
13
14
15
16
17
18
19
20
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
0 2 4 6 8 10 12 14 16 18 20
MgO (wt)
0 10 20 30 40 50
percent of total mass
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
) Mineral proportions
Sp (e)
Ol
CpxPlag
PicriteHigh-MgO basaltTholeiitic basalt
TiO2 (wt)
MgO (wt)
Al2O3 (wt) CaO (wt)
FeO (wt)
MgO(a) (b)
(c) (d)
MgO (wt)MgO (wt)
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
)
Residual liquid ()
1220
1190
Ol + Sp
Plag+Ol+Sp Plag+Cpx+Sp
02 wt H2O
no H2O
Residual liquid and crystallizing phases
(f)
0102030405060708090100
plagioclaseaccumulation
plagioclaseaccumulation
plagioclase
accumulatio
n
10
12
14
16
0 10 20 30 40
Al2O3CaO
SrNd
tholeiitic basalts
(g)
Fig 15 Forward fractional crystallization modeling results for major elements from MELTS (Ghiorso amp Sack 1995) compared with picriticand tholeiitic lavas from the Karmutsen Formation (a^d) Major-element compositions of residual liquids compared with Karmutsen lavacompositions For clarity the results are shown only for one starting composition (open circle) using the composition of the estimated primarymagma for the high-MgO lava series from the modeling technique of Herzberg et al (2007) for sample 4723A13 as shown in Fig 12 The esti-mated primary magmas for samples 4723A3 and 93G171 yield similar results A pressure of 1 kbar was used with no H2O and 02wt H2Ocalculated at QFM black line is anhydrous and grey line is 02wt H2OThe end-point for the anhydrous result is at 78 fractionation and11308C and the end-point for the result with 02wt H2O is 75 fractionation and11108C (e) Mineral proportions of crystallized phases fromone MELTrun (sample 4723A13) with no H2O (f) Per cent residual liquid vs temperature labelled with intervals of crystallizing phases (g)Al2O3CaO vs SrNd for the tholeiitic basalts
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
32
in a single eruptive phase (c 231^225 Ma) as an emergentbasalt sequence during a deeper-water shallow-water andsubaerial stage and subsided below sea level relativelysoon after its formation (Fig 16) The deeper-water stagecontains pillowed and massive submarine flows (3000mthick) that built up the submarine volcanic edifice from theseafloor The lowest pillow basalts were emplaced onunconsolidated and lithified fine-grained siliciclastic andcarbonaceous sediments containing Daonella beds(c 235^232 Ma) that are intruded by abundantKarmutsen mafic sills (Fig 16a) The different pillowedand massive submarine flows resulted primarily from dif-ferent effusive rates and local topography Some of themassive submarine flows commonly recognized by radiat-ing columnar jointing represent master tubes for the deliv-ery of lava to distal parts of submarine flow fields Thepillowed flows which commonly have uneven topography
in modern sequences may have served as levees to themassive submarine flows rather than allowing unob-structed spreading and inflation typical of the massive sub-aerial flows that form on shallow slopes in continentalflood basalt provinces (eg Self et al 1997) The rapid suc-cession of flows is evident by the near absence of sedimentbetween submarine flows Minor volumes of picritic pillowlavas occur in the upper part of the pillowed flow sequencein areas of the plateau exposed on northern VancouverIsland (Fig 16 Nixon et al 2008)The growth of the plateau in shallow water led to
emplacement of an increasing proportion of volcaniclastics(pillow breccias and hyaloclastites) that accumulated up to1500m thick in some areas (Fig 16b Nixon et al 2008)Horizons of pillowed flows occur locally within the volca-niclastic stratigraphy Sedimentary structures (eg gradedbedding fluidization structures) and petrographic textures
Late Devonian-Mississippian arc
Daonella beds
Middle Triassic
deeper water
Late Devonian-Mississippian arc
pillowed flows massive submarine
flows
pillow brecciahyaloclastiteshallow water
~4 km
~6 km
Late Triassic limestoneinterflow limestonelenses
Karmutsen Formation
mafic intrusions sediment-sill complex
Karmutsen Formation
(a) (b)
(c) (d)
subaerial flows
Late Devonian-Mississippian arc Late Devonian-Mississippian arc
picritic pillow basalts
Miss to Permian marine sediments
sills
Fig 16 Growth of the Wrangellia oceanic plateau onVancouver Island (a) Deep-water eruptive phase of pillowed and massive submarineflows with almost no intervening sediments onto MiddleTriassic shale containing Daonella beds (c 235^232 Ma) The basement of the plateauis composed of Devonian to Mississippian arc sequences overlain by Mississippian to Permian carbonates and siliciclastic sediments(b) Shallow-water flow emplacement with increasing proportion of volcaniclastics resulted primarily from cooling^contraction granulationPicritic pillow basalts occur primarily near the transition between pillowed submarine flows and volcaniclastic units Pillowed and massiveflows occur locally within the volcaniclastic unit (c) Emergent stage of growth consists of overlapping subaerial flow fields similar to continen-tal flood basalts with pahoehoe flow structures Multi-tiered columnar jointing occurrs locally in response to ponding of flows (eg Bondre et al2004) Interflow carbonate deposits and plagioclase megacrystic flows developed locally during the waning stage of volcanism (d) Post-volcanicsubsidence and deposition of Late Carnian to Norian (c 212^223 Ma) platform carbonates on top of the plateau Units are labeled in the dia-grams Thicknesses of the underlying units are not to scale
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
33
within the volcaniclastic rocks indicate formation primar-ily via cooling^contraction granulation magma^water^steam interaction autobrecciation resedimentation andmass-wasting rather than pyroclastic eruption (Nixonet al 2008)A broad subaerial platform was constructed by emplace-
ment of subaerial lava flows during an emergent subaerialstage (Fig 16c) The subaerial unit predominantly com-prises well-layered sheet flows that formed in a similarway to continental flood basalts as hundred(s) of pahoehoeflow fields erupted on shallow slopes (Self et al 1997) how-ever volcaniclastic and pillowed flow units are locally pre-sent mostly in the upper parts of the stratigraphyLocalized ponding of flows led to the formation of multi-tiered columnar jointing in some subaerial flows (egBondre et al 2004) As volcanism waned local interflowcarbonate deposits developed and plagioclase megacrysticflows were erupted Subsidence of the plateau is recordedby the interflow lenses and deposits of platform carbonates(40^1500m thick) on top of the Karmutsen basalts thatcontain Late Carnian to Early Norian fossils (Fig 16d c216^227 Ma Nixon et al 2008)The evolution of the internal architecture of the
Wrangellia oceanic plateau was primarily related to theconditions of emplacement (ie water depth and eruptionenvironment) There does not appear to have been a sys-tematic coincident evolution or shift of lava compositionsonVancouver Island as has been recognized in most con-tinental flood basalts and parts of the Wrangellia oceanicplateau in Alaska and Yukon (Greene et al 2008 2009)In general with the exception of the picritic pillow lavasin the submarine stratigraphy and local variations betweenhigh-MgO and tholeiitic lavas there are no significantcompositional differences between lavas erupted duringthe different stages of growth of the plateau (ie submarineand subaerial flows) nor are substantial compositional dif-ferences generally found between massive submarine flowsin contact with pillowed flows tholeiitic basalt flows nearthe base of the Karmutsen Formation are compositionallysimilar to flows sampled near the top of the volcanicsequence Where it was possible to sample continuous sec-tions of volcanic stratigraphy greater than several hun-dreds of meters (eg Keogh^Maynard Lake area west ofthe Karmutsen Range Fig 2) there is no clear relationshipbetween stratigraphic position and compositionalvariation
CONCLUSIONSThe Karmutsen Formation covers extensive areas ofVancouver Island (20 000 km2) and forms part of a largeoceanic plateau erupted over a geologically short interval(c 225^231 Ma) The tripartite volcanic stratigraphyexposed on Vancouver Island is at least 6 km thick andcomprises a succession of submarine flows volcaniclastic
deposits and massive subaerial flowsThese volcanologicaldifferences are primarily related to the eruption environ-ment (deep-water shallow-water or subaerial) The rapidgrowth of the plateau prevented significant accumulationsof intervening sediment except in the uppermost stratigra-phy where isolated limestone lenses commonly associatedwith pillow basalt and hyaloclastite deposits preserve arecord of the final stages of subsidence of the plateauThe Wrangellia plateau on Vancouver Island was con-
structed dominantly of tholeiitic basalt with restrictedmajor- and trace-element and isotopic compositional var-iations Minor volumes of picritic pillow basalts eruptedlate in the submarine phase The Karmutsen basalts wereemplaced onto older Paleozoic arc volcanic sequences andlow NbLa for the picrites may indicate a contributionfrom the sub-arc lithospheric mantle Modeling of the pet-rogenesis of the Karmutsen picrites indicates melting ofanomalously hot mantle (15008C) and extensive degreesof partial melting (23^27) and is consistent with aplume initiation model Combined Sr^Nd^Hf^Pb isotopicsystematics indicate that the picritic and tholeiitic lavaswere derived from a depleted plume-type mantle sourcedistinct from the source of MORB that has compositionalsimilarities to the source of the lavas forming theCaribbean Plateau Picrites were generated by melting ofspinel lherzolite whereas the tholeiitic basalts formed asaggregate melts of both garnet and spinel lherzolite Thetholeiitic basalts underwent extensive low-pressure fractio-nation (52^3 kbar) and seismic studies indicate that someof theWrangellia crust beneathVancouver Island may cor-respond to the plutonic residues of the Karmutsen floodbasalts
ACKNOWLEDGEMENTSWe are grateful to Nick Arndt for helping us get this pro-ject started and Nick Massey for insights into VancouverIsland geology We greatly appreciate the assistance withPRIMELT modeling from Claude Herzberg We thankMikkel Schau for his insight and enthusiasm during field-work Jane Barling assisted with analyses by MC-ICP-MSThoughtful and thorough reviews of the manuscript byJohn Mahoney Fred Frey Wendy Bohrson DougalJerram and Teal Riley are greatly appreciated Fundingwas generously provided by the Rocks to Riches Programadministered by the BC amp Yukon Chamber of Mines in2004 by the BC Geological Survey and by NSERCDiscovery Grants to JSS and DW ARG was supportedby a University Graduate Fellowship at UBC
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
34
REFERENCESBarker F Brown A S Budahn J R amp Plafker G (1989) Back-arc
with frontal-arc component origin of Triassic Karmutsen basaltBritish Columbia Canada Chemical Geology 75 81^102
Blichert-Toft JWeis D Maerschalk C Agranier A amp Albare de F(2003) Hawaiian hot spot dynamics as inferred from the Hf and Pbisotope evolution of Mauna Kea volcano Geochemistry GeophysicsGeosystems 4(2) 1^27 doi1010292002GC000340
Bondre N R Duraiswami R A amp Dole G (2004) Morphologyand emplacement of flows from the Deccan Volcanic ProvinceIndia Bulletin of Volcanology 66 29^45
Carlisle D (1963) Pillow breccias and their aquagene tuffs QuadraIsland British Columbia Journal of Geology 71 48^71
Carlisle D (1972) Late Paleozoic to Mid-Triassic sedimentary-volcanic sequence on Northeastern Vancouver Island Geological
Society of Canada Paper Report of Activities 72-1 Part B 24^30Carlisle D amp SuzukiT (1974) Emergent basalt and submergent car-
bonate^clastic sequences including the Upper Triassic Dilleri andWelleri zones on Vancouver Island Canadian Journal of Earth
Sciences 11 254^279Chauvel C amp Blichert-Toft J (2001) A hafnium isotope and trace
element perspective on melting of the depleted mantle Earth and
Planetary Science Letters 190 137^151Chayes F (1954)The theory of thin-section analysis Journal of Geology
62 92^101Cho M Liou J G amp Maruyama S (1987) Transition from the zeo-
lite to prehnite^pumpellyite facies in the Karmutsen metabasitesVancouver Island British Columbia Journal of Petrology 27(2)467^494
Clowes R M Zelt C A Amor J R amp Ellis R M (1995)Lithospheric structure in the southern Canadian Cordillera froma network of seismic refraction lines Canadian Journal of Earth
Sciences 32(10) 1485^1513Coffin M F amp Eldholm O (1994) Large igneous provinces Crustal
structure dimensions and external consequences Reviews of
Geophysics 32(1) 1^36Cox K G (1980) A model for flood basalt volcanism Journal of
Petrology 21 629^650Eldholm O amp Coffin M F (2000) Large igneous provinces
and plate tectonics In Richards M A Gordon R G amp vander Hilst R D (eds) The History and Dynamics of Global
Plate Motions American Geophysical Union Geophysical Monograph 121309^326
Elliott T R Hawkesworth C J amp Grolaquo nvold K (1991) Dynamicmelting of the Iceland plume Nature 351 201^206
Escuder-Viruete J Pecurren rez-Estaucurren n A Contreras F Joubert MWeis D Ullrich T D amp Spadea P (2007) Plume mantle sourceheterogeneity through time Insights from the Duarte ComplexHispaniola northeastern Caribbean Journal of Geophysical Research112(B04203) doi1010292006JB004323
Farnetani C G Richards M A amp Ghiorso M S (1996)Petrological models of magma evolution and deep crustal structurebeneath hotspots and flood basalt provinces Earth and Planetary
Science Letters 143 81^94Fitton J G amp Godard M (2004) Origin and evolution of magmas
on the Ontong Java Plateau In Fitton J G Mahoney J JWallace P J amp Saunders A D (eds) Origin and Evolution of the
Ontong Java Plateau Geological Society London Special Publications 229151^178
Fitton J G Saunders A D Norry M J Hardarson B S ampTaylor R N (1997) Thermal and chemical structure of theIceland plume Earth and Planetary Science Letters 153 197^208
Fitton J G Saunders A D Kempton P D amp Hardarson B S(2003) Does depleted mantle form an intrinsic part of the Icelandplume Geochemistry Geophysics Geosystems 4(3) 1032 doi1010292002GC000424
Frey F A Huang S Blichert-Toft J Regelous M amp Boyet M(2005) Origin of depleted components in basalt related to theHawaiian hotspot Evidence from isotopic and incompatible ele-ment ratios Geochemistry Geophysics Geosystems 6(1) 23 doi1010292004GC000757
Galer S J G amp Abouchami W (1998) Practical application of leadtriple spiking for correction of instrumental mass discriminationMineralogical Magazine 62A 491^492
Ghiorso M S amp Sack R O (1995) Chemical mass transfer in mag-matic processes IV A revised and internally consistent thermody-namic model for the interpolation and extrapolation of liquid^solidequilibria in magmatic systems at elevated temperatures and pres-sures Contributions to Mineralogy and Petrology 119 197^212
Greene A R Scoates J S amp Weis D (2005)WrangelliaTerrane onVancouver Island Distribution of flood basalts with implicationsfor potential Ni^Cu^PGE mineralization In Grant B (ed)Geological Fieldwork 2004 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2005-1 209^220Greene A R Scoates J S Nixon G T amp Weis D (2006) Picritic
lavas and basal sills in the Karmutsen flood basalt provinceWrangellia northern Vancouver Island In Grant B (ed)Geological Fieldwork 2005 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2006-1 39^52Greene A R Scoates J S amp Weis D (2008) Wrangellia flood
basalts in Alaska A record of plume^lithosphere interaction in aLate Triassic accreted oceanic plateau Geochemistry Geophysics
Geosystems 9(Q12004) doi1010292008GC002092Greene A R Scoates J S Weis D amp Israel S (2009)
Geochemistry of triassic flood basalts from the Yukon (Canada)segment of the accreted Wrangellia oceanic plateau Lithosdoi101016jlithos200811010
Hauff F Hoernle K Tilton G Graham D W amp Kerr A C(2000) Large volume recycling of oceanic lithosphere over shorttime scales geochemical constraints from the Caribbean LargeIgneous Province Earth and Planetary Science Letters 174 247^263
Hauff F Hoernle K amp Schmidt A (2003) Sr^Nd^Pb compositionof Mesozoic Pacific oceanic crust (Site 1149 and 801 ODP Leg185)Implications for alteration of ocean crust and the input into theIzu^Bonin^Mariana subduction system Geochemistry Geophysics
Geosystems 4(8) doi1010292002GC000421Herzberg C (2004) Partial melting below the Ontong Java Plateau
In Fitton J G Mahoney J J Wallace P J amp Saunders A D(eds) Origin and Evolution of the Ontong Java Plateau Geological SocietyLondon Special Publications 229 179^183
Herzberg C (2006) Petrology and thermal structure of the Hawaiianplume from Mauna Kea volcano Nature 444(7119) 605
Herzberg C amp Asimow P D (2008) Petrology of some oceanicisland basalts PRIMELT2XLS software for primary magma cal-culation Geochemistry Geophysics Geosystems 9(Q09001) doi1010292008GC002057
Herzberg C amp OrsquoHara M J (2002) Plume-associated ultramaficmagmas of Phanerozoic age Journal of Petrology 43(10) 1857^1883
Herzberg C Asimow P D Arndt N Niu Y Lesher C MFitton J G Cheadle M J amp Saunders A D (2007)Temperatures in ambient mantle and plumes Constraints frombasalts picrites and komatiites Geochemistry Geophysics Geosystems8(Q02006) doi1010292006GC001390
Huang S amp Frey F A (2005) Trace element abundances of MaunaKea basalt from phase 2 of the Hawaii Scientific Drilling Project
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
35
Petrogenetic implications of correlations with major element con-tent and isotopic ratios Geochemistry Geophysics Geosystems 4(6)doi1010292002GC000322
Hussong D M Wipperman L K amp Kroenke L W (1979) Thecrustal structure of the Ontong Java and Manihiki OceanicPlateaus Journal of Geophysical Research 84(B11) 6003^6010
Jerram D A amp Widdowson M (2005) The anatomy of ContinentalFlood Basalt Provinces geological constraints on the processes andproducts of flood volcanism Lithos 79(3^4) 385^405
Jones D L Silberling N J amp Hillhouse J (1977)Wrangellia a dis-placed terrane in northwestern North America CanadianJournal ofEarth Sciences 14(11) 2565^2577
Kerr A C (2003) Oceanic Plateaus In Rudnick R (ed) The Crust
Treatise on GeochemistryVol 3 Oxford Elsevier Science pp 537^565Kerr A C amp Mahoney J J (2007) Oceanic plateaus Problematic
plumes potential paradigms Chemical Geology 241 332^353Kerr A CTarney J Marriner G F Klaver GT Saunders A D
amp Thirlwall M F (1996) The geochemistry and petrogenesis ofthe Late-Cretaceous picrites and basalts of Curacao NetherlandsAntilles a remnant of an oceanic plateau Contributions to
Mineralogy and Petrology 124(1) 29^43Kerr A C Marriner G F Tarney J Nivia A Saunders A D
Thirlwall M F amp Sinton A C (1997) Cretaceous basaltic ter-ranes in Western Colombia Elemental chronological and Sr-Ndisotopic constraints on petrogenesis Journal of Petrology 38(6)677^702
Kinzler R J (1997) Melting of mantle peridotite at pressuresapproaching the spinel to garnet transition Application to mid-ocean ridge basalt petrogenesis Journal of Geophysical Research
102(B1) 853^874Kinzler R J amp Grove T L (1992) Primary magmas of mid-ocean
ridge basalts 1 Experiments and methods Journal of Geophysical
Research 97(B5) 6885^6906Korenaga J (2005) Why did not the Ontong Java Plateau form sub-
aerially Earth and Planetary Science Letters 234 385^399Kuniyoshi S (1972) Petrology of the Karmutsen (Volcanic) Group
northeastern Vancouver Island British Columbia PhD disserta-tion University of California Los Angeles 242 pp
Lassiter J C DePaolo D J amp Mahoney J J (1995) Geochemistry ofthe Wrangellia flood basalt province Implications for the role ofcontinental and oceanic lithosphere in flood basalt genesis Journalof Petrology 36(4) 983^1009
LeBas M J (2000) IUGS reclassification of the high-Mg and picriticvolcanic rocks Journal of Petrology 41(10) 1467^1470
Longhi J (2002) Some phase equilibrium systematics of lherzolitemelting I Geochemistry Geophysics Geosystems 3(3) doi1010292001GC000204
MacDonald G A amp Katsura T (1964) Chemical composition ofHawaiian lavas Journal of Petrology 5 82^133
Mahoney J J (1987) An isotopic survey of Pacific oceanic plateausimplications for their nature and origin In Keating B HFryer P Batiza R amp Boehlert GW (eds) Seamounts Islands andAtolls American Geophysical Union Geophysical Monograph 43 207^220
Mahoney J LeRoex A P Peng Z Fisher R L amp Natland J H(1992) Southwestern limits of Indian Ocean Ridge mantle and theorigin of low 206Pb204Pb mid-ocean ridge basalt isotope systema-tics of the central Southwest Indian Ridge (178^508E) Journal ofGeophysical Research 97(B13) 19771^19790
Mahoney J J Sinton J M Kurz M D MacDougall J DSpencer K J amp Lugmair GW (1994) Isotope and trace elementcharacteristics of a super-fast spreading ridge East Pacific rise13^238S Earth and Planetary Science Letters 121 173^193
Massey N W D (1995a) Geology and mineral resources of theAlberni^Nanaimo Lakes sheet Vancouver Island 92F1W 92F2Eand part of 92F7E British Columbia Ministry of Energy Mines and
Petroleum Resources Paper 1992-2 132 ppMassey N W D (1995b) Geology and mineral resources of the
Cowichan Lake sheet Vancouver Island 92C16 British Columbia
Ministry of Energy Mines and Petroleum Resources Paper 1992-3 112 ppMassey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005a) Digital Geology Map of British Columbia Tile NM9 MidCoast BC British Columbia Ministry of Energy and Mines Geofile
2005-2Massey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005b) Digital Geology Map of British Columbia Tile NM10Southwest BC British Columbia Ministry of Energy and Mines Geofile
2005-3McDonough W F amp Sun S (1995) The composition of the Earth
Chemical Geology 120 223^253Monger JW H amp Journeay J M (1994) Basement geology and tec-
tonic evolution of theVancouver region In Monger JW H (ed)Geology and Geological Hazards of the Vancouver Region Southwestern
British Columbia Geological Survey of Canada Bulletin 481 3^25Muller J E (1977) Geology of Vancouver Island Geological Survey of
Canada Open File Map 463Muller J E (1980)The Paleozoic Sicker Group of Vancouver Island British
Columbia Geological Survey of Canada Paper 79-30 22 ppMuller J E Northcote K E amp Carlisle D (1974) Geology and
mineral deposits of Alert Bay^Cape Scott map area VancouverIsland British Columbia Geological Survey of Canada Paper 74-8 77
Neal C R Mahoney J J Kroenke L W Duncan R A ampPetterson M G (1997) The Ontong^Java Plateau InMahoney J J amp Coffin M F (eds) Large Igneous Provinces
Continental Oceanic and Planetary FloodVolcanism American Geophysical
Union Geophysical Monograph 100 183^216Niu Y Collerson K D Batiza R Wendt J I amp Regelous M
(1999) Origin of enriched-typed mid-ocean ridge basalt at ridgesfar from mantle plumes The East Pacific Rise at 11820rsquoN Journalof Geophysical Research 104 7067^7088
Nixon G T amp Orr A J (2007) Recent revisions to the EarlyMesozoic stratigraphy of Northern Vancouver Island (NTS 102I092L) and metallogenic implicatinos British Columbia InGrant B (ed) Geological Fieldwork 2006 British Columbia Ministry of
Energy Mines and Petroleum Resources Paper 2007-1 163^177Nixon GT Hammack J L KoyanagiV M Payie G J Haggart
JW Orchard M JTozerT Friedman R M Archibald D APalfy J amp Cordey F (2006a) Geology of the Quatsino^PortMcNeill area northernVancouver Island British Columbia Ministry
of Energy Mines and Petroleum Resources Geoscience Map 2006-2 scale150 000
Nixon G T Kelman M C Stevenson D Stokes L A ampJohnston K A (2006b) Preliminary geology of the Nimpkishmap area (NTS 092L07) northern Vancouver Island BritishColumbia In Grant B amp Newell J M (eds) Geological Fieldwork2005 British Columbia Ministry of Energy Mines and Petroleum Resources
Paper 2006-1 135^152Nixon G T Laroque J Pals A Styan J Greene A R amp
Scoates J S (2008) High-Mg lavas in the Karmutsen flood basaltsnorthernVancouver Island (NTS 092L) Stratigraphic setting andmetallogenic significance In Grant B (ed) Geological Fieldwork
2007 British Columbia Ministry of Energy Mines and Petroleum
Resources Paper 2008-1 175^190Nowell G M Kempton P D Noble S R Fitton J G
Saunders A Mahoney J J amp Taylor R N (1998) High precisionHf isotope measurements of MORB and OIB by thermal ionisation
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
36
mass spectrometry insights into the depleted mantle Chemical
Geology 149 211^233Parrish R R amp McNicoll V J (1992) U^Pb age determinations
from the southern Vancouver Island area British Columbia InRadiogenic Age and Isotopic Studies Report 5 Geological Survey of
Canada Paper 91-2 79^86Peate DW (1997) The Parana^Etendeka Province In Coffin M F
amp Mahoney J (eds) Large Igneous Provinces Continental Oceanic andPlanetary Flood Volcanism American Geophysical Union Geophysical
Monograph 100 217^245Perfit M R Fornari D J Ridley W I Kirk P D Casey J
Kastens K A Reynolds J R Edwards M Desonie DShuster R amp Paradis S (1996) Recent volcanism in the Siqueirostransform fault picritic basalts and implications for MORBmagma genesis Earth and Planetary Science Letters 141(1^4) 91^108
Pretorius W Weis D Williams G Hanano D Kieffer B ampScoates J S (2006) Complete trace elemental characterization ofgranitoid (USGSG-2GSP-2) reference materials by high resolutioninductively coupled plasma-mass spectrometry Geostandards and
Geoanalytical Research 30(1) 39^54Regelous M Niu Y Wendt J I Batiza R Greig A amp
Collerson K D (1999) Variations in the geochemistry of magma-tism on the East Pacific Rise at 10830rsquoN since 800 ka Earth and
Planetary Science Letters 168 45^63Recurren villon S Chauvel C Arndt N T Pik R Martineau F
Fourcade S amp Marty B (2002) Heterogeneity of the Caribbeanplateau mantle source Sr O and He isotopic compositions of oli-vine and clinopyroxene from Gorgona Island Earth and Planetary
Science Letters 205(1^2) 91Rhodes J M (1996) Geochemical stratigraphy of lava flows samples
by the Hawaii Scientific Drilling Project Journal of Geophysical
Research 101(B5) 11729^11746Rhodes J M amp Vollinger M J (2004) Composition of basaltic lavas
sampled by phase-2 of the Hawaii Scientific Drilling ProjectGeochemical stratigraphy and magma types Geochemistry
Geophysics Geosystems 5(3) doi1010292002GC000434Richards M A Jones D L Duncan R A amp DePaolo D J (1991)
A mantle plume initiation model for the Wrangellia flood basaltand other oceanic plateaus Science 254 263^267
Salters V J M amp Hart S R (1991) The mantle sources of oceanridges islands and arcs the Hf-isotope connection Earth and
Planetary Science Letters 104 364^380Salters V J M amp Stracke A (2004) Composition of the depleted
SaltersV J amp WhiteW (1998) Hf isotope constraints on mantle evo-lution Chemical Geology 145 447^460
SanoT amp Yamashita S (2004) Experimental petrology of basementlavas from Ocean Drilling Program Leg 192 implications for dif-ferentiation processes in Ontong Java Plateau magmas InFitton J G Mahoney J JWallace P J amp Saunders A D (eds)Origin and Evolution of the Ontong Java Plateau Geological Society
London Special Publications 229 185^218Saunders A D (2005) Large igneous provinces Origin and environ-
mental consequences Elements 1 259^263Self S ThordarsonT amp Keszthely L (1997) Emplacement of conti-
nental flood basalt lava flows In Mahoney J J amp Coffin M F(eds) Large Igneous Provinces Continental Oceanic and Planetary Flood
Volcanism American Geophysical Union Geophysical Monograph 100381^410
Shaw D M (2000) Continuous (dynamic) melting theory revisitedCanadian Mineralogist 38(5) 1041^1063
Sluggett C L (2003) Uranium^lead age and geochemical constraintson Paleozoic and Early Mesozoic magmatism in WrangelliaTerrane Saltspring Island British Columbia BSc thesisUniversity of British ColumbiaVancouver 84 pp
Storey M Mahoney J J amp Saunders A D (1997) Cretaceousbasalts in Madagascar and the transition between plume and con-tinental lithospheric mantle sources In Mahoney J J ampCoffin M F (eds) Large Igneous Provinces Continental Oceanic and
Planetary Flood Volcanism Aamerican Geophysical Union Geophysical
Monograph 100 95^122Surdam R C (1967) Low-grade metamorphism of the Karmutsen
Group PhD dissertation University of California Los Angeles288 pp
Sutherland-Brown A Yorath C J Anderson R G amp Dom K(1986) Geological maps of southern Vancouver IslandLITHOPROBE I 92C10 11 14 16 92F1 2 7 8 Geological Survey ofCanada Open File 1272
Tejada M L G Mahoney J J Castillo P R Ingle S P Sheth HC amp Weis D (2004) Pin-pricking the elephant evidence on theorigin of the Ontong Java Plateau from Pb^Sr^Hf^Nd isotopiccharacteristics of ODP Leg 192 basalts In Fitton J GMahoney J J Wallace P J amp Saunders A D (eds) Origin and
Evolution of the Ontong Java Plateau Geological Society London Special
Publications 229 133^150Thompson P M E Kempton P D amp Kerr A C (2008) Evaluation
of the effects of alteration and leaching on Sm^Nd and Lu^Hf sys-tematics in submarine mafic rocks Lithos 104(1^4) 164^176
Thompson P M E Kempton P D White R V Kerr A CTarney J Saunders A D Fitton J G amp McBirney A (2003)Hf-Nd isotope constraints on the origin of the CretaceousCaribbean plateau and its relationship to the Galapagos plumeEarth and Planetary Science Letters 217(1^2) 59^75
Vervoort J D Patchett P Blichert-Toft J amp Albare de F (1999)Relationships between Lu^Hf and Sm^Nd isotopic systems in theglobal sedimentary system Earth and Planetary Science Letters 16879^99
Walter M J (1998) Melting of garnet peridotite and the origin ofkomatiite and depleted lithosphere Journal of Petrology 39(1) 29^60
Weis D Kieffer B Maerschalk C Barling J de Jong JWilliams G A Hanano D Mattielli N Scoates J SGoolaerts A Friedman R A amp Mahoney J B (2006) High-pre-cision isotopic characterization of USGS reference materials byTIMS and MC-ICP-MS Geochemistry Geophysics Geosystems
7(Q08006) doi1010292006GC001283Weis D Kieffer B Hanano D Silva I N Barling J
Pretorius W Maerschalk C amp Mattielli N (2007) Hf isotopecompositions of US Geological Survey reference materialsGeochemistry Geophysics Geosystems 8(Q06006) doi1010292006GC001473
Wheeler J O amp McFeely P (1991) Tectonic assemblage map of theCanadian Cordillera and adjacent part of the United States ofAmerica Geological Survey of Canada Map 1712A
WhiteW M Albare de F amp Tecurren louk P (2000) High-precision analy-sis of Pb isotope ratios by multi-collector ICP-MS Chemical Geology167 257^270
Yorath C J Sutherland Brown A amp Massey N W D (1999)LITHOPROBE southern Vancouver Island British Columbia Geological
Survey of Canada Bulletin 498 145 ppZou H (1998) Trace element fractionation during modal and
non-model dynamic melting and open-system melting Amathematical treatment Geochimica et Cosmochimica Acta 62(11)1937^1945
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
37
Zou H amp Reid M R (2001) Quantitative modeling of trace elementfractionation during incongruent dynamic melting Geochimica et
Cosmochimica Acta 65(1) 153^162
APPENDIX SAMPLEPREPARAT ION ANDANALYT ICAL METHODSOnly the freshest rocks in the field were sampled and onlythe least altered samples were selected for chemical andisotopic analysis based on thorough petrographic inspec-tion Fifty-nine of the 129 collected samples were crushed(400 g) into small pieces 52mm in diameter in aRocklabs hydraulic piston crusher between WC plates tominimize contamination The coarse-crush was mixedand 100 g was powdered in a planetary mill using agatejars and balls cleaned with quartz sand between samples
ActLabs analytical methodsThe major- and trace-element compositions of the whole-rock powders were determined at Activation LaboratoriesLtd (Actlabs) in Ancaster Ontario Analytical techniquesand detection limits are also available from Actlabs (httpwwwactlabscommethsub_code4erehtm) The analyticalmethod for each of the elements analyzed is indicated inTable 2 For the major elements a 02 g sample was mixedwith a lithium metaborate^lithium tetraborate mixtureand fused in a graphite crucible The molten mixture waspoured into a 5 HNO3 solution and shaken until dis-solved (30min) The samples were analyzed for majoroxides and selected trace elements by inductively coupledplasma optical emission spectrometry (ICP-OES) on acombination simultaneous^sequential Thermo Jarrell-AshEnviro II system Internal calibration was achieved usinga variety of international reference materials (eg W-2BIR-1 DNC-1) and independent control samplesAdditional trace elements were analyzed by both instru-mental neutron activation analysis (INAA) and induc-tively couple plasma mass spectrometry (ICP-MS) ForINAA15^25 g of sample was weighed into small polyeth-ylene vials and irradiated with a control international ref-erence material CANMET WMS-1 and NiCr flux wires ata thermal neutron flux of 71012 neutronscm2 per s in theMcMaster Nuclear Reactor Following a 7 day waitingperiod the samples were measured on an Ortec high-purity Ge detector linked to a Canberra Series 95 multi-channel analyzer Activities for each element weredecay- and weight-corrected and compared with a detectorcalibration developed from multiple international certifiedreference materials For ICP-MS 025 g of sample wasdigested in HF followed by HNO3^HClO4 heated andtaken to dryness The samples were brought back into solu-tion with HCl Samples were analyzed using a PerkinElmer Optima 3000 ICP In-lab standards or certified ref-erence materials (eg W-2 BIR-1 DNC-1) were used for
quality control A total of 15 blind duplicates were ana-lyzed from the same sample powders to assessreproducibility
University of Massachusetts analyticalmethodsFourteen sample duplicate powders taken from the same100 g sample aliquot (high-MgO basalts and picrites)were also analyzed at the Ronald B Gilmore X-RayFluorescence (XRF) Laboratory at the University ofMassachusetts Major elements were measured on a fusedLa-bearing lithium borate glass disc using a SiemensMRS-400 spectrometer Trace element concentrations(Rb Sr Ba Ce Nb Zr Y Pb Zn Ga Ni Cr V) weremeasured on a separate powder pellet using a PhilipsPW2400 sequential spectrometer using a Rh tube LOIand ferrous iron measurements were made as describedby Rhodes amp Vollinger (2004) Precision and accuracy esti-mates for the data have been described by Rhodes (1996)and Rhodes amp Vollinger (2004) Results for each sampleare the average of two separate analyses and are shown inElectronic Appendix 2 for a comparison of analyses ofduplicate powders using XRFand ICP-MS
PCIGR analytical methodsA subset of 19 samples was selected for high-precisiontrace-element analysis and Sr Nd Pb and Hf isotopicanalysis at the Pacific Centre for Isotopic andGeochemical Research (PCIGR) at the University ofBritish Columbia (UBC) Samples were selected from the59 samples analyzed for whole-rock chemistry at ActLabsbased on major- and trace-element chemistry alteration(low LOI and petrographic alteration index) and samplelocation Samples were prepared for trace-element analysisat the PCIGR by the technique described by Pretorius et al(2006) on unleached rock powders Sample powders(100mg) were weighed in 7ml screw-top Savillex bea-kers and dissolved in 1ml 14N HNO3 and 5ml 48 HFon a hotplate for 48 h at 1308C with periodic ultrasonica-tion Samples were dried and redissolved in 6ml 6N HClon a hotplate for 24 h and then dried and redissolved in1ml concentrated HNO3 for 24 h before final dryingTrace element abundances were measured by high-resolution inductively coupled plasma-mass spectrometry(HR-ICP-MS) with aThermo Finnigan Element2 systemfollowing the procedures described by Pretorius et al(2006) within 24 h of redissolution HFSE and LILE weremeasured in medium-resolution mode at 2000 dilutionusing a PFATeflon spray chamber washed with aqua regiafor 3min between samples REE were measured in high-resolution mode and U and Pb in low-resolution modeat 2000 dilution using a glass spray chamber washedwith 2 HNO3 between samples Total procedural blanksand reference materials (BCR-2 BHVO-2) were analyzedwith the batch of samples Indium was used as an internal
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
38
standard in all samples and standard solutionsBackground and standard solutions were analyzed afterevery five samples to detect memory effects and massdrift Results for PCIGR trace-element analyses areshown in Electronic Appendix 3 for the full set of analyzedtrace elementsSample digestion for purification of Sr Nd Hf and Pb
for column chemistry began by weighing each samplepowder (400^500mg) prior to leaching All samples wereleached with 6N HCl and placed in an ultrasonic bath for15min Samples were rinsed twice with 18 M-cm H2Obetween each leaching step (15 total) until the supernatantliquid was clear [following the method of Mahoney (1987)]Samples were then dried on a hotplate for 24 h andweighed again Sample solutions were then prepared bydissolving 100^250mg of the leached powder dissolvedin 1ml 14N HNO3 and 10ml 48 HF on a hotplate for48 h at 1308C with periodic ultrasonication Samples weredried and redissolved in 6ml 6N HCl on a hotplate for24 h and then dried Pb was separated using anion exhangecolumns and the discard was used for Sr REE and Hfseparation Nd was separated from the other REE and Hfrequired two additional purification steps Detailed proce-dures for column chemistry for separating Sr Nd and Pbat the PCIGR have been described byWeis et al (2006) andHf purification has been described byWeis et al (2007) Srand Nd isotope ratios were measured by thermal ioniza-tion mass spectrometry (TIMS) on a Thermo FinniganTriton system in static mode with relay matrix rotation ona single Ta and double Re^Ta filament respectively Fourto five filaments per barrel of 21 were occupied by stan-dards (NBS 987 for Sr and LaJolla for Nd) for each barrelwhere samples were run Sample Sr and Nd isotopic com-positions were corrected for mass fractionation using86Sr88Srfrac14 01194 and 146Nd144Ndfrac14 07219 Each samplewas then normalized using the barrel average of the refer-ence material relative to the values of143Nd144Ndfrac14 0511858 and 87Sr86Srfrac14 0710248 (Weiset al 2006) During the course of the Vancouver Islandanalyses the La Jolla Nd standard gave an average valueof 05118566 (nfrac14 7) and NBS987 standard gave an aver-age of 07102408 (nfrac1411) (2 error is reported as times106) 147Sm144Nd errors are 15 or 0006 USGeological Survey (USGS) reference material BHVO-2was processed with the samples and yielded Sr and Nd iso-topic ratios of 07034607 and 05129786 respectivelyThese are in agreement with the published values of070347920 and 051298411 respectively (Weis et al2006)
Pb and Hf isotopic compositions were analyzed by mul-tiple collector-inductively coupled plasma-mass spectrome-try (MC-ICP-MS) on a static multi-collection on a NuPlasma (Nu Instruments) system The detailed analyticalprocedure for Pb isotopic analyses on the Nu Plasma atthe PCIGR has been described by Weis et al (2006) Theconfiguration for Pb analyses allows for collection of PbTl and Hg togetherTl and Hg are used to monitor instru-mental mass discrimination and isobaric overlap repec-tively All sample solutions were analyzed withapproximately the same PbTl ratio (4) as the referencematerial NIST SRM 981 To accomplish this a small ali-quot of each sample solution from the Pb columns was ana-lyzed on the Element2 to determine the precise amount ofPb available for analysis on the Nu Plasma The SRM 981standard was run after every two samples on the NuPlasma During the time samples were run analyses ofthe SRM 981 Pb reference material gave values of206Pb204Pbfrac1416940322 207Pb204Pbfrac1415495823and 208Pb204Pbfrac14 36713164 (nfrac14 61 2 error is reportedas times 104) these values are in excellent agreement withreported TIMS triple-spike values of Galer amp Abouchami(1998) Fractionation-corrected Pb isotopic ratios were fur-ther corrected by the sample-standard bracketing methodor the ln^ln correction method described by White et al(2000) and Blichert-Toft et al (2003) Leached powders ofUSGS reference material BHVO-2 yielded Pb isotopicratios of 206Pb204Pbfrac14186454 8 207Pb204Pbfrac14 1549105 and 208Pb204Pbfrac14 38222514 and BCR-2yielded 206Pb204Pbfrac141880466 207Pb204Pbfrac14 1562518 and 208Pb204Pbfrac14 3883496 (2 error isreported as times 104) These values are in agreement withthose for leached residues of BHVO-2 and BCR-2 fromWeis et al (2006)Hf isotopic compositions were analyzed following the
procedures detailed byWeis et al (2007) The configurationfor Hf analyses monitored Lu mass 175 andYb mass 172 toallow for interference correction to masses 174 and 176 Hfisotopic ratios were normalized internally for mass fractio-nation to a 179Hf177Hf ratio of 07325 using an exponentialcorrection Standards were run after every two samplesand sample results were normalized to the ratio of the in-run daily average and a 176Hf177Hf ratio for JMC-475 of0282160 During the course of analyses the Hf standardJMC-475 gave an average value 02821679 (nfrac14130)USGS reference materials BCR-2 and BHVO-2 were pro-cessed with the samples and yielded Hf isotopic ratios of02828675 and 02831005 respectively Publishedvalues for BCR-2 and BHVO-2 are 02828717 and0283104 8 respectively (Weis et al 2007)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
39
require high-degree partial melting of the mantle source(eg Herzberg amp OrsquoHara 2002) The Keogh Lake picritesfrom northernVancouver Island are the best candidates foridentifying the least modified partial melts of the mantleplume source despite having accumulated olivine pheno-crysts and can be used to estimate conditions of meltingand the composition of the primary magmas for theKarmutsen FormationPrimary magma compositions mantle potential tem-
peratures and source melt fractions for the picritic lavas ofthe Karmutsen Formation were calculated from primitivewhole-rock compositions using PRIMELT1XLS software(Herzberg et al 2007) and are presented in Fig 12 andTable 6 A detailed discussion of the computationalmethod has been given by Herzberg amp OrsquoHara (2002)and Herzberg et al (2007) key features of the approachare outlined belowPrimary magma composition is calculated for a primi-
tive lava by incremental addition of olivine and compari-son with primary melts of mantle peridotite Lavas thathad experienced plagioclase andor clinopyroxene fractio-nation are excluded from this analysis All calculated pri-mary magma compositions are assumed to be derived byaccumulated fractional melting of fertile peridotite KR-4003 Uncertainties in fertile peridotite composition donot propagate to significant variations in melt fractionand mantle potential temperature melting of depletedperidotite propagates to calculated melt fractions that aretoo high but with a negligible error in mantle potential
temperature PRIMELT1XLS calculates the olivine liqui-dus temperature at 1atm which should be similar to theactual eruption temperature of the primary magmaand is accurate to 308C at the 2 level of confidenceMantle potential temperature is model dependent witha precision that is similar to the accuracy of the olivineliquidus temperature (C Herzberg personal communica-tion 2008)With regard to the evidence for accumulated olivine in
the Keogh Lake picrites it should not matter whether themodeled lava composition is aphyric or laden with accu-mulated olivine phenocrysts PRIMELT1 computes theprimary magma composition by adding olivine to a lavathat has experienced olivine fractionation in this way itinverts the effects of fractional crystallization The onlyrequirement is that the olivine phenocrysts are not acci-dental or exotic to the lava compositions Support for thisprocedure can be seen in the calculated olivine liquid-line-of-descent shown by the curved arrays with 5 olivineaddition increments (Fig 12c) Fractional crystallization ofolivine from model primary magmas having 85^95wt FeO and 153^175wt MgO will yield arrays of deriva-tive liquids and randomly sorted olivines that in most butnot all cases connect the picrites and high-MgO basalts Anewer version of the PRIMELT software (Herzberg ampAsimow 2008) can perform olivine addition to and sub-traction from lavas with highly variably olivine phenocrystcontents and yields results that converge with those shownin Fig 12
Fig 11 Relationship between abundance of olivine phenocrysts and whole-rock MgO contents for the Keogh Lake picrites (a) Tracings ofolivine grains (gray) in six high-MgO samples made from high resolution (2400 dpi) scans of petrographic thin-sections Scale bars represent10mm sample numbers are indicated at the upper left (b) Modal olivine vs whole-rock MgO Continuous line is the best-fit line for 10 high-MgO samplesThe areal proportion of olivine was calculated from raster images of olivine grains using ImageJ image analysis software whichprovides an acceptable estimation of the modal abundance (eg Chayes 1954)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
25
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
Gray lines = final melting pressure
L + Ol + Opx
07
08
09
Pressure (GPa)
10
3
4
5
6
1
SOLIDUS
01
04
0506
L + Ol
2
03
02
PeridotiteKR-4003
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
Thick black lines = initial melting pressure
Dashed lines = melt fraction
Melt Fraction
2
3 4 5 6
L+Ol+Opx
SolidusSpinel
SolidusGarnet Peridotite
7
L+Ol
0 10 20 30 4040
45
50
55
60
MgO (wt)
SiO2
(wt)
PeridotiteKR-4003
Melt Fraction
0 10 20 305
10
15
CaO(wt)
MgO (wt)
3
4 5 6 7
2
46
2
Peridotite Partial Melts
Thick black line = initial melting pressureGray lines = final melting pressure
Peridotite
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
02
Pressure (GPa)
0302
04
Pressure (GPa)
Picrite
High-MgO basalt
Tholeiitic basalt
(c)
(d)
Karmutsen flood basaltsOlivine addition modelOlivine addition (5 increment)Primary magma
PicriteHigh-MgO basaltTholeiitic basalt
(a)
(b)
800 850 900
910
920
930
940
950
Karmutsen flood basalts
Olivine addition modelOlivine addition (5 increment)Primary magma
Gray lines = Fo contents of olivine
Fig 12 Estimated primary magma compositions for three Keogh Lake picrites and high-MgO basalts (samples 4723A4 4723A135616A7) using the forward and inverse modeling technique of Herzberg et al (2007) Karmutsen compositions and modeling results areoverlain on diagrams provided by C Herzberg (a) Whole-rock FeO vs MgO for Karmutsen samples from this study Total iron estimatedto be FeO is 090 Gray lines show olivine compositions that would precipitate from liquid of a given MgO^FeO composition Black lineswith crosses show results of olivine addition (inverse model) using PRIMELT1 (Herzberg et al 2007) Gray field highlights olivinecompositions with Fo contents of 91^92 predicted to precipitate (b) FeO vs MgO showing Karmutsen lava compositions and results ofa forward model for accumulated fractional melting of fertile peridotite KR-4003 (Kettle River Peridotite Walter 1998) (c) SiO2 vsMgO for Karmtusen lavas and model results (d) CaO vs MgO showing Karmtusen lava compositions and model results To brieflysummarize the technique [see Herzberg et al (2007) for complete description] potential parental magma compositions for the high-MgOlava series were selected (highest MgO and appropriate CaO) and using PRIMELT1 software olivine was incrementally added tothe selected compositions to show an array of potential primary magma compositions (inverse model) Then using PRIMELT1 theresults from the inverse model were compared with a range of accumulated fractional melts for fertile peridotite derived fromparameterization of the experimental results for KR-4003 from Walter (1998) forward model Herzberg amp OrsquoHara (2002) A melt fractionwas sought that was unique to both the inverse and forward models (Herzberg et al 2007) A unique solution was found when there was a com-mon melt fraction for both models in FeO^MgO and CaO^MgO^Al2O3^SiO2 (CMAS) projection space This modeling assumes thatolivine was the only phase crystallizing and ignores chromite precipitation and possible augite fractionation in the mantle (Herzbergamp OrsquoHara 2002) Results are best for a residue of spinel lherzolite (not pyroxenite) The presence of accumulated olivine in samples ofthe Keogh Lake picrites used as starting compositions does not significantly affect the results because we are modeling addition of olivine(see text for further description) The tholeiitic basalts cannot be used for modeling because they are all saturated in plagthorn cpxthornolThick black line for initial melting pressure at 4 GPa is not shown in (c) for clarity Gray area in (a) indicates parental magmas thatwill crystallize olivine with Fo 91^92 at the surface Dashed lines in (c) indicate melt fraction Solidi for spinel and garnet peridotite arelabelled in (c)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
The PRIMELT1 results indicate that if the Karmutsenpicritic lavas were derived from accumulated fractionalmelting of fertile peridotite the primary magmas wouldhave contained 15^17wt MgO and 10wt CaO(Fig 12 Table 6) formed from 23^27 partial meltingand would have crystallized olivine with a Fo content of 91 (Fig 12 Table 6) Calculated mantle temperatures forthe Karmutsen picrites (14908C) indicate melting ofanomalously hot mantle (100^2008C above ambient) com-pared with ambient mantle that produces mid-ocean ridgebasalt (MORB 1280^14008C Herzberg et al 2007)which initiated between 3 and 4 GPa Several picritesappear to be near-primary melts and required very littleaddition of olivine (eg sample 93G171 with measured158wt MgO) These temperature and melting esti-mates of the Karmutsen picrites are consistent withdecompression of hot mantle peridotite in an actively con-vecting plume head (ie plume initiation model) Threesamples (picrite 5614A1 and high-MgO basalts 5614A3and 5614A5) are lower in iron than the other Karmutsenlavas with FeO in the 65^80wt range (Fig 12c) andhave MgO and FeO contents that are similar to primitiveMORB from the Sequeiros fracture zone of the EastPacific Rise (EPR Perfit et al 1996) These samples
however differ from EPR MORB in having higher SiO2
and lower Al2O3 and they are restricted geographicallyto one area (Sara Lake Fig 2)We therefore have evidence for primary magmas with
MgO contents that range from a possible low of 110wt to as high as 174wt with inferred mantle potential tem-peratures from 1340 to 15208C This is similar to the rangedisplayed by lavas from the Ontong Java Plateau deter-mined by Herzberg (2004) who found that primarymagma compositions assuming a peridotite sourcewould contain 17wt MgO and 10wt CaO(Table 6) and that these magmas would have formed by27 melting at a mantle potential temperature of15008C (first melting occurring at 36 GPa) Fitton ampGodard (2004) estimated 30 partial melting for theOntong Java lavas based on the Zr contents of primarymagmas of Kroenke-type basalt calculated by incrementaladdition of equilibrium olivine However if a considerableamount of eclogite was involved in the formation of theOntong Java plateau magmas the excess temperatures esti-mated by Herzberg et al (2007) may not be required(Korenaga 2005) The variability in mantle potential tem-perature for the Keogh Lake picrites is also similar to thewide range of temperatures that have been calculated for
Table 6 Estimated primary magma compositions for Karmutsen basalts and other oceanic plateaus or islands
Sample 4723A4 4723A13 5616A7 93G171 Average OJP Mauna Keay Gorgonaz
Ontong Java primary magma composition for accumulated fraction melting (AFM) from Herzberg (2004)yMauna Kea primary magma composition is average of four samples of Herzberg (2006 table 1)zGorgona primary magma composition for AFM for 1F fertile source from Herzberg amp OrsquoHara (2002 table 4)Total iron estimated to be FeO is 090
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
27
lavas from single volcanoes in the Galapagos Hawaii andelsewhere by Herzberg amp Asimow (2008) Those workersinterpreted the range to reflect the transport of magmasfrom both the cool periphery and the hotter axis of aplume A thermally heterogeneous plume will yield pri-mary magmas with different MgO contents and the rangeof primary magma compositions and their inferred mantlepotential temperatures for Karmutsen lavas may reflectmelting from different parts of the plume
Source of the Karmutsen Formation lavasThe Sr^Nd^Hf^Pb isotopic compositions of theKarmutsen volcanic rocks on Vancouver Island indicatethat they formed from plume-type mantle containing adepleted component that is compositionally similar to butdistinct from other oceanic plateaus that formed in thePacific Ocean (eg Ontong Java and Caribbean Fig 13)Trace element and isotopic constraints can be used to testwhether the depleted component of the KarmutsenFormation was intrinsic to the mantle plume and distinctfrom the source of MORB or the result of mixing or invol-vement of ambient depleted MORB mantle with plume-type mantle The Karmutsen tholeiitic basalts have initialeNd and
87Sr86Sr ratios that fall within the range of basaltsfrom the Caribbean Plateau (eg Kerr et al 1996 1997Hauff et al 2000 Kerr 2003) and a range of Hawaiian iso-tope data (eg Frey et al 2005) and lie within and abovethe range for the Ontong Java Plateau (Tejada et al 2004Fig 13a) Karmutsen tholeiitic basalts with the highest ini-tial eNd and lowest initial 87Sr86Sr lie within and justbelow the field for age-corrected northern EPR MORB at230 Ma (eg Niu et al1999 Regelous et al1999) Initial Hfand Nd isotopic compositions for the Karmutsen samplesare noticeably offset to the low side of the ocean islandbasalt (OIB) array (Vervoort et al 1999) and lie belowand slightly overlap the low eHf end of fields for Hawaiiand the Caribbean Plateau (Fig 13c) the Karmutsen sam-ples mostly fall between and slightly overlap a field forage-corrected Pacific MORB at 230 Ma and Ontong Java(Mahoney et al 1992 1994 Nowell et al 1998 Chauvel ampBlichert-Toft 2001) Karmutsen lava Pb isotope ratios over-lap the broad range for the Caribbean Plateau and thelinear trends for the Karmutsen samples intersect thefields for EPR MORB Hawaii and Ontong Java in208Pb^206Pb space but do not intersect these fields in207Pb^206Pb space (Fig 13d and e) The more radiogenic207Pb204Pb for a given 206Pb204Pb of the Karmutsenbasalts requires high UPb ratios at an ancient time when235U was abundant similar to the Caribbean Plateau andenriched in comparison with EPR MORB Hawaii andOntong JavaThe low eHf^low eNd component of the Karmutsen
basalts that places them below the OIB mantle array andpartly within the field for age-corrected Pacific MORBcan form from low ratios of LuHf and high SmNd over
time (eg Salters amp White 1998) MORB will develop aHf^Nd isotopic signature over time that falls below themantle array Low LuHf can also lead to low 176Hf177Hffrom remelting of residues produced by melting in theabsence of garnet (eg Salters amp Hart 1991 Salters ampWhite 1998) The Hf^Nd isotopic compositions of theKarmutsen Formation indicate the possible involvementof depleted MORB mantle however similar to Iceland(Fitton et al 2003) Nb^Zr^Y variations provide a way todistinguish between a depleted component within theplume and a MORB-like component The Karmutsenbasalts and picrites fall within the Iceland array and donot overlap the MORB field in a logarithmic plot of NbYvs ZrY (Fig 13b) using the fields established by Fittonet al (1997) Similar to the depleted component within theCaribbean Plateau (Thompson et al 2003) the trend ofHf^Nd isotopic compositions towards Pacific MORB-likecompositions is not followed by MORB values in Nb^Zr^Y systematics and this suggests that the depleted compo-nent in the source of the Karmutsen basalts is distinctfrom the source of MORB and probably an intrinsic partof the plume The relatively radiogenic Pb isotopic ratiosand REE patterns distinct from MORB also support thisinterpretationTrace-element characteristics suggest the possible invol-
vement of sub-arc lithospheric mantle in the generation ofthe Karmutsen picrites Lassiter et al (1995) suggested thatmixing of a plume-type source with eNdthorn 6 to thorn 7 witharc material with low NbTh could reproduce the varia-tions observed in the Karmutsen basalts however theyconcluded that the absence of low NbLa ratios in theirsample of tholeiitic basalts restricts the amount of arc litho-sphere involved The study of Lassiter et al (1995) wasbased on major and trace element and Sr Nd and Pb iso-topic data for a suite of 29 samples from Buttle Lake in theStrathcona Provincial Park on central Vancouver Island(Fig 1) Whereas there are no clear HFSE depletions inthe Karmutsen tholeiitic basalts the low NbLa (and TaLa) of the Karmutsen picritic lavas in this study indicatethe possible involvement of Paleozoic arc lithosphere onVancouver Island (Fig 8)
REE modeling dynamic melting andsource mineralogyThe combined isotopic and trace-element variations of thepicritic and tholeiitic Karmutsen lavas indicate that differ-ences in trace elements may have originated from differentmelting histories For example the LuHf ratios of theKarmutsen picritic lavas (mean 008) are considerablyhigher than those in the tholeiitic basalts (mean 002)however the small range of overlapping initial eHf valuesfor the two lava suites suggests that these differences devel-oped during melting within the plume (Fig 9b) Below wemodel the effects of source mineralogy and depth of
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
28
melting on differences in trace-element concentrations inthe tholeiitic basalts and picritesDynamic melting models simulate progressive decom-
pression melting where some of the melt fraction isretained in the residue and only when the degree of partialmelting is greater than the critical mass porosity and the
source becomes permeable is excess melt extracted fromthe residue (Zou1998) For the Karmutsen lavas evolutionof trace element concentrations was simulated using theincongruent dynamic melting model developed by Zou ampReid (2001) Melting of the mantle at pressures greater than1 GPa involves incongruent melting (eg Longhi 2002)
OIB array
PacificMORB(230 Ma)
(0 Ma)
EPR(230 Ma)
-4
-2
0
2
4
8
10
12
14
16
18
20
22
-4 -2 0 2 4 6 8 10 12 14
minus2
0
2
4
6
8
10
12
14
0702 0703 0704 0705 0706
IndianMORB
87Sr 86Sr (initial)
Hf (initial)
(a) (c)
Nd (initial)
Karmutsen tholeiitic basalt
HawaiiOntong JavaCaribbean Plateau
Karmutsen high-MgO lava
high-MgO lavasKarmutsen
1540
1545
1550
1555
1560
1565
175 180 185 190 195 200375
380
385
390
395
175 180 185 190 195 200
(d) (e)
EPR
EPR
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb
Karmutsen Formation (initial)
HawaiiOntong JavaCaribbean Plateau
East Pacific Rise
Karmutsen Formation (measured)
Juan de FucaGorda
ε Nd
(initial)
Karmutsen Formation (different age-correction for 3 picrites)
(0 Ma)
NbY
001
01
1
1 10
ZrY
OIB
NMORB
(b)
Iceland array
6
Fig 13 Comparison of age-corrected (230 Ma) Sr^Nd^Hf^Pb isotopic compositions for Karmutsen flood basalts onVancouver Island withage-corrected OIB and MORB and Nb^Zr^Y variation (a) Initial eNd vs 87Sr86Sr (b) NbYand ZrY variation (after Fitton et al 1997)(c) Initial eHf vs eNd (d) Measured and initial 207Pb204Pb vs 206Pb204Pb (e) Measured and initial 208Pb204Pb vs 206Pb204Pb References fordata sources are listed in Electronic Appendix 4 Most of the compiled data were extracted from the GEOROC database (httpgeorocmpch-mainzgwdgdegeoroc) OIB array line in (c) is fromVervoort et al (1999) EPR East Pacific Rise Several high-MgO samples affected bysecondary alteration were not plotted for clarity in Pb isotope plots Estimation of fields for age-corrected Pacific MORB and EPR at 230 Mawere made using parent^daughter concentrations (in ppm) from Salters amp Stracke (2004) of Lufrac14 0063 Hffrac14 0202 Smfrac14 027 Ndfrac14 071Rbfrac14 0086 and Srfrac14 836 In the NbYvs ZrYplot the normal (N)-MORB and OIB fields not including Icelandic OIB are from data com-pilations of Fitton et al (1997 2003) The OIB field is data from 780 basalts (MgO45wt ) from major ocean islands given by Fitton et al(2003) The N-MORB field is based on data from the Reykjanes Ridge EPR and Southwest Indian Ridge from Fitton et al (1997) The twoparallel gray lines in (b) (Iceland array) are the limits of data for Icelandic rocks
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
29
with olivine being produced during the reactioncpxthornopxthorn spfrac14meltthornol for spinel peridotites The mod-eling used coefficients for incongruent melting reactionsbased on experiments on lherzolite melting (eg Kinzleramp Grove 1992 Longhi 2002) and was separated into (1)melting of spinel lherzolite (2) two combined intervals ofmelting for a garnet lherzolite (gt lherz) and spinel lher-zolite (sp lherz) source (3) melting of garnet lherzoliteThe model solutions are non-unique and a range indegree of melting partition coefficients melt reactioncoefficients and proportion of mixed source componentsis possible to achieve acceptable solutions (see Fig 14 cap-tion for description of modeling parameters anduncertainties)The LREE-depleted high-MgO lavas require melting of
a depleted spinel lherzolite source (Fig 14a) The modelingresults indicate a high degree of melting (22^25) similarto the results from PRIMELT1 based on major-elementvariations (discussed above) of a LREE-depleted sourceThe high-MgO lavas lack a residual garnet signature Theenriched tholeiitic basalts involved melting of both garnetand spinel lherzolite and represent aggregate melts pro-duced from continuous melting throughout the depth ofthe melting column (Fig 14b) Trace-element concentra-tions in the tholeiitic and picritic lavas are not consistentwith low-degree melting of garnet lherzolite alone(Fig 14c) The modeling results indicate that the aggregatemelts that formed the tholeiitic basalts would haveinvolved lesser proportions of enriched small-degreemelts (1^6 melting) generated at depth and greateramounts of depleted high-degree melts (12^25) gener-ated at lower pressure (a ratio of garnet lherzolite tospinel lherzolite melt of 14 provides the best fits seeFig 14b and description in figure caption) The totaldegree of melting for the tholeiitic basalts was high(23^27) similar to the degree of partial melting in thespinel lherzolite facies for the primary melts that eruptedas the Keogh Lake picrites of a source that was also prob-ably depleted in LREE
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
(c)
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
Melting of garnet lherzolite
High-MgO lavas
Melting of spinel lherzolite
source (07DM+03PM)
source (07DM+03PM)
6 melting
30 melting
(b)
(a)
30 melting
Sam
ple
Cho
ndrit
eS
ampl
eC
hond
rite 20 melting
1 melting
Melting of garnet lherzoliteand spinel lherzolite
x gt lherz + 4x sp lherz
5 meltingx gt lherz + 4x sp lherz
Sam
ple
Cho
ndrit
e
Tholeiitic lavas
Tholeiitic lavas
High-MgO lavas
source (07DM+03PM)
Fig 14 Trace-element modeling results for incongruent dynamicmantle melting for picritic and tholeiitic lavas from the KarmutsenFormation Three steps of modeling are shown (a) melting of spinellherzolite compared with high-MgO lavas (b) melting of garnet lher-zolite and spinel lherzolite compared with tholeiitic lavas (c) meltingof garnet lherzolite compared with tholeiitic and picritic lavasShaded fields in each panel for tholeiitic basalts and picrites arelabeled Patterns with symbols in each panel are modeling results(patterns are 5 melting increments in (a) and (b) and 1 incre-ments in (c) PM primitive mantle DM depleted MORB mantle gtlherz garnet lherzolite sp lherz spinel lherzolite In (b) the ratios ofper cent melting for garnet and spinel lherzolite are indicated (x gtlherzthorn 4x sp lherz means that for 5 melting 1 melt in garnetfacies and 4 melt in spinel facies where x is per cent melting in theinterval of modeling) The ratio of the respective melt proportions inthe aggregate melt (x gt lherzthorn 4x sp lherz) was kept constant andthis ratio of melt contribution provided the best fit to the data Arange of proportion of source components (07DMthorn 03PM to
09DMthorn 01PM) can reproduce the variation in the high-MgOlavas Melting modeling uses the formulation of Zou amp Reid (2001)an example calculation is shown in their Appendix PM fromMcDonough amp Sun (1995) and DM from Salters amp Stracke (2004)Melt reaction coefficients of spinel lherzolite from Kinzler amp Grove(1992) and garnet lherzolite fromWalter (1998) Partition coefficientsfrom Salters amp Stracke (2004) and Shaw (2000) were kept constantSource mineralogy for spinel lherzolite (018cpx027opx052ol003sp) from Kinzler (1997) and for garnet lherzolite(034cpx008opx053ol005gt) from Salters amp Stracke (2004) Arange of source compositions for melting of spinel lherzolite(07DMthorn 03PM to 03DMthorn 07PM) reproduce REE patternssimilar to those of the high-MgO lavas with best fits for 22ndash25melting and 07DMthorn 03PM source composition Concentrationsof tholeiitic basalts cannot be reproduced from an entirelyprimitive or depleted source
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
30
The uncertainty in the composition of the source in thespinel lherzolite melting model for the high-MgO lavasprecludes determining whether the source was moredepleted in incompatible trace elements than the source ofthe tholeiitic lavas (see Fig 14 caption for description) It ispossible that early high-pressure low-degree melting left aresidue within parts of the plume (possibly the hotter inte-rior portion) that was depleted in trace elements (egElliott et al 1991) further decompression and high-degreemelting of such depleted regions could then have generatedthe depleted high-MgO Karmutsen lavas Shallow high-degree melting would preferentially sample depletedregions whereas deeper low-degree melting may notsample more refractory depleted regions (eg CaribbeanEscuder-Viruete et al 2007) The picrites lie at the highend of the range of Nd isotopic compositions and havelow NbZr similar to depleted Icelandic basalts (Fittonet al 2003) therefore the picrites may represent the partsof the mantle plume that were the most depleted prior tothe initiation of melting As Fitton et al (2003) suggestedthese depleted melts may only rarely be erupted withoutcombining with more voluminous less depleted melts andthey may be detected only as undifferentiated small-volume flows like the Keogh Lake picrites within theKarmutsen Formation on Vancouver IslandDecompression melting within the mantle plume initiatedwithin the garnet stability field (35^25 GPa) at highmantle potential temperatures (Tp414508C) and pro-ceeded beneath oceanic arc lithosphere within the spinelstability field (25^09 GPa) where more extensivedegrees of melting could occur
Magmatic evolution of Karmutsentholeiitic basaltsPrimary magmas in large igneous provinces leave exten-sive coarse-grained cumulate residues within or below thecrust these mafic and ultramafic plutonic sequences repre-sent a significant proportion of the original magma thatpartially crystallized at depth (eg Farnetani et al 1996)For example seismic and petrological studies of OntongJava (eg Farnetani et al 1996) combined with the use ofMELTS (Ghiorso amp Sack 1995) indicate that the volcanicsequence may represent only 40 of the total magmareaching the Moho Neal et al (1997) estimated that30^45 fractional crystallization took place for theOntong Java magmas and produced cumulate thicknessesof 7^14 km corresponding to 11^22 km of flood basaltsdepending on the presence of pre-existing oceanic crustand the total crustal thickness (25^35 km) Interpretationsof seismic velocity measurements suggest the presence ofpyroxene and gabbroic cumulates 9^16 km thick beneaththe Ontong Java Plateau (eg Hussong et al 1979)Wrangellia is characterized by high crustal velocities in
seismic refraction lines which is indicative of mafic
plutonic rocks extending to depth beneath VancouverIsland (Clowes et al 1995)Wrangellia crust is 25^30 kmthick beneath Vancouver Island and is underlain by astrongly reflective zone of high velocity and density thathas been interpreted as a major shear zone where lowerWrangellia lithosphere was detached (Clowes et al 1995)The vast majority of the Karmutsen flows are evolved tho-leiitic lavas (eg low MgO high FeOMgO) indicating animportant role for partial crystallization of melts at lowpressure and a significant portion of the crust beneathVancouver Island has seismic properties that are consistentwith crystalline residues from partially crystallizedKarmutsen magmas To test the proportion of the primarymagma that fractionated within the crust MELTS(Ghiorso amp Sack 1995) was used to simulate fractionalcrystallization using several estimated primary magmacompositions from the PRIMELT1 modeling results Themajor-element composition of the primary magmas couldnot be estimated for the tholeiitic basalts because of exten-sive plagthorn cpxthornol fractionation and thus the MELTSmodeling of the picrites is used as a proxy for the evolutionof major elements in the volcanic sequence A pressure of 1kbar was used with variable water contents (anhydrousand 02wt H2O) calculated at the quartz^fayalite^magnetite (QFM) oxygen buffer Experimental resultsfrom Ontong Java indicate that most crystallizationoccurs in magma chambers shallower than 6 km deep(52 kbar Sano amp Yamashita 2004) however some crys-tallization may take place at greater depths (3^5 kbarFarnetani et al 1996)A selection of the MELTS results are shown in Fig 15
Olivine and spinel crystallize at high temperature until15^20wt of the liquid mass has fractionated and theresidual magma contains 9^10wt MgO (Fig 15f) At1235^12258C plagioclase begins crystallizing and between1235 and 11908C olivine ceases crystallizing and clinopyr-oxene saturates (Fig 15f) As expected the addition of clin-opyroxene and plagioclase to the crystallization sequencecauses substantial changes in the composition of the resid-ual liquid FeO and TiO2 increase and Al2O3 and CaOdecrease while the decrease in MgO lessens considerably(Fig 15) The near-vertical trends for the Karmutsen tho-leiitic basalts especially for CaO do not match the calcu-lated liquid-line-of-descent very well which misses manyof the high-CaO high-Al2O3 low-FeO basalts A positivecorrelation between Al2O3CaO and SrNd indicates thataccumulation of plagioclase played a major role in the evo-lution of the tholeiitic basalts (Fig 15g) Increasing thecrystallization pressure slightly and the water content ofthe melts does not systematically improve the match ofthe MELTS models to the observed dataThe MELTS results indicate that a significant propor-
tion of the crystallization takes place before plagioclase(15^20 crystallization) and clinopyroxene (35^45
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
31
crystallization) join the crystallization sequence Theseresults suggest that445 crystallization of the primarymagmas occurred (estimated from the residual liquid percent in Fig 15f) and if the original flood basalt stratigra-phy on Vancouver Island was 6 km thick at least anequivalent amount of complementary plutonic rocksshould have crystallized The Karmutsen basalts and theirplutonic residues thus represent a significant addition ofcrust (perhaps 412 km thickness) previously thickenedduring the formation of the underlying Paleozoic arc
Growth of the Wrangellia oceanic plateauon Vancouver IslandThe extrusive part of the Wrangellia oceanic plateau onVancouver Island was constructed in a three-layersequence of mostly tholeiitic basalts with a restrictedrange of composition (Fig 16) The plateau overliesDevonian to Mississippian arc volcanic sequences andmarine sediments of Mississippian to Permian ages thatare intruded by gabbroic rocks related to the flood basaltsThe volcanic stratigraphy of theWrangellia plateau formed
00
05
10
15
20
25
30
35
40
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
16
0 2 4 6 8 10 12 14 16 18 20
10
11
12
13
14
15
16
17
18
19
20
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
0 2 4 6 8 10 12 14 16 18 20
MgO (wt)
0 10 20 30 40 50
percent of total mass
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
) Mineral proportions
Sp (e)
Ol
CpxPlag
PicriteHigh-MgO basaltTholeiitic basalt
TiO2 (wt)
MgO (wt)
Al2O3 (wt) CaO (wt)
FeO (wt)
MgO(a) (b)
(c) (d)
MgO (wt)MgO (wt)
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
)
Residual liquid ()
1220
1190
Ol + Sp
Plag+Ol+Sp Plag+Cpx+Sp
02 wt H2O
no H2O
Residual liquid and crystallizing phases
(f)
0102030405060708090100
plagioclaseaccumulation
plagioclaseaccumulation
plagioclase
accumulatio
n
10
12
14
16
0 10 20 30 40
Al2O3CaO
SrNd
tholeiitic basalts
(g)
Fig 15 Forward fractional crystallization modeling results for major elements from MELTS (Ghiorso amp Sack 1995) compared with picriticand tholeiitic lavas from the Karmutsen Formation (a^d) Major-element compositions of residual liquids compared with Karmutsen lavacompositions For clarity the results are shown only for one starting composition (open circle) using the composition of the estimated primarymagma for the high-MgO lava series from the modeling technique of Herzberg et al (2007) for sample 4723A13 as shown in Fig 12 The esti-mated primary magmas for samples 4723A3 and 93G171 yield similar results A pressure of 1 kbar was used with no H2O and 02wt H2Ocalculated at QFM black line is anhydrous and grey line is 02wt H2OThe end-point for the anhydrous result is at 78 fractionation and11308C and the end-point for the result with 02wt H2O is 75 fractionation and11108C (e) Mineral proportions of crystallized phases fromone MELTrun (sample 4723A13) with no H2O (f) Per cent residual liquid vs temperature labelled with intervals of crystallizing phases (g)Al2O3CaO vs SrNd for the tholeiitic basalts
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
32
in a single eruptive phase (c 231^225 Ma) as an emergentbasalt sequence during a deeper-water shallow-water andsubaerial stage and subsided below sea level relativelysoon after its formation (Fig 16) The deeper-water stagecontains pillowed and massive submarine flows (3000mthick) that built up the submarine volcanic edifice from theseafloor The lowest pillow basalts were emplaced onunconsolidated and lithified fine-grained siliciclastic andcarbonaceous sediments containing Daonella beds(c 235^232 Ma) that are intruded by abundantKarmutsen mafic sills (Fig 16a) The different pillowedand massive submarine flows resulted primarily from dif-ferent effusive rates and local topography Some of themassive submarine flows commonly recognized by radiat-ing columnar jointing represent master tubes for the deliv-ery of lava to distal parts of submarine flow fields Thepillowed flows which commonly have uneven topography
in modern sequences may have served as levees to themassive submarine flows rather than allowing unob-structed spreading and inflation typical of the massive sub-aerial flows that form on shallow slopes in continentalflood basalt provinces (eg Self et al 1997) The rapid suc-cession of flows is evident by the near absence of sedimentbetween submarine flows Minor volumes of picritic pillowlavas occur in the upper part of the pillowed flow sequencein areas of the plateau exposed on northern VancouverIsland (Fig 16 Nixon et al 2008)The growth of the plateau in shallow water led to
emplacement of an increasing proportion of volcaniclastics(pillow breccias and hyaloclastites) that accumulated up to1500m thick in some areas (Fig 16b Nixon et al 2008)Horizons of pillowed flows occur locally within the volca-niclastic stratigraphy Sedimentary structures (eg gradedbedding fluidization structures) and petrographic textures
Late Devonian-Mississippian arc
Daonella beds
Middle Triassic
deeper water
Late Devonian-Mississippian arc
pillowed flows massive submarine
flows
pillow brecciahyaloclastiteshallow water
~4 km
~6 km
Late Triassic limestoneinterflow limestonelenses
Karmutsen Formation
mafic intrusions sediment-sill complex
Karmutsen Formation
(a) (b)
(c) (d)
subaerial flows
Late Devonian-Mississippian arc Late Devonian-Mississippian arc
picritic pillow basalts
Miss to Permian marine sediments
sills
Fig 16 Growth of the Wrangellia oceanic plateau onVancouver Island (a) Deep-water eruptive phase of pillowed and massive submarineflows with almost no intervening sediments onto MiddleTriassic shale containing Daonella beds (c 235^232 Ma) The basement of the plateauis composed of Devonian to Mississippian arc sequences overlain by Mississippian to Permian carbonates and siliciclastic sediments(b) Shallow-water flow emplacement with increasing proportion of volcaniclastics resulted primarily from cooling^contraction granulationPicritic pillow basalts occur primarily near the transition between pillowed submarine flows and volcaniclastic units Pillowed and massiveflows occur locally within the volcaniclastic unit (c) Emergent stage of growth consists of overlapping subaerial flow fields similar to continen-tal flood basalts with pahoehoe flow structures Multi-tiered columnar jointing occurrs locally in response to ponding of flows (eg Bondre et al2004) Interflow carbonate deposits and plagioclase megacrystic flows developed locally during the waning stage of volcanism (d) Post-volcanicsubsidence and deposition of Late Carnian to Norian (c 212^223 Ma) platform carbonates on top of the plateau Units are labeled in the dia-grams Thicknesses of the underlying units are not to scale
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
33
within the volcaniclastic rocks indicate formation primar-ily via cooling^contraction granulation magma^water^steam interaction autobrecciation resedimentation andmass-wasting rather than pyroclastic eruption (Nixonet al 2008)A broad subaerial platform was constructed by emplace-
ment of subaerial lava flows during an emergent subaerialstage (Fig 16c) The subaerial unit predominantly com-prises well-layered sheet flows that formed in a similarway to continental flood basalts as hundred(s) of pahoehoeflow fields erupted on shallow slopes (Self et al 1997) how-ever volcaniclastic and pillowed flow units are locally pre-sent mostly in the upper parts of the stratigraphyLocalized ponding of flows led to the formation of multi-tiered columnar jointing in some subaerial flows (egBondre et al 2004) As volcanism waned local interflowcarbonate deposits developed and plagioclase megacrysticflows were erupted Subsidence of the plateau is recordedby the interflow lenses and deposits of platform carbonates(40^1500m thick) on top of the Karmutsen basalts thatcontain Late Carnian to Early Norian fossils (Fig 16d c216^227 Ma Nixon et al 2008)The evolution of the internal architecture of the
Wrangellia oceanic plateau was primarily related to theconditions of emplacement (ie water depth and eruptionenvironment) There does not appear to have been a sys-tematic coincident evolution or shift of lava compositionsonVancouver Island as has been recognized in most con-tinental flood basalts and parts of the Wrangellia oceanicplateau in Alaska and Yukon (Greene et al 2008 2009)In general with the exception of the picritic pillow lavasin the submarine stratigraphy and local variations betweenhigh-MgO and tholeiitic lavas there are no significantcompositional differences between lavas erupted duringthe different stages of growth of the plateau (ie submarineand subaerial flows) nor are substantial compositional dif-ferences generally found between massive submarine flowsin contact with pillowed flows tholeiitic basalt flows nearthe base of the Karmutsen Formation are compositionallysimilar to flows sampled near the top of the volcanicsequence Where it was possible to sample continuous sec-tions of volcanic stratigraphy greater than several hun-dreds of meters (eg Keogh^Maynard Lake area west ofthe Karmutsen Range Fig 2) there is no clear relationshipbetween stratigraphic position and compositionalvariation
CONCLUSIONSThe Karmutsen Formation covers extensive areas ofVancouver Island (20 000 km2) and forms part of a largeoceanic plateau erupted over a geologically short interval(c 225^231 Ma) The tripartite volcanic stratigraphyexposed on Vancouver Island is at least 6 km thick andcomprises a succession of submarine flows volcaniclastic
deposits and massive subaerial flowsThese volcanologicaldifferences are primarily related to the eruption environ-ment (deep-water shallow-water or subaerial) The rapidgrowth of the plateau prevented significant accumulationsof intervening sediment except in the uppermost stratigra-phy where isolated limestone lenses commonly associatedwith pillow basalt and hyaloclastite deposits preserve arecord of the final stages of subsidence of the plateauThe Wrangellia plateau on Vancouver Island was con-
structed dominantly of tholeiitic basalt with restrictedmajor- and trace-element and isotopic compositional var-iations Minor volumes of picritic pillow basalts eruptedlate in the submarine phase The Karmutsen basalts wereemplaced onto older Paleozoic arc volcanic sequences andlow NbLa for the picrites may indicate a contributionfrom the sub-arc lithospheric mantle Modeling of the pet-rogenesis of the Karmutsen picrites indicates melting ofanomalously hot mantle (15008C) and extensive degreesof partial melting (23^27) and is consistent with aplume initiation model Combined Sr^Nd^Hf^Pb isotopicsystematics indicate that the picritic and tholeiitic lavaswere derived from a depleted plume-type mantle sourcedistinct from the source of MORB that has compositionalsimilarities to the source of the lavas forming theCaribbean Plateau Picrites were generated by melting ofspinel lherzolite whereas the tholeiitic basalts formed asaggregate melts of both garnet and spinel lherzolite Thetholeiitic basalts underwent extensive low-pressure fractio-nation (52^3 kbar) and seismic studies indicate that someof theWrangellia crust beneathVancouver Island may cor-respond to the plutonic residues of the Karmutsen floodbasalts
ACKNOWLEDGEMENTSWe are grateful to Nick Arndt for helping us get this pro-ject started and Nick Massey for insights into VancouverIsland geology We greatly appreciate the assistance withPRIMELT modeling from Claude Herzberg We thankMikkel Schau for his insight and enthusiasm during field-work Jane Barling assisted with analyses by MC-ICP-MSThoughtful and thorough reviews of the manuscript byJohn Mahoney Fred Frey Wendy Bohrson DougalJerram and Teal Riley are greatly appreciated Fundingwas generously provided by the Rocks to Riches Programadministered by the BC amp Yukon Chamber of Mines in2004 by the BC Geological Survey and by NSERCDiscovery Grants to JSS and DW ARG was supportedby a University Graduate Fellowship at UBC
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
34
REFERENCESBarker F Brown A S Budahn J R amp Plafker G (1989) Back-arc
with frontal-arc component origin of Triassic Karmutsen basaltBritish Columbia Canada Chemical Geology 75 81^102
Blichert-Toft JWeis D Maerschalk C Agranier A amp Albare de F(2003) Hawaiian hot spot dynamics as inferred from the Hf and Pbisotope evolution of Mauna Kea volcano Geochemistry GeophysicsGeosystems 4(2) 1^27 doi1010292002GC000340
Bondre N R Duraiswami R A amp Dole G (2004) Morphologyand emplacement of flows from the Deccan Volcanic ProvinceIndia Bulletin of Volcanology 66 29^45
Carlisle D (1963) Pillow breccias and their aquagene tuffs QuadraIsland British Columbia Journal of Geology 71 48^71
Carlisle D (1972) Late Paleozoic to Mid-Triassic sedimentary-volcanic sequence on Northeastern Vancouver Island Geological
Society of Canada Paper Report of Activities 72-1 Part B 24^30Carlisle D amp SuzukiT (1974) Emergent basalt and submergent car-
bonate^clastic sequences including the Upper Triassic Dilleri andWelleri zones on Vancouver Island Canadian Journal of Earth
Sciences 11 254^279Chauvel C amp Blichert-Toft J (2001) A hafnium isotope and trace
element perspective on melting of the depleted mantle Earth and
Planetary Science Letters 190 137^151Chayes F (1954)The theory of thin-section analysis Journal of Geology
62 92^101Cho M Liou J G amp Maruyama S (1987) Transition from the zeo-
lite to prehnite^pumpellyite facies in the Karmutsen metabasitesVancouver Island British Columbia Journal of Petrology 27(2)467^494
Clowes R M Zelt C A Amor J R amp Ellis R M (1995)Lithospheric structure in the southern Canadian Cordillera froma network of seismic refraction lines Canadian Journal of Earth
Sciences 32(10) 1485^1513Coffin M F amp Eldholm O (1994) Large igneous provinces Crustal
structure dimensions and external consequences Reviews of
Geophysics 32(1) 1^36Cox K G (1980) A model for flood basalt volcanism Journal of
Petrology 21 629^650Eldholm O amp Coffin M F (2000) Large igneous provinces
and plate tectonics In Richards M A Gordon R G amp vander Hilst R D (eds) The History and Dynamics of Global
Plate Motions American Geophysical Union Geophysical Monograph 121309^326
Elliott T R Hawkesworth C J amp Grolaquo nvold K (1991) Dynamicmelting of the Iceland plume Nature 351 201^206
Escuder-Viruete J Pecurren rez-Estaucurren n A Contreras F Joubert MWeis D Ullrich T D amp Spadea P (2007) Plume mantle sourceheterogeneity through time Insights from the Duarte ComplexHispaniola northeastern Caribbean Journal of Geophysical Research112(B04203) doi1010292006JB004323
Farnetani C G Richards M A amp Ghiorso M S (1996)Petrological models of magma evolution and deep crustal structurebeneath hotspots and flood basalt provinces Earth and Planetary
Science Letters 143 81^94Fitton J G amp Godard M (2004) Origin and evolution of magmas
on the Ontong Java Plateau In Fitton J G Mahoney J JWallace P J amp Saunders A D (eds) Origin and Evolution of the
Ontong Java Plateau Geological Society London Special Publications 229151^178
Fitton J G Saunders A D Norry M J Hardarson B S ampTaylor R N (1997) Thermal and chemical structure of theIceland plume Earth and Planetary Science Letters 153 197^208
Fitton J G Saunders A D Kempton P D amp Hardarson B S(2003) Does depleted mantle form an intrinsic part of the Icelandplume Geochemistry Geophysics Geosystems 4(3) 1032 doi1010292002GC000424
Frey F A Huang S Blichert-Toft J Regelous M amp Boyet M(2005) Origin of depleted components in basalt related to theHawaiian hotspot Evidence from isotopic and incompatible ele-ment ratios Geochemistry Geophysics Geosystems 6(1) 23 doi1010292004GC000757
Galer S J G amp Abouchami W (1998) Practical application of leadtriple spiking for correction of instrumental mass discriminationMineralogical Magazine 62A 491^492
Ghiorso M S amp Sack R O (1995) Chemical mass transfer in mag-matic processes IV A revised and internally consistent thermody-namic model for the interpolation and extrapolation of liquid^solidequilibria in magmatic systems at elevated temperatures and pres-sures Contributions to Mineralogy and Petrology 119 197^212
Greene A R Scoates J S amp Weis D (2005)WrangelliaTerrane onVancouver Island Distribution of flood basalts with implicationsfor potential Ni^Cu^PGE mineralization In Grant B (ed)Geological Fieldwork 2004 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2005-1 209^220Greene A R Scoates J S Nixon G T amp Weis D (2006) Picritic
lavas and basal sills in the Karmutsen flood basalt provinceWrangellia northern Vancouver Island In Grant B (ed)Geological Fieldwork 2005 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2006-1 39^52Greene A R Scoates J S amp Weis D (2008) Wrangellia flood
basalts in Alaska A record of plume^lithosphere interaction in aLate Triassic accreted oceanic plateau Geochemistry Geophysics
Geosystems 9(Q12004) doi1010292008GC002092Greene A R Scoates J S Weis D amp Israel S (2009)
Geochemistry of triassic flood basalts from the Yukon (Canada)segment of the accreted Wrangellia oceanic plateau Lithosdoi101016jlithos200811010
Hauff F Hoernle K Tilton G Graham D W amp Kerr A C(2000) Large volume recycling of oceanic lithosphere over shorttime scales geochemical constraints from the Caribbean LargeIgneous Province Earth and Planetary Science Letters 174 247^263
Hauff F Hoernle K amp Schmidt A (2003) Sr^Nd^Pb compositionof Mesozoic Pacific oceanic crust (Site 1149 and 801 ODP Leg185)Implications for alteration of ocean crust and the input into theIzu^Bonin^Mariana subduction system Geochemistry Geophysics
Geosystems 4(8) doi1010292002GC000421Herzberg C (2004) Partial melting below the Ontong Java Plateau
In Fitton J G Mahoney J J Wallace P J amp Saunders A D(eds) Origin and Evolution of the Ontong Java Plateau Geological SocietyLondon Special Publications 229 179^183
Herzberg C (2006) Petrology and thermal structure of the Hawaiianplume from Mauna Kea volcano Nature 444(7119) 605
Herzberg C amp Asimow P D (2008) Petrology of some oceanicisland basalts PRIMELT2XLS software for primary magma cal-culation Geochemistry Geophysics Geosystems 9(Q09001) doi1010292008GC002057
Herzberg C amp OrsquoHara M J (2002) Plume-associated ultramaficmagmas of Phanerozoic age Journal of Petrology 43(10) 1857^1883
Herzberg C Asimow P D Arndt N Niu Y Lesher C MFitton J G Cheadle M J amp Saunders A D (2007)Temperatures in ambient mantle and plumes Constraints frombasalts picrites and komatiites Geochemistry Geophysics Geosystems8(Q02006) doi1010292006GC001390
Huang S amp Frey F A (2005) Trace element abundances of MaunaKea basalt from phase 2 of the Hawaii Scientific Drilling Project
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
35
Petrogenetic implications of correlations with major element con-tent and isotopic ratios Geochemistry Geophysics Geosystems 4(6)doi1010292002GC000322
Hussong D M Wipperman L K amp Kroenke L W (1979) Thecrustal structure of the Ontong Java and Manihiki OceanicPlateaus Journal of Geophysical Research 84(B11) 6003^6010
Jerram D A amp Widdowson M (2005) The anatomy of ContinentalFlood Basalt Provinces geological constraints on the processes andproducts of flood volcanism Lithos 79(3^4) 385^405
Jones D L Silberling N J amp Hillhouse J (1977)Wrangellia a dis-placed terrane in northwestern North America CanadianJournal ofEarth Sciences 14(11) 2565^2577
Kerr A C (2003) Oceanic Plateaus In Rudnick R (ed) The Crust
Treatise on GeochemistryVol 3 Oxford Elsevier Science pp 537^565Kerr A C amp Mahoney J J (2007) Oceanic plateaus Problematic
plumes potential paradigms Chemical Geology 241 332^353Kerr A CTarney J Marriner G F Klaver GT Saunders A D
amp Thirlwall M F (1996) The geochemistry and petrogenesis ofthe Late-Cretaceous picrites and basalts of Curacao NetherlandsAntilles a remnant of an oceanic plateau Contributions to
Mineralogy and Petrology 124(1) 29^43Kerr A C Marriner G F Tarney J Nivia A Saunders A D
Thirlwall M F amp Sinton A C (1997) Cretaceous basaltic ter-ranes in Western Colombia Elemental chronological and Sr-Ndisotopic constraints on petrogenesis Journal of Petrology 38(6)677^702
Kinzler R J (1997) Melting of mantle peridotite at pressuresapproaching the spinel to garnet transition Application to mid-ocean ridge basalt petrogenesis Journal of Geophysical Research
102(B1) 853^874Kinzler R J amp Grove T L (1992) Primary magmas of mid-ocean
ridge basalts 1 Experiments and methods Journal of Geophysical
Research 97(B5) 6885^6906Korenaga J (2005) Why did not the Ontong Java Plateau form sub-
aerially Earth and Planetary Science Letters 234 385^399Kuniyoshi S (1972) Petrology of the Karmutsen (Volcanic) Group
northeastern Vancouver Island British Columbia PhD disserta-tion University of California Los Angeles 242 pp
Lassiter J C DePaolo D J amp Mahoney J J (1995) Geochemistry ofthe Wrangellia flood basalt province Implications for the role ofcontinental and oceanic lithosphere in flood basalt genesis Journalof Petrology 36(4) 983^1009
LeBas M J (2000) IUGS reclassification of the high-Mg and picriticvolcanic rocks Journal of Petrology 41(10) 1467^1470
Longhi J (2002) Some phase equilibrium systematics of lherzolitemelting I Geochemistry Geophysics Geosystems 3(3) doi1010292001GC000204
MacDonald G A amp Katsura T (1964) Chemical composition ofHawaiian lavas Journal of Petrology 5 82^133
Mahoney J J (1987) An isotopic survey of Pacific oceanic plateausimplications for their nature and origin In Keating B HFryer P Batiza R amp Boehlert GW (eds) Seamounts Islands andAtolls American Geophysical Union Geophysical Monograph 43 207^220
Mahoney J LeRoex A P Peng Z Fisher R L amp Natland J H(1992) Southwestern limits of Indian Ocean Ridge mantle and theorigin of low 206Pb204Pb mid-ocean ridge basalt isotope systema-tics of the central Southwest Indian Ridge (178^508E) Journal ofGeophysical Research 97(B13) 19771^19790
Mahoney J J Sinton J M Kurz M D MacDougall J DSpencer K J amp Lugmair GW (1994) Isotope and trace elementcharacteristics of a super-fast spreading ridge East Pacific rise13^238S Earth and Planetary Science Letters 121 173^193
Massey N W D (1995a) Geology and mineral resources of theAlberni^Nanaimo Lakes sheet Vancouver Island 92F1W 92F2Eand part of 92F7E British Columbia Ministry of Energy Mines and
Petroleum Resources Paper 1992-2 132 ppMassey N W D (1995b) Geology and mineral resources of the
Cowichan Lake sheet Vancouver Island 92C16 British Columbia
Ministry of Energy Mines and Petroleum Resources Paper 1992-3 112 ppMassey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005a) Digital Geology Map of British Columbia Tile NM9 MidCoast BC British Columbia Ministry of Energy and Mines Geofile
2005-2Massey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005b) Digital Geology Map of British Columbia Tile NM10Southwest BC British Columbia Ministry of Energy and Mines Geofile
2005-3McDonough W F amp Sun S (1995) The composition of the Earth
Chemical Geology 120 223^253Monger JW H amp Journeay J M (1994) Basement geology and tec-
tonic evolution of theVancouver region In Monger JW H (ed)Geology and Geological Hazards of the Vancouver Region Southwestern
British Columbia Geological Survey of Canada Bulletin 481 3^25Muller J E (1977) Geology of Vancouver Island Geological Survey of
Canada Open File Map 463Muller J E (1980)The Paleozoic Sicker Group of Vancouver Island British
Columbia Geological Survey of Canada Paper 79-30 22 ppMuller J E Northcote K E amp Carlisle D (1974) Geology and
mineral deposits of Alert Bay^Cape Scott map area VancouverIsland British Columbia Geological Survey of Canada Paper 74-8 77
Neal C R Mahoney J J Kroenke L W Duncan R A ampPetterson M G (1997) The Ontong^Java Plateau InMahoney J J amp Coffin M F (eds) Large Igneous Provinces
Continental Oceanic and Planetary FloodVolcanism American Geophysical
Union Geophysical Monograph 100 183^216Niu Y Collerson K D Batiza R Wendt J I amp Regelous M
(1999) Origin of enriched-typed mid-ocean ridge basalt at ridgesfar from mantle plumes The East Pacific Rise at 11820rsquoN Journalof Geophysical Research 104 7067^7088
Nixon G T amp Orr A J (2007) Recent revisions to the EarlyMesozoic stratigraphy of Northern Vancouver Island (NTS 102I092L) and metallogenic implicatinos British Columbia InGrant B (ed) Geological Fieldwork 2006 British Columbia Ministry of
Energy Mines and Petroleum Resources Paper 2007-1 163^177Nixon GT Hammack J L KoyanagiV M Payie G J Haggart
JW Orchard M JTozerT Friedman R M Archibald D APalfy J amp Cordey F (2006a) Geology of the Quatsino^PortMcNeill area northernVancouver Island British Columbia Ministry
of Energy Mines and Petroleum Resources Geoscience Map 2006-2 scale150 000
Nixon G T Kelman M C Stevenson D Stokes L A ampJohnston K A (2006b) Preliminary geology of the Nimpkishmap area (NTS 092L07) northern Vancouver Island BritishColumbia In Grant B amp Newell J M (eds) Geological Fieldwork2005 British Columbia Ministry of Energy Mines and Petroleum Resources
Paper 2006-1 135^152Nixon G T Laroque J Pals A Styan J Greene A R amp
Scoates J S (2008) High-Mg lavas in the Karmutsen flood basaltsnorthernVancouver Island (NTS 092L) Stratigraphic setting andmetallogenic significance In Grant B (ed) Geological Fieldwork
2007 British Columbia Ministry of Energy Mines and Petroleum
Resources Paper 2008-1 175^190Nowell G M Kempton P D Noble S R Fitton J G
Saunders A Mahoney J J amp Taylor R N (1998) High precisionHf isotope measurements of MORB and OIB by thermal ionisation
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
36
mass spectrometry insights into the depleted mantle Chemical
Geology 149 211^233Parrish R R amp McNicoll V J (1992) U^Pb age determinations
from the southern Vancouver Island area British Columbia InRadiogenic Age and Isotopic Studies Report 5 Geological Survey of
Canada Paper 91-2 79^86Peate DW (1997) The Parana^Etendeka Province In Coffin M F
amp Mahoney J (eds) Large Igneous Provinces Continental Oceanic andPlanetary Flood Volcanism American Geophysical Union Geophysical
Monograph 100 217^245Perfit M R Fornari D J Ridley W I Kirk P D Casey J
Kastens K A Reynolds J R Edwards M Desonie DShuster R amp Paradis S (1996) Recent volcanism in the Siqueirostransform fault picritic basalts and implications for MORBmagma genesis Earth and Planetary Science Letters 141(1^4) 91^108
Pretorius W Weis D Williams G Hanano D Kieffer B ampScoates J S (2006) Complete trace elemental characterization ofgranitoid (USGSG-2GSP-2) reference materials by high resolutioninductively coupled plasma-mass spectrometry Geostandards and
Geoanalytical Research 30(1) 39^54Regelous M Niu Y Wendt J I Batiza R Greig A amp
Collerson K D (1999) Variations in the geochemistry of magma-tism on the East Pacific Rise at 10830rsquoN since 800 ka Earth and
Planetary Science Letters 168 45^63Recurren villon S Chauvel C Arndt N T Pik R Martineau F
Fourcade S amp Marty B (2002) Heterogeneity of the Caribbeanplateau mantle source Sr O and He isotopic compositions of oli-vine and clinopyroxene from Gorgona Island Earth and Planetary
Science Letters 205(1^2) 91Rhodes J M (1996) Geochemical stratigraphy of lava flows samples
by the Hawaii Scientific Drilling Project Journal of Geophysical
Research 101(B5) 11729^11746Rhodes J M amp Vollinger M J (2004) Composition of basaltic lavas
sampled by phase-2 of the Hawaii Scientific Drilling ProjectGeochemical stratigraphy and magma types Geochemistry
Geophysics Geosystems 5(3) doi1010292002GC000434Richards M A Jones D L Duncan R A amp DePaolo D J (1991)
A mantle plume initiation model for the Wrangellia flood basaltand other oceanic plateaus Science 254 263^267
Salters V J M amp Hart S R (1991) The mantle sources of oceanridges islands and arcs the Hf-isotope connection Earth and
Planetary Science Letters 104 364^380Salters V J M amp Stracke A (2004) Composition of the depleted
SaltersV J amp WhiteW (1998) Hf isotope constraints on mantle evo-lution Chemical Geology 145 447^460
SanoT amp Yamashita S (2004) Experimental petrology of basementlavas from Ocean Drilling Program Leg 192 implications for dif-ferentiation processes in Ontong Java Plateau magmas InFitton J G Mahoney J JWallace P J amp Saunders A D (eds)Origin and Evolution of the Ontong Java Plateau Geological Society
London Special Publications 229 185^218Saunders A D (2005) Large igneous provinces Origin and environ-
mental consequences Elements 1 259^263Self S ThordarsonT amp Keszthely L (1997) Emplacement of conti-
nental flood basalt lava flows In Mahoney J J amp Coffin M F(eds) Large Igneous Provinces Continental Oceanic and Planetary Flood
Volcanism American Geophysical Union Geophysical Monograph 100381^410
Shaw D M (2000) Continuous (dynamic) melting theory revisitedCanadian Mineralogist 38(5) 1041^1063
Sluggett C L (2003) Uranium^lead age and geochemical constraintson Paleozoic and Early Mesozoic magmatism in WrangelliaTerrane Saltspring Island British Columbia BSc thesisUniversity of British ColumbiaVancouver 84 pp
Storey M Mahoney J J amp Saunders A D (1997) Cretaceousbasalts in Madagascar and the transition between plume and con-tinental lithospheric mantle sources In Mahoney J J ampCoffin M F (eds) Large Igneous Provinces Continental Oceanic and
Planetary Flood Volcanism Aamerican Geophysical Union Geophysical
Monograph 100 95^122Surdam R C (1967) Low-grade metamorphism of the Karmutsen
Group PhD dissertation University of California Los Angeles288 pp
Sutherland-Brown A Yorath C J Anderson R G amp Dom K(1986) Geological maps of southern Vancouver IslandLITHOPROBE I 92C10 11 14 16 92F1 2 7 8 Geological Survey ofCanada Open File 1272
Tejada M L G Mahoney J J Castillo P R Ingle S P Sheth HC amp Weis D (2004) Pin-pricking the elephant evidence on theorigin of the Ontong Java Plateau from Pb^Sr^Hf^Nd isotopiccharacteristics of ODP Leg 192 basalts In Fitton J GMahoney J J Wallace P J amp Saunders A D (eds) Origin and
Evolution of the Ontong Java Plateau Geological Society London Special
Publications 229 133^150Thompson P M E Kempton P D amp Kerr A C (2008) Evaluation
of the effects of alteration and leaching on Sm^Nd and Lu^Hf sys-tematics in submarine mafic rocks Lithos 104(1^4) 164^176
Thompson P M E Kempton P D White R V Kerr A CTarney J Saunders A D Fitton J G amp McBirney A (2003)Hf-Nd isotope constraints on the origin of the CretaceousCaribbean plateau and its relationship to the Galapagos plumeEarth and Planetary Science Letters 217(1^2) 59^75
Vervoort J D Patchett P Blichert-Toft J amp Albare de F (1999)Relationships between Lu^Hf and Sm^Nd isotopic systems in theglobal sedimentary system Earth and Planetary Science Letters 16879^99
Walter M J (1998) Melting of garnet peridotite and the origin ofkomatiite and depleted lithosphere Journal of Petrology 39(1) 29^60
Weis D Kieffer B Maerschalk C Barling J de Jong JWilliams G A Hanano D Mattielli N Scoates J SGoolaerts A Friedman R A amp Mahoney J B (2006) High-pre-cision isotopic characterization of USGS reference materials byTIMS and MC-ICP-MS Geochemistry Geophysics Geosystems
7(Q08006) doi1010292006GC001283Weis D Kieffer B Hanano D Silva I N Barling J
Pretorius W Maerschalk C amp Mattielli N (2007) Hf isotopecompositions of US Geological Survey reference materialsGeochemistry Geophysics Geosystems 8(Q06006) doi1010292006GC001473
Wheeler J O amp McFeely P (1991) Tectonic assemblage map of theCanadian Cordillera and adjacent part of the United States ofAmerica Geological Survey of Canada Map 1712A
WhiteW M Albare de F amp Tecurren louk P (2000) High-precision analy-sis of Pb isotope ratios by multi-collector ICP-MS Chemical Geology167 257^270
Yorath C J Sutherland Brown A amp Massey N W D (1999)LITHOPROBE southern Vancouver Island British Columbia Geological
Survey of Canada Bulletin 498 145 ppZou H (1998) Trace element fractionation during modal and
non-model dynamic melting and open-system melting Amathematical treatment Geochimica et Cosmochimica Acta 62(11)1937^1945
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
37
Zou H amp Reid M R (2001) Quantitative modeling of trace elementfractionation during incongruent dynamic melting Geochimica et
Cosmochimica Acta 65(1) 153^162
APPENDIX SAMPLEPREPARAT ION ANDANALYT ICAL METHODSOnly the freshest rocks in the field were sampled and onlythe least altered samples were selected for chemical andisotopic analysis based on thorough petrographic inspec-tion Fifty-nine of the 129 collected samples were crushed(400 g) into small pieces 52mm in diameter in aRocklabs hydraulic piston crusher between WC plates tominimize contamination The coarse-crush was mixedand 100 g was powdered in a planetary mill using agatejars and balls cleaned with quartz sand between samples
ActLabs analytical methodsThe major- and trace-element compositions of the whole-rock powders were determined at Activation LaboratoriesLtd (Actlabs) in Ancaster Ontario Analytical techniquesand detection limits are also available from Actlabs (httpwwwactlabscommethsub_code4erehtm) The analyticalmethod for each of the elements analyzed is indicated inTable 2 For the major elements a 02 g sample was mixedwith a lithium metaborate^lithium tetraborate mixtureand fused in a graphite crucible The molten mixture waspoured into a 5 HNO3 solution and shaken until dis-solved (30min) The samples were analyzed for majoroxides and selected trace elements by inductively coupledplasma optical emission spectrometry (ICP-OES) on acombination simultaneous^sequential Thermo Jarrell-AshEnviro II system Internal calibration was achieved usinga variety of international reference materials (eg W-2BIR-1 DNC-1) and independent control samplesAdditional trace elements were analyzed by both instru-mental neutron activation analysis (INAA) and induc-tively couple plasma mass spectrometry (ICP-MS) ForINAA15^25 g of sample was weighed into small polyeth-ylene vials and irradiated with a control international ref-erence material CANMET WMS-1 and NiCr flux wires ata thermal neutron flux of 71012 neutronscm2 per s in theMcMaster Nuclear Reactor Following a 7 day waitingperiod the samples were measured on an Ortec high-purity Ge detector linked to a Canberra Series 95 multi-channel analyzer Activities for each element weredecay- and weight-corrected and compared with a detectorcalibration developed from multiple international certifiedreference materials For ICP-MS 025 g of sample wasdigested in HF followed by HNO3^HClO4 heated andtaken to dryness The samples were brought back into solu-tion with HCl Samples were analyzed using a PerkinElmer Optima 3000 ICP In-lab standards or certified ref-erence materials (eg W-2 BIR-1 DNC-1) were used for
quality control A total of 15 blind duplicates were ana-lyzed from the same sample powders to assessreproducibility
University of Massachusetts analyticalmethodsFourteen sample duplicate powders taken from the same100 g sample aliquot (high-MgO basalts and picrites)were also analyzed at the Ronald B Gilmore X-RayFluorescence (XRF) Laboratory at the University ofMassachusetts Major elements were measured on a fusedLa-bearing lithium borate glass disc using a SiemensMRS-400 spectrometer Trace element concentrations(Rb Sr Ba Ce Nb Zr Y Pb Zn Ga Ni Cr V) weremeasured on a separate powder pellet using a PhilipsPW2400 sequential spectrometer using a Rh tube LOIand ferrous iron measurements were made as describedby Rhodes amp Vollinger (2004) Precision and accuracy esti-mates for the data have been described by Rhodes (1996)and Rhodes amp Vollinger (2004) Results for each sampleare the average of two separate analyses and are shown inElectronic Appendix 2 for a comparison of analyses ofduplicate powders using XRFand ICP-MS
PCIGR analytical methodsA subset of 19 samples was selected for high-precisiontrace-element analysis and Sr Nd Pb and Hf isotopicanalysis at the Pacific Centre for Isotopic andGeochemical Research (PCIGR) at the University ofBritish Columbia (UBC) Samples were selected from the59 samples analyzed for whole-rock chemistry at ActLabsbased on major- and trace-element chemistry alteration(low LOI and petrographic alteration index) and samplelocation Samples were prepared for trace-element analysisat the PCIGR by the technique described by Pretorius et al(2006) on unleached rock powders Sample powders(100mg) were weighed in 7ml screw-top Savillex bea-kers and dissolved in 1ml 14N HNO3 and 5ml 48 HFon a hotplate for 48 h at 1308C with periodic ultrasonica-tion Samples were dried and redissolved in 6ml 6N HClon a hotplate for 24 h and then dried and redissolved in1ml concentrated HNO3 for 24 h before final dryingTrace element abundances were measured by high-resolution inductively coupled plasma-mass spectrometry(HR-ICP-MS) with aThermo Finnigan Element2 systemfollowing the procedures described by Pretorius et al(2006) within 24 h of redissolution HFSE and LILE weremeasured in medium-resolution mode at 2000 dilutionusing a PFATeflon spray chamber washed with aqua regiafor 3min between samples REE were measured in high-resolution mode and U and Pb in low-resolution modeat 2000 dilution using a glass spray chamber washedwith 2 HNO3 between samples Total procedural blanksand reference materials (BCR-2 BHVO-2) were analyzedwith the batch of samples Indium was used as an internal
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
38
standard in all samples and standard solutionsBackground and standard solutions were analyzed afterevery five samples to detect memory effects and massdrift Results for PCIGR trace-element analyses areshown in Electronic Appendix 3 for the full set of analyzedtrace elementsSample digestion for purification of Sr Nd Hf and Pb
for column chemistry began by weighing each samplepowder (400^500mg) prior to leaching All samples wereleached with 6N HCl and placed in an ultrasonic bath for15min Samples were rinsed twice with 18 M-cm H2Obetween each leaching step (15 total) until the supernatantliquid was clear [following the method of Mahoney (1987)]Samples were then dried on a hotplate for 24 h andweighed again Sample solutions were then prepared bydissolving 100^250mg of the leached powder dissolvedin 1ml 14N HNO3 and 10ml 48 HF on a hotplate for48 h at 1308C with periodic ultrasonication Samples weredried and redissolved in 6ml 6N HCl on a hotplate for24 h and then dried Pb was separated using anion exhangecolumns and the discard was used for Sr REE and Hfseparation Nd was separated from the other REE and Hfrequired two additional purification steps Detailed proce-dures for column chemistry for separating Sr Nd and Pbat the PCIGR have been described byWeis et al (2006) andHf purification has been described byWeis et al (2007) Srand Nd isotope ratios were measured by thermal ioniza-tion mass spectrometry (TIMS) on a Thermo FinniganTriton system in static mode with relay matrix rotation ona single Ta and double Re^Ta filament respectively Fourto five filaments per barrel of 21 were occupied by stan-dards (NBS 987 for Sr and LaJolla for Nd) for each barrelwhere samples were run Sample Sr and Nd isotopic com-positions were corrected for mass fractionation using86Sr88Srfrac14 01194 and 146Nd144Ndfrac14 07219 Each samplewas then normalized using the barrel average of the refer-ence material relative to the values of143Nd144Ndfrac14 0511858 and 87Sr86Srfrac14 0710248 (Weiset al 2006) During the course of the Vancouver Islandanalyses the La Jolla Nd standard gave an average valueof 05118566 (nfrac14 7) and NBS987 standard gave an aver-age of 07102408 (nfrac1411) (2 error is reported as times106) 147Sm144Nd errors are 15 or 0006 USGeological Survey (USGS) reference material BHVO-2was processed with the samples and yielded Sr and Nd iso-topic ratios of 07034607 and 05129786 respectivelyThese are in agreement with the published values of070347920 and 051298411 respectively (Weis et al2006)
Pb and Hf isotopic compositions were analyzed by mul-tiple collector-inductively coupled plasma-mass spectrome-try (MC-ICP-MS) on a static multi-collection on a NuPlasma (Nu Instruments) system The detailed analyticalprocedure for Pb isotopic analyses on the Nu Plasma atthe PCIGR has been described by Weis et al (2006) Theconfiguration for Pb analyses allows for collection of PbTl and Hg togetherTl and Hg are used to monitor instru-mental mass discrimination and isobaric overlap repec-tively All sample solutions were analyzed withapproximately the same PbTl ratio (4) as the referencematerial NIST SRM 981 To accomplish this a small ali-quot of each sample solution from the Pb columns was ana-lyzed on the Element2 to determine the precise amount ofPb available for analysis on the Nu Plasma The SRM 981standard was run after every two samples on the NuPlasma During the time samples were run analyses ofthe SRM 981 Pb reference material gave values of206Pb204Pbfrac1416940322 207Pb204Pbfrac1415495823and 208Pb204Pbfrac14 36713164 (nfrac14 61 2 error is reportedas times 104) these values are in excellent agreement withreported TIMS triple-spike values of Galer amp Abouchami(1998) Fractionation-corrected Pb isotopic ratios were fur-ther corrected by the sample-standard bracketing methodor the ln^ln correction method described by White et al(2000) and Blichert-Toft et al (2003) Leached powders ofUSGS reference material BHVO-2 yielded Pb isotopicratios of 206Pb204Pbfrac14186454 8 207Pb204Pbfrac14 1549105 and 208Pb204Pbfrac14 38222514 and BCR-2yielded 206Pb204Pbfrac141880466 207Pb204Pbfrac14 1562518 and 208Pb204Pbfrac14 3883496 (2 error isreported as times 104) These values are in agreement withthose for leached residues of BHVO-2 and BCR-2 fromWeis et al (2006)Hf isotopic compositions were analyzed following the
procedures detailed byWeis et al (2007) The configurationfor Hf analyses monitored Lu mass 175 andYb mass 172 toallow for interference correction to masses 174 and 176 Hfisotopic ratios were normalized internally for mass fractio-nation to a 179Hf177Hf ratio of 07325 using an exponentialcorrection Standards were run after every two samplesand sample results were normalized to the ratio of the in-run daily average and a 176Hf177Hf ratio for JMC-475 of0282160 During the course of analyses the Hf standardJMC-475 gave an average value 02821679 (nfrac14130)USGS reference materials BCR-2 and BHVO-2 were pro-cessed with the samples and yielded Hf isotopic ratios of02828675 and 02831005 respectively Publishedvalues for BCR-2 and BHVO-2 are 02828717 and0283104 8 respectively (Weis et al 2007)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
39
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
Gray lines = final melting pressure
L + Ol + Opx
07
08
09
Pressure (GPa)
10
3
4
5
6
1
SOLIDUS
01
04
0506
L + Ol
2
03
02
PeridotiteKR-4003
0 10 20 30 40MgO (wt)
4
6
8
10
12
14
FeO(wt)
Thick black lines = initial melting pressure
Dashed lines = melt fraction
Melt Fraction
2
3 4 5 6
L+Ol+Opx
SolidusSpinel
SolidusGarnet Peridotite
7
L+Ol
0 10 20 30 4040
45
50
55
60
MgO (wt)
SiO2
(wt)
PeridotiteKR-4003
Melt Fraction
0 10 20 305
10
15
CaO(wt)
MgO (wt)
3
4 5 6 7
2
46
2
Peridotite Partial Melts
Thick black line = initial melting pressureGray lines = final melting pressure
Peridotite
SOLIDUS
Liquid compositions Fertile peridotite source
Accumulated Fractional Melting model
02
Pressure (GPa)
0302
04
Pressure (GPa)
Picrite
High-MgO basalt
Tholeiitic basalt
(c)
(d)
Karmutsen flood basaltsOlivine addition modelOlivine addition (5 increment)Primary magma
PicriteHigh-MgO basaltTholeiitic basalt
(a)
(b)
800 850 900
910
920
930
940
950
Karmutsen flood basalts
Olivine addition modelOlivine addition (5 increment)Primary magma
Gray lines = Fo contents of olivine
Fig 12 Estimated primary magma compositions for three Keogh Lake picrites and high-MgO basalts (samples 4723A4 4723A135616A7) using the forward and inverse modeling technique of Herzberg et al (2007) Karmutsen compositions and modeling results areoverlain on diagrams provided by C Herzberg (a) Whole-rock FeO vs MgO for Karmutsen samples from this study Total iron estimatedto be FeO is 090 Gray lines show olivine compositions that would precipitate from liquid of a given MgO^FeO composition Black lineswith crosses show results of olivine addition (inverse model) using PRIMELT1 (Herzberg et al 2007) Gray field highlights olivinecompositions with Fo contents of 91^92 predicted to precipitate (b) FeO vs MgO showing Karmutsen lava compositions and results ofa forward model for accumulated fractional melting of fertile peridotite KR-4003 (Kettle River Peridotite Walter 1998) (c) SiO2 vsMgO for Karmtusen lavas and model results (d) CaO vs MgO showing Karmtusen lava compositions and model results To brieflysummarize the technique [see Herzberg et al (2007) for complete description] potential parental magma compositions for the high-MgOlava series were selected (highest MgO and appropriate CaO) and using PRIMELT1 software olivine was incrementally added tothe selected compositions to show an array of potential primary magma compositions (inverse model) Then using PRIMELT1 theresults from the inverse model were compared with a range of accumulated fractional melts for fertile peridotite derived fromparameterization of the experimental results for KR-4003 from Walter (1998) forward model Herzberg amp OrsquoHara (2002) A melt fractionwas sought that was unique to both the inverse and forward models (Herzberg et al 2007) A unique solution was found when there was a com-mon melt fraction for both models in FeO^MgO and CaO^MgO^Al2O3^SiO2 (CMAS) projection space This modeling assumes thatolivine was the only phase crystallizing and ignores chromite precipitation and possible augite fractionation in the mantle (Herzbergamp OrsquoHara 2002) Results are best for a residue of spinel lherzolite (not pyroxenite) The presence of accumulated olivine in samples ofthe Keogh Lake picrites used as starting compositions does not significantly affect the results because we are modeling addition of olivine(see text for further description) The tholeiitic basalts cannot be used for modeling because they are all saturated in plagthorn cpxthornolThick black line for initial melting pressure at 4 GPa is not shown in (c) for clarity Gray area in (a) indicates parental magmas thatwill crystallize olivine with Fo 91^92 at the surface Dashed lines in (c) indicate melt fraction Solidi for spinel and garnet peridotite arelabelled in (c)
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
The PRIMELT1 results indicate that if the Karmutsenpicritic lavas were derived from accumulated fractionalmelting of fertile peridotite the primary magmas wouldhave contained 15^17wt MgO and 10wt CaO(Fig 12 Table 6) formed from 23^27 partial meltingand would have crystallized olivine with a Fo content of 91 (Fig 12 Table 6) Calculated mantle temperatures forthe Karmutsen picrites (14908C) indicate melting ofanomalously hot mantle (100^2008C above ambient) com-pared with ambient mantle that produces mid-ocean ridgebasalt (MORB 1280^14008C Herzberg et al 2007)which initiated between 3 and 4 GPa Several picritesappear to be near-primary melts and required very littleaddition of olivine (eg sample 93G171 with measured158wt MgO) These temperature and melting esti-mates of the Karmutsen picrites are consistent withdecompression of hot mantle peridotite in an actively con-vecting plume head (ie plume initiation model) Threesamples (picrite 5614A1 and high-MgO basalts 5614A3and 5614A5) are lower in iron than the other Karmutsenlavas with FeO in the 65^80wt range (Fig 12c) andhave MgO and FeO contents that are similar to primitiveMORB from the Sequeiros fracture zone of the EastPacific Rise (EPR Perfit et al 1996) These samples
however differ from EPR MORB in having higher SiO2
and lower Al2O3 and they are restricted geographicallyto one area (Sara Lake Fig 2)We therefore have evidence for primary magmas with
MgO contents that range from a possible low of 110wt to as high as 174wt with inferred mantle potential tem-peratures from 1340 to 15208C This is similar to the rangedisplayed by lavas from the Ontong Java Plateau deter-mined by Herzberg (2004) who found that primarymagma compositions assuming a peridotite sourcewould contain 17wt MgO and 10wt CaO(Table 6) and that these magmas would have formed by27 melting at a mantle potential temperature of15008C (first melting occurring at 36 GPa) Fitton ampGodard (2004) estimated 30 partial melting for theOntong Java lavas based on the Zr contents of primarymagmas of Kroenke-type basalt calculated by incrementaladdition of equilibrium olivine However if a considerableamount of eclogite was involved in the formation of theOntong Java plateau magmas the excess temperatures esti-mated by Herzberg et al (2007) may not be required(Korenaga 2005) The variability in mantle potential tem-perature for the Keogh Lake picrites is also similar to thewide range of temperatures that have been calculated for
Table 6 Estimated primary magma compositions for Karmutsen basalts and other oceanic plateaus or islands
Sample 4723A4 4723A13 5616A7 93G171 Average OJP Mauna Keay Gorgonaz
Ontong Java primary magma composition for accumulated fraction melting (AFM) from Herzberg (2004)yMauna Kea primary magma composition is average of four samples of Herzberg (2006 table 1)zGorgona primary magma composition for AFM for 1F fertile source from Herzberg amp OrsquoHara (2002 table 4)Total iron estimated to be FeO is 090
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
27
lavas from single volcanoes in the Galapagos Hawaii andelsewhere by Herzberg amp Asimow (2008) Those workersinterpreted the range to reflect the transport of magmasfrom both the cool periphery and the hotter axis of aplume A thermally heterogeneous plume will yield pri-mary magmas with different MgO contents and the rangeof primary magma compositions and their inferred mantlepotential temperatures for Karmutsen lavas may reflectmelting from different parts of the plume
Source of the Karmutsen Formation lavasThe Sr^Nd^Hf^Pb isotopic compositions of theKarmutsen volcanic rocks on Vancouver Island indicatethat they formed from plume-type mantle containing adepleted component that is compositionally similar to butdistinct from other oceanic plateaus that formed in thePacific Ocean (eg Ontong Java and Caribbean Fig 13)Trace element and isotopic constraints can be used to testwhether the depleted component of the KarmutsenFormation was intrinsic to the mantle plume and distinctfrom the source of MORB or the result of mixing or invol-vement of ambient depleted MORB mantle with plume-type mantle The Karmutsen tholeiitic basalts have initialeNd and
87Sr86Sr ratios that fall within the range of basaltsfrom the Caribbean Plateau (eg Kerr et al 1996 1997Hauff et al 2000 Kerr 2003) and a range of Hawaiian iso-tope data (eg Frey et al 2005) and lie within and abovethe range for the Ontong Java Plateau (Tejada et al 2004Fig 13a) Karmutsen tholeiitic basalts with the highest ini-tial eNd and lowest initial 87Sr86Sr lie within and justbelow the field for age-corrected northern EPR MORB at230 Ma (eg Niu et al1999 Regelous et al1999) Initial Hfand Nd isotopic compositions for the Karmutsen samplesare noticeably offset to the low side of the ocean islandbasalt (OIB) array (Vervoort et al 1999) and lie belowand slightly overlap the low eHf end of fields for Hawaiiand the Caribbean Plateau (Fig 13c) the Karmutsen sam-ples mostly fall between and slightly overlap a field forage-corrected Pacific MORB at 230 Ma and Ontong Java(Mahoney et al 1992 1994 Nowell et al 1998 Chauvel ampBlichert-Toft 2001) Karmutsen lava Pb isotope ratios over-lap the broad range for the Caribbean Plateau and thelinear trends for the Karmutsen samples intersect thefields for EPR MORB Hawaii and Ontong Java in208Pb^206Pb space but do not intersect these fields in207Pb^206Pb space (Fig 13d and e) The more radiogenic207Pb204Pb for a given 206Pb204Pb of the Karmutsenbasalts requires high UPb ratios at an ancient time when235U was abundant similar to the Caribbean Plateau andenriched in comparison with EPR MORB Hawaii andOntong JavaThe low eHf^low eNd component of the Karmutsen
basalts that places them below the OIB mantle array andpartly within the field for age-corrected Pacific MORBcan form from low ratios of LuHf and high SmNd over
time (eg Salters amp White 1998) MORB will develop aHf^Nd isotopic signature over time that falls below themantle array Low LuHf can also lead to low 176Hf177Hffrom remelting of residues produced by melting in theabsence of garnet (eg Salters amp Hart 1991 Salters ampWhite 1998) The Hf^Nd isotopic compositions of theKarmutsen Formation indicate the possible involvementof depleted MORB mantle however similar to Iceland(Fitton et al 2003) Nb^Zr^Y variations provide a way todistinguish between a depleted component within theplume and a MORB-like component The Karmutsenbasalts and picrites fall within the Iceland array and donot overlap the MORB field in a logarithmic plot of NbYvs ZrY (Fig 13b) using the fields established by Fittonet al (1997) Similar to the depleted component within theCaribbean Plateau (Thompson et al 2003) the trend ofHf^Nd isotopic compositions towards Pacific MORB-likecompositions is not followed by MORB values in Nb^Zr^Y systematics and this suggests that the depleted compo-nent in the source of the Karmutsen basalts is distinctfrom the source of MORB and probably an intrinsic partof the plume The relatively radiogenic Pb isotopic ratiosand REE patterns distinct from MORB also support thisinterpretationTrace-element characteristics suggest the possible invol-
vement of sub-arc lithospheric mantle in the generation ofthe Karmutsen picrites Lassiter et al (1995) suggested thatmixing of a plume-type source with eNdthorn 6 to thorn 7 witharc material with low NbTh could reproduce the varia-tions observed in the Karmutsen basalts however theyconcluded that the absence of low NbLa ratios in theirsample of tholeiitic basalts restricts the amount of arc litho-sphere involved The study of Lassiter et al (1995) wasbased on major and trace element and Sr Nd and Pb iso-topic data for a suite of 29 samples from Buttle Lake in theStrathcona Provincial Park on central Vancouver Island(Fig 1) Whereas there are no clear HFSE depletions inthe Karmutsen tholeiitic basalts the low NbLa (and TaLa) of the Karmutsen picritic lavas in this study indicatethe possible involvement of Paleozoic arc lithosphere onVancouver Island (Fig 8)
REE modeling dynamic melting andsource mineralogyThe combined isotopic and trace-element variations of thepicritic and tholeiitic Karmutsen lavas indicate that differ-ences in trace elements may have originated from differentmelting histories For example the LuHf ratios of theKarmutsen picritic lavas (mean 008) are considerablyhigher than those in the tholeiitic basalts (mean 002)however the small range of overlapping initial eHf valuesfor the two lava suites suggests that these differences devel-oped during melting within the plume (Fig 9b) Below wemodel the effects of source mineralogy and depth of
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
28
melting on differences in trace-element concentrations inthe tholeiitic basalts and picritesDynamic melting models simulate progressive decom-
pression melting where some of the melt fraction isretained in the residue and only when the degree of partialmelting is greater than the critical mass porosity and the
source becomes permeable is excess melt extracted fromthe residue (Zou1998) For the Karmutsen lavas evolutionof trace element concentrations was simulated using theincongruent dynamic melting model developed by Zou ampReid (2001) Melting of the mantle at pressures greater than1 GPa involves incongruent melting (eg Longhi 2002)
OIB array
PacificMORB(230 Ma)
(0 Ma)
EPR(230 Ma)
-4
-2
0
2
4
8
10
12
14
16
18
20
22
-4 -2 0 2 4 6 8 10 12 14
minus2
0
2
4
6
8
10
12
14
0702 0703 0704 0705 0706
IndianMORB
87Sr 86Sr (initial)
Hf (initial)
(a) (c)
Nd (initial)
Karmutsen tholeiitic basalt
HawaiiOntong JavaCaribbean Plateau
Karmutsen high-MgO lava
high-MgO lavasKarmutsen
1540
1545
1550
1555
1560
1565
175 180 185 190 195 200375
380
385
390
395
175 180 185 190 195 200
(d) (e)
EPR
EPR
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb
Karmutsen Formation (initial)
HawaiiOntong JavaCaribbean Plateau
East Pacific Rise
Karmutsen Formation (measured)
Juan de FucaGorda
ε Nd
(initial)
Karmutsen Formation (different age-correction for 3 picrites)
(0 Ma)
NbY
001
01
1
1 10
ZrY
OIB
NMORB
(b)
Iceland array
6
Fig 13 Comparison of age-corrected (230 Ma) Sr^Nd^Hf^Pb isotopic compositions for Karmutsen flood basalts onVancouver Island withage-corrected OIB and MORB and Nb^Zr^Y variation (a) Initial eNd vs 87Sr86Sr (b) NbYand ZrY variation (after Fitton et al 1997)(c) Initial eHf vs eNd (d) Measured and initial 207Pb204Pb vs 206Pb204Pb (e) Measured and initial 208Pb204Pb vs 206Pb204Pb References fordata sources are listed in Electronic Appendix 4 Most of the compiled data were extracted from the GEOROC database (httpgeorocmpch-mainzgwdgdegeoroc) OIB array line in (c) is fromVervoort et al (1999) EPR East Pacific Rise Several high-MgO samples affected bysecondary alteration were not plotted for clarity in Pb isotope plots Estimation of fields for age-corrected Pacific MORB and EPR at 230 Mawere made using parent^daughter concentrations (in ppm) from Salters amp Stracke (2004) of Lufrac14 0063 Hffrac14 0202 Smfrac14 027 Ndfrac14 071Rbfrac14 0086 and Srfrac14 836 In the NbYvs ZrYplot the normal (N)-MORB and OIB fields not including Icelandic OIB are from data com-pilations of Fitton et al (1997 2003) The OIB field is data from 780 basalts (MgO45wt ) from major ocean islands given by Fitton et al(2003) The N-MORB field is based on data from the Reykjanes Ridge EPR and Southwest Indian Ridge from Fitton et al (1997) The twoparallel gray lines in (b) (Iceland array) are the limits of data for Icelandic rocks
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
29
with olivine being produced during the reactioncpxthornopxthorn spfrac14meltthornol for spinel peridotites The mod-eling used coefficients for incongruent melting reactionsbased on experiments on lherzolite melting (eg Kinzleramp Grove 1992 Longhi 2002) and was separated into (1)melting of spinel lherzolite (2) two combined intervals ofmelting for a garnet lherzolite (gt lherz) and spinel lher-zolite (sp lherz) source (3) melting of garnet lherzoliteThe model solutions are non-unique and a range indegree of melting partition coefficients melt reactioncoefficients and proportion of mixed source componentsis possible to achieve acceptable solutions (see Fig 14 cap-tion for description of modeling parameters anduncertainties)The LREE-depleted high-MgO lavas require melting of
a depleted spinel lherzolite source (Fig 14a) The modelingresults indicate a high degree of melting (22^25) similarto the results from PRIMELT1 based on major-elementvariations (discussed above) of a LREE-depleted sourceThe high-MgO lavas lack a residual garnet signature Theenriched tholeiitic basalts involved melting of both garnetand spinel lherzolite and represent aggregate melts pro-duced from continuous melting throughout the depth ofthe melting column (Fig 14b) Trace-element concentra-tions in the tholeiitic and picritic lavas are not consistentwith low-degree melting of garnet lherzolite alone(Fig 14c) The modeling results indicate that the aggregatemelts that formed the tholeiitic basalts would haveinvolved lesser proportions of enriched small-degreemelts (1^6 melting) generated at depth and greateramounts of depleted high-degree melts (12^25) gener-ated at lower pressure (a ratio of garnet lherzolite tospinel lherzolite melt of 14 provides the best fits seeFig 14b and description in figure caption) The totaldegree of melting for the tholeiitic basalts was high(23^27) similar to the degree of partial melting in thespinel lherzolite facies for the primary melts that eruptedas the Keogh Lake picrites of a source that was also prob-ably depleted in LREE
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
(c)
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
Melting of garnet lherzolite
High-MgO lavas
Melting of spinel lherzolite
source (07DM+03PM)
source (07DM+03PM)
6 melting
30 melting
(b)
(a)
30 melting
Sam
ple
Cho
ndrit
eS
ampl
eC
hond
rite 20 melting
1 melting
Melting of garnet lherzoliteand spinel lherzolite
x gt lherz + 4x sp lherz
5 meltingx gt lherz + 4x sp lherz
Sam
ple
Cho
ndrit
e
Tholeiitic lavas
Tholeiitic lavas
High-MgO lavas
source (07DM+03PM)
Fig 14 Trace-element modeling results for incongruent dynamicmantle melting for picritic and tholeiitic lavas from the KarmutsenFormation Three steps of modeling are shown (a) melting of spinellherzolite compared with high-MgO lavas (b) melting of garnet lher-zolite and spinel lherzolite compared with tholeiitic lavas (c) meltingof garnet lherzolite compared with tholeiitic and picritic lavasShaded fields in each panel for tholeiitic basalts and picrites arelabeled Patterns with symbols in each panel are modeling results(patterns are 5 melting increments in (a) and (b) and 1 incre-ments in (c) PM primitive mantle DM depleted MORB mantle gtlherz garnet lherzolite sp lherz spinel lherzolite In (b) the ratios ofper cent melting for garnet and spinel lherzolite are indicated (x gtlherzthorn 4x sp lherz means that for 5 melting 1 melt in garnetfacies and 4 melt in spinel facies where x is per cent melting in theinterval of modeling) The ratio of the respective melt proportions inthe aggregate melt (x gt lherzthorn 4x sp lherz) was kept constant andthis ratio of melt contribution provided the best fit to the data Arange of proportion of source components (07DMthorn 03PM to
09DMthorn 01PM) can reproduce the variation in the high-MgOlavas Melting modeling uses the formulation of Zou amp Reid (2001)an example calculation is shown in their Appendix PM fromMcDonough amp Sun (1995) and DM from Salters amp Stracke (2004)Melt reaction coefficients of spinel lherzolite from Kinzler amp Grove(1992) and garnet lherzolite fromWalter (1998) Partition coefficientsfrom Salters amp Stracke (2004) and Shaw (2000) were kept constantSource mineralogy for spinel lherzolite (018cpx027opx052ol003sp) from Kinzler (1997) and for garnet lherzolite(034cpx008opx053ol005gt) from Salters amp Stracke (2004) Arange of source compositions for melting of spinel lherzolite(07DMthorn 03PM to 03DMthorn 07PM) reproduce REE patternssimilar to those of the high-MgO lavas with best fits for 22ndash25melting and 07DMthorn 03PM source composition Concentrationsof tholeiitic basalts cannot be reproduced from an entirelyprimitive or depleted source
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
30
The uncertainty in the composition of the source in thespinel lherzolite melting model for the high-MgO lavasprecludes determining whether the source was moredepleted in incompatible trace elements than the source ofthe tholeiitic lavas (see Fig 14 caption for description) It ispossible that early high-pressure low-degree melting left aresidue within parts of the plume (possibly the hotter inte-rior portion) that was depleted in trace elements (egElliott et al 1991) further decompression and high-degreemelting of such depleted regions could then have generatedthe depleted high-MgO Karmutsen lavas Shallow high-degree melting would preferentially sample depletedregions whereas deeper low-degree melting may notsample more refractory depleted regions (eg CaribbeanEscuder-Viruete et al 2007) The picrites lie at the highend of the range of Nd isotopic compositions and havelow NbZr similar to depleted Icelandic basalts (Fittonet al 2003) therefore the picrites may represent the partsof the mantle plume that were the most depleted prior tothe initiation of melting As Fitton et al (2003) suggestedthese depleted melts may only rarely be erupted withoutcombining with more voluminous less depleted melts andthey may be detected only as undifferentiated small-volume flows like the Keogh Lake picrites within theKarmutsen Formation on Vancouver IslandDecompression melting within the mantle plume initiatedwithin the garnet stability field (35^25 GPa) at highmantle potential temperatures (Tp414508C) and pro-ceeded beneath oceanic arc lithosphere within the spinelstability field (25^09 GPa) where more extensivedegrees of melting could occur
Magmatic evolution of Karmutsentholeiitic basaltsPrimary magmas in large igneous provinces leave exten-sive coarse-grained cumulate residues within or below thecrust these mafic and ultramafic plutonic sequences repre-sent a significant proportion of the original magma thatpartially crystallized at depth (eg Farnetani et al 1996)For example seismic and petrological studies of OntongJava (eg Farnetani et al 1996) combined with the use ofMELTS (Ghiorso amp Sack 1995) indicate that the volcanicsequence may represent only 40 of the total magmareaching the Moho Neal et al (1997) estimated that30^45 fractional crystallization took place for theOntong Java magmas and produced cumulate thicknessesof 7^14 km corresponding to 11^22 km of flood basaltsdepending on the presence of pre-existing oceanic crustand the total crustal thickness (25^35 km) Interpretationsof seismic velocity measurements suggest the presence ofpyroxene and gabbroic cumulates 9^16 km thick beneaththe Ontong Java Plateau (eg Hussong et al 1979)Wrangellia is characterized by high crustal velocities in
seismic refraction lines which is indicative of mafic
plutonic rocks extending to depth beneath VancouverIsland (Clowes et al 1995)Wrangellia crust is 25^30 kmthick beneath Vancouver Island and is underlain by astrongly reflective zone of high velocity and density thathas been interpreted as a major shear zone where lowerWrangellia lithosphere was detached (Clowes et al 1995)The vast majority of the Karmutsen flows are evolved tho-leiitic lavas (eg low MgO high FeOMgO) indicating animportant role for partial crystallization of melts at lowpressure and a significant portion of the crust beneathVancouver Island has seismic properties that are consistentwith crystalline residues from partially crystallizedKarmutsen magmas To test the proportion of the primarymagma that fractionated within the crust MELTS(Ghiorso amp Sack 1995) was used to simulate fractionalcrystallization using several estimated primary magmacompositions from the PRIMELT1 modeling results Themajor-element composition of the primary magmas couldnot be estimated for the tholeiitic basalts because of exten-sive plagthorn cpxthornol fractionation and thus the MELTSmodeling of the picrites is used as a proxy for the evolutionof major elements in the volcanic sequence A pressure of 1kbar was used with variable water contents (anhydrousand 02wt H2O) calculated at the quartz^fayalite^magnetite (QFM) oxygen buffer Experimental resultsfrom Ontong Java indicate that most crystallizationoccurs in magma chambers shallower than 6 km deep(52 kbar Sano amp Yamashita 2004) however some crys-tallization may take place at greater depths (3^5 kbarFarnetani et al 1996)A selection of the MELTS results are shown in Fig 15
Olivine and spinel crystallize at high temperature until15^20wt of the liquid mass has fractionated and theresidual magma contains 9^10wt MgO (Fig 15f) At1235^12258C plagioclase begins crystallizing and between1235 and 11908C olivine ceases crystallizing and clinopyr-oxene saturates (Fig 15f) As expected the addition of clin-opyroxene and plagioclase to the crystallization sequencecauses substantial changes in the composition of the resid-ual liquid FeO and TiO2 increase and Al2O3 and CaOdecrease while the decrease in MgO lessens considerably(Fig 15) The near-vertical trends for the Karmutsen tho-leiitic basalts especially for CaO do not match the calcu-lated liquid-line-of-descent very well which misses manyof the high-CaO high-Al2O3 low-FeO basalts A positivecorrelation between Al2O3CaO and SrNd indicates thataccumulation of plagioclase played a major role in the evo-lution of the tholeiitic basalts (Fig 15g) Increasing thecrystallization pressure slightly and the water content ofthe melts does not systematically improve the match ofthe MELTS models to the observed dataThe MELTS results indicate that a significant propor-
tion of the crystallization takes place before plagioclase(15^20 crystallization) and clinopyroxene (35^45
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
31
crystallization) join the crystallization sequence Theseresults suggest that445 crystallization of the primarymagmas occurred (estimated from the residual liquid percent in Fig 15f) and if the original flood basalt stratigra-phy on Vancouver Island was 6 km thick at least anequivalent amount of complementary plutonic rocksshould have crystallized The Karmutsen basalts and theirplutonic residues thus represent a significant addition ofcrust (perhaps 412 km thickness) previously thickenedduring the formation of the underlying Paleozoic arc
Growth of the Wrangellia oceanic plateauon Vancouver IslandThe extrusive part of the Wrangellia oceanic plateau onVancouver Island was constructed in a three-layersequence of mostly tholeiitic basalts with a restrictedrange of composition (Fig 16) The plateau overliesDevonian to Mississippian arc volcanic sequences andmarine sediments of Mississippian to Permian ages thatare intruded by gabbroic rocks related to the flood basaltsThe volcanic stratigraphy of theWrangellia plateau formed
00
05
10
15
20
25
30
35
40
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
16
0 2 4 6 8 10 12 14 16 18 20
10
11
12
13
14
15
16
17
18
19
20
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
0 2 4 6 8 10 12 14 16 18 20
MgO (wt)
0 10 20 30 40 50
percent of total mass
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
) Mineral proportions
Sp (e)
Ol
CpxPlag
PicriteHigh-MgO basaltTholeiitic basalt
TiO2 (wt)
MgO (wt)
Al2O3 (wt) CaO (wt)
FeO (wt)
MgO(a) (b)
(c) (d)
MgO (wt)MgO (wt)
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
)
Residual liquid ()
1220
1190
Ol + Sp
Plag+Ol+Sp Plag+Cpx+Sp
02 wt H2O
no H2O
Residual liquid and crystallizing phases
(f)
0102030405060708090100
plagioclaseaccumulation
plagioclaseaccumulation
plagioclase
accumulatio
n
10
12
14
16
0 10 20 30 40
Al2O3CaO
SrNd
tholeiitic basalts
(g)
Fig 15 Forward fractional crystallization modeling results for major elements from MELTS (Ghiorso amp Sack 1995) compared with picriticand tholeiitic lavas from the Karmutsen Formation (a^d) Major-element compositions of residual liquids compared with Karmutsen lavacompositions For clarity the results are shown only for one starting composition (open circle) using the composition of the estimated primarymagma for the high-MgO lava series from the modeling technique of Herzberg et al (2007) for sample 4723A13 as shown in Fig 12 The esti-mated primary magmas for samples 4723A3 and 93G171 yield similar results A pressure of 1 kbar was used with no H2O and 02wt H2Ocalculated at QFM black line is anhydrous and grey line is 02wt H2OThe end-point for the anhydrous result is at 78 fractionation and11308C and the end-point for the result with 02wt H2O is 75 fractionation and11108C (e) Mineral proportions of crystallized phases fromone MELTrun (sample 4723A13) with no H2O (f) Per cent residual liquid vs temperature labelled with intervals of crystallizing phases (g)Al2O3CaO vs SrNd for the tholeiitic basalts
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
32
in a single eruptive phase (c 231^225 Ma) as an emergentbasalt sequence during a deeper-water shallow-water andsubaerial stage and subsided below sea level relativelysoon after its formation (Fig 16) The deeper-water stagecontains pillowed and massive submarine flows (3000mthick) that built up the submarine volcanic edifice from theseafloor The lowest pillow basalts were emplaced onunconsolidated and lithified fine-grained siliciclastic andcarbonaceous sediments containing Daonella beds(c 235^232 Ma) that are intruded by abundantKarmutsen mafic sills (Fig 16a) The different pillowedand massive submarine flows resulted primarily from dif-ferent effusive rates and local topography Some of themassive submarine flows commonly recognized by radiat-ing columnar jointing represent master tubes for the deliv-ery of lava to distal parts of submarine flow fields Thepillowed flows which commonly have uneven topography
in modern sequences may have served as levees to themassive submarine flows rather than allowing unob-structed spreading and inflation typical of the massive sub-aerial flows that form on shallow slopes in continentalflood basalt provinces (eg Self et al 1997) The rapid suc-cession of flows is evident by the near absence of sedimentbetween submarine flows Minor volumes of picritic pillowlavas occur in the upper part of the pillowed flow sequencein areas of the plateau exposed on northern VancouverIsland (Fig 16 Nixon et al 2008)The growth of the plateau in shallow water led to
emplacement of an increasing proportion of volcaniclastics(pillow breccias and hyaloclastites) that accumulated up to1500m thick in some areas (Fig 16b Nixon et al 2008)Horizons of pillowed flows occur locally within the volca-niclastic stratigraphy Sedimentary structures (eg gradedbedding fluidization structures) and petrographic textures
Late Devonian-Mississippian arc
Daonella beds
Middle Triassic
deeper water
Late Devonian-Mississippian arc
pillowed flows massive submarine
flows
pillow brecciahyaloclastiteshallow water
~4 km
~6 km
Late Triassic limestoneinterflow limestonelenses
Karmutsen Formation
mafic intrusions sediment-sill complex
Karmutsen Formation
(a) (b)
(c) (d)
subaerial flows
Late Devonian-Mississippian arc Late Devonian-Mississippian arc
picritic pillow basalts
Miss to Permian marine sediments
sills
Fig 16 Growth of the Wrangellia oceanic plateau onVancouver Island (a) Deep-water eruptive phase of pillowed and massive submarineflows with almost no intervening sediments onto MiddleTriassic shale containing Daonella beds (c 235^232 Ma) The basement of the plateauis composed of Devonian to Mississippian arc sequences overlain by Mississippian to Permian carbonates and siliciclastic sediments(b) Shallow-water flow emplacement with increasing proportion of volcaniclastics resulted primarily from cooling^contraction granulationPicritic pillow basalts occur primarily near the transition between pillowed submarine flows and volcaniclastic units Pillowed and massiveflows occur locally within the volcaniclastic unit (c) Emergent stage of growth consists of overlapping subaerial flow fields similar to continen-tal flood basalts with pahoehoe flow structures Multi-tiered columnar jointing occurrs locally in response to ponding of flows (eg Bondre et al2004) Interflow carbonate deposits and plagioclase megacrystic flows developed locally during the waning stage of volcanism (d) Post-volcanicsubsidence and deposition of Late Carnian to Norian (c 212^223 Ma) platform carbonates on top of the plateau Units are labeled in the dia-grams Thicknesses of the underlying units are not to scale
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
33
within the volcaniclastic rocks indicate formation primar-ily via cooling^contraction granulation magma^water^steam interaction autobrecciation resedimentation andmass-wasting rather than pyroclastic eruption (Nixonet al 2008)A broad subaerial platform was constructed by emplace-
ment of subaerial lava flows during an emergent subaerialstage (Fig 16c) The subaerial unit predominantly com-prises well-layered sheet flows that formed in a similarway to continental flood basalts as hundred(s) of pahoehoeflow fields erupted on shallow slopes (Self et al 1997) how-ever volcaniclastic and pillowed flow units are locally pre-sent mostly in the upper parts of the stratigraphyLocalized ponding of flows led to the formation of multi-tiered columnar jointing in some subaerial flows (egBondre et al 2004) As volcanism waned local interflowcarbonate deposits developed and plagioclase megacrysticflows were erupted Subsidence of the plateau is recordedby the interflow lenses and deposits of platform carbonates(40^1500m thick) on top of the Karmutsen basalts thatcontain Late Carnian to Early Norian fossils (Fig 16d c216^227 Ma Nixon et al 2008)The evolution of the internal architecture of the
Wrangellia oceanic plateau was primarily related to theconditions of emplacement (ie water depth and eruptionenvironment) There does not appear to have been a sys-tematic coincident evolution or shift of lava compositionsonVancouver Island as has been recognized in most con-tinental flood basalts and parts of the Wrangellia oceanicplateau in Alaska and Yukon (Greene et al 2008 2009)In general with the exception of the picritic pillow lavasin the submarine stratigraphy and local variations betweenhigh-MgO and tholeiitic lavas there are no significantcompositional differences between lavas erupted duringthe different stages of growth of the plateau (ie submarineand subaerial flows) nor are substantial compositional dif-ferences generally found between massive submarine flowsin contact with pillowed flows tholeiitic basalt flows nearthe base of the Karmutsen Formation are compositionallysimilar to flows sampled near the top of the volcanicsequence Where it was possible to sample continuous sec-tions of volcanic stratigraphy greater than several hun-dreds of meters (eg Keogh^Maynard Lake area west ofthe Karmutsen Range Fig 2) there is no clear relationshipbetween stratigraphic position and compositionalvariation
CONCLUSIONSThe Karmutsen Formation covers extensive areas ofVancouver Island (20 000 km2) and forms part of a largeoceanic plateau erupted over a geologically short interval(c 225^231 Ma) The tripartite volcanic stratigraphyexposed on Vancouver Island is at least 6 km thick andcomprises a succession of submarine flows volcaniclastic
deposits and massive subaerial flowsThese volcanologicaldifferences are primarily related to the eruption environ-ment (deep-water shallow-water or subaerial) The rapidgrowth of the plateau prevented significant accumulationsof intervening sediment except in the uppermost stratigra-phy where isolated limestone lenses commonly associatedwith pillow basalt and hyaloclastite deposits preserve arecord of the final stages of subsidence of the plateauThe Wrangellia plateau on Vancouver Island was con-
structed dominantly of tholeiitic basalt with restrictedmajor- and trace-element and isotopic compositional var-iations Minor volumes of picritic pillow basalts eruptedlate in the submarine phase The Karmutsen basalts wereemplaced onto older Paleozoic arc volcanic sequences andlow NbLa for the picrites may indicate a contributionfrom the sub-arc lithospheric mantle Modeling of the pet-rogenesis of the Karmutsen picrites indicates melting ofanomalously hot mantle (15008C) and extensive degreesof partial melting (23^27) and is consistent with aplume initiation model Combined Sr^Nd^Hf^Pb isotopicsystematics indicate that the picritic and tholeiitic lavaswere derived from a depleted plume-type mantle sourcedistinct from the source of MORB that has compositionalsimilarities to the source of the lavas forming theCaribbean Plateau Picrites were generated by melting ofspinel lherzolite whereas the tholeiitic basalts formed asaggregate melts of both garnet and spinel lherzolite Thetholeiitic basalts underwent extensive low-pressure fractio-nation (52^3 kbar) and seismic studies indicate that someof theWrangellia crust beneathVancouver Island may cor-respond to the plutonic residues of the Karmutsen floodbasalts
ACKNOWLEDGEMENTSWe are grateful to Nick Arndt for helping us get this pro-ject started and Nick Massey for insights into VancouverIsland geology We greatly appreciate the assistance withPRIMELT modeling from Claude Herzberg We thankMikkel Schau for his insight and enthusiasm during field-work Jane Barling assisted with analyses by MC-ICP-MSThoughtful and thorough reviews of the manuscript byJohn Mahoney Fred Frey Wendy Bohrson DougalJerram and Teal Riley are greatly appreciated Fundingwas generously provided by the Rocks to Riches Programadministered by the BC amp Yukon Chamber of Mines in2004 by the BC Geological Survey and by NSERCDiscovery Grants to JSS and DW ARG was supportedby a University Graduate Fellowship at UBC
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
34
REFERENCESBarker F Brown A S Budahn J R amp Plafker G (1989) Back-arc
with frontal-arc component origin of Triassic Karmutsen basaltBritish Columbia Canada Chemical Geology 75 81^102
Blichert-Toft JWeis D Maerschalk C Agranier A amp Albare de F(2003) Hawaiian hot spot dynamics as inferred from the Hf and Pbisotope evolution of Mauna Kea volcano Geochemistry GeophysicsGeosystems 4(2) 1^27 doi1010292002GC000340
Bondre N R Duraiswami R A amp Dole G (2004) Morphologyand emplacement of flows from the Deccan Volcanic ProvinceIndia Bulletin of Volcanology 66 29^45
Carlisle D (1963) Pillow breccias and their aquagene tuffs QuadraIsland British Columbia Journal of Geology 71 48^71
Carlisle D (1972) Late Paleozoic to Mid-Triassic sedimentary-volcanic sequence on Northeastern Vancouver Island Geological
Society of Canada Paper Report of Activities 72-1 Part B 24^30Carlisle D amp SuzukiT (1974) Emergent basalt and submergent car-
bonate^clastic sequences including the Upper Triassic Dilleri andWelleri zones on Vancouver Island Canadian Journal of Earth
Sciences 11 254^279Chauvel C amp Blichert-Toft J (2001) A hafnium isotope and trace
element perspective on melting of the depleted mantle Earth and
Planetary Science Letters 190 137^151Chayes F (1954)The theory of thin-section analysis Journal of Geology
62 92^101Cho M Liou J G amp Maruyama S (1987) Transition from the zeo-
lite to prehnite^pumpellyite facies in the Karmutsen metabasitesVancouver Island British Columbia Journal of Petrology 27(2)467^494
Clowes R M Zelt C A Amor J R amp Ellis R M (1995)Lithospheric structure in the southern Canadian Cordillera froma network of seismic refraction lines Canadian Journal of Earth
Sciences 32(10) 1485^1513Coffin M F amp Eldholm O (1994) Large igneous provinces Crustal
structure dimensions and external consequences Reviews of
Geophysics 32(1) 1^36Cox K G (1980) A model for flood basalt volcanism Journal of
Petrology 21 629^650Eldholm O amp Coffin M F (2000) Large igneous provinces
and plate tectonics In Richards M A Gordon R G amp vander Hilst R D (eds) The History and Dynamics of Global
Plate Motions American Geophysical Union Geophysical Monograph 121309^326
Elliott T R Hawkesworth C J amp Grolaquo nvold K (1991) Dynamicmelting of the Iceland plume Nature 351 201^206
Escuder-Viruete J Pecurren rez-Estaucurren n A Contreras F Joubert MWeis D Ullrich T D amp Spadea P (2007) Plume mantle sourceheterogeneity through time Insights from the Duarte ComplexHispaniola northeastern Caribbean Journal of Geophysical Research112(B04203) doi1010292006JB004323
Farnetani C G Richards M A amp Ghiorso M S (1996)Petrological models of magma evolution and deep crustal structurebeneath hotspots and flood basalt provinces Earth and Planetary
Science Letters 143 81^94Fitton J G amp Godard M (2004) Origin and evolution of magmas
on the Ontong Java Plateau In Fitton J G Mahoney J JWallace P J amp Saunders A D (eds) Origin and Evolution of the
Ontong Java Plateau Geological Society London Special Publications 229151^178
Fitton J G Saunders A D Norry M J Hardarson B S ampTaylor R N (1997) Thermal and chemical structure of theIceland plume Earth and Planetary Science Letters 153 197^208
Fitton J G Saunders A D Kempton P D amp Hardarson B S(2003) Does depleted mantle form an intrinsic part of the Icelandplume Geochemistry Geophysics Geosystems 4(3) 1032 doi1010292002GC000424
Frey F A Huang S Blichert-Toft J Regelous M amp Boyet M(2005) Origin of depleted components in basalt related to theHawaiian hotspot Evidence from isotopic and incompatible ele-ment ratios Geochemistry Geophysics Geosystems 6(1) 23 doi1010292004GC000757
Galer S J G amp Abouchami W (1998) Practical application of leadtriple spiking for correction of instrumental mass discriminationMineralogical Magazine 62A 491^492
Ghiorso M S amp Sack R O (1995) Chemical mass transfer in mag-matic processes IV A revised and internally consistent thermody-namic model for the interpolation and extrapolation of liquid^solidequilibria in magmatic systems at elevated temperatures and pres-sures Contributions to Mineralogy and Petrology 119 197^212
Greene A R Scoates J S amp Weis D (2005)WrangelliaTerrane onVancouver Island Distribution of flood basalts with implicationsfor potential Ni^Cu^PGE mineralization In Grant B (ed)Geological Fieldwork 2004 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2005-1 209^220Greene A R Scoates J S Nixon G T amp Weis D (2006) Picritic
lavas and basal sills in the Karmutsen flood basalt provinceWrangellia northern Vancouver Island In Grant B (ed)Geological Fieldwork 2005 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2006-1 39^52Greene A R Scoates J S amp Weis D (2008) Wrangellia flood
basalts in Alaska A record of plume^lithosphere interaction in aLate Triassic accreted oceanic plateau Geochemistry Geophysics
Geosystems 9(Q12004) doi1010292008GC002092Greene A R Scoates J S Weis D amp Israel S (2009)
Geochemistry of triassic flood basalts from the Yukon (Canada)segment of the accreted Wrangellia oceanic plateau Lithosdoi101016jlithos200811010
Hauff F Hoernle K Tilton G Graham D W amp Kerr A C(2000) Large volume recycling of oceanic lithosphere over shorttime scales geochemical constraints from the Caribbean LargeIgneous Province Earth and Planetary Science Letters 174 247^263
Hauff F Hoernle K amp Schmidt A (2003) Sr^Nd^Pb compositionof Mesozoic Pacific oceanic crust (Site 1149 and 801 ODP Leg185)Implications for alteration of ocean crust and the input into theIzu^Bonin^Mariana subduction system Geochemistry Geophysics
Geosystems 4(8) doi1010292002GC000421Herzberg C (2004) Partial melting below the Ontong Java Plateau
In Fitton J G Mahoney J J Wallace P J amp Saunders A D(eds) Origin and Evolution of the Ontong Java Plateau Geological SocietyLondon Special Publications 229 179^183
Herzberg C (2006) Petrology and thermal structure of the Hawaiianplume from Mauna Kea volcano Nature 444(7119) 605
Herzberg C amp Asimow P D (2008) Petrology of some oceanicisland basalts PRIMELT2XLS software for primary magma cal-culation Geochemistry Geophysics Geosystems 9(Q09001) doi1010292008GC002057
Herzberg C amp OrsquoHara M J (2002) Plume-associated ultramaficmagmas of Phanerozoic age Journal of Petrology 43(10) 1857^1883
Herzberg C Asimow P D Arndt N Niu Y Lesher C MFitton J G Cheadle M J amp Saunders A D (2007)Temperatures in ambient mantle and plumes Constraints frombasalts picrites and komatiites Geochemistry Geophysics Geosystems8(Q02006) doi1010292006GC001390
Huang S amp Frey F A (2005) Trace element abundances of MaunaKea basalt from phase 2 of the Hawaii Scientific Drilling Project
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
35
Petrogenetic implications of correlations with major element con-tent and isotopic ratios Geochemistry Geophysics Geosystems 4(6)doi1010292002GC000322
Hussong D M Wipperman L K amp Kroenke L W (1979) Thecrustal structure of the Ontong Java and Manihiki OceanicPlateaus Journal of Geophysical Research 84(B11) 6003^6010
Jerram D A amp Widdowson M (2005) The anatomy of ContinentalFlood Basalt Provinces geological constraints on the processes andproducts of flood volcanism Lithos 79(3^4) 385^405
Jones D L Silberling N J amp Hillhouse J (1977)Wrangellia a dis-placed terrane in northwestern North America CanadianJournal ofEarth Sciences 14(11) 2565^2577
Kerr A C (2003) Oceanic Plateaus In Rudnick R (ed) The Crust
Treatise on GeochemistryVol 3 Oxford Elsevier Science pp 537^565Kerr A C amp Mahoney J J (2007) Oceanic plateaus Problematic
plumes potential paradigms Chemical Geology 241 332^353Kerr A CTarney J Marriner G F Klaver GT Saunders A D
amp Thirlwall M F (1996) The geochemistry and petrogenesis ofthe Late-Cretaceous picrites and basalts of Curacao NetherlandsAntilles a remnant of an oceanic plateau Contributions to
Mineralogy and Petrology 124(1) 29^43Kerr A C Marriner G F Tarney J Nivia A Saunders A D
Thirlwall M F amp Sinton A C (1997) Cretaceous basaltic ter-ranes in Western Colombia Elemental chronological and Sr-Ndisotopic constraints on petrogenesis Journal of Petrology 38(6)677^702
Kinzler R J (1997) Melting of mantle peridotite at pressuresapproaching the spinel to garnet transition Application to mid-ocean ridge basalt petrogenesis Journal of Geophysical Research
102(B1) 853^874Kinzler R J amp Grove T L (1992) Primary magmas of mid-ocean
ridge basalts 1 Experiments and methods Journal of Geophysical
Research 97(B5) 6885^6906Korenaga J (2005) Why did not the Ontong Java Plateau form sub-
aerially Earth and Planetary Science Letters 234 385^399Kuniyoshi S (1972) Petrology of the Karmutsen (Volcanic) Group
northeastern Vancouver Island British Columbia PhD disserta-tion University of California Los Angeles 242 pp
Lassiter J C DePaolo D J amp Mahoney J J (1995) Geochemistry ofthe Wrangellia flood basalt province Implications for the role ofcontinental and oceanic lithosphere in flood basalt genesis Journalof Petrology 36(4) 983^1009
LeBas M J (2000) IUGS reclassification of the high-Mg and picriticvolcanic rocks Journal of Petrology 41(10) 1467^1470
Longhi J (2002) Some phase equilibrium systematics of lherzolitemelting I Geochemistry Geophysics Geosystems 3(3) doi1010292001GC000204
MacDonald G A amp Katsura T (1964) Chemical composition ofHawaiian lavas Journal of Petrology 5 82^133
Mahoney J J (1987) An isotopic survey of Pacific oceanic plateausimplications for their nature and origin In Keating B HFryer P Batiza R amp Boehlert GW (eds) Seamounts Islands andAtolls American Geophysical Union Geophysical Monograph 43 207^220
Mahoney J LeRoex A P Peng Z Fisher R L amp Natland J H(1992) Southwestern limits of Indian Ocean Ridge mantle and theorigin of low 206Pb204Pb mid-ocean ridge basalt isotope systema-tics of the central Southwest Indian Ridge (178^508E) Journal ofGeophysical Research 97(B13) 19771^19790
Mahoney J J Sinton J M Kurz M D MacDougall J DSpencer K J amp Lugmair GW (1994) Isotope and trace elementcharacteristics of a super-fast spreading ridge East Pacific rise13^238S Earth and Planetary Science Letters 121 173^193
Massey N W D (1995a) Geology and mineral resources of theAlberni^Nanaimo Lakes sheet Vancouver Island 92F1W 92F2Eand part of 92F7E British Columbia Ministry of Energy Mines and
Petroleum Resources Paper 1992-2 132 ppMassey N W D (1995b) Geology and mineral resources of the
Cowichan Lake sheet Vancouver Island 92C16 British Columbia
Ministry of Energy Mines and Petroleum Resources Paper 1992-3 112 ppMassey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005a) Digital Geology Map of British Columbia Tile NM9 MidCoast BC British Columbia Ministry of Energy and Mines Geofile
2005-2Massey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005b) Digital Geology Map of British Columbia Tile NM10Southwest BC British Columbia Ministry of Energy and Mines Geofile
2005-3McDonough W F amp Sun S (1995) The composition of the Earth
Chemical Geology 120 223^253Monger JW H amp Journeay J M (1994) Basement geology and tec-
tonic evolution of theVancouver region In Monger JW H (ed)Geology and Geological Hazards of the Vancouver Region Southwestern
British Columbia Geological Survey of Canada Bulletin 481 3^25Muller J E (1977) Geology of Vancouver Island Geological Survey of
Canada Open File Map 463Muller J E (1980)The Paleozoic Sicker Group of Vancouver Island British
Columbia Geological Survey of Canada Paper 79-30 22 ppMuller J E Northcote K E amp Carlisle D (1974) Geology and
mineral deposits of Alert Bay^Cape Scott map area VancouverIsland British Columbia Geological Survey of Canada Paper 74-8 77
Neal C R Mahoney J J Kroenke L W Duncan R A ampPetterson M G (1997) The Ontong^Java Plateau InMahoney J J amp Coffin M F (eds) Large Igneous Provinces
Continental Oceanic and Planetary FloodVolcanism American Geophysical
Union Geophysical Monograph 100 183^216Niu Y Collerson K D Batiza R Wendt J I amp Regelous M
(1999) Origin of enriched-typed mid-ocean ridge basalt at ridgesfar from mantle plumes The East Pacific Rise at 11820rsquoN Journalof Geophysical Research 104 7067^7088
Nixon G T amp Orr A J (2007) Recent revisions to the EarlyMesozoic stratigraphy of Northern Vancouver Island (NTS 102I092L) and metallogenic implicatinos British Columbia InGrant B (ed) Geological Fieldwork 2006 British Columbia Ministry of
Energy Mines and Petroleum Resources Paper 2007-1 163^177Nixon GT Hammack J L KoyanagiV M Payie G J Haggart
JW Orchard M JTozerT Friedman R M Archibald D APalfy J amp Cordey F (2006a) Geology of the Quatsino^PortMcNeill area northernVancouver Island British Columbia Ministry
of Energy Mines and Petroleum Resources Geoscience Map 2006-2 scale150 000
Nixon G T Kelman M C Stevenson D Stokes L A ampJohnston K A (2006b) Preliminary geology of the Nimpkishmap area (NTS 092L07) northern Vancouver Island BritishColumbia In Grant B amp Newell J M (eds) Geological Fieldwork2005 British Columbia Ministry of Energy Mines and Petroleum Resources
Paper 2006-1 135^152Nixon G T Laroque J Pals A Styan J Greene A R amp
Scoates J S (2008) High-Mg lavas in the Karmutsen flood basaltsnorthernVancouver Island (NTS 092L) Stratigraphic setting andmetallogenic significance In Grant B (ed) Geological Fieldwork
2007 British Columbia Ministry of Energy Mines and Petroleum
Resources Paper 2008-1 175^190Nowell G M Kempton P D Noble S R Fitton J G
Saunders A Mahoney J J amp Taylor R N (1998) High precisionHf isotope measurements of MORB and OIB by thermal ionisation
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
36
mass spectrometry insights into the depleted mantle Chemical
Geology 149 211^233Parrish R R amp McNicoll V J (1992) U^Pb age determinations
from the southern Vancouver Island area British Columbia InRadiogenic Age and Isotopic Studies Report 5 Geological Survey of
Canada Paper 91-2 79^86Peate DW (1997) The Parana^Etendeka Province In Coffin M F
amp Mahoney J (eds) Large Igneous Provinces Continental Oceanic andPlanetary Flood Volcanism American Geophysical Union Geophysical
Monograph 100 217^245Perfit M R Fornari D J Ridley W I Kirk P D Casey J
Kastens K A Reynolds J R Edwards M Desonie DShuster R amp Paradis S (1996) Recent volcanism in the Siqueirostransform fault picritic basalts and implications for MORBmagma genesis Earth and Planetary Science Letters 141(1^4) 91^108
Pretorius W Weis D Williams G Hanano D Kieffer B ampScoates J S (2006) Complete trace elemental characterization ofgranitoid (USGSG-2GSP-2) reference materials by high resolutioninductively coupled plasma-mass spectrometry Geostandards and
Geoanalytical Research 30(1) 39^54Regelous M Niu Y Wendt J I Batiza R Greig A amp
Collerson K D (1999) Variations in the geochemistry of magma-tism on the East Pacific Rise at 10830rsquoN since 800 ka Earth and
Planetary Science Letters 168 45^63Recurren villon S Chauvel C Arndt N T Pik R Martineau F
Fourcade S amp Marty B (2002) Heterogeneity of the Caribbeanplateau mantle source Sr O and He isotopic compositions of oli-vine and clinopyroxene from Gorgona Island Earth and Planetary
Science Letters 205(1^2) 91Rhodes J M (1996) Geochemical stratigraphy of lava flows samples
by the Hawaii Scientific Drilling Project Journal of Geophysical
Research 101(B5) 11729^11746Rhodes J M amp Vollinger M J (2004) Composition of basaltic lavas
sampled by phase-2 of the Hawaii Scientific Drilling ProjectGeochemical stratigraphy and magma types Geochemistry
Geophysics Geosystems 5(3) doi1010292002GC000434Richards M A Jones D L Duncan R A amp DePaolo D J (1991)
A mantle plume initiation model for the Wrangellia flood basaltand other oceanic plateaus Science 254 263^267
Salters V J M amp Hart S R (1991) The mantle sources of oceanridges islands and arcs the Hf-isotope connection Earth and
Planetary Science Letters 104 364^380Salters V J M amp Stracke A (2004) Composition of the depleted
SaltersV J amp WhiteW (1998) Hf isotope constraints on mantle evo-lution Chemical Geology 145 447^460
SanoT amp Yamashita S (2004) Experimental petrology of basementlavas from Ocean Drilling Program Leg 192 implications for dif-ferentiation processes in Ontong Java Plateau magmas InFitton J G Mahoney J JWallace P J amp Saunders A D (eds)Origin and Evolution of the Ontong Java Plateau Geological Society
London Special Publications 229 185^218Saunders A D (2005) Large igneous provinces Origin and environ-
mental consequences Elements 1 259^263Self S ThordarsonT amp Keszthely L (1997) Emplacement of conti-
nental flood basalt lava flows In Mahoney J J amp Coffin M F(eds) Large Igneous Provinces Continental Oceanic and Planetary Flood
Volcanism American Geophysical Union Geophysical Monograph 100381^410
Shaw D M (2000) Continuous (dynamic) melting theory revisitedCanadian Mineralogist 38(5) 1041^1063
Sluggett C L (2003) Uranium^lead age and geochemical constraintson Paleozoic and Early Mesozoic magmatism in WrangelliaTerrane Saltspring Island British Columbia BSc thesisUniversity of British ColumbiaVancouver 84 pp
Storey M Mahoney J J amp Saunders A D (1997) Cretaceousbasalts in Madagascar and the transition between plume and con-tinental lithospheric mantle sources In Mahoney J J ampCoffin M F (eds) Large Igneous Provinces Continental Oceanic and
Planetary Flood Volcanism Aamerican Geophysical Union Geophysical
Monograph 100 95^122Surdam R C (1967) Low-grade metamorphism of the Karmutsen
Group PhD dissertation University of California Los Angeles288 pp
Sutherland-Brown A Yorath C J Anderson R G amp Dom K(1986) Geological maps of southern Vancouver IslandLITHOPROBE I 92C10 11 14 16 92F1 2 7 8 Geological Survey ofCanada Open File 1272
Tejada M L G Mahoney J J Castillo P R Ingle S P Sheth HC amp Weis D (2004) Pin-pricking the elephant evidence on theorigin of the Ontong Java Plateau from Pb^Sr^Hf^Nd isotopiccharacteristics of ODP Leg 192 basalts In Fitton J GMahoney J J Wallace P J amp Saunders A D (eds) Origin and
Evolution of the Ontong Java Plateau Geological Society London Special
Publications 229 133^150Thompson P M E Kempton P D amp Kerr A C (2008) Evaluation
of the effects of alteration and leaching on Sm^Nd and Lu^Hf sys-tematics in submarine mafic rocks Lithos 104(1^4) 164^176
Thompson P M E Kempton P D White R V Kerr A CTarney J Saunders A D Fitton J G amp McBirney A (2003)Hf-Nd isotope constraints on the origin of the CretaceousCaribbean plateau and its relationship to the Galapagos plumeEarth and Planetary Science Letters 217(1^2) 59^75
Vervoort J D Patchett P Blichert-Toft J amp Albare de F (1999)Relationships between Lu^Hf and Sm^Nd isotopic systems in theglobal sedimentary system Earth and Planetary Science Letters 16879^99
Walter M J (1998) Melting of garnet peridotite and the origin ofkomatiite and depleted lithosphere Journal of Petrology 39(1) 29^60
Weis D Kieffer B Maerschalk C Barling J de Jong JWilliams G A Hanano D Mattielli N Scoates J SGoolaerts A Friedman R A amp Mahoney J B (2006) High-pre-cision isotopic characterization of USGS reference materials byTIMS and MC-ICP-MS Geochemistry Geophysics Geosystems
7(Q08006) doi1010292006GC001283Weis D Kieffer B Hanano D Silva I N Barling J
Pretorius W Maerschalk C amp Mattielli N (2007) Hf isotopecompositions of US Geological Survey reference materialsGeochemistry Geophysics Geosystems 8(Q06006) doi1010292006GC001473
Wheeler J O amp McFeely P (1991) Tectonic assemblage map of theCanadian Cordillera and adjacent part of the United States ofAmerica Geological Survey of Canada Map 1712A
WhiteW M Albare de F amp Tecurren louk P (2000) High-precision analy-sis of Pb isotope ratios by multi-collector ICP-MS Chemical Geology167 257^270
Yorath C J Sutherland Brown A amp Massey N W D (1999)LITHOPROBE southern Vancouver Island British Columbia Geological
Survey of Canada Bulletin 498 145 ppZou H (1998) Trace element fractionation during modal and
non-model dynamic melting and open-system melting Amathematical treatment Geochimica et Cosmochimica Acta 62(11)1937^1945
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
37
Zou H amp Reid M R (2001) Quantitative modeling of trace elementfractionation during incongruent dynamic melting Geochimica et
Cosmochimica Acta 65(1) 153^162
APPENDIX SAMPLEPREPARAT ION ANDANALYT ICAL METHODSOnly the freshest rocks in the field were sampled and onlythe least altered samples were selected for chemical andisotopic analysis based on thorough petrographic inspec-tion Fifty-nine of the 129 collected samples were crushed(400 g) into small pieces 52mm in diameter in aRocklabs hydraulic piston crusher between WC plates tominimize contamination The coarse-crush was mixedand 100 g was powdered in a planetary mill using agatejars and balls cleaned with quartz sand between samples
ActLabs analytical methodsThe major- and trace-element compositions of the whole-rock powders were determined at Activation LaboratoriesLtd (Actlabs) in Ancaster Ontario Analytical techniquesand detection limits are also available from Actlabs (httpwwwactlabscommethsub_code4erehtm) The analyticalmethod for each of the elements analyzed is indicated inTable 2 For the major elements a 02 g sample was mixedwith a lithium metaborate^lithium tetraborate mixtureand fused in a graphite crucible The molten mixture waspoured into a 5 HNO3 solution and shaken until dis-solved (30min) The samples were analyzed for majoroxides and selected trace elements by inductively coupledplasma optical emission spectrometry (ICP-OES) on acombination simultaneous^sequential Thermo Jarrell-AshEnviro II system Internal calibration was achieved usinga variety of international reference materials (eg W-2BIR-1 DNC-1) and independent control samplesAdditional trace elements were analyzed by both instru-mental neutron activation analysis (INAA) and induc-tively couple plasma mass spectrometry (ICP-MS) ForINAA15^25 g of sample was weighed into small polyeth-ylene vials and irradiated with a control international ref-erence material CANMET WMS-1 and NiCr flux wires ata thermal neutron flux of 71012 neutronscm2 per s in theMcMaster Nuclear Reactor Following a 7 day waitingperiod the samples were measured on an Ortec high-purity Ge detector linked to a Canberra Series 95 multi-channel analyzer Activities for each element weredecay- and weight-corrected and compared with a detectorcalibration developed from multiple international certifiedreference materials For ICP-MS 025 g of sample wasdigested in HF followed by HNO3^HClO4 heated andtaken to dryness The samples were brought back into solu-tion with HCl Samples were analyzed using a PerkinElmer Optima 3000 ICP In-lab standards or certified ref-erence materials (eg W-2 BIR-1 DNC-1) were used for
quality control A total of 15 blind duplicates were ana-lyzed from the same sample powders to assessreproducibility
University of Massachusetts analyticalmethodsFourteen sample duplicate powders taken from the same100 g sample aliquot (high-MgO basalts and picrites)were also analyzed at the Ronald B Gilmore X-RayFluorescence (XRF) Laboratory at the University ofMassachusetts Major elements were measured on a fusedLa-bearing lithium borate glass disc using a SiemensMRS-400 spectrometer Trace element concentrations(Rb Sr Ba Ce Nb Zr Y Pb Zn Ga Ni Cr V) weremeasured on a separate powder pellet using a PhilipsPW2400 sequential spectrometer using a Rh tube LOIand ferrous iron measurements were made as describedby Rhodes amp Vollinger (2004) Precision and accuracy esti-mates for the data have been described by Rhodes (1996)and Rhodes amp Vollinger (2004) Results for each sampleare the average of two separate analyses and are shown inElectronic Appendix 2 for a comparison of analyses ofduplicate powders using XRFand ICP-MS
PCIGR analytical methodsA subset of 19 samples was selected for high-precisiontrace-element analysis and Sr Nd Pb and Hf isotopicanalysis at the Pacific Centre for Isotopic andGeochemical Research (PCIGR) at the University ofBritish Columbia (UBC) Samples were selected from the59 samples analyzed for whole-rock chemistry at ActLabsbased on major- and trace-element chemistry alteration(low LOI and petrographic alteration index) and samplelocation Samples were prepared for trace-element analysisat the PCIGR by the technique described by Pretorius et al(2006) on unleached rock powders Sample powders(100mg) were weighed in 7ml screw-top Savillex bea-kers and dissolved in 1ml 14N HNO3 and 5ml 48 HFon a hotplate for 48 h at 1308C with periodic ultrasonica-tion Samples were dried and redissolved in 6ml 6N HClon a hotplate for 24 h and then dried and redissolved in1ml concentrated HNO3 for 24 h before final dryingTrace element abundances were measured by high-resolution inductively coupled plasma-mass spectrometry(HR-ICP-MS) with aThermo Finnigan Element2 systemfollowing the procedures described by Pretorius et al(2006) within 24 h of redissolution HFSE and LILE weremeasured in medium-resolution mode at 2000 dilutionusing a PFATeflon spray chamber washed with aqua regiafor 3min between samples REE were measured in high-resolution mode and U and Pb in low-resolution modeat 2000 dilution using a glass spray chamber washedwith 2 HNO3 between samples Total procedural blanksand reference materials (BCR-2 BHVO-2) were analyzedwith the batch of samples Indium was used as an internal
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
38
standard in all samples and standard solutionsBackground and standard solutions were analyzed afterevery five samples to detect memory effects and massdrift Results for PCIGR trace-element analyses areshown in Electronic Appendix 3 for the full set of analyzedtrace elementsSample digestion for purification of Sr Nd Hf and Pb
for column chemistry began by weighing each samplepowder (400^500mg) prior to leaching All samples wereleached with 6N HCl and placed in an ultrasonic bath for15min Samples were rinsed twice with 18 M-cm H2Obetween each leaching step (15 total) until the supernatantliquid was clear [following the method of Mahoney (1987)]Samples were then dried on a hotplate for 24 h andweighed again Sample solutions were then prepared bydissolving 100^250mg of the leached powder dissolvedin 1ml 14N HNO3 and 10ml 48 HF on a hotplate for48 h at 1308C with periodic ultrasonication Samples weredried and redissolved in 6ml 6N HCl on a hotplate for24 h and then dried Pb was separated using anion exhangecolumns and the discard was used for Sr REE and Hfseparation Nd was separated from the other REE and Hfrequired two additional purification steps Detailed proce-dures for column chemistry for separating Sr Nd and Pbat the PCIGR have been described byWeis et al (2006) andHf purification has been described byWeis et al (2007) Srand Nd isotope ratios were measured by thermal ioniza-tion mass spectrometry (TIMS) on a Thermo FinniganTriton system in static mode with relay matrix rotation ona single Ta and double Re^Ta filament respectively Fourto five filaments per barrel of 21 were occupied by stan-dards (NBS 987 for Sr and LaJolla for Nd) for each barrelwhere samples were run Sample Sr and Nd isotopic com-positions were corrected for mass fractionation using86Sr88Srfrac14 01194 and 146Nd144Ndfrac14 07219 Each samplewas then normalized using the barrel average of the refer-ence material relative to the values of143Nd144Ndfrac14 0511858 and 87Sr86Srfrac14 0710248 (Weiset al 2006) During the course of the Vancouver Islandanalyses the La Jolla Nd standard gave an average valueof 05118566 (nfrac14 7) and NBS987 standard gave an aver-age of 07102408 (nfrac1411) (2 error is reported as times106) 147Sm144Nd errors are 15 or 0006 USGeological Survey (USGS) reference material BHVO-2was processed with the samples and yielded Sr and Nd iso-topic ratios of 07034607 and 05129786 respectivelyThese are in agreement with the published values of070347920 and 051298411 respectively (Weis et al2006)
Pb and Hf isotopic compositions were analyzed by mul-tiple collector-inductively coupled plasma-mass spectrome-try (MC-ICP-MS) on a static multi-collection on a NuPlasma (Nu Instruments) system The detailed analyticalprocedure for Pb isotopic analyses on the Nu Plasma atthe PCIGR has been described by Weis et al (2006) Theconfiguration for Pb analyses allows for collection of PbTl and Hg togetherTl and Hg are used to monitor instru-mental mass discrimination and isobaric overlap repec-tively All sample solutions were analyzed withapproximately the same PbTl ratio (4) as the referencematerial NIST SRM 981 To accomplish this a small ali-quot of each sample solution from the Pb columns was ana-lyzed on the Element2 to determine the precise amount ofPb available for analysis on the Nu Plasma The SRM 981standard was run after every two samples on the NuPlasma During the time samples were run analyses ofthe SRM 981 Pb reference material gave values of206Pb204Pbfrac1416940322 207Pb204Pbfrac1415495823and 208Pb204Pbfrac14 36713164 (nfrac14 61 2 error is reportedas times 104) these values are in excellent agreement withreported TIMS triple-spike values of Galer amp Abouchami(1998) Fractionation-corrected Pb isotopic ratios were fur-ther corrected by the sample-standard bracketing methodor the ln^ln correction method described by White et al(2000) and Blichert-Toft et al (2003) Leached powders ofUSGS reference material BHVO-2 yielded Pb isotopicratios of 206Pb204Pbfrac14186454 8 207Pb204Pbfrac14 1549105 and 208Pb204Pbfrac14 38222514 and BCR-2yielded 206Pb204Pbfrac141880466 207Pb204Pbfrac14 1562518 and 208Pb204Pbfrac14 3883496 (2 error isreported as times 104) These values are in agreement withthose for leached residues of BHVO-2 and BCR-2 fromWeis et al (2006)Hf isotopic compositions were analyzed following the
procedures detailed byWeis et al (2007) The configurationfor Hf analyses monitored Lu mass 175 andYb mass 172 toallow for interference correction to masses 174 and 176 Hfisotopic ratios were normalized internally for mass fractio-nation to a 179Hf177Hf ratio of 07325 using an exponentialcorrection Standards were run after every two samplesand sample results were normalized to the ratio of the in-run daily average and a 176Hf177Hf ratio for JMC-475 of0282160 During the course of analyses the Hf standardJMC-475 gave an average value 02821679 (nfrac14130)USGS reference materials BCR-2 and BHVO-2 were pro-cessed with the samples and yielded Hf isotopic ratios of02828675 and 02831005 respectively Publishedvalues for BCR-2 and BHVO-2 are 02828717 and0283104 8 respectively (Weis et al 2007)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
39
The PRIMELT1 results indicate that if the Karmutsenpicritic lavas were derived from accumulated fractionalmelting of fertile peridotite the primary magmas wouldhave contained 15^17wt MgO and 10wt CaO(Fig 12 Table 6) formed from 23^27 partial meltingand would have crystallized olivine with a Fo content of 91 (Fig 12 Table 6) Calculated mantle temperatures forthe Karmutsen picrites (14908C) indicate melting ofanomalously hot mantle (100^2008C above ambient) com-pared with ambient mantle that produces mid-ocean ridgebasalt (MORB 1280^14008C Herzberg et al 2007)which initiated between 3 and 4 GPa Several picritesappear to be near-primary melts and required very littleaddition of olivine (eg sample 93G171 with measured158wt MgO) These temperature and melting esti-mates of the Karmutsen picrites are consistent withdecompression of hot mantle peridotite in an actively con-vecting plume head (ie plume initiation model) Threesamples (picrite 5614A1 and high-MgO basalts 5614A3and 5614A5) are lower in iron than the other Karmutsenlavas with FeO in the 65^80wt range (Fig 12c) andhave MgO and FeO contents that are similar to primitiveMORB from the Sequeiros fracture zone of the EastPacific Rise (EPR Perfit et al 1996) These samples
however differ from EPR MORB in having higher SiO2
and lower Al2O3 and they are restricted geographicallyto one area (Sara Lake Fig 2)We therefore have evidence for primary magmas with
MgO contents that range from a possible low of 110wt to as high as 174wt with inferred mantle potential tem-peratures from 1340 to 15208C This is similar to the rangedisplayed by lavas from the Ontong Java Plateau deter-mined by Herzberg (2004) who found that primarymagma compositions assuming a peridotite sourcewould contain 17wt MgO and 10wt CaO(Table 6) and that these magmas would have formed by27 melting at a mantle potential temperature of15008C (first melting occurring at 36 GPa) Fitton ampGodard (2004) estimated 30 partial melting for theOntong Java lavas based on the Zr contents of primarymagmas of Kroenke-type basalt calculated by incrementaladdition of equilibrium olivine However if a considerableamount of eclogite was involved in the formation of theOntong Java plateau magmas the excess temperatures esti-mated by Herzberg et al (2007) may not be required(Korenaga 2005) The variability in mantle potential tem-perature for the Keogh Lake picrites is also similar to thewide range of temperatures that have been calculated for
Table 6 Estimated primary magma compositions for Karmutsen basalts and other oceanic plateaus or islands
Sample 4723A4 4723A13 5616A7 93G171 Average OJP Mauna Keay Gorgonaz
Ontong Java primary magma composition for accumulated fraction melting (AFM) from Herzberg (2004)yMauna Kea primary magma composition is average of four samples of Herzberg (2006 table 1)zGorgona primary magma composition for AFM for 1F fertile source from Herzberg amp OrsquoHara (2002 table 4)Total iron estimated to be FeO is 090
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
27
lavas from single volcanoes in the Galapagos Hawaii andelsewhere by Herzberg amp Asimow (2008) Those workersinterpreted the range to reflect the transport of magmasfrom both the cool periphery and the hotter axis of aplume A thermally heterogeneous plume will yield pri-mary magmas with different MgO contents and the rangeof primary magma compositions and their inferred mantlepotential temperatures for Karmutsen lavas may reflectmelting from different parts of the plume
Source of the Karmutsen Formation lavasThe Sr^Nd^Hf^Pb isotopic compositions of theKarmutsen volcanic rocks on Vancouver Island indicatethat they formed from plume-type mantle containing adepleted component that is compositionally similar to butdistinct from other oceanic plateaus that formed in thePacific Ocean (eg Ontong Java and Caribbean Fig 13)Trace element and isotopic constraints can be used to testwhether the depleted component of the KarmutsenFormation was intrinsic to the mantle plume and distinctfrom the source of MORB or the result of mixing or invol-vement of ambient depleted MORB mantle with plume-type mantle The Karmutsen tholeiitic basalts have initialeNd and
87Sr86Sr ratios that fall within the range of basaltsfrom the Caribbean Plateau (eg Kerr et al 1996 1997Hauff et al 2000 Kerr 2003) and a range of Hawaiian iso-tope data (eg Frey et al 2005) and lie within and abovethe range for the Ontong Java Plateau (Tejada et al 2004Fig 13a) Karmutsen tholeiitic basalts with the highest ini-tial eNd and lowest initial 87Sr86Sr lie within and justbelow the field for age-corrected northern EPR MORB at230 Ma (eg Niu et al1999 Regelous et al1999) Initial Hfand Nd isotopic compositions for the Karmutsen samplesare noticeably offset to the low side of the ocean islandbasalt (OIB) array (Vervoort et al 1999) and lie belowand slightly overlap the low eHf end of fields for Hawaiiand the Caribbean Plateau (Fig 13c) the Karmutsen sam-ples mostly fall between and slightly overlap a field forage-corrected Pacific MORB at 230 Ma and Ontong Java(Mahoney et al 1992 1994 Nowell et al 1998 Chauvel ampBlichert-Toft 2001) Karmutsen lava Pb isotope ratios over-lap the broad range for the Caribbean Plateau and thelinear trends for the Karmutsen samples intersect thefields for EPR MORB Hawaii and Ontong Java in208Pb^206Pb space but do not intersect these fields in207Pb^206Pb space (Fig 13d and e) The more radiogenic207Pb204Pb for a given 206Pb204Pb of the Karmutsenbasalts requires high UPb ratios at an ancient time when235U was abundant similar to the Caribbean Plateau andenriched in comparison with EPR MORB Hawaii andOntong JavaThe low eHf^low eNd component of the Karmutsen
basalts that places them below the OIB mantle array andpartly within the field for age-corrected Pacific MORBcan form from low ratios of LuHf and high SmNd over
time (eg Salters amp White 1998) MORB will develop aHf^Nd isotopic signature over time that falls below themantle array Low LuHf can also lead to low 176Hf177Hffrom remelting of residues produced by melting in theabsence of garnet (eg Salters amp Hart 1991 Salters ampWhite 1998) The Hf^Nd isotopic compositions of theKarmutsen Formation indicate the possible involvementof depleted MORB mantle however similar to Iceland(Fitton et al 2003) Nb^Zr^Y variations provide a way todistinguish between a depleted component within theplume and a MORB-like component The Karmutsenbasalts and picrites fall within the Iceland array and donot overlap the MORB field in a logarithmic plot of NbYvs ZrY (Fig 13b) using the fields established by Fittonet al (1997) Similar to the depleted component within theCaribbean Plateau (Thompson et al 2003) the trend ofHf^Nd isotopic compositions towards Pacific MORB-likecompositions is not followed by MORB values in Nb^Zr^Y systematics and this suggests that the depleted compo-nent in the source of the Karmutsen basalts is distinctfrom the source of MORB and probably an intrinsic partof the plume The relatively radiogenic Pb isotopic ratiosand REE patterns distinct from MORB also support thisinterpretationTrace-element characteristics suggest the possible invol-
vement of sub-arc lithospheric mantle in the generation ofthe Karmutsen picrites Lassiter et al (1995) suggested thatmixing of a plume-type source with eNdthorn 6 to thorn 7 witharc material with low NbTh could reproduce the varia-tions observed in the Karmutsen basalts however theyconcluded that the absence of low NbLa ratios in theirsample of tholeiitic basalts restricts the amount of arc litho-sphere involved The study of Lassiter et al (1995) wasbased on major and trace element and Sr Nd and Pb iso-topic data for a suite of 29 samples from Buttle Lake in theStrathcona Provincial Park on central Vancouver Island(Fig 1) Whereas there are no clear HFSE depletions inthe Karmutsen tholeiitic basalts the low NbLa (and TaLa) of the Karmutsen picritic lavas in this study indicatethe possible involvement of Paleozoic arc lithosphere onVancouver Island (Fig 8)
REE modeling dynamic melting andsource mineralogyThe combined isotopic and trace-element variations of thepicritic and tholeiitic Karmutsen lavas indicate that differ-ences in trace elements may have originated from differentmelting histories For example the LuHf ratios of theKarmutsen picritic lavas (mean 008) are considerablyhigher than those in the tholeiitic basalts (mean 002)however the small range of overlapping initial eHf valuesfor the two lava suites suggests that these differences devel-oped during melting within the plume (Fig 9b) Below wemodel the effects of source mineralogy and depth of
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
28
melting on differences in trace-element concentrations inthe tholeiitic basalts and picritesDynamic melting models simulate progressive decom-
pression melting where some of the melt fraction isretained in the residue and only when the degree of partialmelting is greater than the critical mass porosity and the
source becomes permeable is excess melt extracted fromthe residue (Zou1998) For the Karmutsen lavas evolutionof trace element concentrations was simulated using theincongruent dynamic melting model developed by Zou ampReid (2001) Melting of the mantle at pressures greater than1 GPa involves incongruent melting (eg Longhi 2002)
OIB array
PacificMORB(230 Ma)
(0 Ma)
EPR(230 Ma)
-4
-2
0
2
4
8
10
12
14
16
18
20
22
-4 -2 0 2 4 6 8 10 12 14
minus2
0
2
4
6
8
10
12
14
0702 0703 0704 0705 0706
IndianMORB
87Sr 86Sr (initial)
Hf (initial)
(a) (c)
Nd (initial)
Karmutsen tholeiitic basalt
HawaiiOntong JavaCaribbean Plateau
Karmutsen high-MgO lava
high-MgO lavasKarmutsen
1540
1545
1550
1555
1560
1565
175 180 185 190 195 200375
380
385
390
395
175 180 185 190 195 200
(d) (e)
EPR
EPR
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb
Karmutsen Formation (initial)
HawaiiOntong JavaCaribbean Plateau
East Pacific Rise
Karmutsen Formation (measured)
Juan de FucaGorda
ε Nd
(initial)
Karmutsen Formation (different age-correction for 3 picrites)
(0 Ma)
NbY
001
01
1
1 10
ZrY
OIB
NMORB
(b)
Iceland array
6
Fig 13 Comparison of age-corrected (230 Ma) Sr^Nd^Hf^Pb isotopic compositions for Karmutsen flood basalts onVancouver Island withage-corrected OIB and MORB and Nb^Zr^Y variation (a) Initial eNd vs 87Sr86Sr (b) NbYand ZrY variation (after Fitton et al 1997)(c) Initial eHf vs eNd (d) Measured and initial 207Pb204Pb vs 206Pb204Pb (e) Measured and initial 208Pb204Pb vs 206Pb204Pb References fordata sources are listed in Electronic Appendix 4 Most of the compiled data were extracted from the GEOROC database (httpgeorocmpch-mainzgwdgdegeoroc) OIB array line in (c) is fromVervoort et al (1999) EPR East Pacific Rise Several high-MgO samples affected bysecondary alteration were not plotted for clarity in Pb isotope plots Estimation of fields for age-corrected Pacific MORB and EPR at 230 Mawere made using parent^daughter concentrations (in ppm) from Salters amp Stracke (2004) of Lufrac14 0063 Hffrac14 0202 Smfrac14 027 Ndfrac14 071Rbfrac14 0086 and Srfrac14 836 In the NbYvs ZrYplot the normal (N)-MORB and OIB fields not including Icelandic OIB are from data com-pilations of Fitton et al (1997 2003) The OIB field is data from 780 basalts (MgO45wt ) from major ocean islands given by Fitton et al(2003) The N-MORB field is based on data from the Reykjanes Ridge EPR and Southwest Indian Ridge from Fitton et al (1997) The twoparallel gray lines in (b) (Iceland array) are the limits of data for Icelandic rocks
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
29
with olivine being produced during the reactioncpxthornopxthorn spfrac14meltthornol for spinel peridotites The mod-eling used coefficients for incongruent melting reactionsbased on experiments on lherzolite melting (eg Kinzleramp Grove 1992 Longhi 2002) and was separated into (1)melting of spinel lherzolite (2) two combined intervals ofmelting for a garnet lherzolite (gt lherz) and spinel lher-zolite (sp lherz) source (3) melting of garnet lherzoliteThe model solutions are non-unique and a range indegree of melting partition coefficients melt reactioncoefficients and proportion of mixed source componentsis possible to achieve acceptable solutions (see Fig 14 cap-tion for description of modeling parameters anduncertainties)The LREE-depleted high-MgO lavas require melting of
a depleted spinel lherzolite source (Fig 14a) The modelingresults indicate a high degree of melting (22^25) similarto the results from PRIMELT1 based on major-elementvariations (discussed above) of a LREE-depleted sourceThe high-MgO lavas lack a residual garnet signature Theenriched tholeiitic basalts involved melting of both garnetand spinel lherzolite and represent aggregate melts pro-duced from continuous melting throughout the depth ofthe melting column (Fig 14b) Trace-element concentra-tions in the tholeiitic and picritic lavas are not consistentwith low-degree melting of garnet lherzolite alone(Fig 14c) The modeling results indicate that the aggregatemelts that formed the tholeiitic basalts would haveinvolved lesser proportions of enriched small-degreemelts (1^6 melting) generated at depth and greateramounts of depleted high-degree melts (12^25) gener-ated at lower pressure (a ratio of garnet lherzolite tospinel lherzolite melt of 14 provides the best fits seeFig 14b and description in figure caption) The totaldegree of melting for the tholeiitic basalts was high(23^27) similar to the degree of partial melting in thespinel lherzolite facies for the primary melts that eruptedas the Keogh Lake picrites of a source that was also prob-ably depleted in LREE
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
(c)
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
Melting of garnet lherzolite
High-MgO lavas
Melting of spinel lherzolite
source (07DM+03PM)
source (07DM+03PM)
6 melting
30 melting
(b)
(a)
30 melting
Sam
ple
Cho
ndrit
eS
ampl
eC
hond
rite 20 melting
1 melting
Melting of garnet lherzoliteand spinel lherzolite
x gt lherz + 4x sp lherz
5 meltingx gt lherz + 4x sp lherz
Sam
ple
Cho
ndrit
e
Tholeiitic lavas
Tholeiitic lavas
High-MgO lavas
source (07DM+03PM)
Fig 14 Trace-element modeling results for incongruent dynamicmantle melting for picritic and tholeiitic lavas from the KarmutsenFormation Three steps of modeling are shown (a) melting of spinellherzolite compared with high-MgO lavas (b) melting of garnet lher-zolite and spinel lherzolite compared with tholeiitic lavas (c) meltingof garnet lherzolite compared with tholeiitic and picritic lavasShaded fields in each panel for tholeiitic basalts and picrites arelabeled Patterns with symbols in each panel are modeling results(patterns are 5 melting increments in (a) and (b) and 1 incre-ments in (c) PM primitive mantle DM depleted MORB mantle gtlherz garnet lherzolite sp lherz spinel lherzolite In (b) the ratios ofper cent melting for garnet and spinel lherzolite are indicated (x gtlherzthorn 4x sp lherz means that for 5 melting 1 melt in garnetfacies and 4 melt in spinel facies where x is per cent melting in theinterval of modeling) The ratio of the respective melt proportions inthe aggregate melt (x gt lherzthorn 4x sp lherz) was kept constant andthis ratio of melt contribution provided the best fit to the data Arange of proportion of source components (07DMthorn 03PM to
09DMthorn 01PM) can reproduce the variation in the high-MgOlavas Melting modeling uses the formulation of Zou amp Reid (2001)an example calculation is shown in their Appendix PM fromMcDonough amp Sun (1995) and DM from Salters amp Stracke (2004)Melt reaction coefficients of spinel lherzolite from Kinzler amp Grove(1992) and garnet lherzolite fromWalter (1998) Partition coefficientsfrom Salters amp Stracke (2004) and Shaw (2000) were kept constantSource mineralogy for spinel lherzolite (018cpx027opx052ol003sp) from Kinzler (1997) and for garnet lherzolite(034cpx008opx053ol005gt) from Salters amp Stracke (2004) Arange of source compositions for melting of spinel lherzolite(07DMthorn 03PM to 03DMthorn 07PM) reproduce REE patternssimilar to those of the high-MgO lavas with best fits for 22ndash25melting and 07DMthorn 03PM source composition Concentrationsof tholeiitic basalts cannot be reproduced from an entirelyprimitive or depleted source
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
30
The uncertainty in the composition of the source in thespinel lherzolite melting model for the high-MgO lavasprecludes determining whether the source was moredepleted in incompatible trace elements than the source ofthe tholeiitic lavas (see Fig 14 caption for description) It ispossible that early high-pressure low-degree melting left aresidue within parts of the plume (possibly the hotter inte-rior portion) that was depleted in trace elements (egElliott et al 1991) further decompression and high-degreemelting of such depleted regions could then have generatedthe depleted high-MgO Karmutsen lavas Shallow high-degree melting would preferentially sample depletedregions whereas deeper low-degree melting may notsample more refractory depleted regions (eg CaribbeanEscuder-Viruete et al 2007) The picrites lie at the highend of the range of Nd isotopic compositions and havelow NbZr similar to depleted Icelandic basalts (Fittonet al 2003) therefore the picrites may represent the partsof the mantle plume that were the most depleted prior tothe initiation of melting As Fitton et al (2003) suggestedthese depleted melts may only rarely be erupted withoutcombining with more voluminous less depleted melts andthey may be detected only as undifferentiated small-volume flows like the Keogh Lake picrites within theKarmutsen Formation on Vancouver IslandDecompression melting within the mantle plume initiatedwithin the garnet stability field (35^25 GPa) at highmantle potential temperatures (Tp414508C) and pro-ceeded beneath oceanic arc lithosphere within the spinelstability field (25^09 GPa) where more extensivedegrees of melting could occur
Magmatic evolution of Karmutsentholeiitic basaltsPrimary magmas in large igneous provinces leave exten-sive coarse-grained cumulate residues within or below thecrust these mafic and ultramafic plutonic sequences repre-sent a significant proportion of the original magma thatpartially crystallized at depth (eg Farnetani et al 1996)For example seismic and petrological studies of OntongJava (eg Farnetani et al 1996) combined with the use ofMELTS (Ghiorso amp Sack 1995) indicate that the volcanicsequence may represent only 40 of the total magmareaching the Moho Neal et al (1997) estimated that30^45 fractional crystallization took place for theOntong Java magmas and produced cumulate thicknessesof 7^14 km corresponding to 11^22 km of flood basaltsdepending on the presence of pre-existing oceanic crustand the total crustal thickness (25^35 km) Interpretationsof seismic velocity measurements suggest the presence ofpyroxene and gabbroic cumulates 9^16 km thick beneaththe Ontong Java Plateau (eg Hussong et al 1979)Wrangellia is characterized by high crustal velocities in
seismic refraction lines which is indicative of mafic
plutonic rocks extending to depth beneath VancouverIsland (Clowes et al 1995)Wrangellia crust is 25^30 kmthick beneath Vancouver Island and is underlain by astrongly reflective zone of high velocity and density thathas been interpreted as a major shear zone where lowerWrangellia lithosphere was detached (Clowes et al 1995)The vast majority of the Karmutsen flows are evolved tho-leiitic lavas (eg low MgO high FeOMgO) indicating animportant role for partial crystallization of melts at lowpressure and a significant portion of the crust beneathVancouver Island has seismic properties that are consistentwith crystalline residues from partially crystallizedKarmutsen magmas To test the proportion of the primarymagma that fractionated within the crust MELTS(Ghiorso amp Sack 1995) was used to simulate fractionalcrystallization using several estimated primary magmacompositions from the PRIMELT1 modeling results Themajor-element composition of the primary magmas couldnot be estimated for the tholeiitic basalts because of exten-sive plagthorn cpxthornol fractionation and thus the MELTSmodeling of the picrites is used as a proxy for the evolutionof major elements in the volcanic sequence A pressure of 1kbar was used with variable water contents (anhydrousand 02wt H2O) calculated at the quartz^fayalite^magnetite (QFM) oxygen buffer Experimental resultsfrom Ontong Java indicate that most crystallizationoccurs in magma chambers shallower than 6 km deep(52 kbar Sano amp Yamashita 2004) however some crys-tallization may take place at greater depths (3^5 kbarFarnetani et al 1996)A selection of the MELTS results are shown in Fig 15
Olivine and spinel crystallize at high temperature until15^20wt of the liquid mass has fractionated and theresidual magma contains 9^10wt MgO (Fig 15f) At1235^12258C plagioclase begins crystallizing and between1235 and 11908C olivine ceases crystallizing and clinopyr-oxene saturates (Fig 15f) As expected the addition of clin-opyroxene and plagioclase to the crystallization sequencecauses substantial changes in the composition of the resid-ual liquid FeO and TiO2 increase and Al2O3 and CaOdecrease while the decrease in MgO lessens considerably(Fig 15) The near-vertical trends for the Karmutsen tho-leiitic basalts especially for CaO do not match the calcu-lated liquid-line-of-descent very well which misses manyof the high-CaO high-Al2O3 low-FeO basalts A positivecorrelation between Al2O3CaO and SrNd indicates thataccumulation of plagioclase played a major role in the evo-lution of the tholeiitic basalts (Fig 15g) Increasing thecrystallization pressure slightly and the water content ofthe melts does not systematically improve the match ofthe MELTS models to the observed dataThe MELTS results indicate that a significant propor-
tion of the crystallization takes place before plagioclase(15^20 crystallization) and clinopyroxene (35^45
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
31
crystallization) join the crystallization sequence Theseresults suggest that445 crystallization of the primarymagmas occurred (estimated from the residual liquid percent in Fig 15f) and if the original flood basalt stratigra-phy on Vancouver Island was 6 km thick at least anequivalent amount of complementary plutonic rocksshould have crystallized The Karmutsen basalts and theirplutonic residues thus represent a significant addition ofcrust (perhaps 412 km thickness) previously thickenedduring the formation of the underlying Paleozoic arc
Growth of the Wrangellia oceanic plateauon Vancouver IslandThe extrusive part of the Wrangellia oceanic plateau onVancouver Island was constructed in a three-layersequence of mostly tholeiitic basalts with a restrictedrange of composition (Fig 16) The plateau overliesDevonian to Mississippian arc volcanic sequences andmarine sediments of Mississippian to Permian ages thatare intruded by gabbroic rocks related to the flood basaltsThe volcanic stratigraphy of theWrangellia plateau formed
00
05
10
15
20
25
30
35
40
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
16
0 2 4 6 8 10 12 14 16 18 20
10
11
12
13
14
15
16
17
18
19
20
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
0 2 4 6 8 10 12 14 16 18 20
MgO (wt)
0 10 20 30 40 50
percent of total mass
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
) Mineral proportions
Sp (e)
Ol
CpxPlag
PicriteHigh-MgO basaltTholeiitic basalt
TiO2 (wt)
MgO (wt)
Al2O3 (wt) CaO (wt)
FeO (wt)
MgO(a) (b)
(c) (d)
MgO (wt)MgO (wt)
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
)
Residual liquid ()
1220
1190
Ol + Sp
Plag+Ol+Sp Plag+Cpx+Sp
02 wt H2O
no H2O
Residual liquid and crystallizing phases
(f)
0102030405060708090100
plagioclaseaccumulation
plagioclaseaccumulation
plagioclase
accumulatio
n
10
12
14
16
0 10 20 30 40
Al2O3CaO
SrNd
tholeiitic basalts
(g)
Fig 15 Forward fractional crystallization modeling results for major elements from MELTS (Ghiorso amp Sack 1995) compared with picriticand tholeiitic lavas from the Karmutsen Formation (a^d) Major-element compositions of residual liquids compared with Karmutsen lavacompositions For clarity the results are shown only for one starting composition (open circle) using the composition of the estimated primarymagma for the high-MgO lava series from the modeling technique of Herzberg et al (2007) for sample 4723A13 as shown in Fig 12 The esti-mated primary magmas for samples 4723A3 and 93G171 yield similar results A pressure of 1 kbar was used with no H2O and 02wt H2Ocalculated at QFM black line is anhydrous and grey line is 02wt H2OThe end-point for the anhydrous result is at 78 fractionation and11308C and the end-point for the result with 02wt H2O is 75 fractionation and11108C (e) Mineral proportions of crystallized phases fromone MELTrun (sample 4723A13) with no H2O (f) Per cent residual liquid vs temperature labelled with intervals of crystallizing phases (g)Al2O3CaO vs SrNd for the tholeiitic basalts
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
32
in a single eruptive phase (c 231^225 Ma) as an emergentbasalt sequence during a deeper-water shallow-water andsubaerial stage and subsided below sea level relativelysoon after its formation (Fig 16) The deeper-water stagecontains pillowed and massive submarine flows (3000mthick) that built up the submarine volcanic edifice from theseafloor The lowest pillow basalts were emplaced onunconsolidated and lithified fine-grained siliciclastic andcarbonaceous sediments containing Daonella beds(c 235^232 Ma) that are intruded by abundantKarmutsen mafic sills (Fig 16a) The different pillowedand massive submarine flows resulted primarily from dif-ferent effusive rates and local topography Some of themassive submarine flows commonly recognized by radiat-ing columnar jointing represent master tubes for the deliv-ery of lava to distal parts of submarine flow fields Thepillowed flows which commonly have uneven topography
in modern sequences may have served as levees to themassive submarine flows rather than allowing unob-structed spreading and inflation typical of the massive sub-aerial flows that form on shallow slopes in continentalflood basalt provinces (eg Self et al 1997) The rapid suc-cession of flows is evident by the near absence of sedimentbetween submarine flows Minor volumes of picritic pillowlavas occur in the upper part of the pillowed flow sequencein areas of the plateau exposed on northern VancouverIsland (Fig 16 Nixon et al 2008)The growth of the plateau in shallow water led to
emplacement of an increasing proportion of volcaniclastics(pillow breccias and hyaloclastites) that accumulated up to1500m thick in some areas (Fig 16b Nixon et al 2008)Horizons of pillowed flows occur locally within the volca-niclastic stratigraphy Sedimentary structures (eg gradedbedding fluidization structures) and petrographic textures
Late Devonian-Mississippian arc
Daonella beds
Middle Triassic
deeper water
Late Devonian-Mississippian arc
pillowed flows massive submarine
flows
pillow brecciahyaloclastiteshallow water
~4 km
~6 km
Late Triassic limestoneinterflow limestonelenses
Karmutsen Formation
mafic intrusions sediment-sill complex
Karmutsen Formation
(a) (b)
(c) (d)
subaerial flows
Late Devonian-Mississippian arc Late Devonian-Mississippian arc
picritic pillow basalts
Miss to Permian marine sediments
sills
Fig 16 Growth of the Wrangellia oceanic plateau onVancouver Island (a) Deep-water eruptive phase of pillowed and massive submarineflows with almost no intervening sediments onto MiddleTriassic shale containing Daonella beds (c 235^232 Ma) The basement of the plateauis composed of Devonian to Mississippian arc sequences overlain by Mississippian to Permian carbonates and siliciclastic sediments(b) Shallow-water flow emplacement with increasing proportion of volcaniclastics resulted primarily from cooling^contraction granulationPicritic pillow basalts occur primarily near the transition between pillowed submarine flows and volcaniclastic units Pillowed and massiveflows occur locally within the volcaniclastic unit (c) Emergent stage of growth consists of overlapping subaerial flow fields similar to continen-tal flood basalts with pahoehoe flow structures Multi-tiered columnar jointing occurrs locally in response to ponding of flows (eg Bondre et al2004) Interflow carbonate deposits and plagioclase megacrystic flows developed locally during the waning stage of volcanism (d) Post-volcanicsubsidence and deposition of Late Carnian to Norian (c 212^223 Ma) platform carbonates on top of the plateau Units are labeled in the dia-grams Thicknesses of the underlying units are not to scale
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
33
within the volcaniclastic rocks indicate formation primar-ily via cooling^contraction granulation magma^water^steam interaction autobrecciation resedimentation andmass-wasting rather than pyroclastic eruption (Nixonet al 2008)A broad subaerial platform was constructed by emplace-
ment of subaerial lava flows during an emergent subaerialstage (Fig 16c) The subaerial unit predominantly com-prises well-layered sheet flows that formed in a similarway to continental flood basalts as hundred(s) of pahoehoeflow fields erupted on shallow slopes (Self et al 1997) how-ever volcaniclastic and pillowed flow units are locally pre-sent mostly in the upper parts of the stratigraphyLocalized ponding of flows led to the formation of multi-tiered columnar jointing in some subaerial flows (egBondre et al 2004) As volcanism waned local interflowcarbonate deposits developed and plagioclase megacrysticflows were erupted Subsidence of the plateau is recordedby the interflow lenses and deposits of platform carbonates(40^1500m thick) on top of the Karmutsen basalts thatcontain Late Carnian to Early Norian fossils (Fig 16d c216^227 Ma Nixon et al 2008)The evolution of the internal architecture of the
Wrangellia oceanic plateau was primarily related to theconditions of emplacement (ie water depth and eruptionenvironment) There does not appear to have been a sys-tematic coincident evolution or shift of lava compositionsonVancouver Island as has been recognized in most con-tinental flood basalts and parts of the Wrangellia oceanicplateau in Alaska and Yukon (Greene et al 2008 2009)In general with the exception of the picritic pillow lavasin the submarine stratigraphy and local variations betweenhigh-MgO and tholeiitic lavas there are no significantcompositional differences between lavas erupted duringthe different stages of growth of the plateau (ie submarineand subaerial flows) nor are substantial compositional dif-ferences generally found between massive submarine flowsin contact with pillowed flows tholeiitic basalt flows nearthe base of the Karmutsen Formation are compositionallysimilar to flows sampled near the top of the volcanicsequence Where it was possible to sample continuous sec-tions of volcanic stratigraphy greater than several hun-dreds of meters (eg Keogh^Maynard Lake area west ofthe Karmutsen Range Fig 2) there is no clear relationshipbetween stratigraphic position and compositionalvariation
CONCLUSIONSThe Karmutsen Formation covers extensive areas ofVancouver Island (20 000 km2) and forms part of a largeoceanic plateau erupted over a geologically short interval(c 225^231 Ma) The tripartite volcanic stratigraphyexposed on Vancouver Island is at least 6 km thick andcomprises a succession of submarine flows volcaniclastic
deposits and massive subaerial flowsThese volcanologicaldifferences are primarily related to the eruption environ-ment (deep-water shallow-water or subaerial) The rapidgrowth of the plateau prevented significant accumulationsof intervening sediment except in the uppermost stratigra-phy where isolated limestone lenses commonly associatedwith pillow basalt and hyaloclastite deposits preserve arecord of the final stages of subsidence of the plateauThe Wrangellia plateau on Vancouver Island was con-
structed dominantly of tholeiitic basalt with restrictedmajor- and trace-element and isotopic compositional var-iations Minor volumes of picritic pillow basalts eruptedlate in the submarine phase The Karmutsen basalts wereemplaced onto older Paleozoic arc volcanic sequences andlow NbLa for the picrites may indicate a contributionfrom the sub-arc lithospheric mantle Modeling of the pet-rogenesis of the Karmutsen picrites indicates melting ofanomalously hot mantle (15008C) and extensive degreesof partial melting (23^27) and is consistent with aplume initiation model Combined Sr^Nd^Hf^Pb isotopicsystematics indicate that the picritic and tholeiitic lavaswere derived from a depleted plume-type mantle sourcedistinct from the source of MORB that has compositionalsimilarities to the source of the lavas forming theCaribbean Plateau Picrites were generated by melting ofspinel lherzolite whereas the tholeiitic basalts formed asaggregate melts of both garnet and spinel lherzolite Thetholeiitic basalts underwent extensive low-pressure fractio-nation (52^3 kbar) and seismic studies indicate that someof theWrangellia crust beneathVancouver Island may cor-respond to the plutonic residues of the Karmutsen floodbasalts
ACKNOWLEDGEMENTSWe are grateful to Nick Arndt for helping us get this pro-ject started and Nick Massey for insights into VancouverIsland geology We greatly appreciate the assistance withPRIMELT modeling from Claude Herzberg We thankMikkel Schau for his insight and enthusiasm during field-work Jane Barling assisted with analyses by MC-ICP-MSThoughtful and thorough reviews of the manuscript byJohn Mahoney Fred Frey Wendy Bohrson DougalJerram and Teal Riley are greatly appreciated Fundingwas generously provided by the Rocks to Riches Programadministered by the BC amp Yukon Chamber of Mines in2004 by the BC Geological Survey and by NSERCDiscovery Grants to JSS and DW ARG was supportedby a University Graduate Fellowship at UBC
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
34
REFERENCESBarker F Brown A S Budahn J R amp Plafker G (1989) Back-arc
with frontal-arc component origin of Triassic Karmutsen basaltBritish Columbia Canada Chemical Geology 75 81^102
Blichert-Toft JWeis D Maerschalk C Agranier A amp Albare de F(2003) Hawaiian hot spot dynamics as inferred from the Hf and Pbisotope evolution of Mauna Kea volcano Geochemistry GeophysicsGeosystems 4(2) 1^27 doi1010292002GC000340
Bondre N R Duraiswami R A amp Dole G (2004) Morphologyand emplacement of flows from the Deccan Volcanic ProvinceIndia Bulletin of Volcanology 66 29^45
Carlisle D (1963) Pillow breccias and their aquagene tuffs QuadraIsland British Columbia Journal of Geology 71 48^71
Carlisle D (1972) Late Paleozoic to Mid-Triassic sedimentary-volcanic sequence on Northeastern Vancouver Island Geological
Society of Canada Paper Report of Activities 72-1 Part B 24^30Carlisle D amp SuzukiT (1974) Emergent basalt and submergent car-
bonate^clastic sequences including the Upper Triassic Dilleri andWelleri zones on Vancouver Island Canadian Journal of Earth
Sciences 11 254^279Chauvel C amp Blichert-Toft J (2001) A hafnium isotope and trace
element perspective on melting of the depleted mantle Earth and
Planetary Science Letters 190 137^151Chayes F (1954)The theory of thin-section analysis Journal of Geology
62 92^101Cho M Liou J G amp Maruyama S (1987) Transition from the zeo-
lite to prehnite^pumpellyite facies in the Karmutsen metabasitesVancouver Island British Columbia Journal of Petrology 27(2)467^494
Clowes R M Zelt C A Amor J R amp Ellis R M (1995)Lithospheric structure in the southern Canadian Cordillera froma network of seismic refraction lines Canadian Journal of Earth
Sciences 32(10) 1485^1513Coffin M F amp Eldholm O (1994) Large igneous provinces Crustal
structure dimensions and external consequences Reviews of
Geophysics 32(1) 1^36Cox K G (1980) A model for flood basalt volcanism Journal of
Petrology 21 629^650Eldholm O amp Coffin M F (2000) Large igneous provinces
and plate tectonics In Richards M A Gordon R G amp vander Hilst R D (eds) The History and Dynamics of Global
Plate Motions American Geophysical Union Geophysical Monograph 121309^326
Elliott T R Hawkesworth C J amp Grolaquo nvold K (1991) Dynamicmelting of the Iceland plume Nature 351 201^206
Escuder-Viruete J Pecurren rez-Estaucurren n A Contreras F Joubert MWeis D Ullrich T D amp Spadea P (2007) Plume mantle sourceheterogeneity through time Insights from the Duarte ComplexHispaniola northeastern Caribbean Journal of Geophysical Research112(B04203) doi1010292006JB004323
Farnetani C G Richards M A amp Ghiorso M S (1996)Petrological models of magma evolution and deep crustal structurebeneath hotspots and flood basalt provinces Earth and Planetary
Science Letters 143 81^94Fitton J G amp Godard M (2004) Origin and evolution of magmas
on the Ontong Java Plateau In Fitton J G Mahoney J JWallace P J amp Saunders A D (eds) Origin and Evolution of the
Ontong Java Plateau Geological Society London Special Publications 229151^178
Fitton J G Saunders A D Norry M J Hardarson B S ampTaylor R N (1997) Thermal and chemical structure of theIceland plume Earth and Planetary Science Letters 153 197^208
Fitton J G Saunders A D Kempton P D amp Hardarson B S(2003) Does depleted mantle form an intrinsic part of the Icelandplume Geochemistry Geophysics Geosystems 4(3) 1032 doi1010292002GC000424
Frey F A Huang S Blichert-Toft J Regelous M amp Boyet M(2005) Origin of depleted components in basalt related to theHawaiian hotspot Evidence from isotopic and incompatible ele-ment ratios Geochemistry Geophysics Geosystems 6(1) 23 doi1010292004GC000757
Galer S J G amp Abouchami W (1998) Practical application of leadtriple spiking for correction of instrumental mass discriminationMineralogical Magazine 62A 491^492
Ghiorso M S amp Sack R O (1995) Chemical mass transfer in mag-matic processes IV A revised and internally consistent thermody-namic model for the interpolation and extrapolation of liquid^solidequilibria in magmatic systems at elevated temperatures and pres-sures Contributions to Mineralogy and Petrology 119 197^212
Greene A R Scoates J S amp Weis D (2005)WrangelliaTerrane onVancouver Island Distribution of flood basalts with implicationsfor potential Ni^Cu^PGE mineralization In Grant B (ed)Geological Fieldwork 2004 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2005-1 209^220Greene A R Scoates J S Nixon G T amp Weis D (2006) Picritic
lavas and basal sills in the Karmutsen flood basalt provinceWrangellia northern Vancouver Island In Grant B (ed)Geological Fieldwork 2005 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2006-1 39^52Greene A R Scoates J S amp Weis D (2008) Wrangellia flood
basalts in Alaska A record of plume^lithosphere interaction in aLate Triassic accreted oceanic plateau Geochemistry Geophysics
Geosystems 9(Q12004) doi1010292008GC002092Greene A R Scoates J S Weis D amp Israel S (2009)
Geochemistry of triassic flood basalts from the Yukon (Canada)segment of the accreted Wrangellia oceanic plateau Lithosdoi101016jlithos200811010
Hauff F Hoernle K Tilton G Graham D W amp Kerr A C(2000) Large volume recycling of oceanic lithosphere over shorttime scales geochemical constraints from the Caribbean LargeIgneous Province Earth and Planetary Science Letters 174 247^263
Hauff F Hoernle K amp Schmidt A (2003) Sr^Nd^Pb compositionof Mesozoic Pacific oceanic crust (Site 1149 and 801 ODP Leg185)Implications for alteration of ocean crust and the input into theIzu^Bonin^Mariana subduction system Geochemistry Geophysics
Geosystems 4(8) doi1010292002GC000421Herzberg C (2004) Partial melting below the Ontong Java Plateau
In Fitton J G Mahoney J J Wallace P J amp Saunders A D(eds) Origin and Evolution of the Ontong Java Plateau Geological SocietyLondon Special Publications 229 179^183
Herzberg C (2006) Petrology and thermal structure of the Hawaiianplume from Mauna Kea volcano Nature 444(7119) 605
Herzberg C amp Asimow P D (2008) Petrology of some oceanicisland basalts PRIMELT2XLS software for primary magma cal-culation Geochemistry Geophysics Geosystems 9(Q09001) doi1010292008GC002057
Herzberg C amp OrsquoHara M J (2002) Plume-associated ultramaficmagmas of Phanerozoic age Journal of Petrology 43(10) 1857^1883
Herzberg C Asimow P D Arndt N Niu Y Lesher C MFitton J G Cheadle M J amp Saunders A D (2007)Temperatures in ambient mantle and plumes Constraints frombasalts picrites and komatiites Geochemistry Geophysics Geosystems8(Q02006) doi1010292006GC001390
Huang S amp Frey F A (2005) Trace element abundances of MaunaKea basalt from phase 2 of the Hawaii Scientific Drilling Project
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
35
Petrogenetic implications of correlations with major element con-tent and isotopic ratios Geochemistry Geophysics Geosystems 4(6)doi1010292002GC000322
Hussong D M Wipperman L K amp Kroenke L W (1979) Thecrustal structure of the Ontong Java and Manihiki OceanicPlateaus Journal of Geophysical Research 84(B11) 6003^6010
Jerram D A amp Widdowson M (2005) The anatomy of ContinentalFlood Basalt Provinces geological constraints on the processes andproducts of flood volcanism Lithos 79(3^4) 385^405
Jones D L Silberling N J amp Hillhouse J (1977)Wrangellia a dis-placed terrane in northwestern North America CanadianJournal ofEarth Sciences 14(11) 2565^2577
Kerr A C (2003) Oceanic Plateaus In Rudnick R (ed) The Crust
Treatise on GeochemistryVol 3 Oxford Elsevier Science pp 537^565Kerr A C amp Mahoney J J (2007) Oceanic plateaus Problematic
plumes potential paradigms Chemical Geology 241 332^353Kerr A CTarney J Marriner G F Klaver GT Saunders A D
amp Thirlwall M F (1996) The geochemistry and petrogenesis ofthe Late-Cretaceous picrites and basalts of Curacao NetherlandsAntilles a remnant of an oceanic plateau Contributions to
Mineralogy and Petrology 124(1) 29^43Kerr A C Marriner G F Tarney J Nivia A Saunders A D
Thirlwall M F amp Sinton A C (1997) Cretaceous basaltic ter-ranes in Western Colombia Elemental chronological and Sr-Ndisotopic constraints on petrogenesis Journal of Petrology 38(6)677^702
Kinzler R J (1997) Melting of mantle peridotite at pressuresapproaching the spinel to garnet transition Application to mid-ocean ridge basalt petrogenesis Journal of Geophysical Research
102(B1) 853^874Kinzler R J amp Grove T L (1992) Primary magmas of mid-ocean
ridge basalts 1 Experiments and methods Journal of Geophysical
Research 97(B5) 6885^6906Korenaga J (2005) Why did not the Ontong Java Plateau form sub-
aerially Earth and Planetary Science Letters 234 385^399Kuniyoshi S (1972) Petrology of the Karmutsen (Volcanic) Group
northeastern Vancouver Island British Columbia PhD disserta-tion University of California Los Angeles 242 pp
Lassiter J C DePaolo D J amp Mahoney J J (1995) Geochemistry ofthe Wrangellia flood basalt province Implications for the role ofcontinental and oceanic lithosphere in flood basalt genesis Journalof Petrology 36(4) 983^1009
LeBas M J (2000) IUGS reclassification of the high-Mg and picriticvolcanic rocks Journal of Petrology 41(10) 1467^1470
Longhi J (2002) Some phase equilibrium systematics of lherzolitemelting I Geochemistry Geophysics Geosystems 3(3) doi1010292001GC000204
MacDonald G A amp Katsura T (1964) Chemical composition ofHawaiian lavas Journal of Petrology 5 82^133
Mahoney J J (1987) An isotopic survey of Pacific oceanic plateausimplications for their nature and origin In Keating B HFryer P Batiza R amp Boehlert GW (eds) Seamounts Islands andAtolls American Geophysical Union Geophysical Monograph 43 207^220
Mahoney J LeRoex A P Peng Z Fisher R L amp Natland J H(1992) Southwestern limits of Indian Ocean Ridge mantle and theorigin of low 206Pb204Pb mid-ocean ridge basalt isotope systema-tics of the central Southwest Indian Ridge (178^508E) Journal ofGeophysical Research 97(B13) 19771^19790
Mahoney J J Sinton J M Kurz M D MacDougall J DSpencer K J amp Lugmair GW (1994) Isotope and trace elementcharacteristics of a super-fast spreading ridge East Pacific rise13^238S Earth and Planetary Science Letters 121 173^193
Massey N W D (1995a) Geology and mineral resources of theAlberni^Nanaimo Lakes sheet Vancouver Island 92F1W 92F2Eand part of 92F7E British Columbia Ministry of Energy Mines and
Petroleum Resources Paper 1992-2 132 ppMassey N W D (1995b) Geology and mineral resources of the
Cowichan Lake sheet Vancouver Island 92C16 British Columbia
Ministry of Energy Mines and Petroleum Resources Paper 1992-3 112 ppMassey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005a) Digital Geology Map of British Columbia Tile NM9 MidCoast BC British Columbia Ministry of Energy and Mines Geofile
2005-2Massey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005b) Digital Geology Map of British Columbia Tile NM10Southwest BC British Columbia Ministry of Energy and Mines Geofile
2005-3McDonough W F amp Sun S (1995) The composition of the Earth
Chemical Geology 120 223^253Monger JW H amp Journeay J M (1994) Basement geology and tec-
tonic evolution of theVancouver region In Monger JW H (ed)Geology and Geological Hazards of the Vancouver Region Southwestern
British Columbia Geological Survey of Canada Bulletin 481 3^25Muller J E (1977) Geology of Vancouver Island Geological Survey of
Canada Open File Map 463Muller J E (1980)The Paleozoic Sicker Group of Vancouver Island British
Columbia Geological Survey of Canada Paper 79-30 22 ppMuller J E Northcote K E amp Carlisle D (1974) Geology and
mineral deposits of Alert Bay^Cape Scott map area VancouverIsland British Columbia Geological Survey of Canada Paper 74-8 77
Neal C R Mahoney J J Kroenke L W Duncan R A ampPetterson M G (1997) The Ontong^Java Plateau InMahoney J J amp Coffin M F (eds) Large Igneous Provinces
Continental Oceanic and Planetary FloodVolcanism American Geophysical
Union Geophysical Monograph 100 183^216Niu Y Collerson K D Batiza R Wendt J I amp Regelous M
(1999) Origin of enriched-typed mid-ocean ridge basalt at ridgesfar from mantle plumes The East Pacific Rise at 11820rsquoN Journalof Geophysical Research 104 7067^7088
Nixon G T amp Orr A J (2007) Recent revisions to the EarlyMesozoic stratigraphy of Northern Vancouver Island (NTS 102I092L) and metallogenic implicatinos British Columbia InGrant B (ed) Geological Fieldwork 2006 British Columbia Ministry of
Energy Mines and Petroleum Resources Paper 2007-1 163^177Nixon GT Hammack J L KoyanagiV M Payie G J Haggart
JW Orchard M JTozerT Friedman R M Archibald D APalfy J amp Cordey F (2006a) Geology of the Quatsino^PortMcNeill area northernVancouver Island British Columbia Ministry
of Energy Mines and Petroleum Resources Geoscience Map 2006-2 scale150 000
Nixon G T Kelman M C Stevenson D Stokes L A ampJohnston K A (2006b) Preliminary geology of the Nimpkishmap area (NTS 092L07) northern Vancouver Island BritishColumbia In Grant B amp Newell J M (eds) Geological Fieldwork2005 British Columbia Ministry of Energy Mines and Petroleum Resources
Paper 2006-1 135^152Nixon G T Laroque J Pals A Styan J Greene A R amp
Scoates J S (2008) High-Mg lavas in the Karmutsen flood basaltsnorthernVancouver Island (NTS 092L) Stratigraphic setting andmetallogenic significance In Grant B (ed) Geological Fieldwork
2007 British Columbia Ministry of Energy Mines and Petroleum
Resources Paper 2008-1 175^190Nowell G M Kempton P D Noble S R Fitton J G
Saunders A Mahoney J J amp Taylor R N (1998) High precisionHf isotope measurements of MORB and OIB by thermal ionisation
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
36
mass spectrometry insights into the depleted mantle Chemical
Geology 149 211^233Parrish R R amp McNicoll V J (1992) U^Pb age determinations
from the southern Vancouver Island area British Columbia InRadiogenic Age and Isotopic Studies Report 5 Geological Survey of
Canada Paper 91-2 79^86Peate DW (1997) The Parana^Etendeka Province In Coffin M F
amp Mahoney J (eds) Large Igneous Provinces Continental Oceanic andPlanetary Flood Volcanism American Geophysical Union Geophysical
Monograph 100 217^245Perfit M R Fornari D J Ridley W I Kirk P D Casey J
Kastens K A Reynolds J R Edwards M Desonie DShuster R amp Paradis S (1996) Recent volcanism in the Siqueirostransform fault picritic basalts and implications for MORBmagma genesis Earth and Planetary Science Letters 141(1^4) 91^108
Pretorius W Weis D Williams G Hanano D Kieffer B ampScoates J S (2006) Complete trace elemental characterization ofgranitoid (USGSG-2GSP-2) reference materials by high resolutioninductively coupled plasma-mass spectrometry Geostandards and
Geoanalytical Research 30(1) 39^54Regelous M Niu Y Wendt J I Batiza R Greig A amp
Collerson K D (1999) Variations in the geochemistry of magma-tism on the East Pacific Rise at 10830rsquoN since 800 ka Earth and
Planetary Science Letters 168 45^63Recurren villon S Chauvel C Arndt N T Pik R Martineau F
Fourcade S amp Marty B (2002) Heterogeneity of the Caribbeanplateau mantle source Sr O and He isotopic compositions of oli-vine and clinopyroxene from Gorgona Island Earth and Planetary
Science Letters 205(1^2) 91Rhodes J M (1996) Geochemical stratigraphy of lava flows samples
by the Hawaii Scientific Drilling Project Journal of Geophysical
Research 101(B5) 11729^11746Rhodes J M amp Vollinger M J (2004) Composition of basaltic lavas
sampled by phase-2 of the Hawaii Scientific Drilling ProjectGeochemical stratigraphy and magma types Geochemistry
Geophysics Geosystems 5(3) doi1010292002GC000434Richards M A Jones D L Duncan R A amp DePaolo D J (1991)
A mantle plume initiation model for the Wrangellia flood basaltand other oceanic plateaus Science 254 263^267
Salters V J M amp Hart S R (1991) The mantle sources of oceanridges islands and arcs the Hf-isotope connection Earth and
Planetary Science Letters 104 364^380Salters V J M amp Stracke A (2004) Composition of the depleted
SaltersV J amp WhiteW (1998) Hf isotope constraints on mantle evo-lution Chemical Geology 145 447^460
SanoT amp Yamashita S (2004) Experimental petrology of basementlavas from Ocean Drilling Program Leg 192 implications for dif-ferentiation processes in Ontong Java Plateau magmas InFitton J G Mahoney J JWallace P J amp Saunders A D (eds)Origin and Evolution of the Ontong Java Plateau Geological Society
London Special Publications 229 185^218Saunders A D (2005) Large igneous provinces Origin and environ-
mental consequences Elements 1 259^263Self S ThordarsonT amp Keszthely L (1997) Emplacement of conti-
nental flood basalt lava flows In Mahoney J J amp Coffin M F(eds) Large Igneous Provinces Continental Oceanic and Planetary Flood
Volcanism American Geophysical Union Geophysical Monograph 100381^410
Shaw D M (2000) Continuous (dynamic) melting theory revisitedCanadian Mineralogist 38(5) 1041^1063
Sluggett C L (2003) Uranium^lead age and geochemical constraintson Paleozoic and Early Mesozoic magmatism in WrangelliaTerrane Saltspring Island British Columbia BSc thesisUniversity of British ColumbiaVancouver 84 pp
Storey M Mahoney J J amp Saunders A D (1997) Cretaceousbasalts in Madagascar and the transition between plume and con-tinental lithospheric mantle sources In Mahoney J J ampCoffin M F (eds) Large Igneous Provinces Continental Oceanic and
Planetary Flood Volcanism Aamerican Geophysical Union Geophysical
Monograph 100 95^122Surdam R C (1967) Low-grade metamorphism of the Karmutsen
Group PhD dissertation University of California Los Angeles288 pp
Sutherland-Brown A Yorath C J Anderson R G amp Dom K(1986) Geological maps of southern Vancouver IslandLITHOPROBE I 92C10 11 14 16 92F1 2 7 8 Geological Survey ofCanada Open File 1272
Tejada M L G Mahoney J J Castillo P R Ingle S P Sheth HC amp Weis D (2004) Pin-pricking the elephant evidence on theorigin of the Ontong Java Plateau from Pb^Sr^Hf^Nd isotopiccharacteristics of ODP Leg 192 basalts In Fitton J GMahoney J J Wallace P J amp Saunders A D (eds) Origin and
Evolution of the Ontong Java Plateau Geological Society London Special
Publications 229 133^150Thompson P M E Kempton P D amp Kerr A C (2008) Evaluation
of the effects of alteration and leaching on Sm^Nd and Lu^Hf sys-tematics in submarine mafic rocks Lithos 104(1^4) 164^176
Thompson P M E Kempton P D White R V Kerr A CTarney J Saunders A D Fitton J G amp McBirney A (2003)Hf-Nd isotope constraints on the origin of the CretaceousCaribbean plateau and its relationship to the Galapagos plumeEarth and Planetary Science Letters 217(1^2) 59^75
Vervoort J D Patchett P Blichert-Toft J amp Albare de F (1999)Relationships between Lu^Hf and Sm^Nd isotopic systems in theglobal sedimentary system Earth and Planetary Science Letters 16879^99
Walter M J (1998) Melting of garnet peridotite and the origin ofkomatiite and depleted lithosphere Journal of Petrology 39(1) 29^60
Weis D Kieffer B Maerschalk C Barling J de Jong JWilliams G A Hanano D Mattielli N Scoates J SGoolaerts A Friedman R A amp Mahoney J B (2006) High-pre-cision isotopic characterization of USGS reference materials byTIMS and MC-ICP-MS Geochemistry Geophysics Geosystems
7(Q08006) doi1010292006GC001283Weis D Kieffer B Hanano D Silva I N Barling J
Pretorius W Maerschalk C amp Mattielli N (2007) Hf isotopecompositions of US Geological Survey reference materialsGeochemistry Geophysics Geosystems 8(Q06006) doi1010292006GC001473
Wheeler J O amp McFeely P (1991) Tectonic assemblage map of theCanadian Cordillera and adjacent part of the United States ofAmerica Geological Survey of Canada Map 1712A
WhiteW M Albare de F amp Tecurren louk P (2000) High-precision analy-sis of Pb isotope ratios by multi-collector ICP-MS Chemical Geology167 257^270
Yorath C J Sutherland Brown A amp Massey N W D (1999)LITHOPROBE southern Vancouver Island British Columbia Geological
Survey of Canada Bulletin 498 145 ppZou H (1998) Trace element fractionation during modal and
non-model dynamic melting and open-system melting Amathematical treatment Geochimica et Cosmochimica Acta 62(11)1937^1945
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
37
Zou H amp Reid M R (2001) Quantitative modeling of trace elementfractionation during incongruent dynamic melting Geochimica et
Cosmochimica Acta 65(1) 153^162
APPENDIX SAMPLEPREPARAT ION ANDANALYT ICAL METHODSOnly the freshest rocks in the field were sampled and onlythe least altered samples were selected for chemical andisotopic analysis based on thorough petrographic inspec-tion Fifty-nine of the 129 collected samples were crushed(400 g) into small pieces 52mm in diameter in aRocklabs hydraulic piston crusher between WC plates tominimize contamination The coarse-crush was mixedand 100 g was powdered in a planetary mill using agatejars and balls cleaned with quartz sand between samples
ActLabs analytical methodsThe major- and trace-element compositions of the whole-rock powders were determined at Activation LaboratoriesLtd (Actlabs) in Ancaster Ontario Analytical techniquesand detection limits are also available from Actlabs (httpwwwactlabscommethsub_code4erehtm) The analyticalmethod for each of the elements analyzed is indicated inTable 2 For the major elements a 02 g sample was mixedwith a lithium metaborate^lithium tetraborate mixtureand fused in a graphite crucible The molten mixture waspoured into a 5 HNO3 solution and shaken until dis-solved (30min) The samples were analyzed for majoroxides and selected trace elements by inductively coupledplasma optical emission spectrometry (ICP-OES) on acombination simultaneous^sequential Thermo Jarrell-AshEnviro II system Internal calibration was achieved usinga variety of international reference materials (eg W-2BIR-1 DNC-1) and independent control samplesAdditional trace elements were analyzed by both instru-mental neutron activation analysis (INAA) and induc-tively couple plasma mass spectrometry (ICP-MS) ForINAA15^25 g of sample was weighed into small polyeth-ylene vials and irradiated with a control international ref-erence material CANMET WMS-1 and NiCr flux wires ata thermal neutron flux of 71012 neutronscm2 per s in theMcMaster Nuclear Reactor Following a 7 day waitingperiod the samples were measured on an Ortec high-purity Ge detector linked to a Canberra Series 95 multi-channel analyzer Activities for each element weredecay- and weight-corrected and compared with a detectorcalibration developed from multiple international certifiedreference materials For ICP-MS 025 g of sample wasdigested in HF followed by HNO3^HClO4 heated andtaken to dryness The samples were brought back into solu-tion with HCl Samples were analyzed using a PerkinElmer Optima 3000 ICP In-lab standards or certified ref-erence materials (eg W-2 BIR-1 DNC-1) were used for
quality control A total of 15 blind duplicates were ana-lyzed from the same sample powders to assessreproducibility
University of Massachusetts analyticalmethodsFourteen sample duplicate powders taken from the same100 g sample aliquot (high-MgO basalts and picrites)were also analyzed at the Ronald B Gilmore X-RayFluorescence (XRF) Laboratory at the University ofMassachusetts Major elements were measured on a fusedLa-bearing lithium borate glass disc using a SiemensMRS-400 spectrometer Trace element concentrations(Rb Sr Ba Ce Nb Zr Y Pb Zn Ga Ni Cr V) weremeasured on a separate powder pellet using a PhilipsPW2400 sequential spectrometer using a Rh tube LOIand ferrous iron measurements were made as describedby Rhodes amp Vollinger (2004) Precision and accuracy esti-mates for the data have been described by Rhodes (1996)and Rhodes amp Vollinger (2004) Results for each sampleare the average of two separate analyses and are shown inElectronic Appendix 2 for a comparison of analyses ofduplicate powders using XRFand ICP-MS
PCIGR analytical methodsA subset of 19 samples was selected for high-precisiontrace-element analysis and Sr Nd Pb and Hf isotopicanalysis at the Pacific Centre for Isotopic andGeochemical Research (PCIGR) at the University ofBritish Columbia (UBC) Samples were selected from the59 samples analyzed for whole-rock chemistry at ActLabsbased on major- and trace-element chemistry alteration(low LOI and petrographic alteration index) and samplelocation Samples were prepared for trace-element analysisat the PCIGR by the technique described by Pretorius et al(2006) on unleached rock powders Sample powders(100mg) were weighed in 7ml screw-top Savillex bea-kers and dissolved in 1ml 14N HNO3 and 5ml 48 HFon a hotplate for 48 h at 1308C with periodic ultrasonica-tion Samples were dried and redissolved in 6ml 6N HClon a hotplate for 24 h and then dried and redissolved in1ml concentrated HNO3 for 24 h before final dryingTrace element abundances were measured by high-resolution inductively coupled plasma-mass spectrometry(HR-ICP-MS) with aThermo Finnigan Element2 systemfollowing the procedures described by Pretorius et al(2006) within 24 h of redissolution HFSE and LILE weremeasured in medium-resolution mode at 2000 dilutionusing a PFATeflon spray chamber washed with aqua regiafor 3min between samples REE were measured in high-resolution mode and U and Pb in low-resolution modeat 2000 dilution using a glass spray chamber washedwith 2 HNO3 between samples Total procedural blanksand reference materials (BCR-2 BHVO-2) were analyzedwith the batch of samples Indium was used as an internal
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
38
standard in all samples and standard solutionsBackground and standard solutions were analyzed afterevery five samples to detect memory effects and massdrift Results for PCIGR trace-element analyses areshown in Electronic Appendix 3 for the full set of analyzedtrace elementsSample digestion for purification of Sr Nd Hf and Pb
for column chemistry began by weighing each samplepowder (400^500mg) prior to leaching All samples wereleached with 6N HCl and placed in an ultrasonic bath for15min Samples were rinsed twice with 18 M-cm H2Obetween each leaching step (15 total) until the supernatantliquid was clear [following the method of Mahoney (1987)]Samples were then dried on a hotplate for 24 h andweighed again Sample solutions were then prepared bydissolving 100^250mg of the leached powder dissolvedin 1ml 14N HNO3 and 10ml 48 HF on a hotplate for48 h at 1308C with periodic ultrasonication Samples weredried and redissolved in 6ml 6N HCl on a hotplate for24 h and then dried Pb was separated using anion exhangecolumns and the discard was used for Sr REE and Hfseparation Nd was separated from the other REE and Hfrequired two additional purification steps Detailed proce-dures for column chemistry for separating Sr Nd and Pbat the PCIGR have been described byWeis et al (2006) andHf purification has been described byWeis et al (2007) Srand Nd isotope ratios were measured by thermal ioniza-tion mass spectrometry (TIMS) on a Thermo FinniganTriton system in static mode with relay matrix rotation ona single Ta and double Re^Ta filament respectively Fourto five filaments per barrel of 21 were occupied by stan-dards (NBS 987 for Sr and LaJolla for Nd) for each barrelwhere samples were run Sample Sr and Nd isotopic com-positions were corrected for mass fractionation using86Sr88Srfrac14 01194 and 146Nd144Ndfrac14 07219 Each samplewas then normalized using the barrel average of the refer-ence material relative to the values of143Nd144Ndfrac14 0511858 and 87Sr86Srfrac14 0710248 (Weiset al 2006) During the course of the Vancouver Islandanalyses the La Jolla Nd standard gave an average valueof 05118566 (nfrac14 7) and NBS987 standard gave an aver-age of 07102408 (nfrac1411) (2 error is reported as times106) 147Sm144Nd errors are 15 or 0006 USGeological Survey (USGS) reference material BHVO-2was processed with the samples and yielded Sr and Nd iso-topic ratios of 07034607 and 05129786 respectivelyThese are in agreement with the published values of070347920 and 051298411 respectively (Weis et al2006)
Pb and Hf isotopic compositions were analyzed by mul-tiple collector-inductively coupled plasma-mass spectrome-try (MC-ICP-MS) on a static multi-collection on a NuPlasma (Nu Instruments) system The detailed analyticalprocedure for Pb isotopic analyses on the Nu Plasma atthe PCIGR has been described by Weis et al (2006) Theconfiguration for Pb analyses allows for collection of PbTl and Hg togetherTl and Hg are used to monitor instru-mental mass discrimination and isobaric overlap repec-tively All sample solutions were analyzed withapproximately the same PbTl ratio (4) as the referencematerial NIST SRM 981 To accomplish this a small ali-quot of each sample solution from the Pb columns was ana-lyzed on the Element2 to determine the precise amount ofPb available for analysis on the Nu Plasma The SRM 981standard was run after every two samples on the NuPlasma During the time samples were run analyses ofthe SRM 981 Pb reference material gave values of206Pb204Pbfrac1416940322 207Pb204Pbfrac1415495823and 208Pb204Pbfrac14 36713164 (nfrac14 61 2 error is reportedas times 104) these values are in excellent agreement withreported TIMS triple-spike values of Galer amp Abouchami(1998) Fractionation-corrected Pb isotopic ratios were fur-ther corrected by the sample-standard bracketing methodor the ln^ln correction method described by White et al(2000) and Blichert-Toft et al (2003) Leached powders ofUSGS reference material BHVO-2 yielded Pb isotopicratios of 206Pb204Pbfrac14186454 8 207Pb204Pbfrac14 1549105 and 208Pb204Pbfrac14 38222514 and BCR-2yielded 206Pb204Pbfrac141880466 207Pb204Pbfrac14 1562518 and 208Pb204Pbfrac14 3883496 (2 error isreported as times 104) These values are in agreement withthose for leached residues of BHVO-2 and BCR-2 fromWeis et al (2006)Hf isotopic compositions were analyzed following the
procedures detailed byWeis et al (2007) The configurationfor Hf analyses monitored Lu mass 175 andYb mass 172 toallow for interference correction to masses 174 and 176 Hfisotopic ratios were normalized internally for mass fractio-nation to a 179Hf177Hf ratio of 07325 using an exponentialcorrection Standards were run after every two samplesand sample results were normalized to the ratio of the in-run daily average and a 176Hf177Hf ratio for JMC-475 of0282160 During the course of analyses the Hf standardJMC-475 gave an average value 02821679 (nfrac14130)USGS reference materials BCR-2 and BHVO-2 were pro-cessed with the samples and yielded Hf isotopic ratios of02828675 and 02831005 respectively Publishedvalues for BCR-2 and BHVO-2 are 02828717 and0283104 8 respectively (Weis et al 2007)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
39
lavas from single volcanoes in the Galapagos Hawaii andelsewhere by Herzberg amp Asimow (2008) Those workersinterpreted the range to reflect the transport of magmasfrom both the cool periphery and the hotter axis of aplume A thermally heterogeneous plume will yield pri-mary magmas with different MgO contents and the rangeof primary magma compositions and their inferred mantlepotential temperatures for Karmutsen lavas may reflectmelting from different parts of the plume
Source of the Karmutsen Formation lavasThe Sr^Nd^Hf^Pb isotopic compositions of theKarmutsen volcanic rocks on Vancouver Island indicatethat they formed from plume-type mantle containing adepleted component that is compositionally similar to butdistinct from other oceanic plateaus that formed in thePacific Ocean (eg Ontong Java and Caribbean Fig 13)Trace element and isotopic constraints can be used to testwhether the depleted component of the KarmutsenFormation was intrinsic to the mantle plume and distinctfrom the source of MORB or the result of mixing or invol-vement of ambient depleted MORB mantle with plume-type mantle The Karmutsen tholeiitic basalts have initialeNd and
87Sr86Sr ratios that fall within the range of basaltsfrom the Caribbean Plateau (eg Kerr et al 1996 1997Hauff et al 2000 Kerr 2003) and a range of Hawaiian iso-tope data (eg Frey et al 2005) and lie within and abovethe range for the Ontong Java Plateau (Tejada et al 2004Fig 13a) Karmutsen tholeiitic basalts with the highest ini-tial eNd and lowest initial 87Sr86Sr lie within and justbelow the field for age-corrected northern EPR MORB at230 Ma (eg Niu et al1999 Regelous et al1999) Initial Hfand Nd isotopic compositions for the Karmutsen samplesare noticeably offset to the low side of the ocean islandbasalt (OIB) array (Vervoort et al 1999) and lie belowand slightly overlap the low eHf end of fields for Hawaiiand the Caribbean Plateau (Fig 13c) the Karmutsen sam-ples mostly fall between and slightly overlap a field forage-corrected Pacific MORB at 230 Ma and Ontong Java(Mahoney et al 1992 1994 Nowell et al 1998 Chauvel ampBlichert-Toft 2001) Karmutsen lava Pb isotope ratios over-lap the broad range for the Caribbean Plateau and thelinear trends for the Karmutsen samples intersect thefields for EPR MORB Hawaii and Ontong Java in208Pb^206Pb space but do not intersect these fields in207Pb^206Pb space (Fig 13d and e) The more radiogenic207Pb204Pb for a given 206Pb204Pb of the Karmutsenbasalts requires high UPb ratios at an ancient time when235U was abundant similar to the Caribbean Plateau andenriched in comparison with EPR MORB Hawaii andOntong JavaThe low eHf^low eNd component of the Karmutsen
basalts that places them below the OIB mantle array andpartly within the field for age-corrected Pacific MORBcan form from low ratios of LuHf and high SmNd over
time (eg Salters amp White 1998) MORB will develop aHf^Nd isotopic signature over time that falls below themantle array Low LuHf can also lead to low 176Hf177Hffrom remelting of residues produced by melting in theabsence of garnet (eg Salters amp Hart 1991 Salters ampWhite 1998) The Hf^Nd isotopic compositions of theKarmutsen Formation indicate the possible involvementof depleted MORB mantle however similar to Iceland(Fitton et al 2003) Nb^Zr^Y variations provide a way todistinguish between a depleted component within theplume and a MORB-like component The Karmutsenbasalts and picrites fall within the Iceland array and donot overlap the MORB field in a logarithmic plot of NbYvs ZrY (Fig 13b) using the fields established by Fittonet al (1997) Similar to the depleted component within theCaribbean Plateau (Thompson et al 2003) the trend ofHf^Nd isotopic compositions towards Pacific MORB-likecompositions is not followed by MORB values in Nb^Zr^Y systematics and this suggests that the depleted compo-nent in the source of the Karmutsen basalts is distinctfrom the source of MORB and probably an intrinsic partof the plume The relatively radiogenic Pb isotopic ratiosand REE patterns distinct from MORB also support thisinterpretationTrace-element characteristics suggest the possible invol-
vement of sub-arc lithospheric mantle in the generation ofthe Karmutsen picrites Lassiter et al (1995) suggested thatmixing of a plume-type source with eNdthorn 6 to thorn 7 witharc material with low NbTh could reproduce the varia-tions observed in the Karmutsen basalts however theyconcluded that the absence of low NbLa ratios in theirsample of tholeiitic basalts restricts the amount of arc litho-sphere involved The study of Lassiter et al (1995) wasbased on major and trace element and Sr Nd and Pb iso-topic data for a suite of 29 samples from Buttle Lake in theStrathcona Provincial Park on central Vancouver Island(Fig 1) Whereas there are no clear HFSE depletions inthe Karmutsen tholeiitic basalts the low NbLa (and TaLa) of the Karmutsen picritic lavas in this study indicatethe possible involvement of Paleozoic arc lithosphere onVancouver Island (Fig 8)
REE modeling dynamic melting andsource mineralogyThe combined isotopic and trace-element variations of thepicritic and tholeiitic Karmutsen lavas indicate that differ-ences in trace elements may have originated from differentmelting histories For example the LuHf ratios of theKarmutsen picritic lavas (mean 008) are considerablyhigher than those in the tholeiitic basalts (mean 002)however the small range of overlapping initial eHf valuesfor the two lava suites suggests that these differences devel-oped during melting within the plume (Fig 9b) Below wemodel the effects of source mineralogy and depth of
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
28
melting on differences in trace-element concentrations inthe tholeiitic basalts and picritesDynamic melting models simulate progressive decom-
pression melting where some of the melt fraction isretained in the residue and only when the degree of partialmelting is greater than the critical mass porosity and the
source becomes permeable is excess melt extracted fromthe residue (Zou1998) For the Karmutsen lavas evolutionof trace element concentrations was simulated using theincongruent dynamic melting model developed by Zou ampReid (2001) Melting of the mantle at pressures greater than1 GPa involves incongruent melting (eg Longhi 2002)
OIB array
PacificMORB(230 Ma)
(0 Ma)
EPR(230 Ma)
-4
-2
0
2
4
8
10
12
14
16
18
20
22
-4 -2 0 2 4 6 8 10 12 14
minus2
0
2
4
6
8
10
12
14
0702 0703 0704 0705 0706
IndianMORB
87Sr 86Sr (initial)
Hf (initial)
(a) (c)
Nd (initial)
Karmutsen tholeiitic basalt
HawaiiOntong JavaCaribbean Plateau
Karmutsen high-MgO lava
high-MgO lavasKarmutsen
1540
1545
1550
1555
1560
1565
175 180 185 190 195 200375
380
385
390
395
175 180 185 190 195 200
(d) (e)
EPR
EPR
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb
Karmutsen Formation (initial)
HawaiiOntong JavaCaribbean Plateau
East Pacific Rise
Karmutsen Formation (measured)
Juan de FucaGorda
ε Nd
(initial)
Karmutsen Formation (different age-correction for 3 picrites)
(0 Ma)
NbY
001
01
1
1 10
ZrY
OIB
NMORB
(b)
Iceland array
6
Fig 13 Comparison of age-corrected (230 Ma) Sr^Nd^Hf^Pb isotopic compositions for Karmutsen flood basalts onVancouver Island withage-corrected OIB and MORB and Nb^Zr^Y variation (a) Initial eNd vs 87Sr86Sr (b) NbYand ZrY variation (after Fitton et al 1997)(c) Initial eHf vs eNd (d) Measured and initial 207Pb204Pb vs 206Pb204Pb (e) Measured and initial 208Pb204Pb vs 206Pb204Pb References fordata sources are listed in Electronic Appendix 4 Most of the compiled data were extracted from the GEOROC database (httpgeorocmpch-mainzgwdgdegeoroc) OIB array line in (c) is fromVervoort et al (1999) EPR East Pacific Rise Several high-MgO samples affected bysecondary alteration were not plotted for clarity in Pb isotope plots Estimation of fields for age-corrected Pacific MORB and EPR at 230 Mawere made using parent^daughter concentrations (in ppm) from Salters amp Stracke (2004) of Lufrac14 0063 Hffrac14 0202 Smfrac14 027 Ndfrac14 071Rbfrac14 0086 and Srfrac14 836 In the NbYvs ZrYplot the normal (N)-MORB and OIB fields not including Icelandic OIB are from data com-pilations of Fitton et al (1997 2003) The OIB field is data from 780 basalts (MgO45wt ) from major ocean islands given by Fitton et al(2003) The N-MORB field is based on data from the Reykjanes Ridge EPR and Southwest Indian Ridge from Fitton et al (1997) The twoparallel gray lines in (b) (Iceland array) are the limits of data for Icelandic rocks
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
29
with olivine being produced during the reactioncpxthornopxthorn spfrac14meltthornol for spinel peridotites The mod-eling used coefficients for incongruent melting reactionsbased on experiments on lherzolite melting (eg Kinzleramp Grove 1992 Longhi 2002) and was separated into (1)melting of spinel lherzolite (2) two combined intervals ofmelting for a garnet lherzolite (gt lherz) and spinel lher-zolite (sp lherz) source (3) melting of garnet lherzoliteThe model solutions are non-unique and a range indegree of melting partition coefficients melt reactioncoefficients and proportion of mixed source componentsis possible to achieve acceptable solutions (see Fig 14 cap-tion for description of modeling parameters anduncertainties)The LREE-depleted high-MgO lavas require melting of
a depleted spinel lherzolite source (Fig 14a) The modelingresults indicate a high degree of melting (22^25) similarto the results from PRIMELT1 based on major-elementvariations (discussed above) of a LREE-depleted sourceThe high-MgO lavas lack a residual garnet signature Theenriched tholeiitic basalts involved melting of both garnetand spinel lherzolite and represent aggregate melts pro-duced from continuous melting throughout the depth ofthe melting column (Fig 14b) Trace-element concentra-tions in the tholeiitic and picritic lavas are not consistentwith low-degree melting of garnet lherzolite alone(Fig 14c) The modeling results indicate that the aggregatemelts that formed the tholeiitic basalts would haveinvolved lesser proportions of enriched small-degreemelts (1^6 melting) generated at depth and greateramounts of depleted high-degree melts (12^25) gener-ated at lower pressure (a ratio of garnet lherzolite tospinel lherzolite melt of 14 provides the best fits seeFig 14b and description in figure caption) The totaldegree of melting for the tholeiitic basalts was high(23^27) similar to the degree of partial melting in thespinel lherzolite facies for the primary melts that eruptedas the Keogh Lake picrites of a source that was also prob-ably depleted in LREE
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
(c)
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
Melting of garnet lherzolite
High-MgO lavas
Melting of spinel lherzolite
source (07DM+03PM)
source (07DM+03PM)
6 melting
30 melting
(b)
(a)
30 melting
Sam
ple
Cho
ndrit
eS
ampl
eC
hond
rite 20 melting
1 melting
Melting of garnet lherzoliteand spinel lherzolite
x gt lherz + 4x sp lherz
5 meltingx gt lherz + 4x sp lherz
Sam
ple
Cho
ndrit
e
Tholeiitic lavas
Tholeiitic lavas
High-MgO lavas
source (07DM+03PM)
Fig 14 Trace-element modeling results for incongruent dynamicmantle melting for picritic and tholeiitic lavas from the KarmutsenFormation Three steps of modeling are shown (a) melting of spinellherzolite compared with high-MgO lavas (b) melting of garnet lher-zolite and spinel lherzolite compared with tholeiitic lavas (c) meltingof garnet lherzolite compared with tholeiitic and picritic lavasShaded fields in each panel for tholeiitic basalts and picrites arelabeled Patterns with symbols in each panel are modeling results(patterns are 5 melting increments in (a) and (b) and 1 incre-ments in (c) PM primitive mantle DM depleted MORB mantle gtlherz garnet lherzolite sp lherz spinel lherzolite In (b) the ratios ofper cent melting for garnet and spinel lherzolite are indicated (x gtlherzthorn 4x sp lherz means that for 5 melting 1 melt in garnetfacies and 4 melt in spinel facies where x is per cent melting in theinterval of modeling) The ratio of the respective melt proportions inthe aggregate melt (x gt lherzthorn 4x sp lherz) was kept constant andthis ratio of melt contribution provided the best fit to the data Arange of proportion of source components (07DMthorn 03PM to
09DMthorn 01PM) can reproduce the variation in the high-MgOlavas Melting modeling uses the formulation of Zou amp Reid (2001)an example calculation is shown in their Appendix PM fromMcDonough amp Sun (1995) and DM from Salters amp Stracke (2004)Melt reaction coefficients of spinel lherzolite from Kinzler amp Grove(1992) and garnet lherzolite fromWalter (1998) Partition coefficientsfrom Salters amp Stracke (2004) and Shaw (2000) were kept constantSource mineralogy for spinel lherzolite (018cpx027opx052ol003sp) from Kinzler (1997) and for garnet lherzolite(034cpx008opx053ol005gt) from Salters amp Stracke (2004) Arange of source compositions for melting of spinel lherzolite(07DMthorn 03PM to 03DMthorn 07PM) reproduce REE patternssimilar to those of the high-MgO lavas with best fits for 22ndash25melting and 07DMthorn 03PM source composition Concentrationsof tholeiitic basalts cannot be reproduced from an entirelyprimitive or depleted source
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
30
The uncertainty in the composition of the source in thespinel lherzolite melting model for the high-MgO lavasprecludes determining whether the source was moredepleted in incompatible trace elements than the source ofthe tholeiitic lavas (see Fig 14 caption for description) It ispossible that early high-pressure low-degree melting left aresidue within parts of the plume (possibly the hotter inte-rior portion) that was depleted in trace elements (egElliott et al 1991) further decompression and high-degreemelting of such depleted regions could then have generatedthe depleted high-MgO Karmutsen lavas Shallow high-degree melting would preferentially sample depletedregions whereas deeper low-degree melting may notsample more refractory depleted regions (eg CaribbeanEscuder-Viruete et al 2007) The picrites lie at the highend of the range of Nd isotopic compositions and havelow NbZr similar to depleted Icelandic basalts (Fittonet al 2003) therefore the picrites may represent the partsof the mantle plume that were the most depleted prior tothe initiation of melting As Fitton et al (2003) suggestedthese depleted melts may only rarely be erupted withoutcombining with more voluminous less depleted melts andthey may be detected only as undifferentiated small-volume flows like the Keogh Lake picrites within theKarmutsen Formation on Vancouver IslandDecompression melting within the mantle plume initiatedwithin the garnet stability field (35^25 GPa) at highmantle potential temperatures (Tp414508C) and pro-ceeded beneath oceanic arc lithosphere within the spinelstability field (25^09 GPa) where more extensivedegrees of melting could occur
Magmatic evolution of Karmutsentholeiitic basaltsPrimary magmas in large igneous provinces leave exten-sive coarse-grained cumulate residues within or below thecrust these mafic and ultramafic plutonic sequences repre-sent a significant proportion of the original magma thatpartially crystallized at depth (eg Farnetani et al 1996)For example seismic and petrological studies of OntongJava (eg Farnetani et al 1996) combined with the use ofMELTS (Ghiorso amp Sack 1995) indicate that the volcanicsequence may represent only 40 of the total magmareaching the Moho Neal et al (1997) estimated that30^45 fractional crystallization took place for theOntong Java magmas and produced cumulate thicknessesof 7^14 km corresponding to 11^22 km of flood basaltsdepending on the presence of pre-existing oceanic crustand the total crustal thickness (25^35 km) Interpretationsof seismic velocity measurements suggest the presence ofpyroxene and gabbroic cumulates 9^16 km thick beneaththe Ontong Java Plateau (eg Hussong et al 1979)Wrangellia is characterized by high crustal velocities in
seismic refraction lines which is indicative of mafic
plutonic rocks extending to depth beneath VancouverIsland (Clowes et al 1995)Wrangellia crust is 25^30 kmthick beneath Vancouver Island and is underlain by astrongly reflective zone of high velocity and density thathas been interpreted as a major shear zone where lowerWrangellia lithosphere was detached (Clowes et al 1995)The vast majority of the Karmutsen flows are evolved tho-leiitic lavas (eg low MgO high FeOMgO) indicating animportant role for partial crystallization of melts at lowpressure and a significant portion of the crust beneathVancouver Island has seismic properties that are consistentwith crystalline residues from partially crystallizedKarmutsen magmas To test the proportion of the primarymagma that fractionated within the crust MELTS(Ghiorso amp Sack 1995) was used to simulate fractionalcrystallization using several estimated primary magmacompositions from the PRIMELT1 modeling results Themajor-element composition of the primary magmas couldnot be estimated for the tholeiitic basalts because of exten-sive plagthorn cpxthornol fractionation and thus the MELTSmodeling of the picrites is used as a proxy for the evolutionof major elements in the volcanic sequence A pressure of 1kbar was used with variable water contents (anhydrousand 02wt H2O) calculated at the quartz^fayalite^magnetite (QFM) oxygen buffer Experimental resultsfrom Ontong Java indicate that most crystallizationoccurs in magma chambers shallower than 6 km deep(52 kbar Sano amp Yamashita 2004) however some crys-tallization may take place at greater depths (3^5 kbarFarnetani et al 1996)A selection of the MELTS results are shown in Fig 15
Olivine and spinel crystallize at high temperature until15^20wt of the liquid mass has fractionated and theresidual magma contains 9^10wt MgO (Fig 15f) At1235^12258C plagioclase begins crystallizing and between1235 and 11908C olivine ceases crystallizing and clinopyr-oxene saturates (Fig 15f) As expected the addition of clin-opyroxene and plagioclase to the crystallization sequencecauses substantial changes in the composition of the resid-ual liquid FeO and TiO2 increase and Al2O3 and CaOdecrease while the decrease in MgO lessens considerably(Fig 15) The near-vertical trends for the Karmutsen tho-leiitic basalts especially for CaO do not match the calcu-lated liquid-line-of-descent very well which misses manyof the high-CaO high-Al2O3 low-FeO basalts A positivecorrelation between Al2O3CaO and SrNd indicates thataccumulation of plagioclase played a major role in the evo-lution of the tholeiitic basalts (Fig 15g) Increasing thecrystallization pressure slightly and the water content ofthe melts does not systematically improve the match ofthe MELTS models to the observed dataThe MELTS results indicate that a significant propor-
tion of the crystallization takes place before plagioclase(15^20 crystallization) and clinopyroxene (35^45
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
31
crystallization) join the crystallization sequence Theseresults suggest that445 crystallization of the primarymagmas occurred (estimated from the residual liquid percent in Fig 15f) and if the original flood basalt stratigra-phy on Vancouver Island was 6 km thick at least anequivalent amount of complementary plutonic rocksshould have crystallized The Karmutsen basalts and theirplutonic residues thus represent a significant addition ofcrust (perhaps 412 km thickness) previously thickenedduring the formation of the underlying Paleozoic arc
Growth of the Wrangellia oceanic plateauon Vancouver IslandThe extrusive part of the Wrangellia oceanic plateau onVancouver Island was constructed in a three-layersequence of mostly tholeiitic basalts with a restrictedrange of composition (Fig 16) The plateau overliesDevonian to Mississippian arc volcanic sequences andmarine sediments of Mississippian to Permian ages thatare intruded by gabbroic rocks related to the flood basaltsThe volcanic stratigraphy of theWrangellia plateau formed
00
05
10
15
20
25
30
35
40
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
16
0 2 4 6 8 10 12 14 16 18 20
10
11
12
13
14
15
16
17
18
19
20
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
0 2 4 6 8 10 12 14 16 18 20
MgO (wt)
0 10 20 30 40 50
percent of total mass
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
) Mineral proportions
Sp (e)
Ol
CpxPlag
PicriteHigh-MgO basaltTholeiitic basalt
TiO2 (wt)
MgO (wt)
Al2O3 (wt) CaO (wt)
FeO (wt)
MgO(a) (b)
(c) (d)
MgO (wt)MgO (wt)
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
)
Residual liquid ()
1220
1190
Ol + Sp
Plag+Ol+Sp Plag+Cpx+Sp
02 wt H2O
no H2O
Residual liquid and crystallizing phases
(f)
0102030405060708090100
plagioclaseaccumulation
plagioclaseaccumulation
plagioclase
accumulatio
n
10
12
14
16
0 10 20 30 40
Al2O3CaO
SrNd
tholeiitic basalts
(g)
Fig 15 Forward fractional crystallization modeling results for major elements from MELTS (Ghiorso amp Sack 1995) compared with picriticand tholeiitic lavas from the Karmutsen Formation (a^d) Major-element compositions of residual liquids compared with Karmutsen lavacompositions For clarity the results are shown only for one starting composition (open circle) using the composition of the estimated primarymagma for the high-MgO lava series from the modeling technique of Herzberg et al (2007) for sample 4723A13 as shown in Fig 12 The esti-mated primary magmas for samples 4723A3 and 93G171 yield similar results A pressure of 1 kbar was used with no H2O and 02wt H2Ocalculated at QFM black line is anhydrous and grey line is 02wt H2OThe end-point for the anhydrous result is at 78 fractionation and11308C and the end-point for the result with 02wt H2O is 75 fractionation and11108C (e) Mineral proportions of crystallized phases fromone MELTrun (sample 4723A13) with no H2O (f) Per cent residual liquid vs temperature labelled with intervals of crystallizing phases (g)Al2O3CaO vs SrNd for the tholeiitic basalts
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
32
in a single eruptive phase (c 231^225 Ma) as an emergentbasalt sequence during a deeper-water shallow-water andsubaerial stage and subsided below sea level relativelysoon after its formation (Fig 16) The deeper-water stagecontains pillowed and massive submarine flows (3000mthick) that built up the submarine volcanic edifice from theseafloor The lowest pillow basalts were emplaced onunconsolidated and lithified fine-grained siliciclastic andcarbonaceous sediments containing Daonella beds(c 235^232 Ma) that are intruded by abundantKarmutsen mafic sills (Fig 16a) The different pillowedand massive submarine flows resulted primarily from dif-ferent effusive rates and local topography Some of themassive submarine flows commonly recognized by radiat-ing columnar jointing represent master tubes for the deliv-ery of lava to distal parts of submarine flow fields Thepillowed flows which commonly have uneven topography
in modern sequences may have served as levees to themassive submarine flows rather than allowing unob-structed spreading and inflation typical of the massive sub-aerial flows that form on shallow slopes in continentalflood basalt provinces (eg Self et al 1997) The rapid suc-cession of flows is evident by the near absence of sedimentbetween submarine flows Minor volumes of picritic pillowlavas occur in the upper part of the pillowed flow sequencein areas of the plateau exposed on northern VancouverIsland (Fig 16 Nixon et al 2008)The growth of the plateau in shallow water led to
emplacement of an increasing proportion of volcaniclastics(pillow breccias and hyaloclastites) that accumulated up to1500m thick in some areas (Fig 16b Nixon et al 2008)Horizons of pillowed flows occur locally within the volca-niclastic stratigraphy Sedimentary structures (eg gradedbedding fluidization structures) and petrographic textures
Late Devonian-Mississippian arc
Daonella beds
Middle Triassic
deeper water
Late Devonian-Mississippian arc
pillowed flows massive submarine
flows
pillow brecciahyaloclastiteshallow water
~4 km
~6 km
Late Triassic limestoneinterflow limestonelenses
Karmutsen Formation
mafic intrusions sediment-sill complex
Karmutsen Formation
(a) (b)
(c) (d)
subaerial flows
Late Devonian-Mississippian arc Late Devonian-Mississippian arc
picritic pillow basalts
Miss to Permian marine sediments
sills
Fig 16 Growth of the Wrangellia oceanic plateau onVancouver Island (a) Deep-water eruptive phase of pillowed and massive submarineflows with almost no intervening sediments onto MiddleTriassic shale containing Daonella beds (c 235^232 Ma) The basement of the plateauis composed of Devonian to Mississippian arc sequences overlain by Mississippian to Permian carbonates and siliciclastic sediments(b) Shallow-water flow emplacement with increasing proportion of volcaniclastics resulted primarily from cooling^contraction granulationPicritic pillow basalts occur primarily near the transition between pillowed submarine flows and volcaniclastic units Pillowed and massiveflows occur locally within the volcaniclastic unit (c) Emergent stage of growth consists of overlapping subaerial flow fields similar to continen-tal flood basalts with pahoehoe flow structures Multi-tiered columnar jointing occurrs locally in response to ponding of flows (eg Bondre et al2004) Interflow carbonate deposits and plagioclase megacrystic flows developed locally during the waning stage of volcanism (d) Post-volcanicsubsidence and deposition of Late Carnian to Norian (c 212^223 Ma) platform carbonates on top of the plateau Units are labeled in the dia-grams Thicknesses of the underlying units are not to scale
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
33
within the volcaniclastic rocks indicate formation primar-ily via cooling^contraction granulation magma^water^steam interaction autobrecciation resedimentation andmass-wasting rather than pyroclastic eruption (Nixonet al 2008)A broad subaerial platform was constructed by emplace-
ment of subaerial lava flows during an emergent subaerialstage (Fig 16c) The subaerial unit predominantly com-prises well-layered sheet flows that formed in a similarway to continental flood basalts as hundred(s) of pahoehoeflow fields erupted on shallow slopes (Self et al 1997) how-ever volcaniclastic and pillowed flow units are locally pre-sent mostly in the upper parts of the stratigraphyLocalized ponding of flows led to the formation of multi-tiered columnar jointing in some subaerial flows (egBondre et al 2004) As volcanism waned local interflowcarbonate deposits developed and plagioclase megacrysticflows were erupted Subsidence of the plateau is recordedby the interflow lenses and deposits of platform carbonates(40^1500m thick) on top of the Karmutsen basalts thatcontain Late Carnian to Early Norian fossils (Fig 16d c216^227 Ma Nixon et al 2008)The evolution of the internal architecture of the
Wrangellia oceanic plateau was primarily related to theconditions of emplacement (ie water depth and eruptionenvironment) There does not appear to have been a sys-tematic coincident evolution or shift of lava compositionsonVancouver Island as has been recognized in most con-tinental flood basalts and parts of the Wrangellia oceanicplateau in Alaska and Yukon (Greene et al 2008 2009)In general with the exception of the picritic pillow lavasin the submarine stratigraphy and local variations betweenhigh-MgO and tholeiitic lavas there are no significantcompositional differences between lavas erupted duringthe different stages of growth of the plateau (ie submarineand subaerial flows) nor are substantial compositional dif-ferences generally found between massive submarine flowsin contact with pillowed flows tholeiitic basalt flows nearthe base of the Karmutsen Formation are compositionallysimilar to flows sampled near the top of the volcanicsequence Where it was possible to sample continuous sec-tions of volcanic stratigraphy greater than several hun-dreds of meters (eg Keogh^Maynard Lake area west ofthe Karmutsen Range Fig 2) there is no clear relationshipbetween stratigraphic position and compositionalvariation
CONCLUSIONSThe Karmutsen Formation covers extensive areas ofVancouver Island (20 000 km2) and forms part of a largeoceanic plateau erupted over a geologically short interval(c 225^231 Ma) The tripartite volcanic stratigraphyexposed on Vancouver Island is at least 6 km thick andcomprises a succession of submarine flows volcaniclastic
deposits and massive subaerial flowsThese volcanologicaldifferences are primarily related to the eruption environ-ment (deep-water shallow-water or subaerial) The rapidgrowth of the plateau prevented significant accumulationsof intervening sediment except in the uppermost stratigra-phy where isolated limestone lenses commonly associatedwith pillow basalt and hyaloclastite deposits preserve arecord of the final stages of subsidence of the plateauThe Wrangellia plateau on Vancouver Island was con-
structed dominantly of tholeiitic basalt with restrictedmajor- and trace-element and isotopic compositional var-iations Minor volumes of picritic pillow basalts eruptedlate in the submarine phase The Karmutsen basalts wereemplaced onto older Paleozoic arc volcanic sequences andlow NbLa for the picrites may indicate a contributionfrom the sub-arc lithospheric mantle Modeling of the pet-rogenesis of the Karmutsen picrites indicates melting ofanomalously hot mantle (15008C) and extensive degreesof partial melting (23^27) and is consistent with aplume initiation model Combined Sr^Nd^Hf^Pb isotopicsystematics indicate that the picritic and tholeiitic lavaswere derived from a depleted plume-type mantle sourcedistinct from the source of MORB that has compositionalsimilarities to the source of the lavas forming theCaribbean Plateau Picrites were generated by melting ofspinel lherzolite whereas the tholeiitic basalts formed asaggregate melts of both garnet and spinel lherzolite Thetholeiitic basalts underwent extensive low-pressure fractio-nation (52^3 kbar) and seismic studies indicate that someof theWrangellia crust beneathVancouver Island may cor-respond to the plutonic residues of the Karmutsen floodbasalts
ACKNOWLEDGEMENTSWe are grateful to Nick Arndt for helping us get this pro-ject started and Nick Massey for insights into VancouverIsland geology We greatly appreciate the assistance withPRIMELT modeling from Claude Herzberg We thankMikkel Schau for his insight and enthusiasm during field-work Jane Barling assisted with analyses by MC-ICP-MSThoughtful and thorough reviews of the manuscript byJohn Mahoney Fred Frey Wendy Bohrson DougalJerram and Teal Riley are greatly appreciated Fundingwas generously provided by the Rocks to Riches Programadministered by the BC amp Yukon Chamber of Mines in2004 by the BC Geological Survey and by NSERCDiscovery Grants to JSS and DW ARG was supportedby a University Graduate Fellowship at UBC
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
34
REFERENCESBarker F Brown A S Budahn J R amp Plafker G (1989) Back-arc
with frontal-arc component origin of Triassic Karmutsen basaltBritish Columbia Canada Chemical Geology 75 81^102
Blichert-Toft JWeis D Maerschalk C Agranier A amp Albare de F(2003) Hawaiian hot spot dynamics as inferred from the Hf and Pbisotope evolution of Mauna Kea volcano Geochemistry GeophysicsGeosystems 4(2) 1^27 doi1010292002GC000340
Bondre N R Duraiswami R A amp Dole G (2004) Morphologyand emplacement of flows from the Deccan Volcanic ProvinceIndia Bulletin of Volcanology 66 29^45
Carlisle D (1963) Pillow breccias and their aquagene tuffs QuadraIsland British Columbia Journal of Geology 71 48^71
Carlisle D (1972) Late Paleozoic to Mid-Triassic sedimentary-volcanic sequence on Northeastern Vancouver Island Geological
Society of Canada Paper Report of Activities 72-1 Part B 24^30Carlisle D amp SuzukiT (1974) Emergent basalt and submergent car-
bonate^clastic sequences including the Upper Triassic Dilleri andWelleri zones on Vancouver Island Canadian Journal of Earth
Sciences 11 254^279Chauvel C amp Blichert-Toft J (2001) A hafnium isotope and trace
element perspective on melting of the depleted mantle Earth and
Planetary Science Letters 190 137^151Chayes F (1954)The theory of thin-section analysis Journal of Geology
62 92^101Cho M Liou J G amp Maruyama S (1987) Transition from the zeo-
lite to prehnite^pumpellyite facies in the Karmutsen metabasitesVancouver Island British Columbia Journal of Petrology 27(2)467^494
Clowes R M Zelt C A Amor J R amp Ellis R M (1995)Lithospheric structure in the southern Canadian Cordillera froma network of seismic refraction lines Canadian Journal of Earth
Sciences 32(10) 1485^1513Coffin M F amp Eldholm O (1994) Large igneous provinces Crustal
structure dimensions and external consequences Reviews of
Geophysics 32(1) 1^36Cox K G (1980) A model for flood basalt volcanism Journal of
Petrology 21 629^650Eldholm O amp Coffin M F (2000) Large igneous provinces
and plate tectonics In Richards M A Gordon R G amp vander Hilst R D (eds) The History and Dynamics of Global
Plate Motions American Geophysical Union Geophysical Monograph 121309^326
Elliott T R Hawkesworth C J amp Grolaquo nvold K (1991) Dynamicmelting of the Iceland plume Nature 351 201^206
Escuder-Viruete J Pecurren rez-Estaucurren n A Contreras F Joubert MWeis D Ullrich T D amp Spadea P (2007) Plume mantle sourceheterogeneity through time Insights from the Duarte ComplexHispaniola northeastern Caribbean Journal of Geophysical Research112(B04203) doi1010292006JB004323
Farnetani C G Richards M A amp Ghiorso M S (1996)Petrological models of magma evolution and deep crustal structurebeneath hotspots and flood basalt provinces Earth and Planetary
Science Letters 143 81^94Fitton J G amp Godard M (2004) Origin and evolution of magmas
on the Ontong Java Plateau In Fitton J G Mahoney J JWallace P J amp Saunders A D (eds) Origin and Evolution of the
Ontong Java Plateau Geological Society London Special Publications 229151^178
Fitton J G Saunders A D Norry M J Hardarson B S ampTaylor R N (1997) Thermal and chemical structure of theIceland plume Earth and Planetary Science Letters 153 197^208
Fitton J G Saunders A D Kempton P D amp Hardarson B S(2003) Does depleted mantle form an intrinsic part of the Icelandplume Geochemistry Geophysics Geosystems 4(3) 1032 doi1010292002GC000424
Frey F A Huang S Blichert-Toft J Regelous M amp Boyet M(2005) Origin of depleted components in basalt related to theHawaiian hotspot Evidence from isotopic and incompatible ele-ment ratios Geochemistry Geophysics Geosystems 6(1) 23 doi1010292004GC000757
Galer S J G amp Abouchami W (1998) Practical application of leadtriple spiking for correction of instrumental mass discriminationMineralogical Magazine 62A 491^492
Ghiorso M S amp Sack R O (1995) Chemical mass transfer in mag-matic processes IV A revised and internally consistent thermody-namic model for the interpolation and extrapolation of liquid^solidequilibria in magmatic systems at elevated temperatures and pres-sures Contributions to Mineralogy and Petrology 119 197^212
Greene A R Scoates J S amp Weis D (2005)WrangelliaTerrane onVancouver Island Distribution of flood basalts with implicationsfor potential Ni^Cu^PGE mineralization In Grant B (ed)Geological Fieldwork 2004 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2005-1 209^220Greene A R Scoates J S Nixon G T amp Weis D (2006) Picritic
lavas and basal sills in the Karmutsen flood basalt provinceWrangellia northern Vancouver Island In Grant B (ed)Geological Fieldwork 2005 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2006-1 39^52Greene A R Scoates J S amp Weis D (2008) Wrangellia flood
basalts in Alaska A record of plume^lithosphere interaction in aLate Triassic accreted oceanic plateau Geochemistry Geophysics
Geosystems 9(Q12004) doi1010292008GC002092Greene A R Scoates J S Weis D amp Israel S (2009)
Geochemistry of triassic flood basalts from the Yukon (Canada)segment of the accreted Wrangellia oceanic plateau Lithosdoi101016jlithos200811010
Hauff F Hoernle K Tilton G Graham D W amp Kerr A C(2000) Large volume recycling of oceanic lithosphere over shorttime scales geochemical constraints from the Caribbean LargeIgneous Province Earth and Planetary Science Letters 174 247^263
Hauff F Hoernle K amp Schmidt A (2003) Sr^Nd^Pb compositionof Mesozoic Pacific oceanic crust (Site 1149 and 801 ODP Leg185)Implications for alteration of ocean crust and the input into theIzu^Bonin^Mariana subduction system Geochemistry Geophysics
Geosystems 4(8) doi1010292002GC000421Herzberg C (2004) Partial melting below the Ontong Java Plateau
In Fitton J G Mahoney J J Wallace P J amp Saunders A D(eds) Origin and Evolution of the Ontong Java Plateau Geological SocietyLondon Special Publications 229 179^183
Herzberg C (2006) Petrology and thermal structure of the Hawaiianplume from Mauna Kea volcano Nature 444(7119) 605
Herzberg C amp Asimow P D (2008) Petrology of some oceanicisland basalts PRIMELT2XLS software for primary magma cal-culation Geochemistry Geophysics Geosystems 9(Q09001) doi1010292008GC002057
Herzberg C amp OrsquoHara M J (2002) Plume-associated ultramaficmagmas of Phanerozoic age Journal of Petrology 43(10) 1857^1883
Herzberg C Asimow P D Arndt N Niu Y Lesher C MFitton J G Cheadle M J amp Saunders A D (2007)Temperatures in ambient mantle and plumes Constraints frombasalts picrites and komatiites Geochemistry Geophysics Geosystems8(Q02006) doi1010292006GC001390
Huang S amp Frey F A (2005) Trace element abundances of MaunaKea basalt from phase 2 of the Hawaii Scientific Drilling Project
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
35
Petrogenetic implications of correlations with major element con-tent and isotopic ratios Geochemistry Geophysics Geosystems 4(6)doi1010292002GC000322
Hussong D M Wipperman L K amp Kroenke L W (1979) Thecrustal structure of the Ontong Java and Manihiki OceanicPlateaus Journal of Geophysical Research 84(B11) 6003^6010
Jerram D A amp Widdowson M (2005) The anatomy of ContinentalFlood Basalt Provinces geological constraints on the processes andproducts of flood volcanism Lithos 79(3^4) 385^405
Jones D L Silberling N J amp Hillhouse J (1977)Wrangellia a dis-placed terrane in northwestern North America CanadianJournal ofEarth Sciences 14(11) 2565^2577
Kerr A C (2003) Oceanic Plateaus In Rudnick R (ed) The Crust
Treatise on GeochemistryVol 3 Oxford Elsevier Science pp 537^565Kerr A C amp Mahoney J J (2007) Oceanic plateaus Problematic
plumes potential paradigms Chemical Geology 241 332^353Kerr A CTarney J Marriner G F Klaver GT Saunders A D
amp Thirlwall M F (1996) The geochemistry and petrogenesis ofthe Late-Cretaceous picrites and basalts of Curacao NetherlandsAntilles a remnant of an oceanic plateau Contributions to
Mineralogy and Petrology 124(1) 29^43Kerr A C Marriner G F Tarney J Nivia A Saunders A D
Thirlwall M F amp Sinton A C (1997) Cretaceous basaltic ter-ranes in Western Colombia Elemental chronological and Sr-Ndisotopic constraints on petrogenesis Journal of Petrology 38(6)677^702
Kinzler R J (1997) Melting of mantle peridotite at pressuresapproaching the spinel to garnet transition Application to mid-ocean ridge basalt petrogenesis Journal of Geophysical Research
102(B1) 853^874Kinzler R J amp Grove T L (1992) Primary magmas of mid-ocean
ridge basalts 1 Experiments and methods Journal of Geophysical
Research 97(B5) 6885^6906Korenaga J (2005) Why did not the Ontong Java Plateau form sub-
aerially Earth and Planetary Science Letters 234 385^399Kuniyoshi S (1972) Petrology of the Karmutsen (Volcanic) Group
northeastern Vancouver Island British Columbia PhD disserta-tion University of California Los Angeles 242 pp
Lassiter J C DePaolo D J amp Mahoney J J (1995) Geochemistry ofthe Wrangellia flood basalt province Implications for the role ofcontinental and oceanic lithosphere in flood basalt genesis Journalof Petrology 36(4) 983^1009
LeBas M J (2000) IUGS reclassification of the high-Mg and picriticvolcanic rocks Journal of Petrology 41(10) 1467^1470
Longhi J (2002) Some phase equilibrium systematics of lherzolitemelting I Geochemistry Geophysics Geosystems 3(3) doi1010292001GC000204
MacDonald G A amp Katsura T (1964) Chemical composition ofHawaiian lavas Journal of Petrology 5 82^133
Mahoney J J (1987) An isotopic survey of Pacific oceanic plateausimplications for their nature and origin In Keating B HFryer P Batiza R amp Boehlert GW (eds) Seamounts Islands andAtolls American Geophysical Union Geophysical Monograph 43 207^220
Mahoney J LeRoex A P Peng Z Fisher R L amp Natland J H(1992) Southwestern limits of Indian Ocean Ridge mantle and theorigin of low 206Pb204Pb mid-ocean ridge basalt isotope systema-tics of the central Southwest Indian Ridge (178^508E) Journal ofGeophysical Research 97(B13) 19771^19790
Mahoney J J Sinton J M Kurz M D MacDougall J DSpencer K J amp Lugmair GW (1994) Isotope and trace elementcharacteristics of a super-fast spreading ridge East Pacific rise13^238S Earth and Planetary Science Letters 121 173^193
Massey N W D (1995a) Geology and mineral resources of theAlberni^Nanaimo Lakes sheet Vancouver Island 92F1W 92F2Eand part of 92F7E British Columbia Ministry of Energy Mines and
Petroleum Resources Paper 1992-2 132 ppMassey N W D (1995b) Geology and mineral resources of the
Cowichan Lake sheet Vancouver Island 92C16 British Columbia
Ministry of Energy Mines and Petroleum Resources Paper 1992-3 112 ppMassey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005a) Digital Geology Map of British Columbia Tile NM9 MidCoast BC British Columbia Ministry of Energy and Mines Geofile
2005-2Massey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005b) Digital Geology Map of British Columbia Tile NM10Southwest BC British Columbia Ministry of Energy and Mines Geofile
2005-3McDonough W F amp Sun S (1995) The composition of the Earth
Chemical Geology 120 223^253Monger JW H amp Journeay J M (1994) Basement geology and tec-
tonic evolution of theVancouver region In Monger JW H (ed)Geology and Geological Hazards of the Vancouver Region Southwestern
British Columbia Geological Survey of Canada Bulletin 481 3^25Muller J E (1977) Geology of Vancouver Island Geological Survey of
Canada Open File Map 463Muller J E (1980)The Paleozoic Sicker Group of Vancouver Island British
Columbia Geological Survey of Canada Paper 79-30 22 ppMuller J E Northcote K E amp Carlisle D (1974) Geology and
mineral deposits of Alert Bay^Cape Scott map area VancouverIsland British Columbia Geological Survey of Canada Paper 74-8 77
Neal C R Mahoney J J Kroenke L W Duncan R A ampPetterson M G (1997) The Ontong^Java Plateau InMahoney J J amp Coffin M F (eds) Large Igneous Provinces
Continental Oceanic and Planetary FloodVolcanism American Geophysical
Union Geophysical Monograph 100 183^216Niu Y Collerson K D Batiza R Wendt J I amp Regelous M
(1999) Origin of enriched-typed mid-ocean ridge basalt at ridgesfar from mantle plumes The East Pacific Rise at 11820rsquoN Journalof Geophysical Research 104 7067^7088
Nixon G T amp Orr A J (2007) Recent revisions to the EarlyMesozoic stratigraphy of Northern Vancouver Island (NTS 102I092L) and metallogenic implicatinos British Columbia InGrant B (ed) Geological Fieldwork 2006 British Columbia Ministry of
Energy Mines and Petroleum Resources Paper 2007-1 163^177Nixon GT Hammack J L KoyanagiV M Payie G J Haggart
JW Orchard M JTozerT Friedman R M Archibald D APalfy J amp Cordey F (2006a) Geology of the Quatsino^PortMcNeill area northernVancouver Island British Columbia Ministry
of Energy Mines and Petroleum Resources Geoscience Map 2006-2 scale150 000
Nixon G T Kelman M C Stevenson D Stokes L A ampJohnston K A (2006b) Preliminary geology of the Nimpkishmap area (NTS 092L07) northern Vancouver Island BritishColumbia In Grant B amp Newell J M (eds) Geological Fieldwork2005 British Columbia Ministry of Energy Mines and Petroleum Resources
Paper 2006-1 135^152Nixon G T Laroque J Pals A Styan J Greene A R amp
Scoates J S (2008) High-Mg lavas in the Karmutsen flood basaltsnorthernVancouver Island (NTS 092L) Stratigraphic setting andmetallogenic significance In Grant B (ed) Geological Fieldwork
2007 British Columbia Ministry of Energy Mines and Petroleum
Resources Paper 2008-1 175^190Nowell G M Kempton P D Noble S R Fitton J G
Saunders A Mahoney J J amp Taylor R N (1998) High precisionHf isotope measurements of MORB and OIB by thermal ionisation
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
36
mass spectrometry insights into the depleted mantle Chemical
Geology 149 211^233Parrish R R amp McNicoll V J (1992) U^Pb age determinations
from the southern Vancouver Island area British Columbia InRadiogenic Age and Isotopic Studies Report 5 Geological Survey of
Canada Paper 91-2 79^86Peate DW (1997) The Parana^Etendeka Province In Coffin M F
amp Mahoney J (eds) Large Igneous Provinces Continental Oceanic andPlanetary Flood Volcanism American Geophysical Union Geophysical
Monograph 100 217^245Perfit M R Fornari D J Ridley W I Kirk P D Casey J
Kastens K A Reynolds J R Edwards M Desonie DShuster R amp Paradis S (1996) Recent volcanism in the Siqueirostransform fault picritic basalts and implications for MORBmagma genesis Earth and Planetary Science Letters 141(1^4) 91^108
Pretorius W Weis D Williams G Hanano D Kieffer B ampScoates J S (2006) Complete trace elemental characterization ofgranitoid (USGSG-2GSP-2) reference materials by high resolutioninductively coupled plasma-mass spectrometry Geostandards and
Geoanalytical Research 30(1) 39^54Regelous M Niu Y Wendt J I Batiza R Greig A amp
Collerson K D (1999) Variations in the geochemistry of magma-tism on the East Pacific Rise at 10830rsquoN since 800 ka Earth and
Planetary Science Letters 168 45^63Recurren villon S Chauvel C Arndt N T Pik R Martineau F
Fourcade S amp Marty B (2002) Heterogeneity of the Caribbeanplateau mantle source Sr O and He isotopic compositions of oli-vine and clinopyroxene from Gorgona Island Earth and Planetary
Science Letters 205(1^2) 91Rhodes J M (1996) Geochemical stratigraphy of lava flows samples
by the Hawaii Scientific Drilling Project Journal of Geophysical
Research 101(B5) 11729^11746Rhodes J M amp Vollinger M J (2004) Composition of basaltic lavas
sampled by phase-2 of the Hawaii Scientific Drilling ProjectGeochemical stratigraphy and magma types Geochemistry
Geophysics Geosystems 5(3) doi1010292002GC000434Richards M A Jones D L Duncan R A amp DePaolo D J (1991)
A mantle plume initiation model for the Wrangellia flood basaltand other oceanic plateaus Science 254 263^267
Salters V J M amp Hart S R (1991) The mantle sources of oceanridges islands and arcs the Hf-isotope connection Earth and
Planetary Science Letters 104 364^380Salters V J M amp Stracke A (2004) Composition of the depleted
SaltersV J amp WhiteW (1998) Hf isotope constraints on mantle evo-lution Chemical Geology 145 447^460
SanoT amp Yamashita S (2004) Experimental petrology of basementlavas from Ocean Drilling Program Leg 192 implications for dif-ferentiation processes in Ontong Java Plateau magmas InFitton J G Mahoney J JWallace P J amp Saunders A D (eds)Origin and Evolution of the Ontong Java Plateau Geological Society
London Special Publications 229 185^218Saunders A D (2005) Large igneous provinces Origin and environ-
mental consequences Elements 1 259^263Self S ThordarsonT amp Keszthely L (1997) Emplacement of conti-
nental flood basalt lava flows In Mahoney J J amp Coffin M F(eds) Large Igneous Provinces Continental Oceanic and Planetary Flood
Volcanism American Geophysical Union Geophysical Monograph 100381^410
Shaw D M (2000) Continuous (dynamic) melting theory revisitedCanadian Mineralogist 38(5) 1041^1063
Sluggett C L (2003) Uranium^lead age and geochemical constraintson Paleozoic and Early Mesozoic magmatism in WrangelliaTerrane Saltspring Island British Columbia BSc thesisUniversity of British ColumbiaVancouver 84 pp
Storey M Mahoney J J amp Saunders A D (1997) Cretaceousbasalts in Madagascar and the transition between plume and con-tinental lithospheric mantle sources In Mahoney J J ampCoffin M F (eds) Large Igneous Provinces Continental Oceanic and
Planetary Flood Volcanism Aamerican Geophysical Union Geophysical
Monograph 100 95^122Surdam R C (1967) Low-grade metamorphism of the Karmutsen
Group PhD dissertation University of California Los Angeles288 pp
Sutherland-Brown A Yorath C J Anderson R G amp Dom K(1986) Geological maps of southern Vancouver IslandLITHOPROBE I 92C10 11 14 16 92F1 2 7 8 Geological Survey ofCanada Open File 1272
Tejada M L G Mahoney J J Castillo P R Ingle S P Sheth HC amp Weis D (2004) Pin-pricking the elephant evidence on theorigin of the Ontong Java Plateau from Pb^Sr^Hf^Nd isotopiccharacteristics of ODP Leg 192 basalts In Fitton J GMahoney J J Wallace P J amp Saunders A D (eds) Origin and
Evolution of the Ontong Java Plateau Geological Society London Special
Publications 229 133^150Thompson P M E Kempton P D amp Kerr A C (2008) Evaluation
of the effects of alteration and leaching on Sm^Nd and Lu^Hf sys-tematics in submarine mafic rocks Lithos 104(1^4) 164^176
Thompson P M E Kempton P D White R V Kerr A CTarney J Saunders A D Fitton J G amp McBirney A (2003)Hf-Nd isotope constraints on the origin of the CretaceousCaribbean plateau and its relationship to the Galapagos plumeEarth and Planetary Science Letters 217(1^2) 59^75
Vervoort J D Patchett P Blichert-Toft J amp Albare de F (1999)Relationships between Lu^Hf and Sm^Nd isotopic systems in theglobal sedimentary system Earth and Planetary Science Letters 16879^99
Walter M J (1998) Melting of garnet peridotite and the origin ofkomatiite and depleted lithosphere Journal of Petrology 39(1) 29^60
Weis D Kieffer B Maerschalk C Barling J de Jong JWilliams G A Hanano D Mattielli N Scoates J SGoolaerts A Friedman R A amp Mahoney J B (2006) High-pre-cision isotopic characterization of USGS reference materials byTIMS and MC-ICP-MS Geochemistry Geophysics Geosystems
7(Q08006) doi1010292006GC001283Weis D Kieffer B Hanano D Silva I N Barling J
Pretorius W Maerschalk C amp Mattielli N (2007) Hf isotopecompositions of US Geological Survey reference materialsGeochemistry Geophysics Geosystems 8(Q06006) doi1010292006GC001473
Wheeler J O amp McFeely P (1991) Tectonic assemblage map of theCanadian Cordillera and adjacent part of the United States ofAmerica Geological Survey of Canada Map 1712A
WhiteW M Albare de F amp Tecurren louk P (2000) High-precision analy-sis of Pb isotope ratios by multi-collector ICP-MS Chemical Geology167 257^270
Yorath C J Sutherland Brown A amp Massey N W D (1999)LITHOPROBE southern Vancouver Island British Columbia Geological
Survey of Canada Bulletin 498 145 ppZou H (1998) Trace element fractionation during modal and
non-model dynamic melting and open-system melting Amathematical treatment Geochimica et Cosmochimica Acta 62(11)1937^1945
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
37
Zou H amp Reid M R (2001) Quantitative modeling of trace elementfractionation during incongruent dynamic melting Geochimica et
Cosmochimica Acta 65(1) 153^162
APPENDIX SAMPLEPREPARAT ION ANDANALYT ICAL METHODSOnly the freshest rocks in the field were sampled and onlythe least altered samples were selected for chemical andisotopic analysis based on thorough petrographic inspec-tion Fifty-nine of the 129 collected samples were crushed(400 g) into small pieces 52mm in diameter in aRocklabs hydraulic piston crusher between WC plates tominimize contamination The coarse-crush was mixedand 100 g was powdered in a planetary mill using agatejars and balls cleaned with quartz sand between samples
ActLabs analytical methodsThe major- and trace-element compositions of the whole-rock powders were determined at Activation LaboratoriesLtd (Actlabs) in Ancaster Ontario Analytical techniquesand detection limits are also available from Actlabs (httpwwwactlabscommethsub_code4erehtm) The analyticalmethod for each of the elements analyzed is indicated inTable 2 For the major elements a 02 g sample was mixedwith a lithium metaborate^lithium tetraborate mixtureand fused in a graphite crucible The molten mixture waspoured into a 5 HNO3 solution and shaken until dis-solved (30min) The samples were analyzed for majoroxides and selected trace elements by inductively coupledplasma optical emission spectrometry (ICP-OES) on acombination simultaneous^sequential Thermo Jarrell-AshEnviro II system Internal calibration was achieved usinga variety of international reference materials (eg W-2BIR-1 DNC-1) and independent control samplesAdditional trace elements were analyzed by both instru-mental neutron activation analysis (INAA) and induc-tively couple plasma mass spectrometry (ICP-MS) ForINAA15^25 g of sample was weighed into small polyeth-ylene vials and irradiated with a control international ref-erence material CANMET WMS-1 and NiCr flux wires ata thermal neutron flux of 71012 neutronscm2 per s in theMcMaster Nuclear Reactor Following a 7 day waitingperiod the samples were measured on an Ortec high-purity Ge detector linked to a Canberra Series 95 multi-channel analyzer Activities for each element weredecay- and weight-corrected and compared with a detectorcalibration developed from multiple international certifiedreference materials For ICP-MS 025 g of sample wasdigested in HF followed by HNO3^HClO4 heated andtaken to dryness The samples were brought back into solu-tion with HCl Samples were analyzed using a PerkinElmer Optima 3000 ICP In-lab standards or certified ref-erence materials (eg W-2 BIR-1 DNC-1) were used for
quality control A total of 15 blind duplicates were ana-lyzed from the same sample powders to assessreproducibility
University of Massachusetts analyticalmethodsFourteen sample duplicate powders taken from the same100 g sample aliquot (high-MgO basalts and picrites)were also analyzed at the Ronald B Gilmore X-RayFluorescence (XRF) Laboratory at the University ofMassachusetts Major elements were measured on a fusedLa-bearing lithium borate glass disc using a SiemensMRS-400 spectrometer Trace element concentrations(Rb Sr Ba Ce Nb Zr Y Pb Zn Ga Ni Cr V) weremeasured on a separate powder pellet using a PhilipsPW2400 sequential spectrometer using a Rh tube LOIand ferrous iron measurements were made as describedby Rhodes amp Vollinger (2004) Precision and accuracy esti-mates for the data have been described by Rhodes (1996)and Rhodes amp Vollinger (2004) Results for each sampleare the average of two separate analyses and are shown inElectronic Appendix 2 for a comparison of analyses ofduplicate powders using XRFand ICP-MS
PCIGR analytical methodsA subset of 19 samples was selected for high-precisiontrace-element analysis and Sr Nd Pb and Hf isotopicanalysis at the Pacific Centre for Isotopic andGeochemical Research (PCIGR) at the University ofBritish Columbia (UBC) Samples were selected from the59 samples analyzed for whole-rock chemistry at ActLabsbased on major- and trace-element chemistry alteration(low LOI and petrographic alteration index) and samplelocation Samples were prepared for trace-element analysisat the PCIGR by the technique described by Pretorius et al(2006) on unleached rock powders Sample powders(100mg) were weighed in 7ml screw-top Savillex bea-kers and dissolved in 1ml 14N HNO3 and 5ml 48 HFon a hotplate for 48 h at 1308C with periodic ultrasonica-tion Samples were dried and redissolved in 6ml 6N HClon a hotplate for 24 h and then dried and redissolved in1ml concentrated HNO3 for 24 h before final dryingTrace element abundances were measured by high-resolution inductively coupled plasma-mass spectrometry(HR-ICP-MS) with aThermo Finnigan Element2 systemfollowing the procedures described by Pretorius et al(2006) within 24 h of redissolution HFSE and LILE weremeasured in medium-resolution mode at 2000 dilutionusing a PFATeflon spray chamber washed with aqua regiafor 3min between samples REE were measured in high-resolution mode and U and Pb in low-resolution modeat 2000 dilution using a glass spray chamber washedwith 2 HNO3 between samples Total procedural blanksand reference materials (BCR-2 BHVO-2) were analyzedwith the batch of samples Indium was used as an internal
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
38
standard in all samples and standard solutionsBackground and standard solutions were analyzed afterevery five samples to detect memory effects and massdrift Results for PCIGR trace-element analyses areshown in Electronic Appendix 3 for the full set of analyzedtrace elementsSample digestion for purification of Sr Nd Hf and Pb
for column chemistry began by weighing each samplepowder (400^500mg) prior to leaching All samples wereleached with 6N HCl and placed in an ultrasonic bath for15min Samples were rinsed twice with 18 M-cm H2Obetween each leaching step (15 total) until the supernatantliquid was clear [following the method of Mahoney (1987)]Samples were then dried on a hotplate for 24 h andweighed again Sample solutions were then prepared bydissolving 100^250mg of the leached powder dissolvedin 1ml 14N HNO3 and 10ml 48 HF on a hotplate for48 h at 1308C with periodic ultrasonication Samples weredried and redissolved in 6ml 6N HCl on a hotplate for24 h and then dried Pb was separated using anion exhangecolumns and the discard was used for Sr REE and Hfseparation Nd was separated from the other REE and Hfrequired two additional purification steps Detailed proce-dures for column chemistry for separating Sr Nd and Pbat the PCIGR have been described byWeis et al (2006) andHf purification has been described byWeis et al (2007) Srand Nd isotope ratios were measured by thermal ioniza-tion mass spectrometry (TIMS) on a Thermo FinniganTriton system in static mode with relay matrix rotation ona single Ta and double Re^Ta filament respectively Fourto five filaments per barrel of 21 were occupied by stan-dards (NBS 987 for Sr and LaJolla for Nd) for each barrelwhere samples were run Sample Sr and Nd isotopic com-positions were corrected for mass fractionation using86Sr88Srfrac14 01194 and 146Nd144Ndfrac14 07219 Each samplewas then normalized using the barrel average of the refer-ence material relative to the values of143Nd144Ndfrac14 0511858 and 87Sr86Srfrac14 0710248 (Weiset al 2006) During the course of the Vancouver Islandanalyses the La Jolla Nd standard gave an average valueof 05118566 (nfrac14 7) and NBS987 standard gave an aver-age of 07102408 (nfrac1411) (2 error is reported as times106) 147Sm144Nd errors are 15 or 0006 USGeological Survey (USGS) reference material BHVO-2was processed with the samples and yielded Sr and Nd iso-topic ratios of 07034607 and 05129786 respectivelyThese are in agreement with the published values of070347920 and 051298411 respectively (Weis et al2006)
Pb and Hf isotopic compositions were analyzed by mul-tiple collector-inductively coupled plasma-mass spectrome-try (MC-ICP-MS) on a static multi-collection on a NuPlasma (Nu Instruments) system The detailed analyticalprocedure for Pb isotopic analyses on the Nu Plasma atthe PCIGR has been described by Weis et al (2006) Theconfiguration for Pb analyses allows for collection of PbTl and Hg togetherTl and Hg are used to monitor instru-mental mass discrimination and isobaric overlap repec-tively All sample solutions were analyzed withapproximately the same PbTl ratio (4) as the referencematerial NIST SRM 981 To accomplish this a small ali-quot of each sample solution from the Pb columns was ana-lyzed on the Element2 to determine the precise amount ofPb available for analysis on the Nu Plasma The SRM 981standard was run after every two samples on the NuPlasma During the time samples were run analyses ofthe SRM 981 Pb reference material gave values of206Pb204Pbfrac1416940322 207Pb204Pbfrac1415495823and 208Pb204Pbfrac14 36713164 (nfrac14 61 2 error is reportedas times 104) these values are in excellent agreement withreported TIMS triple-spike values of Galer amp Abouchami(1998) Fractionation-corrected Pb isotopic ratios were fur-ther corrected by the sample-standard bracketing methodor the ln^ln correction method described by White et al(2000) and Blichert-Toft et al (2003) Leached powders ofUSGS reference material BHVO-2 yielded Pb isotopicratios of 206Pb204Pbfrac14186454 8 207Pb204Pbfrac14 1549105 and 208Pb204Pbfrac14 38222514 and BCR-2yielded 206Pb204Pbfrac141880466 207Pb204Pbfrac14 1562518 and 208Pb204Pbfrac14 3883496 (2 error isreported as times 104) These values are in agreement withthose for leached residues of BHVO-2 and BCR-2 fromWeis et al (2006)Hf isotopic compositions were analyzed following the
procedures detailed byWeis et al (2007) The configurationfor Hf analyses monitored Lu mass 175 andYb mass 172 toallow for interference correction to masses 174 and 176 Hfisotopic ratios were normalized internally for mass fractio-nation to a 179Hf177Hf ratio of 07325 using an exponentialcorrection Standards were run after every two samplesand sample results were normalized to the ratio of the in-run daily average and a 176Hf177Hf ratio for JMC-475 of0282160 During the course of analyses the Hf standardJMC-475 gave an average value 02821679 (nfrac14130)USGS reference materials BCR-2 and BHVO-2 were pro-cessed with the samples and yielded Hf isotopic ratios of02828675 and 02831005 respectively Publishedvalues for BCR-2 and BHVO-2 are 02828717 and0283104 8 respectively (Weis et al 2007)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
39
melting on differences in trace-element concentrations inthe tholeiitic basalts and picritesDynamic melting models simulate progressive decom-
pression melting where some of the melt fraction isretained in the residue and only when the degree of partialmelting is greater than the critical mass porosity and the
source becomes permeable is excess melt extracted fromthe residue (Zou1998) For the Karmutsen lavas evolutionof trace element concentrations was simulated using theincongruent dynamic melting model developed by Zou ampReid (2001) Melting of the mantle at pressures greater than1 GPa involves incongruent melting (eg Longhi 2002)
OIB array
PacificMORB(230 Ma)
(0 Ma)
EPR(230 Ma)
-4
-2
0
2
4
8
10
12
14
16
18
20
22
-4 -2 0 2 4 6 8 10 12 14
minus2
0
2
4
6
8
10
12
14
0702 0703 0704 0705 0706
IndianMORB
87Sr 86Sr (initial)
Hf (initial)
(a) (c)
Nd (initial)
Karmutsen tholeiitic basalt
HawaiiOntong JavaCaribbean Plateau
Karmutsen high-MgO lava
high-MgO lavasKarmutsen
1540
1545
1550
1555
1560
1565
175 180 185 190 195 200375
380
385
390
395
175 180 185 190 195 200
(d) (e)
EPR
EPR
206Pb204Pb
208Pb204Pb
206Pb204Pb
207Pb204Pb
Karmutsen Formation (initial)
HawaiiOntong JavaCaribbean Plateau
East Pacific Rise
Karmutsen Formation (measured)
Juan de FucaGorda
ε Nd
(initial)
Karmutsen Formation (different age-correction for 3 picrites)
(0 Ma)
NbY
001
01
1
1 10
ZrY
OIB
NMORB
(b)
Iceland array
6
Fig 13 Comparison of age-corrected (230 Ma) Sr^Nd^Hf^Pb isotopic compositions for Karmutsen flood basalts onVancouver Island withage-corrected OIB and MORB and Nb^Zr^Y variation (a) Initial eNd vs 87Sr86Sr (b) NbYand ZrY variation (after Fitton et al 1997)(c) Initial eHf vs eNd (d) Measured and initial 207Pb204Pb vs 206Pb204Pb (e) Measured and initial 208Pb204Pb vs 206Pb204Pb References fordata sources are listed in Electronic Appendix 4 Most of the compiled data were extracted from the GEOROC database (httpgeorocmpch-mainzgwdgdegeoroc) OIB array line in (c) is fromVervoort et al (1999) EPR East Pacific Rise Several high-MgO samples affected bysecondary alteration were not plotted for clarity in Pb isotope plots Estimation of fields for age-corrected Pacific MORB and EPR at 230 Mawere made using parent^daughter concentrations (in ppm) from Salters amp Stracke (2004) of Lufrac14 0063 Hffrac14 0202 Smfrac14 027 Ndfrac14 071Rbfrac14 0086 and Srfrac14 836 In the NbYvs ZrYplot the normal (N)-MORB and OIB fields not including Icelandic OIB are from data com-pilations of Fitton et al (1997 2003) The OIB field is data from 780 basalts (MgO45wt ) from major ocean islands given by Fitton et al(2003) The N-MORB field is based on data from the Reykjanes Ridge EPR and Southwest Indian Ridge from Fitton et al (1997) The twoparallel gray lines in (b) (Iceland array) are the limits of data for Icelandic rocks
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
29
with olivine being produced during the reactioncpxthornopxthorn spfrac14meltthornol for spinel peridotites The mod-eling used coefficients for incongruent melting reactionsbased on experiments on lherzolite melting (eg Kinzleramp Grove 1992 Longhi 2002) and was separated into (1)melting of spinel lherzolite (2) two combined intervals ofmelting for a garnet lherzolite (gt lherz) and spinel lher-zolite (sp lherz) source (3) melting of garnet lherzoliteThe model solutions are non-unique and a range indegree of melting partition coefficients melt reactioncoefficients and proportion of mixed source componentsis possible to achieve acceptable solutions (see Fig 14 cap-tion for description of modeling parameters anduncertainties)The LREE-depleted high-MgO lavas require melting of
a depleted spinel lherzolite source (Fig 14a) The modelingresults indicate a high degree of melting (22^25) similarto the results from PRIMELT1 based on major-elementvariations (discussed above) of a LREE-depleted sourceThe high-MgO lavas lack a residual garnet signature Theenriched tholeiitic basalts involved melting of both garnetand spinel lherzolite and represent aggregate melts pro-duced from continuous melting throughout the depth ofthe melting column (Fig 14b) Trace-element concentra-tions in the tholeiitic and picritic lavas are not consistentwith low-degree melting of garnet lherzolite alone(Fig 14c) The modeling results indicate that the aggregatemelts that formed the tholeiitic basalts would haveinvolved lesser proportions of enriched small-degreemelts (1^6 melting) generated at depth and greateramounts of depleted high-degree melts (12^25) gener-ated at lower pressure (a ratio of garnet lherzolite tospinel lherzolite melt of 14 provides the best fits seeFig 14b and description in figure caption) The totaldegree of melting for the tholeiitic basalts was high(23^27) similar to the degree of partial melting in thespinel lherzolite facies for the primary melts that eruptedas the Keogh Lake picrites of a source that was also prob-ably depleted in LREE
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
(c)
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
Melting of garnet lherzolite
High-MgO lavas
Melting of spinel lherzolite
source (07DM+03PM)
source (07DM+03PM)
6 melting
30 melting
(b)
(a)
30 melting
Sam
ple
Cho
ndrit
eS
ampl
eC
hond
rite 20 melting
1 melting
Melting of garnet lherzoliteand spinel lherzolite
x gt lherz + 4x sp lherz
5 meltingx gt lherz + 4x sp lherz
Sam
ple
Cho
ndrit
e
Tholeiitic lavas
Tholeiitic lavas
High-MgO lavas
source (07DM+03PM)
Fig 14 Trace-element modeling results for incongruent dynamicmantle melting for picritic and tholeiitic lavas from the KarmutsenFormation Three steps of modeling are shown (a) melting of spinellherzolite compared with high-MgO lavas (b) melting of garnet lher-zolite and spinel lherzolite compared with tholeiitic lavas (c) meltingof garnet lherzolite compared with tholeiitic and picritic lavasShaded fields in each panel for tholeiitic basalts and picrites arelabeled Patterns with symbols in each panel are modeling results(patterns are 5 melting increments in (a) and (b) and 1 incre-ments in (c) PM primitive mantle DM depleted MORB mantle gtlherz garnet lherzolite sp lherz spinel lherzolite In (b) the ratios ofper cent melting for garnet and spinel lherzolite are indicated (x gtlherzthorn 4x sp lherz means that for 5 melting 1 melt in garnetfacies and 4 melt in spinel facies where x is per cent melting in theinterval of modeling) The ratio of the respective melt proportions inthe aggregate melt (x gt lherzthorn 4x sp lherz) was kept constant andthis ratio of melt contribution provided the best fit to the data Arange of proportion of source components (07DMthorn 03PM to
09DMthorn 01PM) can reproduce the variation in the high-MgOlavas Melting modeling uses the formulation of Zou amp Reid (2001)an example calculation is shown in their Appendix PM fromMcDonough amp Sun (1995) and DM from Salters amp Stracke (2004)Melt reaction coefficients of spinel lherzolite from Kinzler amp Grove(1992) and garnet lherzolite fromWalter (1998) Partition coefficientsfrom Salters amp Stracke (2004) and Shaw (2000) were kept constantSource mineralogy for spinel lherzolite (018cpx027opx052ol003sp) from Kinzler (1997) and for garnet lherzolite(034cpx008opx053ol005gt) from Salters amp Stracke (2004) Arange of source compositions for melting of spinel lherzolite(07DMthorn 03PM to 03DMthorn 07PM) reproduce REE patternssimilar to those of the high-MgO lavas with best fits for 22ndash25melting and 07DMthorn 03PM source composition Concentrationsof tholeiitic basalts cannot be reproduced from an entirelyprimitive or depleted source
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
30
The uncertainty in the composition of the source in thespinel lherzolite melting model for the high-MgO lavasprecludes determining whether the source was moredepleted in incompatible trace elements than the source ofthe tholeiitic lavas (see Fig 14 caption for description) It ispossible that early high-pressure low-degree melting left aresidue within parts of the plume (possibly the hotter inte-rior portion) that was depleted in trace elements (egElliott et al 1991) further decompression and high-degreemelting of such depleted regions could then have generatedthe depleted high-MgO Karmutsen lavas Shallow high-degree melting would preferentially sample depletedregions whereas deeper low-degree melting may notsample more refractory depleted regions (eg CaribbeanEscuder-Viruete et al 2007) The picrites lie at the highend of the range of Nd isotopic compositions and havelow NbZr similar to depleted Icelandic basalts (Fittonet al 2003) therefore the picrites may represent the partsof the mantle plume that were the most depleted prior tothe initiation of melting As Fitton et al (2003) suggestedthese depleted melts may only rarely be erupted withoutcombining with more voluminous less depleted melts andthey may be detected only as undifferentiated small-volume flows like the Keogh Lake picrites within theKarmutsen Formation on Vancouver IslandDecompression melting within the mantle plume initiatedwithin the garnet stability field (35^25 GPa) at highmantle potential temperatures (Tp414508C) and pro-ceeded beneath oceanic arc lithosphere within the spinelstability field (25^09 GPa) where more extensivedegrees of melting could occur
Magmatic evolution of Karmutsentholeiitic basaltsPrimary magmas in large igneous provinces leave exten-sive coarse-grained cumulate residues within or below thecrust these mafic and ultramafic plutonic sequences repre-sent a significant proportion of the original magma thatpartially crystallized at depth (eg Farnetani et al 1996)For example seismic and petrological studies of OntongJava (eg Farnetani et al 1996) combined with the use ofMELTS (Ghiorso amp Sack 1995) indicate that the volcanicsequence may represent only 40 of the total magmareaching the Moho Neal et al (1997) estimated that30^45 fractional crystallization took place for theOntong Java magmas and produced cumulate thicknessesof 7^14 km corresponding to 11^22 km of flood basaltsdepending on the presence of pre-existing oceanic crustand the total crustal thickness (25^35 km) Interpretationsof seismic velocity measurements suggest the presence ofpyroxene and gabbroic cumulates 9^16 km thick beneaththe Ontong Java Plateau (eg Hussong et al 1979)Wrangellia is characterized by high crustal velocities in
seismic refraction lines which is indicative of mafic
plutonic rocks extending to depth beneath VancouverIsland (Clowes et al 1995)Wrangellia crust is 25^30 kmthick beneath Vancouver Island and is underlain by astrongly reflective zone of high velocity and density thathas been interpreted as a major shear zone where lowerWrangellia lithosphere was detached (Clowes et al 1995)The vast majority of the Karmutsen flows are evolved tho-leiitic lavas (eg low MgO high FeOMgO) indicating animportant role for partial crystallization of melts at lowpressure and a significant portion of the crust beneathVancouver Island has seismic properties that are consistentwith crystalline residues from partially crystallizedKarmutsen magmas To test the proportion of the primarymagma that fractionated within the crust MELTS(Ghiorso amp Sack 1995) was used to simulate fractionalcrystallization using several estimated primary magmacompositions from the PRIMELT1 modeling results Themajor-element composition of the primary magmas couldnot be estimated for the tholeiitic basalts because of exten-sive plagthorn cpxthornol fractionation and thus the MELTSmodeling of the picrites is used as a proxy for the evolutionof major elements in the volcanic sequence A pressure of 1kbar was used with variable water contents (anhydrousand 02wt H2O) calculated at the quartz^fayalite^magnetite (QFM) oxygen buffer Experimental resultsfrom Ontong Java indicate that most crystallizationoccurs in magma chambers shallower than 6 km deep(52 kbar Sano amp Yamashita 2004) however some crys-tallization may take place at greater depths (3^5 kbarFarnetani et al 1996)A selection of the MELTS results are shown in Fig 15
Olivine and spinel crystallize at high temperature until15^20wt of the liquid mass has fractionated and theresidual magma contains 9^10wt MgO (Fig 15f) At1235^12258C plagioclase begins crystallizing and between1235 and 11908C olivine ceases crystallizing and clinopyr-oxene saturates (Fig 15f) As expected the addition of clin-opyroxene and plagioclase to the crystallization sequencecauses substantial changes in the composition of the resid-ual liquid FeO and TiO2 increase and Al2O3 and CaOdecrease while the decrease in MgO lessens considerably(Fig 15) The near-vertical trends for the Karmutsen tho-leiitic basalts especially for CaO do not match the calcu-lated liquid-line-of-descent very well which misses manyof the high-CaO high-Al2O3 low-FeO basalts A positivecorrelation between Al2O3CaO and SrNd indicates thataccumulation of plagioclase played a major role in the evo-lution of the tholeiitic basalts (Fig 15g) Increasing thecrystallization pressure slightly and the water content ofthe melts does not systematically improve the match ofthe MELTS models to the observed dataThe MELTS results indicate that a significant propor-
tion of the crystallization takes place before plagioclase(15^20 crystallization) and clinopyroxene (35^45
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
31
crystallization) join the crystallization sequence Theseresults suggest that445 crystallization of the primarymagmas occurred (estimated from the residual liquid percent in Fig 15f) and if the original flood basalt stratigra-phy on Vancouver Island was 6 km thick at least anequivalent amount of complementary plutonic rocksshould have crystallized The Karmutsen basalts and theirplutonic residues thus represent a significant addition ofcrust (perhaps 412 km thickness) previously thickenedduring the formation of the underlying Paleozoic arc
Growth of the Wrangellia oceanic plateauon Vancouver IslandThe extrusive part of the Wrangellia oceanic plateau onVancouver Island was constructed in a three-layersequence of mostly tholeiitic basalts with a restrictedrange of composition (Fig 16) The plateau overliesDevonian to Mississippian arc volcanic sequences andmarine sediments of Mississippian to Permian ages thatare intruded by gabbroic rocks related to the flood basaltsThe volcanic stratigraphy of theWrangellia plateau formed
00
05
10
15
20
25
30
35
40
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
16
0 2 4 6 8 10 12 14 16 18 20
10
11
12
13
14
15
16
17
18
19
20
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
0 2 4 6 8 10 12 14 16 18 20
MgO (wt)
0 10 20 30 40 50
percent of total mass
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
) Mineral proportions
Sp (e)
Ol
CpxPlag
PicriteHigh-MgO basaltTholeiitic basalt
TiO2 (wt)
MgO (wt)
Al2O3 (wt) CaO (wt)
FeO (wt)
MgO(a) (b)
(c) (d)
MgO (wt)MgO (wt)
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
)
Residual liquid ()
1220
1190
Ol + Sp
Plag+Ol+Sp Plag+Cpx+Sp
02 wt H2O
no H2O
Residual liquid and crystallizing phases
(f)
0102030405060708090100
plagioclaseaccumulation
plagioclaseaccumulation
plagioclase
accumulatio
n
10
12
14
16
0 10 20 30 40
Al2O3CaO
SrNd
tholeiitic basalts
(g)
Fig 15 Forward fractional crystallization modeling results for major elements from MELTS (Ghiorso amp Sack 1995) compared with picriticand tholeiitic lavas from the Karmutsen Formation (a^d) Major-element compositions of residual liquids compared with Karmutsen lavacompositions For clarity the results are shown only for one starting composition (open circle) using the composition of the estimated primarymagma for the high-MgO lava series from the modeling technique of Herzberg et al (2007) for sample 4723A13 as shown in Fig 12 The esti-mated primary magmas for samples 4723A3 and 93G171 yield similar results A pressure of 1 kbar was used with no H2O and 02wt H2Ocalculated at QFM black line is anhydrous and grey line is 02wt H2OThe end-point for the anhydrous result is at 78 fractionation and11308C and the end-point for the result with 02wt H2O is 75 fractionation and11108C (e) Mineral proportions of crystallized phases fromone MELTrun (sample 4723A13) with no H2O (f) Per cent residual liquid vs temperature labelled with intervals of crystallizing phases (g)Al2O3CaO vs SrNd for the tholeiitic basalts
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
32
in a single eruptive phase (c 231^225 Ma) as an emergentbasalt sequence during a deeper-water shallow-water andsubaerial stage and subsided below sea level relativelysoon after its formation (Fig 16) The deeper-water stagecontains pillowed and massive submarine flows (3000mthick) that built up the submarine volcanic edifice from theseafloor The lowest pillow basalts were emplaced onunconsolidated and lithified fine-grained siliciclastic andcarbonaceous sediments containing Daonella beds(c 235^232 Ma) that are intruded by abundantKarmutsen mafic sills (Fig 16a) The different pillowedand massive submarine flows resulted primarily from dif-ferent effusive rates and local topography Some of themassive submarine flows commonly recognized by radiat-ing columnar jointing represent master tubes for the deliv-ery of lava to distal parts of submarine flow fields Thepillowed flows which commonly have uneven topography
in modern sequences may have served as levees to themassive submarine flows rather than allowing unob-structed spreading and inflation typical of the massive sub-aerial flows that form on shallow slopes in continentalflood basalt provinces (eg Self et al 1997) The rapid suc-cession of flows is evident by the near absence of sedimentbetween submarine flows Minor volumes of picritic pillowlavas occur in the upper part of the pillowed flow sequencein areas of the plateau exposed on northern VancouverIsland (Fig 16 Nixon et al 2008)The growth of the plateau in shallow water led to
emplacement of an increasing proportion of volcaniclastics(pillow breccias and hyaloclastites) that accumulated up to1500m thick in some areas (Fig 16b Nixon et al 2008)Horizons of pillowed flows occur locally within the volca-niclastic stratigraphy Sedimentary structures (eg gradedbedding fluidization structures) and petrographic textures
Late Devonian-Mississippian arc
Daonella beds
Middle Triassic
deeper water
Late Devonian-Mississippian arc
pillowed flows massive submarine
flows
pillow brecciahyaloclastiteshallow water
~4 km
~6 km
Late Triassic limestoneinterflow limestonelenses
Karmutsen Formation
mafic intrusions sediment-sill complex
Karmutsen Formation
(a) (b)
(c) (d)
subaerial flows
Late Devonian-Mississippian arc Late Devonian-Mississippian arc
picritic pillow basalts
Miss to Permian marine sediments
sills
Fig 16 Growth of the Wrangellia oceanic plateau onVancouver Island (a) Deep-water eruptive phase of pillowed and massive submarineflows with almost no intervening sediments onto MiddleTriassic shale containing Daonella beds (c 235^232 Ma) The basement of the plateauis composed of Devonian to Mississippian arc sequences overlain by Mississippian to Permian carbonates and siliciclastic sediments(b) Shallow-water flow emplacement with increasing proportion of volcaniclastics resulted primarily from cooling^contraction granulationPicritic pillow basalts occur primarily near the transition between pillowed submarine flows and volcaniclastic units Pillowed and massiveflows occur locally within the volcaniclastic unit (c) Emergent stage of growth consists of overlapping subaerial flow fields similar to continen-tal flood basalts with pahoehoe flow structures Multi-tiered columnar jointing occurrs locally in response to ponding of flows (eg Bondre et al2004) Interflow carbonate deposits and plagioclase megacrystic flows developed locally during the waning stage of volcanism (d) Post-volcanicsubsidence and deposition of Late Carnian to Norian (c 212^223 Ma) platform carbonates on top of the plateau Units are labeled in the dia-grams Thicknesses of the underlying units are not to scale
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
33
within the volcaniclastic rocks indicate formation primar-ily via cooling^contraction granulation magma^water^steam interaction autobrecciation resedimentation andmass-wasting rather than pyroclastic eruption (Nixonet al 2008)A broad subaerial platform was constructed by emplace-
ment of subaerial lava flows during an emergent subaerialstage (Fig 16c) The subaerial unit predominantly com-prises well-layered sheet flows that formed in a similarway to continental flood basalts as hundred(s) of pahoehoeflow fields erupted on shallow slopes (Self et al 1997) how-ever volcaniclastic and pillowed flow units are locally pre-sent mostly in the upper parts of the stratigraphyLocalized ponding of flows led to the formation of multi-tiered columnar jointing in some subaerial flows (egBondre et al 2004) As volcanism waned local interflowcarbonate deposits developed and plagioclase megacrysticflows were erupted Subsidence of the plateau is recordedby the interflow lenses and deposits of platform carbonates(40^1500m thick) on top of the Karmutsen basalts thatcontain Late Carnian to Early Norian fossils (Fig 16d c216^227 Ma Nixon et al 2008)The evolution of the internal architecture of the
Wrangellia oceanic plateau was primarily related to theconditions of emplacement (ie water depth and eruptionenvironment) There does not appear to have been a sys-tematic coincident evolution or shift of lava compositionsonVancouver Island as has been recognized in most con-tinental flood basalts and parts of the Wrangellia oceanicplateau in Alaska and Yukon (Greene et al 2008 2009)In general with the exception of the picritic pillow lavasin the submarine stratigraphy and local variations betweenhigh-MgO and tholeiitic lavas there are no significantcompositional differences between lavas erupted duringthe different stages of growth of the plateau (ie submarineand subaerial flows) nor are substantial compositional dif-ferences generally found between massive submarine flowsin contact with pillowed flows tholeiitic basalt flows nearthe base of the Karmutsen Formation are compositionallysimilar to flows sampled near the top of the volcanicsequence Where it was possible to sample continuous sec-tions of volcanic stratigraphy greater than several hun-dreds of meters (eg Keogh^Maynard Lake area west ofthe Karmutsen Range Fig 2) there is no clear relationshipbetween stratigraphic position and compositionalvariation
CONCLUSIONSThe Karmutsen Formation covers extensive areas ofVancouver Island (20 000 km2) and forms part of a largeoceanic plateau erupted over a geologically short interval(c 225^231 Ma) The tripartite volcanic stratigraphyexposed on Vancouver Island is at least 6 km thick andcomprises a succession of submarine flows volcaniclastic
deposits and massive subaerial flowsThese volcanologicaldifferences are primarily related to the eruption environ-ment (deep-water shallow-water or subaerial) The rapidgrowth of the plateau prevented significant accumulationsof intervening sediment except in the uppermost stratigra-phy where isolated limestone lenses commonly associatedwith pillow basalt and hyaloclastite deposits preserve arecord of the final stages of subsidence of the plateauThe Wrangellia plateau on Vancouver Island was con-
structed dominantly of tholeiitic basalt with restrictedmajor- and trace-element and isotopic compositional var-iations Minor volumes of picritic pillow basalts eruptedlate in the submarine phase The Karmutsen basalts wereemplaced onto older Paleozoic arc volcanic sequences andlow NbLa for the picrites may indicate a contributionfrom the sub-arc lithospheric mantle Modeling of the pet-rogenesis of the Karmutsen picrites indicates melting ofanomalously hot mantle (15008C) and extensive degreesof partial melting (23^27) and is consistent with aplume initiation model Combined Sr^Nd^Hf^Pb isotopicsystematics indicate that the picritic and tholeiitic lavaswere derived from a depleted plume-type mantle sourcedistinct from the source of MORB that has compositionalsimilarities to the source of the lavas forming theCaribbean Plateau Picrites were generated by melting ofspinel lherzolite whereas the tholeiitic basalts formed asaggregate melts of both garnet and spinel lherzolite Thetholeiitic basalts underwent extensive low-pressure fractio-nation (52^3 kbar) and seismic studies indicate that someof theWrangellia crust beneathVancouver Island may cor-respond to the plutonic residues of the Karmutsen floodbasalts
ACKNOWLEDGEMENTSWe are grateful to Nick Arndt for helping us get this pro-ject started and Nick Massey for insights into VancouverIsland geology We greatly appreciate the assistance withPRIMELT modeling from Claude Herzberg We thankMikkel Schau for his insight and enthusiasm during field-work Jane Barling assisted with analyses by MC-ICP-MSThoughtful and thorough reviews of the manuscript byJohn Mahoney Fred Frey Wendy Bohrson DougalJerram and Teal Riley are greatly appreciated Fundingwas generously provided by the Rocks to Riches Programadministered by the BC amp Yukon Chamber of Mines in2004 by the BC Geological Survey and by NSERCDiscovery Grants to JSS and DW ARG was supportedby a University Graduate Fellowship at UBC
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
34
REFERENCESBarker F Brown A S Budahn J R amp Plafker G (1989) Back-arc
with frontal-arc component origin of Triassic Karmutsen basaltBritish Columbia Canada Chemical Geology 75 81^102
Blichert-Toft JWeis D Maerschalk C Agranier A amp Albare de F(2003) Hawaiian hot spot dynamics as inferred from the Hf and Pbisotope evolution of Mauna Kea volcano Geochemistry GeophysicsGeosystems 4(2) 1^27 doi1010292002GC000340
Bondre N R Duraiswami R A amp Dole G (2004) Morphologyand emplacement of flows from the Deccan Volcanic ProvinceIndia Bulletin of Volcanology 66 29^45
Carlisle D (1963) Pillow breccias and their aquagene tuffs QuadraIsland British Columbia Journal of Geology 71 48^71
Carlisle D (1972) Late Paleozoic to Mid-Triassic sedimentary-volcanic sequence on Northeastern Vancouver Island Geological
Society of Canada Paper Report of Activities 72-1 Part B 24^30Carlisle D amp SuzukiT (1974) Emergent basalt and submergent car-
bonate^clastic sequences including the Upper Triassic Dilleri andWelleri zones on Vancouver Island Canadian Journal of Earth
Sciences 11 254^279Chauvel C amp Blichert-Toft J (2001) A hafnium isotope and trace
element perspective on melting of the depleted mantle Earth and
Planetary Science Letters 190 137^151Chayes F (1954)The theory of thin-section analysis Journal of Geology
62 92^101Cho M Liou J G amp Maruyama S (1987) Transition from the zeo-
lite to prehnite^pumpellyite facies in the Karmutsen metabasitesVancouver Island British Columbia Journal of Petrology 27(2)467^494
Clowes R M Zelt C A Amor J R amp Ellis R M (1995)Lithospheric structure in the southern Canadian Cordillera froma network of seismic refraction lines Canadian Journal of Earth
Sciences 32(10) 1485^1513Coffin M F amp Eldholm O (1994) Large igneous provinces Crustal
structure dimensions and external consequences Reviews of
Geophysics 32(1) 1^36Cox K G (1980) A model for flood basalt volcanism Journal of
Petrology 21 629^650Eldholm O amp Coffin M F (2000) Large igneous provinces
and plate tectonics In Richards M A Gordon R G amp vander Hilst R D (eds) The History and Dynamics of Global
Plate Motions American Geophysical Union Geophysical Monograph 121309^326
Elliott T R Hawkesworth C J amp Grolaquo nvold K (1991) Dynamicmelting of the Iceland plume Nature 351 201^206
Escuder-Viruete J Pecurren rez-Estaucurren n A Contreras F Joubert MWeis D Ullrich T D amp Spadea P (2007) Plume mantle sourceheterogeneity through time Insights from the Duarte ComplexHispaniola northeastern Caribbean Journal of Geophysical Research112(B04203) doi1010292006JB004323
Farnetani C G Richards M A amp Ghiorso M S (1996)Petrological models of magma evolution and deep crustal structurebeneath hotspots and flood basalt provinces Earth and Planetary
Science Letters 143 81^94Fitton J G amp Godard M (2004) Origin and evolution of magmas
on the Ontong Java Plateau In Fitton J G Mahoney J JWallace P J amp Saunders A D (eds) Origin and Evolution of the
Ontong Java Plateau Geological Society London Special Publications 229151^178
Fitton J G Saunders A D Norry M J Hardarson B S ampTaylor R N (1997) Thermal and chemical structure of theIceland plume Earth and Planetary Science Letters 153 197^208
Fitton J G Saunders A D Kempton P D amp Hardarson B S(2003) Does depleted mantle form an intrinsic part of the Icelandplume Geochemistry Geophysics Geosystems 4(3) 1032 doi1010292002GC000424
Frey F A Huang S Blichert-Toft J Regelous M amp Boyet M(2005) Origin of depleted components in basalt related to theHawaiian hotspot Evidence from isotopic and incompatible ele-ment ratios Geochemistry Geophysics Geosystems 6(1) 23 doi1010292004GC000757
Galer S J G amp Abouchami W (1998) Practical application of leadtriple spiking for correction of instrumental mass discriminationMineralogical Magazine 62A 491^492
Ghiorso M S amp Sack R O (1995) Chemical mass transfer in mag-matic processes IV A revised and internally consistent thermody-namic model for the interpolation and extrapolation of liquid^solidequilibria in magmatic systems at elevated temperatures and pres-sures Contributions to Mineralogy and Petrology 119 197^212
Greene A R Scoates J S amp Weis D (2005)WrangelliaTerrane onVancouver Island Distribution of flood basalts with implicationsfor potential Ni^Cu^PGE mineralization In Grant B (ed)Geological Fieldwork 2004 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2005-1 209^220Greene A R Scoates J S Nixon G T amp Weis D (2006) Picritic
lavas and basal sills in the Karmutsen flood basalt provinceWrangellia northern Vancouver Island In Grant B (ed)Geological Fieldwork 2005 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2006-1 39^52Greene A R Scoates J S amp Weis D (2008) Wrangellia flood
basalts in Alaska A record of plume^lithosphere interaction in aLate Triassic accreted oceanic plateau Geochemistry Geophysics
Geosystems 9(Q12004) doi1010292008GC002092Greene A R Scoates J S Weis D amp Israel S (2009)
Geochemistry of triassic flood basalts from the Yukon (Canada)segment of the accreted Wrangellia oceanic plateau Lithosdoi101016jlithos200811010
Hauff F Hoernle K Tilton G Graham D W amp Kerr A C(2000) Large volume recycling of oceanic lithosphere over shorttime scales geochemical constraints from the Caribbean LargeIgneous Province Earth and Planetary Science Letters 174 247^263
Hauff F Hoernle K amp Schmidt A (2003) Sr^Nd^Pb compositionof Mesozoic Pacific oceanic crust (Site 1149 and 801 ODP Leg185)Implications for alteration of ocean crust and the input into theIzu^Bonin^Mariana subduction system Geochemistry Geophysics
Geosystems 4(8) doi1010292002GC000421Herzberg C (2004) Partial melting below the Ontong Java Plateau
In Fitton J G Mahoney J J Wallace P J amp Saunders A D(eds) Origin and Evolution of the Ontong Java Plateau Geological SocietyLondon Special Publications 229 179^183
Herzberg C (2006) Petrology and thermal structure of the Hawaiianplume from Mauna Kea volcano Nature 444(7119) 605
Herzberg C amp Asimow P D (2008) Petrology of some oceanicisland basalts PRIMELT2XLS software for primary magma cal-culation Geochemistry Geophysics Geosystems 9(Q09001) doi1010292008GC002057
Herzberg C amp OrsquoHara M J (2002) Plume-associated ultramaficmagmas of Phanerozoic age Journal of Petrology 43(10) 1857^1883
Herzberg C Asimow P D Arndt N Niu Y Lesher C MFitton J G Cheadle M J amp Saunders A D (2007)Temperatures in ambient mantle and plumes Constraints frombasalts picrites and komatiites Geochemistry Geophysics Geosystems8(Q02006) doi1010292006GC001390
Huang S amp Frey F A (2005) Trace element abundances of MaunaKea basalt from phase 2 of the Hawaii Scientific Drilling Project
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
35
Petrogenetic implications of correlations with major element con-tent and isotopic ratios Geochemistry Geophysics Geosystems 4(6)doi1010292002GC000322
Hussong D M Wipperman L K amp Kroenke L W (1979) Thecrustal structure of the Ontong Java and Manihiki OceanicPlateaus Journal of Geophysical Research 84(B11) 6003^6010
Jerram D A amp Widdowson M (2005) The anatomy of ContinentalFlood Basalt Provinces geological constraints on the processes andproducts of flood volcanism Lithos 79(3^4) 385^405
Jones D L Silberling N J amp Hillhouse J (1977)Wrangellia a dis-placed terrane in northwestern North America CanadianJournal ofEarth Sciences 14(11) 2565^2577
Kerr A C (2003) Oceanic Plateaus In Rudnick R (ed) The Crust
Treatise on GeochemistryVol 3 Oxford Elsevier Science pp 537^565Kerr A C amp Mahoney J J (2007) Oceanic plateaus Problematic
plumes potential paradigms Chemical Geology 241 332^353Kerr A CTarney J Marriner G F Klaver GT Saunders A D
amp Thirlwall M F (1996) The geochemistry and petrogenesis ofthe Late-Cretaceous picrites and basalts of Curacao NetherlandsAntilles a remnant of an oceanic plateau Contributions to
Mineralogy and Petrology 124(1) 29^43Kerr A C Marriner G F Tarney J Nivia A Saunders A D
Thirlwall M F amp Sinton A C (1997) Cretaceous basaltic ter-ranes in Western Colombia Elemental chronological and Sr-Ndisotopic constraints on petrogenesis Journal of Petrology 38(6)677^702
Kinzler R J (1997) Melting of mantle peridotite at pressuresapproaching the spinel to garnet transition Application to mid-ocean ridge basalt petrogenesis Journal of Geophysical Research
102(B1) 853^874Kinzler R J amp Grove T L (1992) Primary magmas of mid-ocean
ridge basalts 1 Experiments and methods Journal of Geophysical
Research 97(B5) 6885^6906Korenaga J (2005) Why did not the Ontong Java Plateau form sub-
aerially Earth and Planetary Science Letters 234 385^399Kuniyoshi S (1972) Petrology of the Karmutsen (Volcanic) Group
northeastern Vancouver Island British Columbia PhD disserta-tion University of California Los Angeles 242 pp
Lassiter J C DePaolo D J amp Mahoney J J (1995) Geochemistry ofthe Wrangellia flood basalt province Implications for the role ofcontinental and oceanic lithosphere in flood basalt genesis Journalof Petrology 36(4) 983^1009
LeBas M J (2000) IUGS reclassification of the high-Mg and picriticvolcanic rocks Journal of Petrology 41(10) 1467^1470
Longhi J (2002) Some phase equilibrium systematics of lherzolitemelting I Geochemistry Geophysics Geosystems 3(3) doi1010292001GC000204
MacDonald G A amp Katsura T (1964) Chemical composition ofHawaiian lavas Journal of Petrology 5 82^133
Mahoney J J (1987) An isotopic survey of Pacific oceanic plateausimplications for their nature and origin In Keating B HFryer P Batiza R amp Boehlert GW (eds) Seamounts Islands andAtolls American Geophysical Union Geophysical Monograph 43 207^220
Mahoney J LeRoex A P Peng Z Fisher R L amp Natland J H(1992) Southwestern limits of Indian Ocean Ridge mantle and theorigin of low 206Pb204Pb mid-ocean ridge basalt isotope systema-tics of the central Southwest Indian Ridge (178^508E) Journal ofGeophysical Research 97(B13) 19771^19790
Mahoney J J Sinton J M Kurz M D MacDougall J DSpencer K J amp Lugmair GW (1994) Isotope and trace elementcharacteristics of a super-fast spreading ridge East Pacific rise13^238S Earth and Planetary Science Letters 121 173^193
Massey N W D (1995a) Geology and mineral resources of theAlberni^Nanaimo Lakes sheet Vancouver Island 92F1W 92F2Eand part of 92F7E British Columbia Ministry of Energy Mines and
Petroleum Resources Paper 1992-2 132 ppMassey N W D (1995b) Geology and mineral resources of the
Cowichan Lake sheet Vancouver Island 92C16 British Columbia
Ministry of Energy Mines and Petroleum Resources Paper 1992-3 112 ppMassey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005a) Digital Geology Map of British Columbia Tile NM9 MidCoast BC British Columbia Ministry of Energy and Mines Geofile
2005-2Massey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005b) Digital Geology Map of British Columbia Tile NM10Southwest BC British Columbia Ministry of Energy and Mines Geofile
2005-3McDonough W F amp Sun S (1995) The composition of the Earth
Chemical Geology 120 223^253Monger JW H amp Journeay J M (1994) Basement geology and tec-
tonic evolution of theVancouver region In Monger JW H (ed)Geology and Geological Hazards of the Vancouver Region Southwestern
British Columbia Geological Survey of Canada Bulletin 481 3^25Muller J E (1977) Geology of Vancouver Island Geological Survey of
Canada Open File Map 463Muller J E (1980)The Paleozoic Sicker Group of Vancouver Island British
Columbia Geological Survey of Canada Paper 79-30 22 ppMuller J E Northcote K E amp Carlisle D (1974) Geology and
mineral deposits of Alert Bay^Cape Scott map area VancouverIsland British Columbia Geological Survey of Canada Paper 74-8 77
Neal C R Mahoney J J Kroenke L W Duncan R A ampPetterson M G (1997) The Ontong^Java Plateau InMahoney J J amp Coffin M F (eds) Large Igneous Provinces
Continental Oceanic and Planetary FloodVolcanism American Geophysical
Union Geophysical Monograph 100 183^216Niu Y Collerson K D Batiza R Wendt J I amp Regelous M
(1999) Origin of enriched-typed mid-ocean ridge basalt at ridgesfar from mantle plumes The East Pacific Rise at 11820rsquoN Journalof Geophysical Research 104 7067^7088
Nixon G T amp Orr A J (2007) Recent revisions to the EarlyMesozoic stratigraphy of Northern Vancouver Island (NTS 102I092L) and metallogenic implicatinos British Columbia InGrant B (ed) Geological Fieldwork 2006 British Columbia Ministry of
Energy Mines and Petroleum Resources Paper 2007-1 163^177Nixon GT Hammack J L KoyanagiV M Payie G J Haggart
JW Orchard M JTozerT Friedman R M Archibald D APalfy J amp Cordey F (2006a) Geology of the Quatsino^PortMcNeill area northernVancouver Island British Columbia Ministry
of Energy Mines and Petroleum Resources Geoscience Map 2006-2 scale150 000
Nixon G T Kelman M C Stevenson D Stokes L A ampJohnston K A (2006b) Preliminary geology of the Nimpkishmap area (NTS 092L07) northern Vancouver Island BritishColumbia In Grant B amp Newell J M (eds) Geological Fieldwork2005 British Columbia Ministry of Energy Mines and Petroleum Resources
Paper 2006-1 135^152Nixon G T Laroque J Pals A Styan J Greene A R amp
Scoates J S (2008) High-Mg lavas in the Karmutsen flood basaltsnorthernVancouver Island (NTS 092L) Stratigraphic setting andmetallogenic significance In Grant B (ed) Geological Fieldwork
2007 British Columbia Ministry of Energy Mines and Petroleum
Resources Paper 2008-1 175^190Nowell G M Kempton P D Noble S R Fitton J G
Saunders A Mahoney J J amp Taylor R N (1998) High precisionHf isotope measurements of MORB and OIB by thermal ionisation
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
36
mass spectrometry insights into the depleted mantle Chemical
Geology 149 211^233Parrish R R amp McNicoll V J (1992) U^Pb age determinations
from the southern Vancouver Island area British Columbia InRadiogenic Age and Isotopic Studies Report 5 Geological Survey of
Canada Paper 91-2 79^86Peate DW (1997) The Parana^Etendeka Province In Coffin M F
amp Mahoney J (eds) Large Igneous Provinces Continental Oceanic andPlanetary Flood Volcanism American Geophysical Union Geophysical
Monograph 100 217^245Perfit M R Fornari D J Ridley W I Kirk P D Casey J
Kastens K A Reynolds J R Edwards M Desonie DShuster R amp Paradis S (1996) Recent volcanism in the Siqueirostransform fault picritic basalts and implications for MORBmagma genesis Earth and Planetary Science Letters 141(1^4) 91^108
Pretorius W Weis D Williams G Hanano D Kieffer B ampScoates J S (2006) Complete trace elemental characterization ofgranitoid (USGSG-2GSP-2) reference materials by high resolutioninductively coupled plasma-mass spectrometry Geostandards and
Geoanalytical Research 30(1) 39^54Regelous M Niu Y Wendt J I Batiza R Greig A amp
Collerson K D (1999) Variations in the geochemistry of magma-tism on the East Pacific Rise at 10830rsquoN since 800 ka Earth and
Planetary Science Letters 168 45^63Recurren villon S Chauvel C Arndt N T Pik R Martineau F
Fourcade S amp Marty B (2002) Heterogeneity of the Caribbeanplateau mantle source Sr O and He isotopic compositions of oli-vine and clinopyroxene from Gorgona Island Earth and Planetary
Science Letters 205(1^2) 91Rhodes J M (1996) Geochemical stratigraphy of lava flows samples
by the Hawaii Scientific Drilling Project Journal of Geophysical
Research 101(B5) 11729^11746Rhodes J M amp Vollinger M J (2004) Composition of basaltic lavas
sampled by phase-2 of the Hawaii Scientific Drilling ProjectGeochemical stratigraphy and magma types Geochemistry
Geophysics Geosystems 5(3) doi1010292002GC000434Richards M A Jones D L Duncan R A amp DePaolo D J (1991)
A mantle plume initiation model for the Wrangellia flood basaltand other oceanic plateaus Science 254 263^267
Salters V J M amp Hart S R (1991) The mantle sources of oceanridges islands and arcs the Hf-isotope connection Earth and
Planetary Science Letters 104 364^380Salters V J M amp Stracke A (2004) Composition of the depleted
SaltersV J amp WhiteW (1998) Hf isotope constraints on mantle evo-lution Chemical Geology 145 447^460
SanoT amp Yamashita S (2004) Experimental petrology of basementlavas from Ocean Drilling Program Leg 192 implications for dif-ferentiation processes in Ontong Java Plateau magmas InFitton J G Mahoney J JWallace P J amp Saunders A D (eds)Origin and Evolution of the Ontong Java Plateau Geological Society
London Special Publications 229 185^218Saunders A D (2005) Large igneous provinces Origin and environ-
mental consequences Elements 1 259^263Self S ThordarsonT amp Keszthely L (1997) Emplacement of conti-
nental flood basalt lava flows In Mahoney J J amp Coffin M F(eds) Large Igneous Provinces Continental Oceanic and Planetary Flood
Volcanism American Geophysical Union Geophysical Monograph 100381^410
Shaw D M (2000) Continuous (dynamic) melting theory revisitedCanadian Mineralogist 38(5) 1041^1063
Sluggett C L (2003) Uranium^lead age and geochemical constraintson Paleozoic and Early Mesozoic magmatism in WrangelliaTerrane Saltspring Island British Columbia BSc thesisUniversity of British ColumbiaVancouver 84 pp
Storey M Mahoney J J amp Saunders A D (1997) Cretaceousbasalts in Madagascar and the transition between plume and con-tinental lithospheric mantle sources In Mahoney J J ampCoffin M F (eds) Large Igneous Provinces Continental Oceanic and
Planetary Flood Volcanism Aamerican Geophysical Union Geophysical
Monograph 100 95^122Surdam R C (1967) Low-grade metamorphism of the Karmutsen
Group PhD dissertation University of California Los Angeles288 pp
Sutherland-Brown A Yorath C J Anderson R G amp Dom K(1986) Geological maps of southern Vancouver IslandLITHOPROBE I 92C10 11 14 16 92F1 2 7 8 Geological Survey ofCanada Open File 1272
Tejada M L G Mahoney J J Castillo P R Ingle S P Sheth HC amp Weis D (2004) Pin-pricking the elephant evidence on theorigin of the Ontong Java Plateau from Pb^Sr^Hf^Nd isotopiccharacteristics of ODP Leg 192 basalts In Fitton J GMahoney J J Wallace P J amp Saunders A D (eds) Origin and
Evolution of the Ontong Java Plateau Geological Society London Special
Publications 229 133^150Thompson P M E Kempton P D amp Kerr A C (2008) Evaluation
of the effects of alteration and leaching on Sm^Nd and Lu^Hf sys-tematics in submarine mafic rocks Lithos 104(1^4) 164^176
Thompson P M E Kempton P D White R V Kerr A CTarney J Saunders A D Fitton J G amp McBirney A (2003)Hf-Nd isotope constraints on the origin of the CretaceousCaribbean plateau and its relationship to the Galapagos plumeEarth and Planetary Science Letters 217(1^2) 59^75
Vervoort J D Patchett P Blichert-Toft J amp Albare de F (1999)Relationships between Lu^Hf and Sm^Nd isotopic systems in theglobal sedimentary system Earth and Planetary Science Letters 16879^99
Walter M J (1998) Melting of garnet peridotite and the origin ofkomatiite and depleted lithosphere Journal of Petrology 39(1) 29^60
Weis D Kieffer B Maerschalk C Barling J de Jong JWilliams G A Hanano D Mattielli N Scoates J SGoolaerts A Friedman R A amp Mahoney J B (2006) High-pre-cision isotopic characterization of USGS reference materials byTIMS and MC-ICP-MS Geochemistry Geophysics Geosystems
7(Q08006) doi1010292006GC001283Weis D Kieffer B Hanano D Silva I N Barling J
Pretorius W Maerschalk C amp Mattielli N (2007) Hf isotopecompositions of US Geological Survey reference materialsGeochemistry Geophysics Geosystems 8(Q06006) doi1010292006GC001473
Wheeler J O amp McFeely P (1991) Tectonic assemblage map of theCanadian Cordillera and adjacent part of the United States ofAmerica Geological Survey of Canada Map 1712A
WhiteW M Albare de F amp Tecurren louk P (2000) High-precision analy-sis of Pb isotope ratios by multi-collector ICP-MS Chemical Geology167 257^270
Yorath C J Sutherland Brown A amp Massey N W D (1999)LITHOPROBE southern Vancouver Island British Columbia Geological
Survey of Canada Bulletin 498 145 ppZou H (1998) Trace element fractionation during modal and
non-model dynamic melting and open-system melting Amathematical treatment Geochimica et Cosmochimica Acta 62(11)1937^1945
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
37
Zou H amp Reid M R (2001) Quantitative modeling of trace elementfractionation during incongruent dynamic melting Geochimica et
Cosmochimica Acta 65(1) 153^162
APPENDIX SAMPLEPREPARAT ION ANDANALYT ICAL METHODSOnly the freshest rocks in the field were sampled and onlythe least altered samples were selected for chemical andisotopic analysis based on thorough petrographic inspec-tion Fifty-nine of the 129 collected samples were crushed(400 g) into small pieces 52mm in diameter in aRocklabs hydraulic piston crusher between WC plates tominimize contamination The coarse-crush was mixedand 100 g was powdered in a planetary mill using agatejars and balls cleaned with quartz sand between samples
ActLabs analytical methodsThe major- and trace-element compositions of the whole-rock powders were determined at Activation LaboratoriesLtd (Actlabs) in Ancaster Ontario Analytical techniquesand detection limits are also available from Actlabs (httpwwwactlabscommethsub_code4erehtm) The analyticalmethod for each of the elements analyzed is indicated inTable 2 For the major elements a 02 g sample was mixedwith a lithium metaborate^lithium tetraborate mixtureand fused in a graphite crucible The molten mixture waspoured into a 5 HNO3 solution and shaken until dis-solved (30min) The samples were analyzed for majoroxides and selected trace elements by inductively coupledplasma optical emission spectrometry (ICP-OES) on acombination simultaneous^sequential Thermo Jarrell-AshEnviro II system Internal calibration was achieved usinga variety of international reference materials (eg W-2BIR-1 DNC-1) and independent control samplesAdditional trace elements were analyzed by both instru-mental neutron activation analysis (INAA) and induc-tively couple plasma mass spectrometry (ICP-MS) ForINAA15^25 g of sample was weighed into small polyeth-ylene vials and irradiated with a control international ref-erence material CANMET WMS-1 and NiCr flux wires ata thermal neutron flux of 71012 neutronscm2 per s in theMcMaster Nuclear Reactor Following a 7 day waitingperiod the samples were measured on an Ortec high-purity Ge detector linked to a Canberra Series 95 multi-channel analyzer Activities for each element weredecay- and weight-corrected and compared with a detectorcalibration developed from multiple international certifiedreference materials For ICP-MS 025 g of sample wasdigested in HF followed by HNO3^HClO4 heated andtaken to dryness The samples were brought back into solu-tion with HCl Samples were analyzed using a PerkinElmer Optima 3000 ICP In-lab standards or certified ref-erence materials (eg W-2 BIR-1 DNC-1) were used for
quality control A total of 15 blind duplicates were ana-lyzed from the same sample powders to assessreproducibility
University of Massachusetts analyticalmethodsFourteen sample duplicate powders taken from the same100 g sample aliquot (high-MgO basalts and picrites)were also analyzed at the Ronald B Gilmore X-RayFluorescence (XRF) Laboratory at the University ofMassachusetts Major elements were measured on a fusedLa-bearing lithium borate glass disc using a SiemensMRS-400 spectrometer Trace element concentrations(Rb Sr Ba Ce Nb Zr Y Pb Zn Ga Ni Cr V) weremeasured on a separate powder pellet using a PhilipsPW2400 sequential spectrometer using a Rh tube LOIand ferrous iron measurements were made as describedby Rhodes amp Vollinger (2004) Precision and accuracy esti-mates for the data have been described by Rhodes (1996)and Rhodes amp Vollinger (2004) Results for each sampleare the average of two separate analyses and are shown inElectronic Appendix 2 for a comparison of analyses ofduplicate powders using XRFand ICP-MS
PCIGR analytical methodsA subset of 19 samples was selected for high-precisiontrace-element analysis and Sr Nd Pb and Hf isotopicanalysis at the Pacific Centre for Isotopic andGeochemical Research (PCIGR) at the University ofBritish Columbia (UBC) Samples were selected from the59 samples analyzed for whole-rock chemistry at ActLabsbased on major- and trace-element chemistry alteration(low LOI and petrographic alteration index) and samplelocation Samples were prepared for trace-element analysisat the PCIGR by the technique described by Pretorius et al(2006) on unleached rock powders Sample powders(100mg) were weighed in 7ml screw-top Savillex bea-kers and dissolved in 1ml 14N HNO3 and 5ml 48 HFon a hotplate for 48 h at 1308C with periodic ultrasonica-tion Samples were dried and redissolved in 6ml 6N HClon a hotplate for 24 h and then dried and redissolved in1ml concentrated HNO3 for 24 h before final dryingTrace element abundances were measured by high-resolution inductively coupled plasma-mass spectrometry(HR-ICP-MS) with aThermo Finnigan Element2 systemfollowing the procedures described by Pretorius et al(2006) within 24 h of redissolution HFSE and LILE weremeasured in medium-resolution mode at 2000 dilutionusing a PFATeflon spray chamber washed with aqua regiafor 3min between samples REE were measured in high-resolution mode and U and Pb in low-resolution modeat 2000 dilution using a glass spray chamber washedwith 2 HNO3 between samples Total procedural blanksand reference materials (BCR-2 BHVO-2) were analyzedwith the batch of samples Indium was used as an internal
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
38
standard in all samples and standard solutionsBackground and standard solutions were analyzed afterevery five samples to detect memory effects and massdrift Results for PCIGR trace-element analyses areshown in Electronic Appendix 3 for the full set of analyzedtrace elementsSample digestion for purification of Sr Nd Hf and Pb
for column chemistry began by weighing each samplepowder (400^500mg) prior to leaching All samples wereleached with 6N HCl and placed in an ultrasonic bath for15min Samples were rinsed twice with 18 M-cm H2Obetween each leaching step (15 total) until the supernatantliquid was clear [following the method of Mahoney (1987)]Samples were then dried on a hotplate for 24 h andweighed again Sample solutions were then prepared bydissolving 100^250mg of the leached powder dissolvedin 1ml 14N HNO3 and 10ml 48 HF on a hotplate for48 h at 1308C with periodic ultrasonication Samples weredried and redissolved in 6ml 6N HCl on a hotplate for24 h and then dried Pb was separated using anion exhangecolumns and the discard was used for Sr REE and Hfseparation Nd was separated from the other REE and Hfrequired two additional purification steps Detailed proce-dures for column chemistry for separating Sr Nd and Pbat the PCIGR have been described byWeis et al (2006) andHf purification has been described byWeis et al (2007) Srand Nd isotope ratios were measured by thermal ioniza-tion mass spectrometry (TIMS) on a Thermo FinniganTriton system in static mode with relay matrix rotation ona single Ta and double Re^Ta filament respectively Fourto five filaments per barrel of 21 were occupied by stan-dards (NBS 987 for Sr and LaJolla for Nd) for each barrelwhere samples were run Sample Sr and Nd isotopic com-positions were corrected for mass fractionation using86Sr88Srfrac14 01194 and 146Nd144Ndfrac14 07219 Each samplewas then normalized using the barrel average of the refer-ence material relative to the values of143Nd144Ndfrac14 0511858 and 87Sr86Srfrac14 0710248 (Weiset al 2006) During the course of the Vancouver Islandanalyses the La Jolla Nd standard gave an average valueof 05118566 (nfrac14 7) and NBS987 standard gave an aver-age of 07102408 (nfrac1411) (2 error is reported as times106) 147Sm144Nd errors are 15 or 0006 USGeological Survey (USGS) reference material BHVO-2was processed with the samples and yielded Sr and Nd iso-topic ratios of 07034607 and 05129786 respectivelyThese are in agreement with the published values of070347920 and 051298411 respectively (Weis et al2006)
Pb and Hf isotopic compositions were analyzed by mul-tiple collector-inductively coupled plasma-mass spectrome-try (MC-ICP-MS) on a static multi-collection on a NuPlasma (Nu Instruments) system The detailed analyticalprocedure for Pb isotopic analyses on the Nu Plasma atthe PCIGR has been described by Weis et al (2006) Theconfiguration for Pb analyses allows for collection of PbTl and Hg togetherTl and Hg are used to monitor instru-mental mass discrimination and isobaric overlap repec-tively All sample solutions were analyzed withapproximately the same PbTl ratio (4) as the referencematerial NIST SRM 981 To accomplish this a small ali-quot of each sample solution from the Pb columns was ana-lyzed on the Element2 to determine the precise amount ofPb available for analysis on the Nu Plasma The SRM 981standard was run after every two samples on the NuPlasma During the time samples were run analyses ofthe SRM 981 Pb reference material gave values of206Pb204Pbfrac1416940322 207Pb204Pbfrac1415495823and 208Pb204Pbfrac14 36713164 (nfrac14 61 2 error is reportedas times 104) these values are in excellent agreement withreported TIMS triple-spike values of Galer amp Abouchami(1998) Fractionation-corrected Pb isotopic ratios were fur-ther corrected by the sample-standard bracketing methodor the ln^ln correction method described by White et al(2000) and Blichert-Toft et al (2003) Leached powders ofUSGS reference material BHVO-2 yielded Pb isotopicratios of 206Pb204Pbfrac14186454 8 207Pb204Pbfrac14 1549105 and 208Pb204Pbfrac14 38222514 and BCR-2yielded 206Pb204Pbfrac141880466 207Pb204Pbfrac14 1562518 and 208Pb204Pbfrac14 3883496 (2 error isreported as times 104) These values are in agreement withthose for leached residues of BHVO-2 and BCR-2 fromWeis et al (2006)Hf isotopic compositions were analyzed following the
procedures detailed byWeis et al (2007) The configurationfor Hf analyses monitored Lu mass 175 andYb mass 172 toallow for interference correction to masses 174 and 176 Hfisotopic ratios were normalized internally for mass fractio-nation to a 179Hf177Hf ratio of 07325 using an exponentialcorrection Standards were run after every two samplesand sample results were normalized to the ratio of the in-run daily average and a 176Hf177Hf ratio for JMC-475 of0282160 During the course of analyses the Hf standardJMC-475 gave an average value 02821679 (nfrac14130)USGS reference materials BCR-2 and BHVO-2 were pro-cessed with the samples and yielded Hf isotopic ratios of02828675 and 02831005 respectively Publishedvalues for BCR-2 and BHVO-2 are 02828717 and0283104 8 respectively (Weis et al 2007)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
39
with olivine being produced during the reactioncpxthornopxthorn spfrac14meltthornol for spinel peridotites The mod-eling used coefficients for incongruent melting reactionsbased on experiments on lherzolite melting (eg Kinzleramp Grove 1992 Longhi 2002) and was separated into (1)melting of spinel lherzolite (2) two combined intervals ofmelting for a garnet lherzolite (gt lherz) and spinel lher-zolite (sp lherz) source (3) melting of garnet lherzoliteThe model solutions are non-unique and a range indegree of melting partition coefficients melt reactioncoefficients and proportion of mixed source componentsis possible to achieve acceptable solutions (see Fig 14 cap-tion for description of modeling parameters anduncertainties)The LREE-depleted high-MgO lavas require melting of
a depleted spinel lherzolite source (Fig 14a) The modelingresults indicate a high degree of melting (22^25) similarto the results from PRIMELT1 based on major-elementvariations (discussed above) of a LREE-depleted sourceThe high-MgO lavas lack a residual garnet signature Theenriched tholeiitic basalts involved melting of both garnetand spinel lherzolite and represent aggregate melts pro-duced from continuous melting throughout the depth ofthe melting column (Fig 14b) Trace-element concentra-tions in the tholeiitic and picritic lavas are not consistentwith low-degree melting of garnet lherzolite alone(Fig 14c) The modeling results indicate that the aggregatemelts that formed the tholeiitic basalts would haveinvolved lesser proportions of enriched small-degreemelts (1^6 melting) generated at depth and greateramounts of depleted high-degree melts (12^25) gener-ated at lower pressure (a ratio of garnet lherzolite tospinel lherzolite melt of 14 provides the best fits seeFig 14b and description in figure caption) The totaldegree of melting for the tholeiitic basalts was high(23^27) similar to the degree of partial melting in thespinel lherzolite facies for the primary melts that eruptedas the Keogh Lake picrites of a source that was also prob-ably depleted in LREE
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
(c)
1
10
100
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
Melting of garnet lherzolite
High-MgO lavas
Melting of spinel lherzolite
source (07DM+03PM)
source (07DM+03PM)
6 melting
30 melting
(b)
(a)
30 melting
Sam
ple
Cho
ndrit
eS
ampl
eC
hond
rite 20 melting
1 melting
Melting of garnet lherzoliteand spinel lherzolite
x gt lherz + 4x sp lherz
5 meltingx gt lherz + 4x sp lherz
Sam
ple
Cho
ndrit
e
Tholeiitic lavas
Tholeiitic lavas
High-MgO lavas
source (07DM+03PM)
Fig 14 Trace-element modeling results for incongruent dynamicmantle melting for picritic and tholeiitic lavas from the KarmutsenFormation Three steps of modeling are shown (a) melting of spinellherzolite compared with high-MgO lavas (b) melting of garnet lher-zolite and spinel lherzolite compared with tholeiitic lavas (c) meltingof garnet lherzolite compared with tholeiitic and picritic lavasShaded fields in each panel for tholeiitic basalts and picrites arelabeled Patterns with symbols in each panel are modeling results(patterns are 5 melting increments in (a) and (b) and 1 incre-ments in (c) PM primitive mantle DM depleted MORB mantle gtlherz garnet lherzolite sp lherz spinel lherzolite In (b) the ratios ofper cent melting for garnet and spinel lherzolite are indicated (x gtlherzthorn 4x sp lherz means that for 5 melting 1 melt in garnetfacies and 4 melt in spinel facies where x is per cent melting in theinterval of modeling) The ratio of the respective melt proportions inthe aggregate melt (x gt lherzthorn 4x sp lherz) was kept constant andthis ratio of melt contribution provided the best fit to the data Arange of proportion of source components (07DMthorn 03PM to
09DMthorn 01PM) can reproduce the variation in the high-MgOlavas Melting modeling uses the formulation of Zou amp Reid (2001)an example calculation is shown in their Appendix PM fromMcDonough amp Sun (1995) and DM from Salters amp Stracke (2004)Melt reaction coefficients of spinel lherzolite from Kinzler amp Grove(1992) and garnet lherzolite fromWalter (1998) Partition coefficientsfrom Salters amp Stracke (2004) and Shaw (2000) were kept constantSource mineralogy for spinel lherzolite (018cpx027opx052ol003sp) from Kinzler (1997) and for garnet lherzolite(034cpx008opx053ol005gt) from Salters amp Stracke (2004) Arange of source compositions for melting of spinel lherzolite(07DMthorn 03PM to 03DMthorn 07PM) reproduce REE patternssimilar to those of the high-MgO lavas with best fits for 22ndash25melting and 07DMthorn 03PM source composition Concentrationsof tholeiitic basalts cannot be reproduced from an entirelyprimitive or depleted source
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
30
The uncertainty in the composition of the source in thespinel lherzolite melting model for the high-MgO lavasprecludes determining whether the source was moredepleted in incompatible trace elements than the source ofthe tholeiitic lavas (see Fig 14 caption for description) It ispossible that early high-pressure low-degree melting left aresidue within parts of the plume (possibly the hotter inte-rior portion) that was depleted in trace elements (egElliott et al 1991) further decompression and high-degreemelting of such depleted regions could then have generatedthe depleted high-MgO Karmutsen lavas Shallow high-degree melting would preferentially sample depletedregions whereas deeper low-degree melting may notsample more refractory depleted regions (eg CaribbeanEscuder-Viruete et al 2007) The picrites lie at the highend of the range of Nd isotopic compositions and havelow NbZr similar to depleted Icelandic basalts (Fittonet al 2003) therefore the picrites may represent the partsof the mantle plume that were the most depleted prior tothe initiation of melting As Fitton et al (2003) suggestedthese depleted melts may only rarely be erupted withoutcombining with more voluminous less depleted melts andthey may be detected only as undifferentiated small-volume flows like the Keogh Lake picrites within theKarmutsen Formation on Vancouver IslandDecompression melting within the mantle plume initiatedwithin the garnet stability field (35^25 GPa) at highmantle potential temperatures (Tp414508C) and pro-ceeded beneath oceanic arc lithosphere within the spinelstability field (25^09 GPa) where more extensivedegrees of melting could occur
Magmatic evolution of Karmutsentholeiitic basaltsPrimary magmas in large igneous provinces leave exten-sive coarse-grained cumulate residues within or below thecrust these mafic and ultramafic plutonic sequences repre-sent a significant proportion of the original magma thatpartially crystallized at depth (eg Farnetani et al 1996)For example seismic and petrological studies of OntongJava (eg Farnetani et al 1996) combined with the use ofMELTS (Ghiorso amp Sack 1995) indicate that the volcanicsequence may represent only 40 of the total magmareaching the Moho Neal et al (1997) estimated that30^45 fractional crystallization took place for theOntong Java magmas and produced cumulate thicknessesof 7^14 km corresponding to 11^22 km of flood basaltsdepending on the presence of pre-existing oceanic crustand the total crustal thickness (25^35 km) Interpretationsof seismic velocity measurements suggest the presence ofpyroxene and gabbroic cumulates 9^16 km thick beneaththe Ontong Java Plateau (eg Hussong et al 1979)Wrangellia is characterized by high crustal velocities in
seismic refraction lines which is indicative of mafic
plutonic rocks extending to depth beneath VancouverIsland (Clowes et al 1995)Wrangellia crust is 25^30 kmthick beneath Vancouver Island and is underlain by astrongly reflective zone of high velocity and density thathas been interpreted as a major shear zone where lowerWrangellia lithosphere was detached (Clowes et al 1995)The vast majority of the Karmutsen flows are evolved tho-leiitic lavas (eg low MgO high FeOMgO) indicating animportant role for partial crystallization of melts at lowpressure and a significant portion of the crust beneathVancouver Island has seismic properties that are consistentwith crystalline residues from partially crystallizedKarmutsen magmas To test the proportion of the primarymagma that fractionated within the crust MELTS(Ghiorso amp Sack 1995) was used to simulate fractionalcrystallization using several estimated primary magmacompositions from the PRIMELT1 modeling results Themajor-element composition of the primary magmas couldnot be estimated for the tholeiitic basalts because of exten-sive plagthorn cpxthornol fractionation and thus the MELTSmodeling of the picrites is used as a proxy for the evolutionof major elements in the volcanic sequence A pressure of 1kbar was used with variable water contents (anhydrousand 02wt H2O) calculated at the quartz^fayalite^magnetite (QFM) oxygen buffer Experimental resultsfrom Ontong Java indicate that most crystallizationoccurs in magma chambers shallower than 6 km deep(52 kbar Sano amp Yamashita 2004) however some crys-tallization may take place at greater depths (3^5 kbarFarnetani et al 1996)A selection of the MELTS results are shown in Fig 15
Olivine and spinel crystallize at high temperature until15^20wt of the liquid mass has fractionated and theresidual magma contains 9^10wt MgO (Fig 15f) At1235^12258C plagioclase begins crystallizing and between1235 and 11908C olivine ceases crystallizing and clinopyr-oxene saturates (Fig 15f) As expected the addition of clin-opyroxene and plagioclase to the crystallization sequencecauses substantial changes in the composition of the resid-ual liquid FeO and TiO2 increase and Al2O3 and CaOdecrease while the decrease in MgO lessens considerably(Fig 15) The near-vertical trends for the Karmutsen tho-leiitic basalts especially for CaO do not match the calcu-lated liquid-line-of-descent very well which misses manyof the high-CaO high-Al2O3 low-FeO basalts A positivecorrelation between Al2O3CaO and SrNd indicates thataccumulation of plagioclase played a major role in the evo-lution of the tholeiitic basalts (Fig 15g) Increasing thecrystallization pressure slightly and the water content ofthe melts does not systematically improve the match ofthe MELTS models to the observed dataThe MELTS results indicate that a significant propor-
tion of the crystallization takes place before plagioclase(15^20 crystallization) and clinopyroxene (35^45
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
31
crystallization) join the crystallization sequence Theseresults suggest that445 crystallization of the primarymagmas occurred (estimated from the residual liquid percent in Fig 15f) and if the original flood basalt stratigra-phy on Vancouver Island was 6 km thick at least anequivalent amount of complementary plutonic rocksshould have crystallized The Karmutsen basalts and theirplutonic residues thus represent a significant addition ofcrust (perhaps 412 km thickness) previously thickenedduring the formation of the underlying Paleozoic arc
Growth of the Wrangellia oceanic plateauon Vancouver IslandThe extrusive part of the Wrangellia oceanic plateau onVancouver Island was constructed in a three-layersequence of mostly tholeiitic basalts with a restrictedrange of composition (Fig 16) The plateau overliesDevonian to Mississippian arc volcanic sequences andmarine sediments of Mississippian to Permian ages thatare intruded by gabbroic rocks related to the flood basaltsThe volcanic stratigraphy of theWrangellia plateau formed
00
05
10
15
20
25
30
35
40
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
16
0 2 4 6 8 10 12 14 16 18 20
10
11
12
13
14
15
16
17
18
19
20
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
0 2 4 6 8 10 12 14 16 18 20
MgO (wt)
0 10 20 30 40 50
percent of total mass
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
) Mineral proportions
Sp (e)
Ol
CpxPlag
PicriteHigh-MgO basaltTholeiitic basalt
TiO2 (wt)
MgO (wt)
Al2O3 (wt) CaO (wt)
FeO (wt)
MgO(a) (b)
(c) (d)
MgO (wt)MgO (wt)
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
)
Residual liquid ()
1220
1190
Ol + Sp
Plag+Ol+Sp Plag+Cpx+Sp
02 wt H2O
no H2O
Residual liquid and crystallizing phases
(f)
0102030405060708090100
plagioclaseaccumulation
plagioclaseaccumulation
plagioclase
accumulatio
n
10
12
14
16
0 10 20 30 40
Al2O3CaO
SrNd
tholeiitic basalts
(g)
Fig 15 Forward fractional crystallization modeling results for major elements from MELTS (Ghiorso amp Sack 1995) compared with picriticand tholeiitic lavas from the Karmutsen Formation (a^d) Major-element compositions of residual liquids compared with Karmutsen lavacompositions For clarity the results are shown only for one starting composition (open circle) using the composition of the estimated primarymagma for the high-MgO lava series from the modeling technique of Herzberg et al (2007) for sample 4723A13 as shown in Fig 12 The esti-mated primary magmas for samples 4723A3 and 93G171 yield similar results A pressure of 1 kbar was used with no H2O and 02wt H2Ocalculated at QFM black line is anhydrous and grey line is 02wt H2OThe end-point for the anhydrous result is at 78 fractionation and11308C and the end-point for the result with 02wt H2O is 75 fractionation and11108C (e) Mineral proportions of crystallized phases fromone MELTrun (sample 4723A13) with no H2O (f) Per cent residual liquid vs temperature labelled with intervals of crystallizing phases (g)Al2O3CaO vs SrNd for the tholeiitic basalts
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
32
in a single eruptive phase (c 231^225 Ma) as an emergentbasalt sequence during a deeper-water shallow-water andsubaerial stage and subsided below sea level relativelysoon after its formation (Fig 16) The deeper-water stagecontains pillowed and massive submarine flows (3000mthick) that built up the submarine volcanic edifice from theseafloor The lowest pillow basalts were emplaced onunconsolidated and lithified fine-grained siliciclastic andcarbonaceous sediments containing Daonella beds(c 235^232 Ma) that are intruded by abundantKarmutsen mafic sills (Fig 16a) The different pillowedand massive submarine flows resulted primarily from dif-ferent effusive rates and local topography Some of themassive submarine flows commonly recognized by radiat-ing columnar jointing represent master tubes for the deliv-ery of lava to distal parts of submarine flow fields Thepillowed flows which commonly have uneven topography
in modern sequences may have served as levees to themassive submarine flows rather than allowing unob-structed spreading and inflation typical of the massive sub-aerial flows that form on shallow slopes in continentalflood basalt provinces (eg Self et al 1997) The rapid suc-cession of flows is evident by the near absence of sedimentbetween submarine flows Minor volumes of picritic pillowlavas occur in the upper part of the pillowed flow sequencein areas of the plateau exposed on northern VancouverIsland (Fig 16 Nixon et al 2008)The growth of the plateau in shallow water led to
emplacement of an increasing proportion of volcaniclastics(pillow breccias and hyaloclastites) that accumulated up to1500m thick in some areas (Fig 16b Nixon et al 2008)Horizons of pillowed flows occur locally within the volca-niclastic stratigraphy Sedimentary structures (eg gradedbedding fluidization structures) and petrographic textures
Late Devonian-Mississippian arc
Daonella beds
Middle Triassic
deeper water
Late Devonian-Mississippian arc
pillowed flows massive submarine
flows
pillow brecciahyaloclastiteshallow water
~4 km
~6 km
Late Triassic limestoneinterflow limestonelenses
Karmutsen Formation
mafic intrusions sediment-sill complex
Karmutsen Formation
(a) (b)
(c) (d)
subaerial flows
Late Devonian-Mississippian arc Late Devonian-Mississippian arc
picritic pillow basalts
Miss to Permian marine sediments
sills
Fig 16 Growth of the Wrangellia oceanic plateau onVancouver Island (a) Deep-water eruptive phase of pillowed and massive submarineflows with almost no intervening sediments onto MiddleTriassic shale containing Daonella beds (c 235^232 Ma) The basement of the plateauis composed of Devonian to Mississippian arc sequences overlain by Mississippian to Permian carbonates and siliciclastic sediments(b) Shallow-water flow emplacement with increasing proportion of volcaniclastics resulted primarily from cooling^contraction granulationPicritic pillow basalts occur primarily near the transition between pillowed submarine flows and volcaniclastic units Pillowed and massiveflows occur locally within the volcaniclastic unit (c) Emergent stage of growth consists of overlapping subaerial flow fields similar to continen-tal flood basalts with pahoehoe flow structures Multi-tiered columnar jointing occurrs locally in response to ponding of flows (eg Bondre et al2004) Interflow carbonate deposits and plagioclase megacrystic flows developed locally during the waning stage of volcanism (d) Post-volcanicsubsidence and deposition of Late Carnian to Norian (c 212^223 Ma) platform carbonates on top of the plateau Units are labeled in the dia-grams Thicknesses of the underlying units are not to scale
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
33
within the volcaniclastic rocks indicate formation primar-ily via cooling^contraction granulation magma^water^steam interaction autobrecciation resedimentation andmass-wasting rather than pyroclastic eruption (Nixonet al 2008)A broad subaerial platform was constructed by emplace-
ment of subaerial lava flows during an emergent subaerialstage (Fig 16c) The subaerial unit predominantly com-prises well-layered sheet flows that formed in a similarway to continental flood basalts as hundred(s) of pahoehoeflow fields erupted on shallow slopes (Self et al 1997) how-ever volcaniclastic and pillowed flow units are locally pre-sent mostly in the upper parts of the stratigraphyLocalized ponding of flows led to the formation of multi-tiered columnar jointing in some subaerial flows (egBondre et al 2004) As volcanism waned local interflowcarbonate deposits developed and plagioclase megacrysticflows were erupted Subsidence of the plateau is recordedby the interflow lenses and deposits of platform carbonates(40^1500m thick) on top of the Karmutsen basalts thatcontain Late Carnian to Early Norian fossils (Fig 16d c216^227 Ma Nixon et al 2008)The evolution of the internal architecture of the
Wrangellia oceanic plateau was primarily related to theconditions of emplacement (ie water depth and eruptionenvironment) There does not appear to have been a sys-tematic coincident evolution or shift of lava compositionsonVancouver Island as has been recognized in most con-tinental flood basalts and parts of the Wrangellia oceanicplateau in Alaska and Yukon (Greene et al 2008 2009)In general with the exception of the picritic pillow lavasin the submarine stratigraphy and local variations betweenhigh-MgO and tholeiitic lavas there are no significantcompositional differences between lavas erupted duringthe different stages of growth of the plateau (ie submarineand subaerial flows) nor are substantial compositional dif-ferences generally found between massive submarine flowsin contact with pillowed flows tholeiitic basalt flows nearthe base of the Karmutsen Formation are compositionallysimilar to flows sampled near the top of the volcanicsequence Where it was possible to sample continuous sec-tions of volcanic stratigraphy greater than several hun-dreds of meters (eg Keogh^Maynard Lake area west ofthe Karmutsen Range Fig 2) there is no clear relationshipbetween stratigraphic position and compositionalvariation
CONCLUSIONSThe Karmutsen Formation covers extensive areas ofVancouver Island (20 000 km2) and forms part of a largeoceanic plateau erupted over a geologically short interval(c 225^231 Ma) The tripartite volcanic stratigraphyexposed on Vancouver Island is at least 6 km thick andcomprises a succession of submarine flows volcaniclastic
deposits and massive subaerial flowsThese volcanologicaldifferences are primarily related to the eruption environ-ment (deep-water shallow-water or subaerial) The rapidgrowth of the plateau prevented significant accumulationsof intervening sediment except in the uppermost stratigra-phy where isolated limestone lenses commonly associatedwith pillow basalt and hyaloclastite deposits preserve arecord of the final stages of subsidence of the plateauThe Wrangellia plateau on Vancouver Island was con-
structed dominantly of tholeiitic basalt with restrictedmajor- and trace-element and isotopic compositional var-iations Minor volumes of picritic pillow basalts eruptedlate in the submarine phase The Karmutsen basalts wereemplaced onto older Paleozoic arc volcanic sequences andlow NbLa for the picrites may indicate a contributionfrom the sub-arc lithospheric mantle Modeling of the pet-rogenesis of the Karmutsen picrites indicates melting ofanomalously hot mantle (15008C) and extensive degreesof partial melting (23^27) and is consistent with aplume initiation model Combined Sr^Nd^Hf^Pb isotopicsystematics indicate that the picritic and tholeiitic lavaswere derived from a depleted plume-type mantle sourcedistinct from the source of MORB that has compositionalsimilarities to the source of the lavas forming theCaribbean Plateau Picrites were generated by melting ofspinel lherzolite whereas the tholeiitic basalts formed asaggregate melts of both garnet and spinel lherzolite Thetholeiitic basalts underwent extensive low-pressure fractio-nation (52^3 kbar) and seismic studies indicate that someof theWrangellia crust beneathVancouver Island may cor-respond to the plutonic residues of the Karmutsen floodbasalts
ACKNOWLEDGEMENTSWe are grateful to Nick Arndt for helping us get this pro-ject started and Nick Massey for insights into VancouverIsland geology We greatly appreciate the assistance withPRIMELT modeling from Claude Herzberg We thankMikkel Schau for his insight and enthusiasm during field-work Jane Barling assisted with analyses by MC-ICP-MSThoughtful and thorough reviews of the manuscript byJohn Mahoney Fred Frey Wendy Bohrson DougalJerram and Teal Riley are greatly appreciated Fundingwas generously provided by the Rocks to Riches Programadministered by the BC amp Yukon Chamber of Mines in2004 by the BC Geological Survey and by NSERCDiscovery Grants to JSS and DW ARG was supportedby a University Graduate Fellowship at UBC
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
34
REFERENCESBarker F Brown A S Budahn J R amp Plafker G (1989) Back-arc
with frontal-arc component origin of Triassic Karmutsen basaltBritish Columbia Canada Chemical Geology 75 81^102
Blichert-Toft JWeis D Maerschalk C Agranier A amp Albare de F(2003) Hawaiian hot spot dynamics as inferred from the Hf and Pbisotope evolution of Mauna Kea volcano Geochemistry GeophysicsGeosystems 4(2) 1^27 doi1010292002GC000340
Bondre N R Duraiswami R A amp Dole G (2004) Morphologyand emplacement of flows from the Deccan Volcanic ProvinceIndia Bulletin of Volcanology 66 29^45
Carlisle D (1963) Pillow breccias and their aquagene tuffs QuadraIsland British Columbia Journal of Geology 71 48^71
Carlisle D (1972) Late Paleozoic to Mid-Triassic sedimentary-volcanic sequence on Northeastern Vancouver Island Geological
Society of Canada Paper Report of Activities 72-1 Part B 24^30Carlisle D amp SuzukiT (1974) Emergent basalt and submergent car-
bonate^clastic sequences including the Upper Triassic Dilleri andWelleri zones on Vancouver Island Canadian Journal of Earth
Sciences 11 254^279Chauvel C amp Blichert-Toft J (2001) A hafnium isotope and trace
element perspective on melting of the depleted mantle Earth and
Planetary Science Letters 190 137^151Chayes F (1954)The theory of thin-section analysis Journal of Geology
62 92^101Cho M Liou J G amp Maruyama S (1987) Transition from the zeo-
lite to prehnite^pumpellyite facies in the Karmutsen metabasitesVancouver Island British Columbia Journal of Petrology 27(2)467^494
Clowes R M Zelt C A Amor J R amp Ellis R M (1995)Lithospheric structure in the southern Canadian Cordillera froma network of seismic refraction lines Canadian Journal of Earth
Sciences 32(10) 1485^1513Coffin M F amp Eldholm O (1994) Large igneous provinces Crustal
structure dimensions and external consequences Reviews of
Geophysics 32(1) 1^36Cox K G (1980) A model for flood basalt volcanism Journal of
Petrology 21 629^650Eldholm O amp Coffin M F (2000) Large igneous provinces
and plate tectonics In Richards M A Gordon R G amp vander Hilst R D (eds) The History and Dynamics of Global
Plate Motions American Geophysical Union Geophysical Monograph 121309^326
Elliott T R Hawkesworth C J amp Grolaquo nvold K (1991) Dynamicmelting of the Iceland plume Nature 351 201^206
Escuder-Viruete J Pecurren rez-Estaucurren n A Contreras F Joubert MWeis D Ullrich T D amp Spadea P (2007) Plume mantle sourceheterogeneity through time Insights from the Duarte ComplexHispaniola northeastern Caribbean Journal of Geophysical Research112(B04203) doi1010292006JB004323
Farnetani C G Richards M A amp Ghiorso M S (1996)Petrological models of magma evolution and deep crustal structurebeneath hotspots and flood basalt provinces Earth and Planetary
Science Letters 143 81^94Fitton J G amp Godard M (2004) Origin and evolution of magmas
on the Ontong Java Plateau In Fitton J G Mahoney J JWallace P J amp Saunders A D (eds) Origin and Evolution of the
Ontong Java Plateau Geological Society London Special Publications 229151^178
Fitton J G Saunders A D Norry M J Hardarson B S ampTaylor R N (1997) Thermal and chemical structure of theIceland plume Earth and Planetary Science Letters 153 197^208
Fitton J G Saunders A D Kempton P D amp Hardarson B S(2003) Does depleted mantle form an intrinsic part of the Icelandplume Geochemistry Geophysics Geosystems 4(3) 1032 doi1010292002GC000424
Frey F A Huang S Blichert-Toft J Regelous M amp Boyet M(2005) Origin of depleted components in basalt related to theHawaiian hotspot Evidence from isotopic and incompatible ele-ment ratios Geochemistry Geophysics Geosystems 6(1) 23 doi1010292004GC000757
Galer S J G amp Abouchami W (1998) Practical application of leadtriple spiking for correction of instrumental mass discriminationMineralogical Magazine 62A 491^492
Ghiorso M S amp Sack R O (1995) Chemical mass transfer in mag-matic processes IV A revised and internally consistent thermody-namic model for the interpolation and extrapolation of liquid^solidequilibria in magmatic systems at elevated temperatures and pres-sures Contributions to Mineralogy and Petrology 119 197^212
Greene A R Scoates J S amp Weis D (2005)WrangelliaTerrane onVancouver Island Distribution of flood basalts with implicationsfor potential Ni^Cu^PGE mineralization In Grant B (ed)Geological Fieldwork 2004 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2005-1 209^220Greene A R Scoates J S Nixon G T amp Weis D (2006) Picritic
lavas and basal sills in the Karmutsen flood basalt provinceWrangellia northern Vancouver Island In Grant B (ed)Geological Fieldwork 2005 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2006-1 39^52Greene A R Scoates J S amp Weis D (2008) Wrangellia flood
basalts in Alaska A record of plume^lithosphere interaction in aLate Triassic accreted oceanic plateau Geochemistry Geophysics
Geosystems 9(Q12004) doi1010292008GC002092Greene A R Scoates J S Weis D amp Israel S (2009)
Geochemistry of triassic flood basalts from the Yukon (Canada)segment of the accreted Wrangellia oceanic plateau Lithosdoi101016jlithos200811010
Hauff F Hoernle K Tilton G Graham D W amp Kerr A C(2000) Large volume recycling of oceanic lithosphere over shorttime scales geochemical constraints from the Caribbean LargeIgneous Province Earth and Planetary Science Letters 174 247^263
Hauff F Hoernle K amp Schmidt A (2003) Sr^Nd^Pb compositionof Mesozoic Pacific oceanic crust (Site 1149 and 801 ODP Leg185)Implications for alteration of ocean crust and the input into theIzu^Bonin^Mariana subduction system Geochemistry Geophysics
Geosystems 4(8) doi1010292002GC000421Herzberg C (2004) Partial melting below the Ontong Java Plateau
In Fitton J G Mahoney J J Wallace P J amp Saunders A D(eds) Origin and Evolution of the Ontong Java Plateau Geological SocietyLondon Special Publications 229 179^183
Herzberg C (2006) Petrology and thermal structure of the Hawaiianplume from Mauna Kea volcano Nature 444(7119) 605
Herzberg C amp Asimow P D (2008) Petrology of some oceanicisland basalts PRIMELT2XLS software for primary magma cal-culation Geochemistry Geophysics Geosystems 9(Q09001) doi1010292008GC002057
Herzberg C amp OrsquoHara M J (2002) Plume-associated ultramaficmagmas of Phanerozoic age Journal of Petrology 43(10) 1857^1883
Herzberg C Asimow P D Arndt N Niu Y Lesher C MFitton J G Cheadle M J amp Saunders A D (2007)Temperatures in ambient mantle and plumes Constraints frombasalts picrites and komatiites Geochemistry Geophysics Geosystems8(Q02006) doi1010292006GC001390
Huang S amp Frey F A (2005) Trace element abundances of MaunaKea basalt from phase 2 of the Hawaii Scientific Drilling Project
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
35
Petrogenetic implications of correlations with major element con-tent and isotopic ratios Geochemistry Geophysics Geosystems 4(6)doi1010292002GC000322
Hussong D M Wipperman L K amp Kroenke L W (1979) Thecrustal structure of the Ontong Java and Manihiki OceanicPlateaus Journal of Geophysical Research 84(B11) 6003^6010
Jerram D A amp Widdowson M (2005) The anatomy of ContinentalFlood Basalt Provinces geological constraints on the processes andproducts of flood volcanism Lithos 79(3^4) 385^405
Jones D L Silberling N J amp Hillhouse J (1977)Wrangellia a dis-placed terrane in northwestern North America CanadianJournal ofEarth Sciences 14(11) 2565^2577
Kerr A C (2003) Oceanic Plateaus In Rudnick R (ed) The Crust
Treatise on GeochemistryVol 3 Oxford Elsevier Science pp 537^565Kerr A C amp Mahoney J J (2007) Oceanic plateaus Problematic
plumes potential paradigms Chemical Geology 241 332^353Kerr A CTarney J Marriner G F Klaver GT Saunders A D
amp Thirlwall M F (1996) The geochemistry and petrogenesis ofthe Late-Cretaceous picrites and basalts of Curacao NetherlandsAntilles a remnant of an oceanic plateau Contributions to
Mineralogy and Petrology 124(1) 29^43Kerr A C Marriner G F Tarney J Nivia A Saunders A D
Thirlwall M F amp Sinton A C (1997) Cretaceous basaltic ter-ranes in Western Colombia Elemental chronological and Sr-Ndisotopic constraints on petrogenesis Journal of Petrology 38(6)677^702
Kinzler R J (1997) Melting of mantle peridotite at pressuresapproaching the spinel to garnet transition Application to mid-ocean ridge basalt petrogenesis Journal of Geophysical Research
102(B1) 853^874Kinzler R J amp Grove T L (1992) Primary magmas of mid-ocean
ridge basalts 1 Experiments and methods Journal of Geophysical
Research 97(B5) 6885^6906Korenaga J (2005) Why did not the Ontong Java Plateau form sub-
aerially Earth and Planetary Science Letters 234 385^399Kuniyoshi S (1972) Petrology of the Karmutsen (Volcanic) Group
northeastern Vancouver Island British Columbia PhD disserta-tion University of California Los Angeles 242 pp
Lassiter J C DePaolo D J amp Mahoney J J (1995) Geochemistry ofthe Wrangellia flood basalt province Implications for the role ofcontinental and oceanic lithosphere in flood basalt genesis Journalof Petrology 36(4) 983^1009
LeBas M J (2000) IUGS reclassification of the high-Mg and picriticvolcanic rocks Journal of Petrology 41(10) 1467^1470
Longhi J (2002) Some phase equilibrium systematics of lherzolitemelting I Geochemistry Geophysics Geosystems 3(3) doi1010292001GC000204
MacDonald G A amp Katsura T (1964) Chemical composition ofHawaiian lavas Journal of Petrology 5 82^133
Mahoney J J (1987) An isotopic survey of Pacific oceanic plateausimplications for their nature and origin In Keating B HFryer P Batiza R amp Boehlert GW (eds) Seamounts Islands andAtolls American Geophysical Union Geophysical Monograph 43 207^220
Mahoney J LeRoex A P Peng Z Fisher R L amp Natland J H(1992) Southwestern limits of Indian Ocean Ridge mantle and theorigin of low 206Pb204Pb mid-ocean ridge basalt isotope systema-tics of the central Southwest Indian Ridge (178^508E) Journal ofGeophysical Research 97(B13) 19771^19790
Mahoney J J Sinton J M Kurz M D MacDougall J DSpencer K J amp Lugmair GW (1994) Isotope and trace elementcharacteristics of a super-fast spreading ridge East Pacific rise13^238S Earth and Planetary Science Letters 121 173^193
Massey N W D (1995a) Geology and mineral resources of theAlberni^Nanaimo Lakes sheet Vancouver Island 92F1W 92F2Eand part of 92F7E British Columbia Ministry of Energy Mines and
Petroleum Resources Paper 1992-2 132 ppMassey N W D (1995b) Geology and mineral resources of the
Cowichan Lake sheet Vancouver Island 92C16 British Columbia
Ministry of Energy Mines and Petroleum Resources Paper 1992-3 112 ppMassey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005a) Digital Geology Map of British Columbia Tile NM9 MidCoast BC British Columbia Ministry of Energy and Mines Geofile
2005-2Massey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005b) Digital Geology Map of British Columbia Tile NM10Southwest BC British Columbia Ministry of Energy and Mines Geofile
2005-3McDonough W F amp Sun S (1995) The composition of the Earth
Chemical Geology 120 223^253Monger JW H amp Journeay J M (1994) Basement geology and tec-
tonic evolution of theVancouver region In Monger JW H (ed)Geology and Geological Hazards of the Vancouver Region Southwestern
British Columbia Geological Survey of Canada Bulletin 481 3^25Muller J E (1977) Geology of Vancouver Island Geological Survey of
Canada Open File Map 463Muller J E (1980)The Paleozoic Sicker Group of Vancouver Island British
Columbia Geological Survey of Canada Paper 79-30 22 ppMuller J E Northcote K E amp Carlisle D (1974) Geology and
mineral deposits of Alert Bay^Cape Scott map area VancouverIsland British Columbia Geological Survey of Canada Paper 74-8 77
Neal C R Mahoney J J Kroenke L W Duncan R A ampPetterson M G (1997) The Ontong^Java Plateau InMahoney J J amp Coffin M F (eds) Large Igneous Provinces
Continental Oceanic and Planetary FloodVolcanism American Geophysical
Union Geophysical Monograph 100 183^216Niu Y Collerson K D Batiza R Wendt J I amp Regelous M
(1999) Origin of enriched-typed mid-ocean ridge basalt at ridgesfar from mantle plumes The East Pacific Rise at 11820rsquoN Journalof Geophysical Research 104 7067^7088
Nixon G T amp Orr A J (2007) Recent revisions to the EarlyMesozoic stratigraphy of Northern Vancouver Island (NTS 102I092L) and metallogenic implicatinos British Columbia InGrant B (ed) Geological Fieldwork 2006 British Columbia Ministry of
Energy Mines and Petroleum Resources Paper 2007-1 163^177Nixon GT Hammack J L KoyanagiV M Payie G J Haggart
JW Orchard M JTozerT Friedman R M Archibald D APalfy J amp Cordey F (2006a) Geology of the Quatsino^PortMcNeill area northernVancouver Island British Columbia Ministry
of Energy Mines and Petroleum Resources Geoscience Map 2006-2 scale150 000
Nixon G T Kelman M C Stevenson D Stokes L A ampJohnston K A (2006b) Preliminary geology of the Nimpkishmap area (NTS 092L07) northern Vancouver Island BritishColumbia In Grant B amp Newell J M (eds) Geological Fieldwork2005 British Columbia Ministry of Energy Mines and Petroleum Resources
Paper 2006-1 135^152Nixon G T Laroque J Pals A Styan J Greene A R amp
Scoates J S (2008) High-Mg lavas in the Karmutsen flood basaltsnorthernVancouver Island (NTS 092L) Stratigraphic setting andmetallogenic significance In Grant B (ed) Geological Fieldwork
2007 British Columbia Ministry of Energy Mines and Petroleum
Resources Paper 2008-1 175^190Nowell G M Kempton P D Noble S R Fitton J G
Saunders A Mahoney J J amp Taylor R N (1998) High precisionHf isotope measurements of MORB and OIB by thermal ionisation
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
36
mass spectrometry insights into the depleted mantle Chemical
Geology 149 211^233Parrish R R amp McNicoll V J (1992) U^Pb age determinations
from the southern Vancouver Island area British Columbia InRadiogenic Age and Isotopic Studies Report 5 Geological Survey of
Canada Paper 91-2 79^86Peate DW (1997) The Parana^Etendeka Province In Coffin M F
amp Mahoney J (eds) Large Igneous Provinces Continental Oceanic andPlanetary Flood Volcanism American Geophysical Union Geophysical
Monograph 100 217^245Perfit M R Fornari D J Ridley W I Kirk P D Casey J
Kastens K A Reynolds J R Edwards M Desonie DShuster R amp Paradis S (1996) Recent volcanism in the Siqueirostransform fault picritic basalts and implications for MORBmagma genesis Earth and Planetary Science Letters 141(1^4) 91^108
Pretorius W Weis D Williams G Hanano D Kieffer B ampScoates J S (2006) Complete trace elemental characterization ofgranitoid (USGSG-2GSP-2) reference materials by high resolutioninductively coupled plasma-mass spectrometry Geostandards and
Geoanalytical Research 30(1) 39^54Regelous M Niu Y Wendt J I Batiza R Greig A amp
Collerson K D (1999) Variations in the geochemistry of magma-tism on the East Pacific Rise at 10830rsquoN since 800 ka Earth and
Planetary Science Letters 168 45^63Recurren villon S Chauvel C Arndt N T Pik R Martineau F
Fourcade S amp Marty B (2002) Heterogeneity of the Caribbeanplateau mantle source Sr O and He isotopic compositions of oli-vine and clinopyroxene from Gorgona Island Earth and Planetary
Science Letters 205(1^2) 91Rhodes J M (1996) Geochemical stratigraphy of lava flows samples
by the Hawaii Scientific Drilling Project Journal of Geophysical
Research 101(B5) 11729^11746Rhodes J M amp Vollinger M J (2004) Composition of basaltic lavas
sampled by phase-2 of the Hawaii Scientific Drilling ProjectGeochemical stratigraphy and magma types Geochemistry
Geophysics Geosystems 5(3) doi1010292002GC000434Richards M A Jones D L Duncan R A amp DePaolo D J (1991)
A mantle plume initiation model for the Wrangellia flood basaltand other oceanic plateaus Science 254 263^267
Salters V J M amp Hart S R (1991) The mantle sources of oceanridges islands and arcs the Hf-isotope connection Earth and
Planetary Science Letters 104 364^380Salters V J M amp Stracke A (2004) Composition of the depleted
SaltersV J amp WhiteW (1998) Hf isotope constraints on mantle evo-lution Chemical Geology 145 447^460
SanoT amp Yamashita S (2004) Experimental petrology of basementlavas from Ocean Drilling Program Leg 192 implications for dif-ferentiation processes in Ontong Java Plateau magmas InFitton J G Mahoney J JWallace P J amp Saunders A D (eds)Origin and Evolution of the Ontong Java Plateau Geological Society
London Special Publications 229 185^218Saunders A D (2005) Large igneous provinces Origin and environ-
mental consequences Elements 1 259^263Self S ThordarsonT amp Keszthely L (1997) Emplacement of conti-
nental flood basalt lava flows In Mahoney J J amp Coffin M F(eds) Large Igneous Provinces Continental Oceanic and Planetary Flood
Volcanism American Geophysical Union Geophysical Monograph 100381^410
Shaw D M (2000) Continuous (dynamic) melting theory revisitedCanadian Mineralogist 38(5) 1041^1063
Sluggett C L (2003) Uranium^lead age and geochemical constraintson Paleozoic and Early Mesozoic magmatism in WrangelliaTerrane Saltspring Island British Columbia BSc thesisUniversity of British ColumbiaVancouver 84 pp
Storey M Mahoney J J amp Saunders A D (1997) Cretaceousbasalts in Madagascar and the transition between plume and con-tinental lithospheric mantle sources In Mahoney J J ampCoffin M F (eds) Large Igneous Provinces Continental Oceanic and
Planetary Flood Volcanism Aamerican Geophysical Union Geophysical
Monograph 100 95^122Surdam R C (1967) Low-grade metamorphism of the Karmutsen
Group PhD dissertation University of California Los Angeles288 pp
Sutherland-Brown A Yorath C J Anderson R G amp Dom K(1986) Geological maps of southern Vancouver IslandLITHOPROBE I 92C10 11 14 16 92F1 2 7 8 Geological Survey ofCanada Open File 1272
Tejada M L G Mahoney J J Castillo P R Ingle S P Sheth HC amp Weis D (2004) Pin-pricking the elephant evidence on theorigin of the Ontong Java Plateau from Pb^Sr^Hf^Nd isotopiccharacteristics of ODP Leg 192 basalts In Fitton J GMahoney J J Wallace P J amp Saunders A D (eds) Origin and
Evolution of the Ontong Java Plateau Geological Society London Special
Publications 229 133^150Thompson P M E Kempton P D amp Kerr A C (2008) Evaluation
of the effects of alteration and leaching on Sm^Nd and Lu^Hf sys-tematics in submarine mafic rocks Lithos 104(1^4) 164^176
Thompson P M E Kempton P D White R V Kerr A CTarney J Saunders A D Fitton J G amp McBirney A (2003)Hf-Nd isotope constraints on the origin of the CretaceousCaribbean plateau and its relationship to the Galapagos plumeEarth and Planetary Science Letters 217(1^2) 59^75
Vervoort J D Patchett P Blichert-Toft J amp Albare de F (1999)Relationships between Lu^Hf and Sm^Nd isotopic systems in theglobal sedimentary system Earth and Planetary Science Letters 16879^99
Walter M J (1998) Melting of garnet peridotite and the origin ofkomatiite and depleted lithosphere Journal of Petrology 39(1) 29^60
Weis D Kieffer B Maerschalk C Barling J de Jong JWilliams G A Hanano D Mattielli N Scoates J SGoolaerts A Friedman R A amp Mahoney J B (2006) High-pre-cision isotopic characterization of USGS reference materials byTIMS and MC-ICP-MS Geochemistry Geophysics Geosystems
7(Q08006) doi1010292006GC001283Weis D Kieffer B Hanano D Silva I N Barling J
Pretorius W Maerschalk C amp Mattielli N (2007) Hf isotopecompositions of US Geological Survey reference materialsGeochemistry Geophysics Geosystems 8(Q06006) doi1010292006GC001473
Wheeler J O amp McFeely P (1991) Tectonic assemblage map of theCanadian Cordillera and adjacent part of the United States ofAmerica Geological Survey of Canada Map 1712A
WhiteW M Albare de F amp Tecurren louk P (2000) High-precision analy-sis of Pb isotope ratios by multi-collector ICP-MS Chemical Geology167 257^270
Yorath C J Sutherland Brown A amp Massey N W D (1999)LITHOPROBE southern Vancouver Island British Columbia Geological
Survey of Canada Bulletin 498 145 ppZou H (1998) Trace element fractionation during modal and
non-model dynamic melting and open-system melting Amathematical treatment Geochimica et Cosmochimica Acta 62(11)1937^1945
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
37
Zou H amp Reid M R (2001) Quantitative modeling of trace elementfractionation during incongruent dynamic melting Geochimica et
Cosmochimica Acta 65(1) 153^162
APPENDIX SAMPLEPREPARAT ION ANDANALYT ICAL METHODSOnly the freshest rocks in the field were sampled and onlythe least altered samples were selected for chemical andisotopic analysis based on thorough petrographic inspec-tion Fifty-nine of the 129 collected samples were crushed(400 g) into small pieces 52mm in diameter in aRocklabs hydraulic piston crusher between WC plates tominimize contamination The coarse-crush was mixedand 100 g was powdered in a planetary mill using agatejars and balls cleaned with quartz sand between samples
ActLabs analytical methodsThe major- and trace-element compositions of the whole-rock powders were determined at Activation LaboratoriesLtd (Actlabs) in Ancaster Ontario Analytical techniquesand detection limits are also available from Actlabs (httpwwwactlabscommethsub_code4erehtm) The analyticalmethod for each of the elements analyzed is indicated inTable 2 For the major elements a 02 g sample was mixedwith a lithium metaborate^lithium tetraborate mixtureand fused in a graphite crucible The molten mixture waspoured into a 5 HNO3 solution and shaken until dis-solved (30min) The samples were analyzed for majoroxides and selected trace elements by inductively coupledplasma optical emission spectrometry (ICP-OES) on acombination simultaneous^sequential Thermo Jarrell-AshEnviro II system Internal calibration was achieved usinga variety of international reference materials (eg W-2BIR-1 DNC-1) and independent control samplesAdditional trace elements were analyzed by both instru-mental neutron activation analysis (INAA) and induc-tively couple plasma mass spectrometry (ICP-MS) ForINAA15^25 g of sample was weighed into small polyeth-ylene vials and irradiated with a control international ref-erence material CANMET WMS-1 and NiCr flux wires ata thermal neutron flux of 71012 neutronscm2 per s in theMcMaster Nuclear Reactor Following a 7 day waitingperiod the samples were measured on an Ortec high-purity Ge detector linked to a Canberra Series 95 multi-channel analyzer Activities for each element weredecay- and weight-corrected and compared with a detectorcalibration developed from multiple international certifiedreference materials For ICP-MS 025 g of sample wasdigested in HF followed by HNO3^HClO4 heated andtaken to dryness The samples were brought back into solu-tion with HCl Samples were analyzed using a PerkinElmer Optima 3000 ICP In-lab standards or certified ref-erence materials (eg W-2 BIR-1 DNC-1) were used for
quality control A total of 15 blind duplicates were ana-lyzed from the same sample powders to assessreproducibility
University of Massachusetts analyticalmethodsFourteen sample duplicate powders taken from the same100 g sample aliquot (high-MgO basalts and picrites)were also analyzed at the Ronald B Gilmore X-RayFluorescence (XRF) Laboratory at the University ofMassachusetts Major elements were measured on a fusedLa-bearing lithium borate glass disc using a SiemensMRS-400 spectrometer Trace element concentrations(Rb Sr Ba Ce Nb Zr Y Pb Zn Ga Ni Cr V) weremeasured on a separate powder pellet using a PhilipsPW2400 sequential spectrometer using a Rh tube LOIand ferrous iron measurements were made as describedby Rhodes amp Vollinger (2004) Precision and accuracy esti-mates for the data have been described by Rhodes (1996)and Rhodes amp Vollinger (2004) Results for each sampleare the average of two separate analyses and are shown inElectronic Appendix 2 for a comparison of analyses ofduplicate powders using XRFand ICP-MS
PCIGR analytical methodsA subset of 19 samples was selected for high-precisiontrace-element analysis and Sr Nd Pb and Hf isotopicanalysis at the Pacific Centre for Isotopic andGeochemical Research (PCIGR) at the University ofBritish Columbia (UBC) Samples were selected from the59 samples analyzed for whole-rock chemistry at ActLabsbased on major- and trace-element chemistry alteration(low LOI and petrographic alteration index) and samplelocation Samples were prepared for trace-element analysisat the PCIGR by the technique described by Pretorius et al(2006) on unleached rock powders Sample powders(100mg) were weighed in 7ml screw-top Savillex bea-kers and dissolved in 1ml 14N HNO3 and 5ml 48 HFon a hotplate for 48 h at 1308C with periodic ultrasonica-tion Samples were dried and redissolved in 6ml 6N HClon a hotplate for 24 h and then dried and redissolved in1ml concentrated HNO3 for 24 h before final dryingTrace element abundances were measured by high-resolution inductively coupled plasma-mass spectrometry(HR-ICP-MS) with aThermo Finnigan Element2 systemfollowing the procedures described by Pretorius et al(2006) within 24 h of redissolution HFSE and LILE weremeasured in medium-resolution mode at 2000 dilutionusing a PFATeflon spray chamber washed with aqua regiafor 3min between samples REE were measured in high-resolution mode and U and Pb in low-resolution modeat 2000 dilution using a glass spray chamber washedwith 2 HNO3 between samples Total procedural blanksand reference materials (BCR-2 BHVO-2) were analyzedwith the batch of samples Indium was used as an internal
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
38
standard in all samples and standard solutionsBackground and standard solutions were analyzed afterevery five samples to detect memory effects and massdrift Results for PCIGR trace-element analyses areshown in Electronic Appendix 3 for the full set of analyzedtrace elementsSample digestion for purification of Sr Nd Hf and Pb
for column chemistry began by weighing each samplepowder (400^500mg) prior to leaching All samples wereleached with 6N HCl and placed in an ultrasonic bath for15min Samples were rinsed twice with 18 M-cm H2Obetween each leaching step (15 total) until the supernatantliquid was clear [following the method of Mahoney (1987)]Samples were then dried on a hotplate for 24 h andweighed again Sample solutions were then prepared bydissolving 100^250mg of the leached powder dissolvedin 1ml 14N HNO3 and 10ml 48 HF on a hotplate for48 h at 1308C with periodic ultrasonication Samples weredried and redissolved in 6ml 6N HCl on a hotplate for24 h and then dried Pb was separated using anion exhangecolumns and the discard was used for Sr REE and Hfseparation Nd was separated from the other REE and Hfrequired two additional purification steps Detailed proce-dures for column chemistry for separating Sr Nd and Pbat the PCIGR have been described byWeis et al (2006) andHf purification has been described byWeis et al (2007) Srand Nd isotope ratios were measured by thermal ioniza-tion mass spectrometry (TIMS) on a Thermo FinniganTriton system in static mode with relay matrix rotation ona single Ta and double Re^Ta filament respectively Fourto five filaments per barrel of 21 were occupied by stan-dards (NBS 987 for Sr and LaJolla for Nd) for each barrelwhere samples were run Sample Sr and Nd isotopic com-positions were corrected for mass fractionation using86Sr88Srfrac14 01194 and 146Nd144Ndfrac14 07219 Each samplewas then normalized using the barrel average of the refer-ence material relative to the values of143Nd144Ndfrac14 0511858 and 87Sr86Srfrac14 0710248 (Weiset al 2006) During the course of the Vancouver Islandanalyses the La Jolla Nd standard gave an average valueof 05118566 (nfrac14 7) and NBS987 standard gave an aver-age of 07102408 (nfrac1411) (2 error is reported as times106) 147Sm144Nd errors are 15 or 0006 USGeological Survey (USGS) reference material BHVO-2was processed with the samples and yielded Sr and Nd iso-topic ratios of 07034607 and 05129786 respectivelyThese are in agreement with the published values of070347920 and 051298411 respectively (Weis et al2006)
Pb and Hf isotopic compositions were analyzed by mul-tiple collector-inductively coupled plasma-mass spectrome-try (MC-ICP-MS) on a static multi-collection on a NuPlasma (Nu Instruments) system The detailed analyticalprocedure for Pb isotopic analyses on the Nu Plasma atthe PCIGR has been described by Weis et al (2006) Theconfiguration for Pb analyses allows for collection of PbTl and Hg togetherTl and Hg are used to monitor instru-mental mass discrimination and isobaric overlap repec-tively All sample solutions were analyzed withapproximately the same PbTl ratio (4) as the referencematerial NIST SRM 981 To accomplish this a small ali-quot of each sample solution from the Pb columns was ana-lyzed on the Element2 to determine the precise amount ofPb available for analysis on the Nu Plasma The SRM 981standard was run after every two samples on the NuPlasma During the time samples were run analyses ofthe SRM 981 Pb reference material gave values of206Pb204Pbfrac1416940322 207Pb204Pbfrac1415495823and 208Pb204Pbfrac14 36713164 (nfrac14 61 2 error is reportedas times 104) these values are in excellent agreement withreported TIMS triple-spike values of Galer amp Abouchami(1998) Fractionation-corrected Pb isotopic ratios were fur-ther corrected by the sample-standard bracketing methodor the ln^ln correction method described by White et al(2000) and Blichert-Toft et al (2003) Leached powders ofUSGS reference material BHVO-2 yielded Pb isotopicratios of 206Pb204Pbfrac14186454 8 207Pb204Pbfrac14 1549105 and 208Pb204Pbfrac14 38222514 and BCR-2yielded 206Pb204Pbfrac141880466 207Pb204Pbfrac14 1562518 and 208Pb204Pbfrac14 3883496 (2 error isreported as times 104) These values are in agreement withthose for leached residues of BHVO-2 and BCR-2 fromWeis et al (2006)Hf isotopic compositions were analyzed following the
procedures detailed byWeis et al (2007) The configurationfor Hf analyses monitored Lu mass 175 andYb mass 172 toallow for interference correction to masses 174 and 176 Hfisotopic ratios were normalized internally for mass fractio-nation to a 179Hf177Hf ratio of 07325 using an exponentialcorrection Standards were run after every two samplesand sample results were normalized to the ratio of the in-run daily average and a 176Hf177Hf ratio for JMC-475 of0282160 During the course of analyses the Hf standardJMC-475 gave an average value 02821679 (nfrac14130)USGS reference materials BCR-2 and BHVO-2 were pro-cessed with the samples and yielded Hf isotopic ratios of02828675 and 02831005 respectively Publishedvalues for BCR-2 and BHVO-2 are 02828717 and0283104 8 respectively (Weis et al 2007)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
39
The uncertainty in the composition of the source in thespinel lherzolite melting model for the high-MgO lavasprecludes determining whether the source was moredepleted in incompatible trace elements than the source ofthe tholeiitic lavas (see Fig 14 caption for description) It ispossible that early high-pressure low-degree melting left aresidue within parts of the plume (possibly the hotter inte-rior portion) that was depleted in trace elements (egElliott et al 1991) further decompression and high-degreemelting of such depleted regions could then have generatedthe depleted high-MgO Karmutsen lavas Shallow high-degree melting would preferentially sample depletedregions whereas deeper low-degree melting may notsample more refractory depleted regions (eg CaribbeanEscuder-Viruete et al 2007) The picrites lie at the highend of the range of Nd isotopic compositions and havelow NbZr similar to depleted Icelandic basalts (Fittonet al 2003) therefore the picrites may represent the partsof the mantle plume that were the most depleted prior tothe initiation of melting As Fitton et al (2003) suggestedthese depleted melts may only rarely be erupted withoutcombining with more voluminous less depleted melts andthey may be detected only as undifferentiated small-volume flows like the Keogh Lake picrites within theKarmutsen Formation on Vancouver IslandDecompression melting within the mantle plume initiatedwithin the garnet stability field (35^25 GPa) at highmantle potential temperatures (Tp414508C) and pro-ceeded beneath oceanic arc lithosphere within the spinelstability field (25^09 GPa) where more extensivedegrees of melting could occur
Magmatic evolution of Karmutsentholeiitic basaltsPrimary magmas in large igneous provinces leave exten-sive coarse-grained cumulate residues within or below thecrust these mafic and ultramafic plutonic sequences repre-sent a significant proportion of the original magma thatpartially crystallized at depth (eg Farnetani et al 1996)For example seismic and petrological studies of OntongJava (eg Farnetani et al 1996) combined with the use ofMELTS (Ghiorso amp Sack 1995) indicate that the volcanicsequence may represent only 40 of the total magmareaching the Moho Neal et al (1997) estimated that30^45 fractional crystallization took place for theOntong Java magmas and produced cumulate thicknessesof 7^14 km corresponding to 11^22 km of flood basaltsdepending on the presence of pre-existing oceanic crustand the total crustal thickness (25^35 km) Interpretationsof seismic velocity measurements suggest the presence ofpyroxene and gabbroic cumulates 9^16 km thick beneaththe Ontong Java Plateau (eg Hussong et al 1979)Wrangellia is characterized by high crustal velocities in
seismic refraction lines which is indicative of mafic
plutonic rocks extending to depth beneath VancouverIsland (Clowes et al 1995)Wrangellia crust is 25^30 kmthick beneath Vancouver Island and is underlain by astrongly reflective zone of high velocity and density thathas been interpreted as a major shear zone where lowerWrangellia lithosphere was detached (Clowes et al 1995)The vast majority of the Karmutsen flows are evolved tho-leiitic lavas (eg low MgO high FeOMgO) indicating animportant role for partial crystallization of melts at lowpressure and a significant portion of the crust beneathVancouver Island has seismic properties that are consistentwith crystalline residues from partially crystallizedKarmutsen magmas To test the proportion of the primarymagma that fractionated within the crust MELTS(Ghiorso amp Sack 1995) was used to simulate fractionalcrystallization using several estimated primary magmacompositions from the PRIMELT1 modeling results Themajor-element composition of the primary magmas couldnot be estimated for the tholeiitic basalts because of exten-sive plagthorn cpxthornol fractionation and thus the MELTSmodeling of the picrites is used as a proxy for the evolutionof major elements in the volcanic sequence A pressure of 1kbar was used with variable water contents (anhydrousand 02wt H2O) calculated at the quartz^fayalite^magnetite (QFM) oxygen buffer Experimental resultsfrom Ontong Java indicate that most crystallizationoccurs in magma chambers shallower than 6 km deep(52 kbar Sano amp Yamashita 2004) however some crys-tallization may take place at greater depths (3^5 kbarFarnetani et al 1996)A selection of the MELTS results are shown in Fig 15
Olivine and spinel crystallize at high temperature until15^20wt of the liquid mass has fractionated and theresidual magma contains 9^10wt MgO (Fig 15f) At1235^12258C plagioclase begins crystallizing and between1235 and 11908C olivine ceases crystallizing and clinopyr-oxene saturates (Fig 15f) As expected the addition of clin-opyroxene and plagioclase to the crystallization sequencecauses substantial changes in the composition of the resid-ual liquid FeO and TiO2 increase and Al2O3 and CaOdecrease while the decrease in MgO lessens considerably(Fig 15) The near-vertical trends for the Karmutsen tho-leiitic basalts especially for CaO do not match the calcu-lated liquid-line-of-descent very well which misses manyof the high-CaO high-Al2O3 low-FeO basalts A positivecorrelation between Al2O3CaO and SrNd indicates thataccumulation of plagioclase played a major role in the evo-lution of the tholeiitic basalts (Fig 15g) Increasing thecrystallization pressure slightly and the water content ofthe melts does not systematically improve the match ofthe MELTS models to the observed dataThe MELTS results indicate that a significant propor-
tion of the crystallization takes place before plagioclase(15^20 crystallization) and clinopyroxene (35^45
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
31
crystallization) join the crystallization sequence Theseresults suggest that445 crystallization of the primarymagmas occurred (estimated from the residual liquid percent in Fig 15f) and if the original flood basalt stratigra-phy on Vancouver Island was 6 km thick at least anequivalent amount of complementary plutonic rocksshould have crystallized The Karmutsen basalts and theirplutonic residues thus represent a significant addition ofcrust (perhaps 412 km thickness) previously thickenedduring the formation of the underlying Paleozoic arc
Growth of the Wrangellia oceanic plateauon Vancouver IslandThe extrusive part of the Wrangellia oceanic plateau onVancouver Island was constructed in a three-layersequence of mostly tholeiitic basalts with a restrictedrange of composition (Fig 16) The plateau overliesDevonian to Mississippian arc volcanic sequences andmarine sediments of Mississippian to Permian ages thatare intruded by gabbroic rocks related to the flood basaltsThe volcanic stratigraphy of theWrangellia plateau formed
00
05
10
15
20
25
30
35
40
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
16
0 2 4 6 8 10 12 14 16 18 20
10
11
12
13
14
15
16
17
18
19
20
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
0 2 4 6 8 10 12 14 16 18 20
MgO (wt)
0 10 20 30 40 50
percent of total mass
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
) Mineral proportions
Sp (e)
Ol
CpxPlag
PicriteHigh-MgO basaltTholeiitic basalt
TiO2 (wt)
MgO (wt)
Al2O3 (wt) CaO (wt)
FeO (wt)
MgO(a) (b)
(c) (d)
MgO (wt)MgO (wt)
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
)
Residual liquid ()
1220
1190
Ol + Sp
Plag+Ol+Sp Plag+Cpx+Sp
02 wt H2O
no H2O
Residual liquid and crystallizing phases
(f)
0102030405060708090100
plagioclaseaccumulation
plagioclaseaccumulation
plagioclase
accumulatio
n
10
12
14
16
0 10 20 30 40
Al2O3CaO
SrNd
tholeiitic basalts
(g)
Fig 15 Forward fractional crystallization modeling results for major elements from MELTS (Ghiorso amp Sack 1995) compared with picriticand tholeiitic lavas from the Karmutsen Formation (a^d) Major-element compositions of residual liquids compared with Karmutsen lavacompositions For clarity the results are shown only for one starting composition (open circle) using the composition of the estimated primarymagma for the high-MgO lava series from the modeling technique of Herzberg et al (2007) for sample 4723A13 as shown in Fig 12 The esti-mated primary magmas for samples 4723A3 and 93G171 yield similar results A pressure of 1 kbar was used with no H2O and 02wt H2Ocalculated at QFM black line is anhydrous and grey line is 02wt H2OThe end-point for the anhydrous result is at 78 fractionation and11308C and the end-point for the result with 02wt H2O is 75 fractionation and11108C (e) Mineral proportions of crystallized phases fromone MELTrun (sample 4723A13) with no H2O (f) Per cent residual liquid vs temperature labelled with intervals of crystallizing phases (g)Al2O3CaO vs SrNd for the tholeiitic basalts
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
32
in a single eruptive phase (c 231^225 Ma) as an emergentbasalt sequence during a deeper-water shallow-water andsubaerial stage and subsided below sea level relativelysoon after its formation (Fig 16) The deeper-water stagecontains pillowed and massive submarine flows (3000mthick) that built up the submarine volcanic edifice from theseafloor The lowest pillow basalts were emplaced onunconsolidated and lithified fine-grained siliciclastic andcarbonaceous sediments containing Daonella beds(c 235^232 Ma) that are intruded by abundantKarmutsen mafic sills (Fig 16a) The different pillowedand massive submarine flows resulted primarily from dif-ferent effusive rates and local topography Some of themassive submarine flows commonly recognized by radiat-ing columnar jointing represent master tubes for the deliv-ery of lava to distal parts of submarine flow fields Thepillowed flows which commonly have uneven topography
in modern sequences may have served as levees to themassive submarine flows rather than allowing unob-structed spreading and inflation typical of the massive sub-aerial flows that form on shallow slopes in continentalflood basalt provinces (eg Self et al 1997) The rapid suc-cession of flows is evident by the near absence of sedimentbetween submarine flows Minor volumes of picritic pillowlavas occur in the upper part of the pillowed flow sequencein areas of the plateau exposed on northern VancouverIsland (Fig 16 Nixon et al 2008)The growth of the plateau in shallow water led to
emplacement of an increasing proportion of volcaniclastics(pillow breccias and hyaloclastites) that accumulated up to1500m thick in some areas (Fig 16b Nixon et al 2008)Horizons of pillowed flows occur locally within the volca-niclastic stratigraphy Sedimentary structures (eg gradedbedding fluidization structures) and petrographic textures
Late Devonian-Mississippian arc
Daonella beds
Middle Triassic
deeper water
Late Devonian-Mississippian arc
pillowed flows massive submarine
flows
pillow brecciahyaloclastiteshallow water
~4 km
~6 km
Late Triassic limestoneinterflow limestonelenses
Karmutsen Formation
mafic intrusions sediment-sill complex
Karmutsen Formation
(a) (b)
(c) (d)
subaerial flows
Late Devonian-Mississippian arc Late Devonian-Mississippian arc
picritic pillow basalts
Miss to Permian marine sediments
sills
Fig 16 Growth of the Wrangellia oceanic plateau onVancouver Island (a) Deep-water eruptive phase of pillowed and massive submarineflows with almost no intervening sediments onto MiddleTriassic shale containing Daonella beds (c 235^232 Ma) The basement of the plateauis composed of Devonian to Mississippian arc sequences overlain by Mississippian to Permian carbonates and siliciclastic sediments(b) Shallow-water flow emplacement with increasing proportion of volcaniclastics resulted primarily from cooling^contraction granulationPicritic pillow basalts occur primarily near the transition between pillowed submarine flows and volcaniclastic units Pillowed and massiveflows occur locally within the volcaniclastic unit (c) Emergent stage of growth consists of overlapping subaerial flow fields similar to continen-tal flood basalts with pahoehoe flow structures Multi-tiered columnar jointing occurrs locally in response to ponding of flows (eg Bondre et al2004) Interflow carbonate deposits and plagioclase megacrystic flows developed locally during the waning stage of volcanism (d) Post-volcanicsubsidence and deposition of Late Carnian to Norian (c 212^223 Ma) platform carbonates on top of the plateau Units are labeled in the dia-grams Thicknesses of the underlying units are not to scale
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
33
within the volcaniclastic rocks indicate formation primar-ily via cooling^contraction granulation magma^water^steam interaction autobrecciation resedimentation andmass-wasting rather than pyroclastic eruption (Nixonet al 2008)A broad subaerial platform was constructed by emplace-
ment of subaerial lava flows during an emergent subaerialstage (Fig 16c) The subaerial unit predominantly com-prises well-layered sheet flows that formed in a similarway to continental flood basalts as hundred(s) of pahoehoeflow fields erupted on shallow slopes (Self et al 1997) how-ever volcaniclastic and pillowed flow units are locally pre-sent mostly in the upper parts of the stratigraphyLocalized ponding of flows led to the formation of multi-tiered columnar jointing in some subaerial flows (egBondre et al 2004) As volcanism waned local interflowcarbonate deposits developed and plagioclase megacrysticflows were erupted Subsidence of the plateau is recordedby the interflow lenses and deposits of platform carbonates(40^1500m thick) on top of the Karmutsen basalts thatcontain Late Carnian to Early Norian fossils (Fig 16d c216^227 Ma Nixon et al 2008)The evolution of the internal architecture of the
Wrangellia oceanic plateau was primarily related to theconditions of emplacement (ie water depth and eruptionenvironment) There does not appear to have been a sys-tematic coincident evolution or shift of lava compositionsonVancouver Island as has been recognized in most con-tinental flood basalts and parts of the Wrangellia oceanicplateau in Alaska and Yukon (Greene et al 2008 2009)In general with the exception of the picritic pillow lavasin the submarine stratigraphy and local variations betweenhigh-MgO and tholeiitic lavas there are no significantcompositional differences between lavas erupted duringthe different stages of growth of the plateau (ie submarineand subaerial flows) nor are substantial compositional dif-ferences generally found between massive submarine flowsin contact with pillowed flows tholeiitic basalt flows nearthe base of the Karmutsen Formation are compositionallysimilar to flows sampled near the top of the volcanicsequence Where it was possible to sample continuous sec-tions of volcanic stratigraphy greater than several hun-dreds of meters (eg Keogh^Maynard Lake area west ofthe Karmutsen Range Fig 2) there is no clear relationshipbetween stratigraphic position and compositionalvariation
CONCLUSIONSThe Karmutsen Formation covers extensive areas ofVancouver Island (20 000 km2) and forms part of a largeoceanic plateau erupted over a geologically short interval(c 225^231 Ma) The tripartite volcanic stratigraphyexposed on Vancouver Island is at least 6 km thick andcomprises a succession of submarine flows volcaniclastic
deposits and massive subaerial flowsThese volcanologicaldifferences are primarily related to the eruption environ-ment (deep-water shallow-water or subaerial) The rapidgrowth of the plateau prevented significant accumulationsof intervening sediment except in the uppermost stratigra-phy where isolated limestone lenses commonly associatedwith pillow basalt and hyaloclastite deposits preserve arecord of the final stages of subsidence of the plateauThe Wrangellia plateau on Vancouver Island was con-
structed dominantly of tholeiitic basalt with restrictedmajor- and trace-element and isotopic compositional var-iations Minor volumes of picritic pillow basalts eruptedlate in the submarine phase The Karmutsen basalts wereemplaced onto older Paleozoic arc volcanic sequences andlow NbLa for the picrites may indicate a contributionfrom the sub-arc lithospheric mantle Modeling of the pet-rogenesis of the Karmutsen picrites indicates melting ofanomalously hot mantle (15008C) and extensive degreesof partial melting (23^27) and is consistent with aplume initiation model Combined Sr^Nd^Hf^Pb isotopicsystematics indicate that the picritic and tholeiitic lavaswere derived from a depleted plume-type mantle sourcedistinct from the source of MORB that has compositionalsimilarities to the source of the lavas forming theCaribbean Plateau Picrites were generated by melting ofspinel lherzolite whereas the tholeiitic basalts formed asaggregate melts of both garnet and spinel lherzolite Thetholeiitic basalts underwent extensive low-pressure fractio-nation (52^3 kbar) and seismic studies indicate that someof theWrangellia crust beneathVancouver Island may cor-respond to the plutonic residues of the Karmutsen floodbasalts
ACKNOWLEDGEMENTSWe are grateful to Nick Arndt for helping us get this pro-ject started and Nick Massey for insights into VancouverIsland geology We greatly appreciate the assistance withPRIMELT modeling from Claude Herzberg We thankMikkel Schau for his insight and enthusiasm during field-work Jane Barling assisted with analyses by MC-ICP-MSThoughtful and thorough reviews of the manuscript byJohn Mahoney Fred Frey Wendy Bohrson DougalJerram and Teal Riley are greatly appreciated Fundingwas generously provided by the Rocks to Riches Programadministered by the BC amp Yukon Chamber of Mines in2004 by the BC Geological Survey and by NSERCDiscovery Grants to JSS and DW ARG was supportedby a University Graduate Fellowship at UBC
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
34
REFERENCESBarker F Brown A S Budahn J R amp Plafker G (1989) Back-arc
with frontal-arc component origin of Triassic Karmutsen basaltBritish Columbia Canada Chemical Geology 75 81^102
Blichert-Toft JWeis D Maerschalk C Agranier A amp Albare de F(2003) Hawaiian hot spot dynamics as inferred from the Hf and Pbisotope evolution of Mauna Kea volcano Geochemistry GeophysicsGeosystems 4(2) 1^27 doi1010292002GC000340
Bondre N R Duraiswami R A amp Dole G (2004) Morphologyand emplacement of flows from the Deccan Volcanic ProvinceIndia Bulletin of Volcanology 66 29^45
Carlisle D (1963) Pillow breccias and their aquagene tuffs QuadraIsland British Columbia Journal of Geology 71 48^71
Carlisle D (1972) Late Paleozoic to Mid-Triassic sedimentary-volcanic sequence on Northeastern Vancouver Island Geological
Society of Canada Paper Report of Activities 72-1 Part B 24^30Carlisle D amp SuzukiT (1974) Emergent basalt and submergent car-
bonate^clastic sequences including the Upper Triassic Dilleri andWelleri zones on Vancouver Island Canadian Journal of Earth
Sciences 11 254^279Chauvel C amp Blichert-Toft J (2001) A hafnium isotope and trace
element perspective on melting of the depleted mantle Earth and
Planetary Science Letters 190 137^151Chayes F (1954)The theory of thin-section analysis Journal of Geology
62 92^101Cho M Liou J G amp Maruyama S (1987) Transition from the zeo-
lite to prehnite^pumpellyite facies in the Karmutsen metabasitesVancouver Island British Columbia Journal of Petrology 27(2)467^494
Clowes R M Zelt C A Amor J R amp Ellis R M (1995)Lithospheric structure in the southern Canadian Cordillera froma network of seismic refraction lines Canadian Journal of Earth
Sciences 32(10) 1485^1513Coffin M F amp Eldholm O (1994) Large igneous provinces Crustal
structure dimensions and external consequences Reviews of
Geophysics 32(1) 1^36Cox K G (1980) A model for flood basalt volcanism Journal of
Petrology 21 629^650Eldholm O amp Coffin M F (2000) Large igneous provinces
and plate tectonics In Richards M A Gordon R G amp vander Hilst R D (eds) The History and Dynamics of Global
Plate Motions American Geophysical Union Geophysical Monograph 121309^326
Elliott T R Hawkesworth C J amp Grolaquo nvold K (1991) Dynamicmelting of the Iceland plume Nature 351 201^206
Escuder-Viruete J Pecurren rez-Estaucurren n A Contreras F Joubert MWeis D Ullrich T D amp Spadea P (2007) Plume mantle sourceheterogeneity through time Insights from the Duarte ComplexHispaniola northeastern Caribbean Journal of Geophysical Research112(B04203) doi1010292006JB004323
Farnetani C G Richards M A amp Ghiorso M S (1996)Petrological models of magma evolution and deep crustal structurebeneath hotspots and flood basalt provinces Earth and Planetary
Science Letters 143 81^94Fitton J G amp Godard M (2004) Origin and evolution of magmas
on the Ontong Java Plateau In Fitton J G Mahoney J JWallace P J amp Saunders A D (eds) Origin and Evolution of the
Ontong Java Plateau Geological Society London Special Publications 229151^178
Fitton J G Saunders A D Norry M J Hardarson B S ampTaylor R N (1997) Thermal and chemical structure of theIceland plume Earth and Planetary Science Letters 153 197^208
Fitton J G Saunders A D Kempton P D amp Hardarson B S(2003) Does depleted mantle form an intrinsic part of the Icelandplume Geochemistry Geophysics Geosystems 4(3) 1032 doi1010292002GC000424
Frey F A Huang S Blichert-Toft J Regelous M amp Boyet M(2005) Origin of depleted components in basalt related to theHawaiian hotspot Evidence from isotopic and incompatible ele-ment ratios Geochemistry Geophysics Geosystems 6(1) 23 doi1010292004GC000757
Galer S J G amp Abouchami W (1998) Practical application of leadtriple spiking for correction of instrumental mass discriminationMineralogical Magazine 62A 491^492
Ghiorso M S amp Sack R O (1995) Chemical mass transfer in mag-matic processes IV A revised and internally consistent thermody-namic model for the interpolation and extrapolation of liquid^solidequilibria in magmatic systems at elevated temperatures and pres-sures Contributions to Mineralogy and Petrology 119 197^212
Greene A R Scoates J S amp Weis D (2005)WrangelliaTerrane onVancouver Island Distribution of flood basalts with implicationsfor potential Ni^Cu^PGE mineralization In Grant B (ed)Geological Fieldwork 2004 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2005-1 209^220Greene A R Scoates J S Nixon G T amp Weis D (2006) Picritic
lavas and basal sills in the Karmutsen flood basalt provinceWrangellia northern Vancouver Island In Grant B (ed)Geological Fieldwork 2005 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2006-1 39^52Greene A R Scoates J S amp Weis D (2008) Wrangellia flood
basalts in Alaska A record of plume^lithosphere interaction in aLate Triassic accreted oceanic plateau Geochemistry Geophysics
Geosystems 9(Q12004) doi1010292008GC002092Greene A R Scoates J S Weis D amp Israel S (2009)
Geochemistry of triassic flood basalts from the Yukon (Canada)segment of the accreted Wrangellia oceanic plateau Lithosdoi101016jlithos200811010
Hauff F Hoernle K Tilton G Graham D W amp Kerr A C(2000) Large volume recycling of oceanic lithosphere over shorttime scales geochemical constraints from the Caribbean LargeIgneous Province Earth and Planetary Science Letters 174 247^263
Hauff F Hoernle K amp Schmidt A (2003) Sr^Nd^Pb compositionof Mesozoic Pacific oceanic crust (Site 1149 and 801 ODP Leg185)Implications for alteration of ocean crust and the input into theIzu^Bonin^Mariana subduction system Geochemistry Geophysics
Geosystems 4(8) doi1010292002GC000421Herzberg C (2004) Partial melting below the Ontong Java Plateau
In Fitton J G Mahoney J J Wallace P J amp Saunders A D(eds) Origin and Evolution of the Ontong Java Plateau Geological SocietyLondon Special Publications 229 179^183
Herzberg C (2006) Petrology and thermal structure of the Hawaiianplume from Mauna Kea volcano Nature 444(7119) 605
Herzberg C amp Asimow P D (2008) Petrology of some oceanicisland basalts PRIMELT2XLS software for primary magma cal-culation Geochemistry Geophysics Geosystems 9(Q09001) doi1010292008GC002057
Herzberg C amp OrsquoHara M J (2002) Plume-associated ultramaficmagmas of Phanerozoic age Journal of Petrology 43(10) 1857^1883
Herzberg C Asimow P D Arndt N Niu Y Lesher C MFitton J G Cheadle M J amp Saunders A D (2007)Temperatures in ambient mantle and plumes Constraints frombasalts picrites and komatiites Geochemistry Geophysics Geosystems8(Q02006) doi1010292006GC001390
Huang S amp Frey F A (2005) Trace element abundances of MaunaKea basalt from phase 2 of the Hawaii Scientific Drilling Project
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
35
Petrogenetic implications of correlations with major element con-tent and isotopic ratios Geochemistry Geophysics Geosystems 4(6)doi1010292002GC000322
Hussong D M Wipperman L K amp Kroenke L W (1979) Thecrustal structure of the Ontong Java and Manihiki OceanicPlateaus Journal of Geophysical Research 84(B11) 6003^6010
Jerram D A amp Widdowson M (2005) The anatomy of ContinentalFlood Basalt Provinces geological constraints on the processes andproducts of flood volcanism Lithos 79(3^4) 385^405
Jones D L Silberling N J amp Hillhouse J (1977)Wrangellia a dis-placed terrane in northwestern North America CanadianJournal ofEarth Sciences 14(11) 2565^2577
Kerr A C (2003) Oceanic Plateaus In Rudnick R (ed) The Crust
Treatise on GeochemistryVol 3 Oxford Elsevier Science pp 537^565Kerr A C amp Mahoney J J (2007) Oceanic plateaus Problematic
plumes potential paradigms Chemical Geology 241 332^353Kerr A CTarney J Marriner G F Klaver GT Saunders A D
amp Thirlwall M F (1996) The geochemistry and petrogenesis ofthe Late-Cretaceous picrites and basalts of Curacao NetherlandsAntilles a remnant of an oceanic plateau Contributions to
Mineralogy and Petrology 124(1) 29^43Kerr A C Marriner G F Tarney J Nivia A Saunders A D
Thirlwall M F amp Sinton A C (1997) Cretaceous basaltic ter-ranes in Western Colombia Elemental chronological and Sr-Ndisotopic constraints on petrogenesis Journal of Petrology 38(6)677^702
Kinzler R J (1997) Melting of mantle peridotite at pressuresapproaching the spinel to garnet transition Application to mid-ocean ridge basalt petrogenesis Journal of Geophysical Research
102(B1) 853^874Kinzler R J amp Grove T L (1992) Primary magmas of mid-ocean
ridge basalts 1 Experiments and methods Journal of Geophysical
Research 97(B5) 6885^6906Korenaga J (2005) Why did not the Ontong Java Plateau form sub-
aerially Earth and Planetary Science Letters 234 385^399Kuniyoshi S (1972) Petrology of the Karmutsen (Volcanic) Group
northeastern Vancouver Island British Columbia PhD disserta-tion University of California Los Angeles 242 pp
Lassiter J C DePaolo D J amp Mahoney J J (1995) Geochemistry ofthe Wrangellia flood basalt province Implications for the role ofcontinental and oceanic lithosphere in flood basalt genesis Journalof Petrology 36(4) 983^1009
LeBas M J (2000) IUGS reclassification of the high-Mg and picriticvolcanic rocks Journal of Petrology 41(10) 1467^1470
Longhi J (2002) Some phase equilibrium systematics of lherzolitemelting I Geochemistry Geophysics Geosystems 3(3) doi1010292001GC000204
MacDonald G A amp Katsura T (1964) Chemical composition ofHawaiian lavas Journal of Petrology 5 82^133
Mahoney J J (1987) An isotopic survey of Pacific oceanic plateausimplications for their nature and origin In Keating B HFryer P Batiza R amp Boehlert GW (eds) Seamounts Islands andAtolls American Geophysical Union Geophysical Monograph 43 207^220
Mahoney J LeRoex A P Peng Z Fisher R L amp Natland J H(1992) Southwestern limits of Indian Ocean Ridge mantle and theorigin of low 206Pb204Pb mid-ocean ridge basalt isotope systema-tics of the central Southwest Indian Ridge (178^508E) Journal ofGeophysical Research 97(B13) 19771^19790
Mahoney J J Sinton J M Kurz M D MacDougall J DSpencer K J amp Lugmair GW (1994) Isotope and trace elementcharacteristics of a super-fast spreading ridge East Pacific rise13^238S Earth and Planetary Science Letters 121 173^193
Massey N W D (1995a) Geology and mineral resources of theAlberni^Nanaimo Lakes sheet Vancouver Island 92F1W 92F2Eand part of 92F7E British Columbia Ministry of Energy Mines and
Petroleum Resources Paper 1992-2 132 ppMassey N W D (1995b) Geology and mineral resources of the
Cowichan Lake sheet Vancouver Island 92C16 British Columbia
Ministry of Energy Mines and Petroleum Resources Paper 1992-3 112 ppMassey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005a) Digital Geology Map of British Columbia Tile NM9 MidCoast BC British Columbia Ministry of Energy and Mines Geofile
2005-2Massey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005b) Digital Geology Map of British Columbia Tile NM10Southwest BC British Columbia Ministry of Energy and Mines Geofile
2005-3McDonough W F amp Sun S (1995) The composition of the Earth
Chemical Geology 120 223^253Monger JW H amp Journeay J M (1994) Basement geology and tec-
tonic evolution of theVancouver region In Monger JW H (ed)Geology and Geological Hazards of the Vancouver Region Southwestern
British Columbia Geological Survey of Canada Bulletin 481 3^25Muller J E (1977) Geology of Vancouver Island Geological Survey of
Canada Open File Map 463Muller J E (1980)The Paleozoic Sicker Group of Vancouver Island British
Columbia Geological Survey of Canada Paper 79-30 22 ppMuller J E Northcote K E amp Carlisle D (1974) Geology and
mineral deposits of Alert Bay^Cape Scott map area VancouverIsland British Columbia Geological Survey of Canada Paper 74-8 77
Neal C R Mahoney J J Kroenke L W Duncan R A ampPetterson M G (1997) The Ontong^Java Plateau InMahoney J J amp Coffin M F (eds) Large Igneous Provinces
Continental Oceanic and Planetary FloodVolcanism American Geophysical
Union Geophysical Monograph 100 183^216Niu Y Collerson K D Batiza R Wendt J I amp Regelous M
(1999) Origin of enriched-typed mid-ocean ridge basalt at ridgesfar from mantle plumes The East Pacific Rise at 11820rsquoN Journalof Geophysical Research 104 7067^7088
Nixon G T amp Orr A J (2007) Recent revisions to the EarlyMesozoic stratigraphy of Northern Vancouver Island (NTS 102I092L) and metallogenic implicatinos British Columbia InGrant B (ed) Geological Fieldwork 2006 British Columbia Ministry of
Energy Mines and Petroleum Resources Paper 2007-1 163^177Nixon GT Hammack J L KoyanagiV M Payie G J Haggart
JW Orchard M JTozerT Friedman R M Archibald D APalfy J amp Cordey F (2006a) Geology of the Quatsino^PortMcNeill area northernVancouver Island British Columbia Ministry
of Energy Mines and Petroleum Resources Geoscience Map 2006-2 scale150 000
Nixon G T Kelman M C Stevenson D Stokes L A ampJohnston K A (2006b) Preliminary geology of the Nimpkishmap area (NTS 092L07) northern Vancouver Island BritishColumbia In Grant B amp Newell J M (eds) Geological Fieldwork2005 British Columbia Ministry of Energy Mines and Petroleum Resources
Paper 2006-1 135^152Nixon G T Laroque J Pals A Styan J Greene A R amp
Scoates J S (2008) High-Mg lavas in the Karmutsen flood basaltsnorthernVancouver Island (NTS 092L) Stratigraphic setting andmetallogenic significance In Grant B (ed) Geological Fieldwork
2007 British Columbia Ministry of Energy Mines and Petroleum
Resources Paper 2008-1 175^190Nowell G M Kempton P D Noble S R Fitton J G
Saunders A Mahoney J J amp Taylor R N (1998) High precisionHf isotope measurements of MORB and OIB by thermal ionisation
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
36
mass spectrometry insights into the depleted mantle Chemical
Geology 149 211^233Parrish R R amp McNicoll V J (1992) U^Pb age determinations
from the southern Vancouver Island area British Columbia InRadiogenic Age and Isotopic Studies Report 5 Geological Survey of
Canada Paper 91-2 79^86Peate DW (1997) The Parana^Etendeka Province In Coffin M F
amp Mahoney J (eds) Large Igneous Provinces Continental Oceanic andPlanetary Flood Volcanism American Geophysical Union Geophysical
Monograph 100 217^245Perfit M R Fornari D J Ridley W I Kirk P D Casey J
Kastens K A Reynolds J R Edwards M Desonie DShuster R amp Paradis S (1996) Recent volcanism in the Siqueirostransform fault picritic basalts and implications for MORBmagma genesis Earth and Planetary Science Letters 141(1^4) 91^108
Pretorius W Weis D Williams G Hanano D Kieffer B ampScoates J S (2006) Complete trace elemental characterization ofgranitoid (USGSG-2GSP-2) reference materials by high resolutioninductively coupled plasma-mass spectrometry Geostandards and
Geoanalytical Research 30(1) 39^54Regelous M Niu Y Wendt J I Batiza R Greig A amp
Collerson K D (1999) Variations in the geochemistry of magma-tism on the East Pacific Rise at 10830rsquoN since 800 ka Earth and
Planetary Science Letters 168 45^63Recurren villon S Chauvel C Arndt N T Pik R Martineau F
Fourcade S amp Marty B (2002) Heterogeneity of the Caribbeanplateau mantle source Sr O and He isotopic compositions of oli-vine and clinopyroxene from Gorgona Island Earth and Planetary
Science Letters 205(1^2) 91Rhodes J M (1996) Geochemical stratigraphy of lava flows samples
by the Hawaii Scientific Drilling Project Journal of Geophysical
Research 101(B5) 11729^11746Rhodes J M amp Vollinger M J (2004) Composition of basaltic lavas
sampled by phase-2 of the Hawaii Scientific Drilling ProjectGeochemical stratigraphy and magma types Geochemistry
Geophysics Geosystems 5(3) doi1010292002GC000434Richards M A Jones D L Duncan R A amp DePaolo D J (1991)
A mantle plume initiation model for the Wrangellia flood basaltand other oceanic plateaus Science 254 263^267
Salters V J M amp Hart S R (1991) The mantle sources of oceanridges islands and arcs the Hf-isotope connection Earth and
Planetary Science Letters 104 364^380Salters V J M amp Stracke A (2004) Composition of the depleted
SaltersV J amp WhiteW (1998) Hf isotope constraints on mantle evo-lution Chemical Geology 145 447^460
SanoT amp Yamashita S (2004) Experimental petrology of basementlavas from Ocean Drilling Program Leg 192 implications for dif-ferentiation processes in Ontong Java Plateau magmas InFitton J G Mahoney J JWallace P J amp Saunders A D (eds)Origin and Evolution of the Ontong Java Plateau Geological Society
London Special Publications 229 185^218Saunders A D (2005) Large igneous provinces Origin and environ-
mental consequences Elements 1 259^263Self S ThordarsonT amp Keszthely L (1997) Emplacement of conti-
nental flood basalt lava flows In Mahoney J J amp Coffin M F(eds) Large Igneous Provinces Continental Oceanic and Planetary Flood
Volcanism American Geophysical Union Geophysical Monograph 100381^410
Shaw D M (2000) Continuous (dynamic) melting theory revisitedCanadian Mineralogist 38(5) 1041^1063
Sluggett C L (2003) Uranium^lead age and geochemical constraintson Paleozoic and Early Mesozoic magmatism in WrangelliaTerrane Saltspring Island British Columbia BSc thesisUniversity of British ColumbiaVancouver 84 pp
Storey M Mahoney J J amp Saunders A D (1997) Cretaceousbasalts in Madagascar and the transition between plume and con-tinental lithospheric mantle sources In Mahoney J J ampCoffin M F (eds) Large Igneous Provinces Continental Oceanic and
Planetary Flood Volcanism Aamerican Geophysical Union Geophysical
Monograph 100 95^122Surdam R C (1967) Low-grade metamorphism of the Karmutsen
Group PhD dissertation University of California Los Angeles288 pp
Sutherland-Brown A Yorath C J Anderson R G amp Dom K(1986) Geological maps of southern Vancouver IslandLITHOPROBE I 92C10 11 14 16 92F1 2 7 8 Geological Survey ofCanada Open File 1272
Tejada M L G Mahoney J J Castillo P R Ingle S P Sheth HC amp Weis D (2004) Pin-pricking the elephant evidence on theorigin of the Ontong Java Plateau from Pb^Sr^Hf^Nd isotopiccharacteristics of ODP Leg 192 basalts In Fitton J GMahoney J J Wallace P J amp Saunders A D (eds) Origin and
Evolution of the Ontong Java Plateau Geological Society London Special
Publications 229 133^150Thompson P M E Kempton P D amp Kerr A C (2008) Evaluation
of the effects of alteration and leaching on Sm^Nd and Lu^Hf sys-tematics in submarine mafic rocks Lithos 104(1^4) 164^176
Thompson P M E Kempton P D White R V Kerr A CTarney J Saunders A D Fitton J G amp McBirney A (2003)Hf-Nd isotope constraints on the origin of the CretaceousCaribbean plateau and its relationship to the Galapagos plumeEarth and Planetary Science Letters 217(1^2) 59^75
Vervoort J D Patchett P Blichert-Toft J amp Albare de F (1999)Relationships between Lu^Hf and Sm^Nd isotopic systems in theglobal sedimentary system Earth and Planetary Science Letters 16879^99
Walter M J (1998) Melting of garnet peridotite and the origin ofkomatiite and depleted lithosphere Journal of Petrology 39(1) 29^60
Weis D Kieffer B Maerschalk C Barling J de Jong JWilliams G A Hanano D Mattielli N Scoates J SGoolaerts A Friedman R A amp Mahoney J B (2006) High-pre-cision isotopic characterization of USGS reference materials byTIMS and MC-ICP-MS Geochemistry Geophysics Geosystems
7(Q08006) doi1010292006GC001283Weis D Kieffer B Hanano D Silva I N Barling J
Pretorius W Maerschalk C amp Mattielli N (2007) Hf isotopecompositions of US Geological Survey reference materialsGeochemistry Geophysics Geosystems 8(Q06006) doi1010292006GC001473
Wheeler J O amp McFeely P (1991) Tectonic assemblage map of theCanadian Cordillera and adjacent part of the United States ofAmerica Geological Survey of Canada Map 1712A
WhiteW M Albare de F amp Tecurren louk P (2000) High-precision analy-sis of Pb isotope ratios by multi-collector ICP-MS Chemical Geology167 257^270
Yorath C J Sutherland Brown A amp Massey N W D (1999)LITHOPROBE southern Vancouver Island British Columbia Geological
Survey of Canada Bulletin 498 145 ppZou H (1998) Trace element fractionation during modal and
non-model dynamic melting and open-system melting Amathematical treatment Geochimica et Cosmochimica Acta 62(11)1937^1945
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
37
Zou H amp Reid M R (2001) Quantitative modeling of trace elementfractionation during incongruent dynamic melting Geochimica et
Cosmochimica Acta 65(1) 153^162
APPENDIX SAMPLEPREPARAT ION ANDANALYT ICAL METHODSOnly the freshest rocks in the field were sampled and onlythe least altered samples were selected for chemical andisotopic analysis based on thorough petrographic inspec-tion Fifty-nine of the 129 collected samples were crushed(400 g) into small pieces 52mm in diameter in aRocklabs hydraulic piston crusher between WC plates tominimize contamination The coarse-crush was mixedand 100 g was powdered in a planetary mill using agatejars and balls cleaned with quartz sand between samples
ActLabs analytical methodsThe major- and trace-element compositions of the whole-rock powders were determined at Activation LaboratoriesLtd (Actlabs) in Ancaster Ontario Analytical techniquesand detection limits are also available from Actlabs (httpwwwactlabscommethsub_code4erehtm) The analyticalmethod for each of the elements analyzed is indicated inTable 2 For the major elements a 02 g sample was mixedwith a lithium metaborate^lithium tetraborate mixtureand fused in a graphite crucible The molten mixture waspoured into a 5 HNO3 solution and shaken until dis-solved (30min) The samples were analyzed for majoroxides and selected trace elements by inductively coupledplasma optical emission spectrometry (ICP-OES) on acombination simultaneous^sequential Thermo Jarrell-AshEnviro II system Internal calibration was achieved usinga variety of international reference materials (eg W-2BIR-1 DNC-1) and independent control samplesAdditional trace elements were analyzed by both instru-mental neutron activation analysis (INAA) and induc-tively couple plasma mass spectrometry (ICP-MS) ForINAA15^25 g of sample was weighed into small polyeth-ylene vials and irradiated with a control international ref-erence material CANMET WMS-1 and NiCr flux wires ata thermal neutron flux of 71012 neutronscm2 per s in theMcMaster Nuclear Reactor Following a 7 day waitingperiod the samples were measured on an Ortec high-purity Ge detector linked to a Canberra Series 95 multi-channel analyzer Activities for each element weredecay- and weight-corrected and compared with a detectorcalibration developed from multiple international certifiedreference materials For ICP-MS 025 g of sample wasdigested in HF followed by HNO3^HClO4 heated andtaken to dryness The samples were brought back into solu-tion with HCl Samples were analyzed using a PerkinElmer Optima 3000 ICP In-lab standards or certified ref-erence materials (eg W-2 BIR-1 DNC-1) were used for
quality control A total of 15 blind duplicates were ana-lyzed from the same sample powders to assessreproducibility
University of Massachusetts analyticalmethodsFourteen sample duplicate powders taken from the same100 g sample aliquot (high-MgO basalts and picrites)were also analyzed at the Ronald B Gilmore X-RayFluorescence (XRF) Laboratory at the University ofMassachusetts Major elements were measured on a fusedLa-bearing lithium borate glass disc using a SiemensMRS-400 spectrometer Trace element concentrations(Rb Sr Ba Ce Nb Zr Y Pb Zn Ga Ni Cr V) weremeasured on a separate powder pellet using a PhilipsPW2400 sequential spectrometer using a Rh tube LOIand ferrous iron measurements were made as describedby Rhodes amp Vollinger (2004) Precision and accuracy esti-mates for the data have been described by Rhodes (1996)and Rhodes amp Vollinger (2004) Results for each sampleare the average of two separate analyses and are shown inElectronic Appendix 2 for a comparison of analyses ofduplicate powders using XRFand ICP-MS
PCIGR analytical methodsA subset of 19 samples was selected for high-precisiontrace-element analysis and Sr Nd Pb and Hf isotopicanalysis at the Pacific Centre for Isotopic andGeochemical Research (PCIGR) at the University ofBritish Columbia (UBC) Samples were selected from the59 samples analyzed for whole-rock chemistry at ActLabsbased on major- and trace-element chemistry alteration(low LOI and petrographic alteration index) and samplelocation Samples were prepared for trace-element analysisat the PCIGR by the technique described by Pretorius et al(2006) on unleached rock powders Sample powders(100mg) were weighed in 7ml screw-top Savillex bea-kers and dissolved in 1ml 14N HNO3 and 5ml 48 HFon a hotplate for 48 h at 1308C with periodic ultrasonica-tion Samples were dried and redissolved in 6ml 6N HClon a hotplate for 24 h and then dried and redissolved in1ml concentrated HNO3 for 24 h before final dryingTrace element abundances were measured by high-resolution inductively coupled plasma-mass spectrometry(HR-ICP-MS) with aThermo Finnigan Element2 systemfollowing the procedures described by Pretorius et al(2006) within 24 h of redissolution HFSE and LILE weremeasured in medium-resolution mode at 2000 dilutionusing a PFATeflon spray chamber washed with aqua regiafor 3min between samples REE were measured in high-resolution mode and U and Pb in low-resolution modeat 2000 dilution using a glass spray chamber washedwith 2 HNO3 between samples Total procedural blanksand reference materials (BCR-2 BHVO-2) were analyzedwith the batch of samples Indium was used as an internal
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
38
standard in all samples and standard solutionsBackground and standard solutions were analyzed afterevery five samples to detect memory effects and massdrift Results for PCIGR trace-element analyses areshown in Electronic Appendix 3 for the full set of analyzedtrace elementsSample digestion for purification of Sr Nd Hf and Pb
for column chemistry began by weighing each samplepowder (400^500mg) prior to leaching All samples wereleached with 6N HCl and placed in an ultrasonic bath for15min Samples were rinsed twice with 18 M-cm H2Obetween each leaching step (15 total) until the supernatantliquid was clear [following the method of Mahoney (1987)]Samples were then dried on a hotplate for 24 h andweighed again Sample solutions were then prepared bydissolving 100^250mg of the leached powder dissolvedin 1ml 14N HNO3 and 10ml 48 HF on a hotplate for48 h at 1308C with periodic ultrasonication Samples weredried and redissolved in 6ml 6N HCl on a hotplate for24 h and then dried Pb was separated using anion exhangecolumns and the discard was used for Sr REE and Hfseparation Nd was separated from the other REE and Hfrequired two additional purification steps Detailed proce-dures for column chemistry for separating Sr Nd and Pbat the PCIGR have been described byWeis et al (2006) andHf purification has been described byWeis et al (2007) Srand Nd isotope ratios were measured by thermal ioniza-tion mass spectrometry (TIMS) on a Thermo FinniganTriton system in static mode with relay matrix rotation ona single Ta and double Re^Ta filament respectively Fourto five filaments per barrel of 21 were occupied by stan-dards (NBS 987 for Sr and LaJolla for Nd) for each barrelwhere samples were run Sample Sr and Nd isotopic com-positions were corrected for mass fractionation using86Sr88Srfrac14 01194 and 146Nd144Ndfrac14 07219 Each samplewas then normalized using the barrel average of the refer-ence material relative to the values of143Nd144Ndfrac14 0511858 and 87Sr86Srfrac14 0710248 (Weiset al 2006) During the course of the Vancouver Islandanalyses the La Jolla Nd standard gave an average valueof 05118566 (nfrac14 7) and NBS987 standard gave an aver-age of 07102408 (nfrac1411) (2 error is reported as times106) 147Sm144Nd errors are 15 or 0006 USGeological Survey (USGS) reference material BHVO-2was processed with the samples and yielded Sr and Nd iso-topic ratios of 07034607 and 05129786 respectivelyThese are in agreement with the published values of070347920 and 051298411 respectively (Weis et al2006)
Pb and Hf isotopic compositions were analyzed by mul-tiple collector-inductively coupled plasma-mass spectrome-try (MC-ICP-MS) on a static multi-collection on a NuPlasma (Nu Instruments) system The detailed analyticalprocedure for Pb isotopic analyses on the Nu Plasma atthe PCIGR has been described by Weis et al (2006) Theconfiguration for Pb analyses allows for collection of PbTl and Hg togetherTl and Hg are used to monitor instru-mental mass discrimination and isobaric overlap repec-tively All sample solutions were analyzed withapproximately the same PbTl ratio (4) as the referencematerial NIST SRM 981 To accomplish this a small ali-quot of each sample solution from the Pb columns was ana-lyzed on the Element2 to determine the precise amount ofPb available for analysis on the Nu Plasma The SRM 981standard was run after every two samples on the NuPlasma During the time samples were run analyses ofthe SRM 981 Pb reference material gave values of206Pb204Pbfrac1416940322 207Pb204Pbfrac1415495823and 208Pb204Pbfrac14 36713164 (nfrac14 61 2 error is reportedas times 104) these values are in excellent agreement withreported TIMS triple-spike values of Galer amp Abouchami(1998) Fractionation-corrected Pb isotopic ratios were fur-ther corrected by the sample-standard bracketing methodor the ln^ln correction method described by White et al(2000) and Blichert-Toft et al (2003) Leached powders ofUSGS reference material BHVO-2 yielded Pb isotopicratios of 206Pb204Pbfrac14186454 8 207Pb204Pbfrac14 1549105 and 208Pb204Pbfrac14 38222514 and BCR-2yielded 206Pb204Pbfrac141880466 207Pb204Pbfrac14 1562518 and 208Pb204Pbfrac14 3883496 (2 error isreported as times 104) These values are in agreement withthose for leached residues of BHVO-2 and BCR-2 fromWeis et al (2006)Hf isotopic compositions were analyzed following the
procedures detailed byWeis et al (2007) The configurationfor Hf analyses monitored Lu mass 175 andYb mass 172 toallow for interference correction to masses 174 and 176 Hfisotopic ratios were normalized internally for mass fractio-nation to a 179Hf177Hf ratio of 07325 using an exponentialcorrection Standards were run after every two samplesand sample results were normalized to the ratio of the in-run daily average and a 176Hf177Hf ratio for JMC-475 of0282160 During the course of analyses the Hf standardJMC-475 gave an average value 02821679 (nfrac14130)USGS reference materials BCR-2 and BHVO-2 were pro-cessed with the samples and yielded Hf isotopic ratios of02828675 and 02831005 respectively Publishedvalues for BCR-2 and BHVO-2 are 02828717 and0283104 8 respectively (Weis et al 2007)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
39
crystallization) join the crystallization sequence Theseresults suggest that445 crystallization of the primarymagmas occurred (estimated from the residual liquid percent in Fig 15f) and if the original flood basalt stratigra-phy on Vancouver Island was 6 km thick at least anequivalent amount of complementary plutonic rocksshould have crystallized The Karmutsen basalts and theirplutonic residues thus represent a significant addition ofcrust (perhaps 412 km thickness) previously thickenedduring the formation of the underlying Paleozoic arc
Growth of the Wrangellia oceanic plateauon Vancouver IslandThe extrusive part of the Wrangellia oceanic plateau onVancouver Island was constructed in a three-layersequence of mostly tholeiitic basalts with a restrictedrange of composition (Fig 16) The plateau overliesDevonian to Mississippian arc volcanic sequences andmarine sediments of Mississippian to Permian ages thatare intruded by gabbroic rocks related to the flood basaltsThe volcanic stratigraphy of theWrangellia plateau formed
00
05
10
15
20
25
30
35
40
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
16
0 2 4 6 8 10 12 14 16 18 20
10
11
12
13
14
15
16
17
18
19
20
0 2 4 6 8 10 12 14 16 18 206
7
8
9
10
11
12
13
14
15
0 2 4 6 8 10 12 14 16 18 20
MgO (wt)
0 10 20 30 40 50
percent of total mass
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
) Mineral proportions
Sp (e)
Ol
CpxPlag
PicriteHigh-MgO basaltTholeiitic basalt
TiO2 (wt)
MgO (wt)
Al2O3 (wt) CaO (wt)
FeO (wt)
MgO(a) (b)
(c) (d)
MgO (wt)MgO (wt)
1400
1300
1200
1100
1000
Tem
pera
ture
(degC
)
Residual liquid ()
1220
1190
Ol + Sp
Plag+Ol+Sp Plag+Cpx+Sp
02 wt H2O
no H2O
Residual liquid and crystallizing phases
(f)
0102030405060708090100
plagioclaseaccumulation
plagioclaseaccumulation
plagioclase
accumulatio
n
10
12
14
16
0 10 20 30 40
Al2O3CaO
SrNd
tholeiitic basalts
(g)
Fig 15 Forward fractional crystallization modeling results for major elements from MELTS (Ghiorso amp Sack 1995) compared with picriticand tholeiitic lavas from the Karmutsen Formation (a^d) Major-element compositions of residual liquids compared with Karmutsen lavacompositions For clarity the results are shown only for one starting composition (open circle) using the composition of the estimated primarymagma for the high-MgO lava series from the modeling technique of Herzberg et al (2007) for sample 4723A13 as shown in Fig 12 The esti-mated primary magmas for samples 4723A3 and 93G171 yield similar results A pressure of 1 kbar was used with no H2O and 02wt H2Ocalculated at QFM black line is anhydrous and grey line is 02wt H2OThe end-point for the anhydrous result is at 78 fractionation and11308C and the end-point for the result with 02wt H2O is 75 fractionation and11108C (e) Mineral proportions of crystallized phases fromone MELTrun (sample 4723A13) with no H2O (f) Per cent residual liquid vs temperature labelled with intervals of crystallizing phases (g)Al2O3CaO vs SrNd for the tholeiitic basalts
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
32
in a single eruptive phase (c 231^225 Ma) as an emergentbasalt sequence during a deeper-water shallow-water andsubaerial stage and subsided below sea level relativelysoon after its formation (Fig 16) The deeper-water stagecontains pillowed and massive submarine flows (3000mthick) that built up the submarine volcanic edifice from theseafloor The lowest pillow basalts were emplaced onunconsolidated and lithified fine-grained siliciclastic andcarbonaceous sediments containing Daonella beds(c 235^232 Ma) that are intruded by abundantKarmutsen mafic sills (Fig 16a) The different pillowedand massive submarine flows resulted primarily from dif-ferent effusive rates and local topography Some of themassive submarine flows commonly recognized by radiat-ing columnar jointing represent master tubes for the deliv-ery of lava to distal parts of submarine flow fields Thepillowed flows which commonly have uneven topography
in modern sequences may have served as levees to themassive submarine flows rather than allowing unob-structed spreading and inflation typical of the massive sub-aerial flows that form on shallow slopes in continentalflood basalt provinces (eg Self et al 1997) The rapid suc-cession of flows is evident by the near absence of sedimentbetween submarine flows Minor volumes of picritic pillowlavas occur in the upper part of the pillowed flow sequencein areas of the plateau exposed on northern VancouverIsland (Fig 16 Nixon et al 2008)The growth of the plateau in shallow water led to
emplacement of an increasing proportion of volcaniclastics(pillow breccias and hyaloclastites) that accumulated up to1500m thick in some areas (Fig 16b Nixon et al 2008)Horizons of pillowed flows occur locally within the volca-niclastic stratigraphy Sedimentary structures (eg gradedbedding fluidization structures) and petrographic textures
Late Devonian-Mississippian arc
Daonella beds
Middle Triassic
deeper water
Late Devonian-Mississippian arc
pillowed flows massive submarine
flows
pillow brecciahyaloclastiteshallow water
~4 km
~6 km
Late Triassic limestoneinterflow limestonelenses
Karmutsen Formation
mafic intrusions sediment-sill complex
Karmutsen Formation
(a) (b)
(c) (d)
subaerial flows
Late Devonian-Mississippian arc Late Devonian-Mississippian arc
picritic pillow basalts
Miss to Permian marine sediments
sills
Fig 16 Growth of the Wrangellia oceanic plateau onVancouver Island (a) Deep-water eruptive phase of pillowed and massive submarineflows with almost no intervening sediments onto MiddleTriassic shale containing Daonella beds (c 235^232 Ma) The basement of the plateauis composed of Devonian to Mississippian arc sequences overlain by Mississippian to Permian carbonates and siliciclastic sediments(b) Shallow-water flow emplacement with increasing proportion of volcaniclastics resulted primarily from cooling^contraction granulationPicritic pillow basalts occur primarily near the transition between pillowed submarine flows and volcaniclastic units Pillowed and massiveflows occur locally within the volcaniclastic unit (c) Emergent stage of growth consists of overlapping subaerial flow fields similar to continen-tal flood basalts with pahoehoe flow structures Multi-tiered columnar jointing occurrs locally in response to ponding of flows (eg Bondre et al2004) Interflow carbonate deposits and plagioclase megacrystic flows developed locally during the waning stage of volcanism (d) Post-volcanicsubsidence and deposition of Late Carnian to Norian (c 212^223 Ma) platform carbonates on top of the plateau Units are labeled in the dia-grams Thicknesses of the underlying units are not to scale
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
33
within the volcaniclastic rocks indicate formation primar-ily via cooling^contraction granulation magma^water^steam interaction autobrecciation resedimentation andmass-wasting rather than pyroclastic eruption (Nixonet al 2008)A broad subaerial platform was constructed by emplace-
ment of subaerial lava flows during an emergent subaerialstage (Fig 16c) The subaerial unit predominantly com-prises well-layered sheet flows that formed in a similarway to continental flood basalts as hundred(s) of pahoehoeflow fields erupted on shallow slopes (Self et al 1997) how-ever volcaniclastic and pillowed flow units are locally pre-sent mostly in the upper parts of the stratigraphyLocalized ponding of flows led to the formation of multi-tiered columnar jointing in some subaerial flows (egBondre et al 2004) As volcanism waned local interflowcarbonate deposits developed and plagioclase megacrysticflows were erupted Subsidence of the plateau is recordedby the interflow lenses and deposits of platform carbonates(40^1500m thick) on top of the Karmutsen basalts thatcontain Late Carnian to Early Norian fossils (Fig 16d c216^227 Ma Nixon et al 2008)The evolution of the internal architecture of the
Wrangellia oceanic plateau was primarily related to theconditions of emplacement (ie water depth and eruptionenvironment) There does not appear to have been a sys-tematic coincident evolution or shift of lava compositionsonVancouver Island as has been recognized in most con-tinental flood basalts and parts of the Wrangellia oceanicplateau in Alaska and Yukon (Greene et al 2008 2009)In general with the exception of the picritic pillow lavasin the submarine stratigraphy and local variations betweenhigh-MgO and tholeiitic lavas there are no significantcompositional differences between lavas erupted duringthe different stages of growth of the plateau (ie submarineand subaerial flows) nor are substantial compositional dif-ferences generally found between massive submarine flowsin contact with pillowed flows tholeiitic basalt flows nearthe base of the Karmutsen Formation are compositionallysimilar to flows sampled near the top of the volcanicsequence Where it was possible to sample continuous sec-tions of volcanic stratigraphy greater than several hun-dreds of meters (eg Keogh^Maynard Lake area west ofthe Karmutsen Range Fig 2) there is no clear relationshipbetween stratigraphic position and compositionalvariation
CONCLUSIONSThe Karmutsen Formation covers extensive areas ofVancouver Island (20 000 km2) and forms part of a largeoceanic plateau erupted over a geologically short interval(c 225^231 Ma) The tripartite volcanic stratigraphyexposed on Vancouver Island is at least 6 km thick andcomprises a succession of submarine flows volcaniclastic
deposits and massive subaerial flowsThese volcanologicaldifferences are primarily related to the eruption environ-ment (deep-water shallow-water or subaerial) The rapidgrowth of the plateau prevented significant accumulationsof intervening sediment except in the uppermost stratigra-phy where isolated limestone lenses commonly associatedwith pillow basalt and hyaloclastite deposits preserve arecord of the final stages of subsidence of the plateauThe Wrangellia plateau on Vancouver Island was con-
structed dominantly of tholeiitic basalt with restrictedmajor- and trace-element and isotopic compositional var-iations Minor volumes of picritic pillow basalts eruptedlate in the submarine phase The Karmutsen basalts wereemplaced onto older Paleozoic arc volcanic sequences andlow NbLa for the picrites may indicate a contributionfrom the sub-arc lithospheric mantle Modeling of the pet-rogenesis of the Karmutsen picrites indicates melting ofanomalously hot mantle (15008C) and extensive degreesof partial melting (23^27) and is consistent with aplume initiation model Combined Sr^Nd^Hf^Pb isotopicsystematics indicate that the picritic and tholeiitic lavaswere derived from a depleted plume-type mantle sourcedistinct from the source of MORB that has compositionalsimilarities to the source of the lavas forming theCaribbean Plateau Picrites were generated by melting ofspinel lherzolite whereas the tholeiitic basalts formed asaggregate melts of both garnet and spinel lherzolite Thetholeiitic basalts underwent extensive low-pressure fractio-nation (52^3 kbar) and seismic studies indicate that someof theWrangellia crust beneathVancouver Island may cor-respond to the plutonic residues of the Karmutsen floodbasalts
ACKNOWLEDGEMENTSWe are grateful to Nick Arndt for helping us get this pro-ject started and Nick Massey for insights into VancouverIsland geology We greatly appreciate the assistance withPRIMELT modeling from Claude Herzberg We thankMikkel Schau for his insight and enthusiasm during field-work Jane Barling assisted with analyses by MC-ICP-MSThoughtful and thorough reviews of the manuscript byJohn Mahoney Fred Frey Wendy Bohrson DougalJerram and Teal Riley are greatly appreciated Fundingwas generously provided by the Rocks to Riches Programadministered by the BC amp Yukon Chamber of Mines in2004 by the BC Geological Survey and by NSERCDiscovery Grants to JSS and DW ARG was supportedby a University Graduate Fellowship at UBC
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
34
REFERENCESBarker F Brown A S Budahn J R amp Plafker G (1989) Back-arc
with frontal-arc component origin of Triassic Karmutsen basaltBritish Columbia Canada Chemical Geology 75 81^102
Blichert-Toft JWeis D Maerschalk C Agranier A amp Albare de F(2003) Hawaiian hot spot dynamics as inferred from the Hf and Pbisotope evolution of Mauna Kea volcano Geochemistry GeophysicsGeosystems 4(2) 1^27 doi1010292002GC000340
Bondre N R Duraiswami R A amp Dole G (2004) Morphologyand emplacement of flows from the Deccan Volcanic ProvinceIndia Bulletin of Volcanology 66 29^45
Carlisle D (1963) Pillow breccias and their aquagene tuffs QuadraIsland British Columbia Journal of Geology 71 48^71
Carlisle D (1972) Late Paleozoic to Mid-Triassic sedimentary-volcanic sequence on Northeastern Vancouver Island Geological
Society of Canada Paper Report of Activities 72-1 Part B 24^30Carlisle D amp SuzukiT (1974) Emergent basalt and submergent car-
bonate^clastic sequences including the Upper Triassic Dilleri andWelleri zones on Vancouver Island Canadian Journal of Earth
Sciences 11 254^279Chauvel C amp Blichert-Toft J (2001) A hafnium isotope and trace
element perspective on melting of the depleted mantle Earth and
Planetary Science Letters 190 137^151Chayes F (1954)The theory of thin-section analysis Journal of Geology
62 92^101Cho M Liou J G amp Maruyama S (1987) Transition from the zeo-
lite to prehnite^pumpellyite facies in the Karmutsen metabasitesVancouver Island British Columbia Journal of Petrology 27(2)467^494
Clowes R M Zelt C A Amor J R amp Ellis R M (1995)Lithospheric structure in the southern Canadian Cordillera froma network of seismic refraction lines Canadian Journal of Earth
Sciences 32(10) 1485^1513Coffin M F amp Eldholm O (1994) Large igneous provinces Crustal
structure dimensions and external consequences Reviews of
Geophysics 32(1) 1^36Cox K G (1980) A model for flood basalt volcanism Journal of
Petrology 21 629^650Eldholm O amp Coffin M F (2000) Large igneous provinces
and plate tectonics In Richards M A Gordon R G amp vander Hilst R D (eds) The History and Dynamics of Global
Plate Motions American Geophysical Union Geophysical Monograph 121309^326
Elliott T R Hawkesworth C J amp Grolaquo nvold K (1991) Dynamicmelting of the Iceland plume Nature 351 201^206
Escuder-Viruete J Pecurren rez-Estaucurren n A Contreras F Joubert MWeis D Ullrich T D amp Spadea P (2007) Plume mantle sourceheterogeneity through time Insights from the Duarte ComplexHispaniola northeastern Caribbean Journal of Geophysical Research112(B04203) doi1010292006JB004323
Farnetani C G Richards M A amp Ghiorso M S (1996)Petrological models of magma evolution and deep crustal structurebeneath hotspots and flood basalt provinces Earth and Planetary
Science Letters 143 81^94Fitton J G amp Godard M (2004) Origin and evolution of magmas
on the Ontong Java Plateau In Fitton J G Mahoney J JWallace P J amp Saunders A D (eds) Origin and Evolution of the
Ontong Java Plateau Geological Society London Special Publications 229151^178
Fitton J G Saunders A D Norry M J Hardarson B S ampTaylor R N (1997) Thermal and chemical structure of theIceland plume Earth and Planetary Science Letters 153 197^208
Fitton J G Saunders A D Kempton P D amp Hardarson B S(2003) Does depleted mantle form an intrinsic part of the Icelandplume Geochemistry Geophysics Geosystems 4(3) 1032 doi1010292002GC000424
Frey F A Huang S Blichert-Toft J Regelous M amp Boyet M(2005) Origin of depleted components in basalt related to theHawaiian hotspot Evidence from isotopic and incompatible ele-ment ratios Geochemistry Geophysics Geosystems 6(1) 23 doi1010292004GC000757
Galer S J G amp Abouchami W (1998) Practical application of leadtriple spiking for correction of instrumental mass discriminationMineralogical Magazine 62A 491^492
Ghiorso M S amp Sack R O (1995) Chemical mass transfer in mag-matic processes IV A revised and internally consistent thermody-namic model for the interpolation and extrapolation of liquid^solidequilibria in magmatic systems at elevated temperatures and pres-sures Contributions to Mineralogy and Petrology 119 197^212
Greene A R Scoates J S amp Weis D (2005)WrangelliaTerrane onVancouver Island Distribution of flood basalts with implicationsfor potential Ni^Cu^PGE mineralization In Grant B (ed)Geological Fieldwork 2004 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2005-1 209^220Greene A R Scoates J S Nixon G T amp Weis D (2006) Picritic
lavas and basal sills in the Karmutsen flood basalt provinceWrangellia northern Vancouver Island In Grant B (ed)Geological Fieldwork 2005 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2006-1 39^52Greene A R Scoates J S amp Weis D (2008) Wrangellia flood
basalts in Alaska A record of plume^lithosphere interaction in aLate Triassic accreted oceanic plateau Geochemistry Geophysics
Geosystems 9(Q12004) doi1010292008GC002092Greene A R Scoates J S Weis D amp Israel S (2009)
Geochemistry of triassic flood basalts from the Yukon (Canada)segment of the accreted Wrangellia oceanic plateau Lithosdoi101016jlithos200811010
Hauff F Hoernle K Tilton G Graham D W amp Kerr A C(2000) Large volume recycling of oceanic lithosphere over shorttime scales geochemical constraints from the Caribbean LargeIgneous Province Earth and Planetary Science Letters 174 247^263
Hauff F Hoernle K amp Schmidt A (2003) Sr^Nd^Pb compositionof Mesozoic Pacific oceanic crust (Site 1149 and 801 ODP Leg185)Implications for alteration of ocean crust and the input into theIzu^Bonin^Mariana subduction system Geochemistry Geophysics
Geosystems 4(8) doi1010292002GC000421Herzberg C (2004) Partial melting below the Ontong Java Plateau
In Fitton J G Mahoney J J Wallace P J amp Saunders A D(eds) Origin and Evolution of the Ontong Java Plateau Geological SocietyLondon Special Publications 229 179^183
Herzberg C (2006) Petrology and thermal structure of the Hawaiianplume from Mauna Kea volcano Nature 444(7119) 605
Herzberg C amp Asimow P D (2008) Petrology of some oceanicisland basalts PRIMELT2XLS software for primary magma cal-culation Geochemistry Geophysics Geosystems 9(Q09001) doi1010292008GC002057
Herzberg C amp OrsquoHara M J (2002) Plume-associated ultramaficmagmas of Phanerozoic age Journal of Petrology 43(10) 1857^1883
Herzberg C Asimow P D Arndt N Niu Y Lesher C MFitton J G Cheadle M J amp Saunders A D (2007)Temperatures in ambient mantle and plumes Constraints frombasalts picrites and komatiites Geochemistry Geophysics Geosystems8(Q02006) doi1010292006GC001390
Huang S amp Frey F A (2005) Trace element abundances of MaunaKea basalt from phase 2 of the Hawaii Scientific Drilling Project
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
35
Petrogenetic implications of correlations with major element con-tent and isotopic ratios Geochemistry Geophysics Geosystems 4(6)doi1010292002GC000322
Hussong D M Wipperman L K amp Kroenke L W (1979) Thecrustal structure of the Ontong Java and Manihiki OceanicPlateaus Journal of Geophysical Research 84(B11) 6003^6010
Jerram D A amp Widdowson M (2005) The anatomy of ContinentalFlood Basalt Provinces geological constraints on the processes andproducts of flood volcanism Lithos 79(3^4) 385^405
Jones D L Silberling N J amp Hillhouse J (1977)Wrangellia a dis-placed terrane in northwestern North America CanadianJournal ofEarth Sciences 14(11) 2565^2577
Kerr A C (2003) Oceanic Plateaus In Rudnick R (ed) The Crust
Treatise on GeochemistryVol 3 Oxford Elsevier Science pp 537^565Kerr A C amp Mahoney J J (2007) Oceanic plateaus Problematic
plumes potential paradigms Chemical Geology 241 332^353Kerr A CTarney J Marriner G F Klaver GT Saunders A D
amp Thirlwall M F (1996) The geochemistry and petrogenesis ofthe Late-Cretaceous picrites and basalts of Curacao NetherlandsAntilles a remnant of an oceanic plateau Contributions to
Mineralogy and Petrology 124(1) 29^43Kerr A C Marriner G F Tarney J Nivia A Saunders A D
Thirlwall M F amp Sinton A C (1997) Cretaceous basaltic ter-ranes in Western Colombia Elemental chronological and Sr-Ndisotopic constraints on petrogenesis Journal of Petrology 38(6)677^702
Kinzler R J (1997) Melting of mantle peridotite at pressuresapproaching the spinel to garnet transition Application to mid-ocean ridge basalt petrogenesis Journal of Geophysical Research
102(B1) 853^874Kinzler R J amp Grove T L (1992) Primary magmas of mid-ocean
ridge basalts 1 Experiments and methods Journal of Geophysical
Research 97(B5) 6885^6906Korenaga J (2005) Why did not the Ontong Java Plateau form sub-
aerially Earth and Planetary Science Letters 234 385^399Kuniyoshi S (1972) Petrology of the Karmutsen (Volcanic) Group
northeastern Vancouver Island British Columbia PhD disserta-tion University of California Los Angeles 242 pp
Lassiter J C DePaolo D J amp Mahoney J J (1995) Geochemistry ofthe Wrangellia flood basalt province Implications for the role ofcontinental and oceanic lithosphere in flood basalt genesis Journalof Petrology 36(4) 983^1009
LeBas M J (2000) IUGS reclassification of the high-Mg and picriticvolcanic rocks Journal of Petrology 41(10) 1467^1470
Longhi J (2002) Some phase equilibrium systematics of lherzolitemelting I Geochemistry Geophysics Geosystems 3(3) doi1010292001GC000204
MacDonald G A amp Katsura T (1964) Chemical composition ofHawaiian lavas Journal of Petrology 5 82^133
Mahoney J J (1987) An isotopic survey of Pacific oceanic plateausimplications for their nature and origin In Keating B HFryer P Batiza R amp Boehlert GW (eds) Seamounts Islands andAtolls American Geophysical Union Geophysical Monograph 43 207^220
Mahoney J LeRoex A P Peng Z Fisher R L amp Natland J H(1992) Southwestern limits of Indian Ocean Ridge mantle and theorigin of low 206Pb204Pb mid-ocean ridge basalt isotope systema-tics of the central Southwest Indian Ridge (178^508E) Journal ofGeophysical Research 97(B13) 19771^19790
Mahoney J J Sinton J M Kurz M D MacDougall J DSpencer K J amp Lugmair GW (1994) Isotope and trace elementcharacteristics of a super-fast spreading ridge East Pacific rise13^238S Earth and Planetary Science Letters 121 173^193
Massey N W D (1995a) Geology and mineral resources of theAlberni^Nanaimo Lakes sheet Vancouver Island 92F1W 92F2Eand part of 92F7E British Columbia Ministry of Energy Mines and
Petroleum Resources Paper 1992-2 132 ppMassey N W D (1995b) Geology and mineral resources of the
Cowichan Lake sheet Vancouver Island 92C16 British Columbia
Ministry of Energy Mines and Petroleum Resources Paper 1992-3 112 ppMassey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005a) Digital Geology Map of British Columbia Tile NM9 MidCoast BC British Columbia Ministry of Energy and Mines Geofile
2005-2Massey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005b) Digital Geology Map of British Columbia Tile NM10Southwest BC British Columbia Ministry of Energy and Mines Geofile
2005-3McDonough W F amp Sun S (1995) The composition of the Earth
Chemical Geology 120 223^253Monger JW H amp Journeay J M (1994) Basement geology and tec-
tonic evolution of theVancouver region In Monger JW H (ed)Geology and Geological Hazards of the Vancouver Region Southwestern
British Columbia Geological Survey of Canada Bulletin 481 3^25Muller J E (1977) Geology of Vancouver Island Geological Survey of
Canada Open File Map 463Muller J E (1980)The Paleozoic Sicker Group of Vancouver Island British
Columbia Geological Survey of Canada Paper 79-30 22 ppMuller J E Northcote K E amp Carlisle D (1974) Geology and
mineral deposits of Alert Bay^Cape Scott map area VancouverIsland British Columbia Geological Survey of Canada Paper 74-8 77
Neal C R Mahoney J J Kroenke L W Duncan R A ampPetterson M G (1997) The Ontong^Java Plateau InMahoney J J amp Coffin M F (eds) Large Igneous Provinces
Continental Oceanic and Planetary FloodVolcanism American Geophysical
Union Geophysical Monograph 100 183^216Niu Y Collerson K D Batiza R Wendt J I amp Regelous M
(1999) Origin of enriched-typed mid-ocean ridge basalt at ridgesfar from mantle plumes The East Pacific Rise at 11820rsquoN Journalof Geophysical Research 104 7067^7088
Nixon G T amp Orr A J (2007) Recent revisions to the EarlyMesozoic stratigraphy of Northern Vancouver Island (NTS 102I092L) and metallogenic implicatinos British Columbia InGrant B (ed) Geological Fieldwork 2006 British Columbia Ministry of
Energy Mines and Petroleum Resources Paper 2007-1 163^177Nixon GT Hammack J L KoyanagiV M Payie G J Haggart
JW Orchard M JTozerT Friedman R M Archibald D APalfy J amp Cordey F (2006a) Geology of the Quatsino^PortMcNeill area northernVancouver Island British Columbia Ministry
of Energy Mines and Petroleum Resources Geoscience Map 2006-2 scale150 000
Nixon G T Kelman M C Stevenson D Stokes L A ampJohnston K A (2006b) Preliminary geology of the Nimpkishmap area (NTS 092L07) northern Vancouver Island BritishColumbia In Grant B amp Newell J M (eds) Geological Fieldwork2005 British Columbia Ministry of Energy Mines and Petroleum Resources
Paper 2006-1 135^152Nixon G T Laroque J Pals A Styan J Greene A R amp
Scoates J S (2008) High-Mg lavas in the Karmutsen flood basaltsnorthernVancouver Island (NTS 092L) Stratigraphic setting andmetallogenic significance In Grant B (ed) Geological Fieldwork
2007 British Columbia Ministry of Energy Mines and Petroleum
Resources Paper 2008-1 175^190Nowell G M Kempton P D Noble S R Fitton J G
Saunders A Mahoney J J amp Taylor R N (1998) High precisionHf isotope measurements of MORB and OIB by thermal ionisation
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
36
mass spectrometry insights into the depleted mantle Chemical
Geology 149 211^233Parrish R R amp McNicoll V J (1992) U^Pb age determinations
from the southern Vancouver Island area British Columbia InRadiogenic Age and Isotopic Studies Report 5 Geological Survey of
Canada Paper 91-2 79^86Peate DW (1997) The Parana^Etendeka Province In Coffin M F
amp Mahoney J (eds) Large Igneous Provinces Continental Oceanic andPlanetary Flood Volcanism American Geophysical Union Geophysical
Monograph 100 217^245Perfit M R Fornari D J Ridley W I Kirk P D Casey J
Kastens K A Reynolds J R Edwards M Desonie DShuster R amp Paradis S (1996) Recent volcanism in the Siqueirostransform fault picritic basalts and implications for MORBmagma genesis Earth and Planetary Science Letters 141(1^4) 91^108
Pretorius W Weis D Williams G Hanano D Kieffer B ampScoates J S (2006) Complete trace elemental characterization ofgranitoid (USGSG-2GSP-2) reference materials by high resolutioninductively coupled plasma-mass spectrometry Geostandards and
Geoanalytical Research 30(1) 39^54Regelous M Niu Y Wendt J I Batiza R Greig A amp
Collerson K D (1999) Variations in the geochemistry of magma-tism on the East Pacific Rise at 10830rsquoN since 800 ka Earth and
Planetary Science Letters 168 45^63Recurren villon S Chauvel C Arndt N T Pik R Martineau F
Fourcade S amp Marty B (2002) Heterogeneity of the Caribbeanplateau mantle source Sr O and He isotopic compositions of oli-vine and clinopyroxene from Gorgona Island Earth and Planetary
Science Letters 205(1^2) 91Rhodes J M (1996) Geochemical stratigraphy of lava flows samples
by the Hawaii Scientific Drilling Project Journal of Geophysical
Research 101(B5) 11729^11746Rhodes J M amp Vollinger M J (2004) Composition of basaltic lavas
sampled by phase-2 of the Hawaii Scientific Drilling ProjectGeochemical stratigraphy and magma types Geochemistry
Geophysics Geosystems 5(3) doi1010292002GC000434Richards M A Jones D L Duncan R A amp DePaolo D J (1991)
A mantle plume initiation model for the Wrangellia flood basaltand other oceanic plateaus Science 254 263^267
Salters V J M amp Hart S R (1991) The mantle sources of oceanridges islands and arcs the Hf-isotope connection Earth and
Planetary Science Letters 104 364^380Salters V J M amp Stracke A (2004) Composition of the depleted
SaltersV J amp WhiteW (1998) Hf isotope constraints on mantle evo-lution Chemical Geology 145 447^460
SanoT amp Yamashita S (2004) Experimental petrology of basementlavas from Ocean Drilling Program Leg 192 implications for dif-ferentiation processes in Ontong Java Plateau magmas InFitton J G Mahoney J JWallace P J amp Saunders A D (eds)Origin and Evolution of the Ontong Java Plateau Geological Society
London Special Publications 229 185^218Saunders A D (2005) Large igneous provinces Origin and environ-
mental consequences Elements 1 259^263Self S ThordarsonT amp Keszthely L (1997) Emplacement of conti-
nental flood basalt lava flows In Mahoney J J amp Coffin M F(eds) Large Igneous Provinces Continental Oceanic and Planetary Flood
Volcanism American Geophysical Union Geophysical Monograph 100381^410
Shaw D M (2000) Continuous (dynamic) melting theory revisitedCanadian Mineralogist 38(5) 1041^1063
Sluggett C L (2003) Uranium^lead age and geochemical constraintson Paleozoic and Early Mesozoic magmatism in WrangelliaTerrane Saltspring Island British Columbia BSc thesisUniversity of British ColumbiaVancouver 84 pp
Storey M Mahoney J J amp Saunders A D (1997) Cretaceousbasalts in Madagascar and the transition between plume and con-tinental lithospheric mantle sources In Mahoney J J ampCoffin M F (eds) Large Igneous Provinces Continental Oceanic and
Planetary Flood Volcanism Aamerican Geophysical Union Geophysical
Monograph 100 95^122Surdam R C (1967) Low-grade metamorphism of the Karmutsen
Group PhD dissertation University of California Los Angeles288 pp
Sutherland-Brown A Yorath C J Anderson R G amp Dom K(1986) Geological maps of southern Vancouver IslandLITHOPROBE I 92C10 11 14 16 92F1 2 7 8 Geological Survey ofCanada Open File 1272
Tejada M L G Mahoney J J Castillo P R Ingle S P Sheth HC amp Weis D (2004) Pin-pricking the elephant evidence on theorigin of the Ontong Java Plateau from Pb^Sr^Hf^Nd isotopiccharacteristics of ODP Leg 192 basalts In Fitton J GMahoney J J Wallace P J amp Saunders A D (eds) Origin and
Evolution of the Ontong Java Plateau Geological Society London Special
Publications 229 133^150Thompson P M E Kempton P D amp Kerr A C (2008) Evaluation
of the effects of alteration and leaching on Sm^Nd and Lu^Hf sys-tematics in submarine mafic rocks Lithos 104(1^4) 164^176
Thompson P M E Kempton P D White R V Kerr A CTarney J Saunders A D Fitton J G amp McBirney A (2003)Hf-Nd isotope constraints on the origin of the CretaceousCaribbean plateau and its relationship to the Galapagos plumeEarth and Planetary Science Letters 217(1^2) 59^75
Vervoort J D Patchett P Blichert-Toft J amp Albare de F (1999)Relationships between Lu^Hf and Sm^Nd isotopic systems in theglobal sedimentary system Earth and Planetary Science Letters 16879^99
Walter M J (1998) Melting of garnet peridotite and the origin ofkomatiite and depleted lithosphere Journal of Petrology 39(1) 29^60
Weis D Kieffer B Maerschalk C Barling J de Jong JWilliams G A Hanano D Mattielli N Scoates J SGoolaerts A Friedman R A amp Mahoney J B (2006) High-pre-cision isotopic characterization of USGS reference materials byTIMS and MC-ICP-MS Geochemistry Geophysics Geosystems
7(Q08006) doi1010292006GC001283Weis D Kieffer B Hanano D Silva I N Barling J
Pretorius W Maerschalk C amp Mattielli N (2007) Hf isotopecompositions of US Geological Survey reference materialsGeochemistry Geophysics Geosystems 8(Q06006) doi1010292006GC001473
Wheeler J O amp McFeely P (1991) Tectonic assemblage map of theCanadian Cordillera and adjacent part of the United States ofAmerica Geological Survey of Canada Map 1712A
WhiteW M Albare de F amp Tecurren louk P (2000) High-precision analy-sis of Pb isotope ratios by multi-collector ICP-MS Chemical Geology167 257^270
Yorath C J Sutherland Brown A amp Massey N W D (1999)LITHOPROBE southern Vancouver Island British Columbia Geological
Survey of Canada Bulletin 498 145 ppZou H (1998) Trace element fractionation during modal and
non-model dynamic melting and open-system melting Amathematical treatment Geochimica et Cosmochimica Acta 62(11)1937^1945
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
37
Zou H amp Reid M R (2001) Quantitative modeling of trace elementfractionation during incongruent dynamic melting Geochimica et
Cosmochimica Acta 65(1) 153^162
APPENDIX SAMPLEPREPARAT ION ANDANALYT ICAL METHODSOnly the freshest rocks in the field were sampled and onlythe least altered samples were selected for chemical andisotopic analysis based on thorough petrographic inspec-tion Fifty-nine of the 129 collected samples were crushed(400 g) into small pieces 52mm in diameter in aRocklabs hydraulic piston crusher between WC plates tominimize contamination The coarse-crush was mixedand 100 g was powdered in a planetary mill using agatejars and balls cleaned with quartz sand between samples
ActLabs analytical methodsThe major- and trace-element compositions of the whole-rock powders were determined at Activation LaboratoriesLtd (Actlabs) in Ancaster Ontario Analytical techniquesand detection limits are also available from Actlabs (httpwwwactlabscommethsub_code4erehtm) The analyticalmethod for each of the elements analyzed is indicated inTable 2 For the major elements a 02 g sample was mixedwith a lithium metaborate^lithium tetraborate mixtureand fused in a graphite crucible The molten mixture waspoured into a 5 HNO3 solution and shaken until dis-solved (30min) The samples were analyzed for majoroxides and selected trace elements by inductively coupledplasma optical emission spectrometry (ICP-OES) on acombination simultaneous^sequential Thermo Jarrell-AshEnviro II system Internal calibration was achieved usinga variety of international reference materials (eg W-2BIR-1 DNC-1) and independent control samplesAdditional trace elements were analyzed by both instru-mental neutron activation analysis (INAA) and induc-tively couple plasma mass spectrometry (ICP-MS) ForINAA15^25 g of sample was weighed into small polyeth-ylene vials and irradiated with a control international ref-erence material CANMET WMS-1 and NiCr flux wires ata thermal neutron flux of 71012 neutronscm2 per s in theMcMaster Nuclear Reactor Following a 7 day waitingperiod the samples were measured on an Ortec high-purity Ge detector linked to a Canberra Series 95 multi-channel analyzer Activities for each element weredecay- and weight-corrected and compared with a detectorcalibration developed from multiple international certifiedreference materials For ICP-MS 025 g of sample wasdigested in HF followed by HNO3^HClO4 heated andtaken to dryness The samples were brought back into solu-tion with HCl Samples were analyzed using a PerkinElmer Optima 3000 ICP In-lab standards or certified ref-erence materials (eg W-2 BIR-1 DNC-1) were used for
quality control A total of 15 blind duplicates were ana-lyzed from the same sample powders to assessreproducibility
University of Massachusetts analyticalmethodsFourteen sample duplicate powders taken from the same100 g sample aliquot (high-MgO basalts and picrites)were also analyzed at the Ronald B Gilmore X-RayFluorescence (XRF) Laboratory at the University ofMassachusetts Major elements were measured on a fusedLa-bearing lithium borate glass disc using a SiemensMRS-400 spectrometer Trace element concentrations(Rb Sr Ba Ce Nb Zr Y Pb Zn Ga Ni Cr V) weremeasured on a separate powder pellet using a PhilipsPW2400 sequential spectrometer using a Rh tube LOIand ferrous iron measurements were made as describedby Rhodes amp Vollinger (2004) Precision and accuracy esti-mates for the data have been described by Rhodes (1996)and Rhodes amp Vollinger (2004) Results for each sampleare the average of two separate analyses and are shown inElectronic Appendix 2 for a comparison of analyses ofduplicate powders using XRFand ICP-MS
PCIGR analytical methodsA subset of 19 samples was selected for high-precisiontrace-element analysis and Sr Nd Pb and Hf isotopicanalysis at the Pacific Centre for Isotopic andGeochemical Research (PCIGR) at the University ofBritish Columbia (UBC) Samples were selected from the59 samples analyzed for whole-rock chemistry at ActLabsbased on major- and trace-element chemistry alteration(low LOI and petrographic alteration index) and samplelocation Samples were prepared for trace-element analysisat the PCIGR by the technique described by Pretorius et al(2006) on unleached rock powders Sample powders(100mg) were weighed in 7ml screw-top Savillex bea-kers and dissolved in 1ml 14N HNO3 and 5ml 48 HFon a hotplate for 48 h at 1308C with periodic ultrasonica-tion Samples were dried and redissolved in 6ml 6N HClon a hotplate for 24 h and then dried and redissolved in1ml concentrated HNO3 for 24 h before final dryingTrace element abundances were measured by high-resolution inductively coupled plasma-mass spectrometry(HR-ICP-MS) with aThermo Finnigan Element2 systemfollowing the procedures described by Pretorius et al(2006) within 24 h of redissolution HFSE and LILE weremeasured in medium-resolution mode at 2000 dilutionusing a PFATeflon spray chamber washed with aqua regiafor 3min between samples REE were measured in high-resolution mode and U and Pb in low-resolution modeat 2000 dilution using a glass spray chamber washedwith 2 HNO3 between samples Total procedural blanksand reference materials (BCR-2 BHVO-2) were analyzedwith the batch of samples Indium was used as an internal
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
38
standard in all samples and standard solutionsBackground and standard solutions were analyzed afterevery five samples to detect memory effects and massdrift Results for PCIGR trace-element analyses areshown in Electronic Appendix 3 for the full set of analyzedtrace elementsSample digestion for purification of Sr Nd Hf and Pb
for column chemistry began by weighing each samplepowder (400^500mg) prior to leaching All samples wereleached with 6N HCl and placed in an ultrasonic bath for15min Samples were rinsed twice with 18 M-cm H2Obetween each leaching step (15 total) until the supernatantliquid was clear [following the method of Mahoney (1987)]Samples were then dried on a hotplate for 24 h andweighed again Sample solutions were then prepared bydissolving 100^250mg of the leached powder dissolvedin 1ml 14N HNO3 and 10ml 48 HF on a hotplate for48 h at 1308C with periodic ultrasonication Samples weredried and redissolved in 6ml 6N HCl on a hotplate for24 h and then dried Pb was separated using anion exhangecolumns and the discard was used for Sr REE and Hfseparation Nd was separated from the other REE and Hfrequired two additional purification steps Detailed proce-dures for column chemistry for separating Sr Nd and Pbat the PCIGR have been described byWeis et al (2006) andHf purification has been described byWeis et al (2007) Srand Nd isotope ratios were measured by thermal ioniza-tion mass spectrometry (TIMS) on a Thermo FinniganTriton system in static mode with relay matrix rotation ona single Ta and double Re^Ta filament respectively Fourto five filaments per barrel of 21 were occupied by stan-dards (NBS 987 for Sr and LaJolla for Nd) for each barrelwhere samples were run Sample Sr and Nd isotopic com-positions were corrected for mass fractionation using86Sr88Srfrac14 01194 and 146Nd144Ndfrac14 07219 Each samplewas then normalized using the barrel average of the refer-ence material relative to the values of143Nd144Ndfrac14 0511858 and 87Sr86Srfrac14 0710248 (Weiset al 2006) During the course of the Vancouver Islandanalyses the La Jolla Nd standard gave an average valueof 05118566 (nfrac14 7) and NBS987 standard gave an aver-age of 07102408 (nfrac1411) (2 error is reported as times106) 147Sm144Nd errors are 15 or 0006 USGeological Survey (USGS) reference material BHVO-2was processed with the samples and yielded Sr and Nd iso-topic ratios of 07034607 and 05129786 respectivelyThese are in agreement with the published values of070347920 and 051298411 respectively (Weis et al2006)
Pb and Hf isotopic compositions were analyzed by mul-tiple collector-inductively coupled plasma-mass spectrome-try (MC-ICP-MS) on a static multi-collection on a NuPlasma (Nu Instruments) system The detailed analyticalprocedure for Pb isotopic analyses on the Nu Plasma atthe PCIGR has been described by Weis et al (2006) Theconfiguration for Pb analyses allows for collection of PbTl and Hg togetherTl and Hg are used to monitor instru-mental mass discrimination and isobaric overlap repec-tively All sample solutions were analyzed withapproximately the same PbTl ratio (4) as the referencematerial NIST SRM 981 To accomplish this a small ali-quot of each sample solution from the Pb columns was ana-lyzed on the Element2 to determine the precise amount ofPb available for analysis on the Nu Plasma The SRM 981standard was run after every two samples on the NuPlasma During the time samples were run analyses ofthe SRM 981 Pb reference material gave values of206Pb204Pbfrac1416940322 207Pb204Pbfrac1415495823and 208Pb204Pbfrac14 36713164 (nfrac14 61 2 error is reportedas times 104) these values are in excellent agreement withreported TIMS triple-spike values of Galer amp Abouchami(1998) Fractionation-corrected Pb isotopic ratios were fur-ther corrected by the sample-standard bracketing methodor the ln^ln correction method described by White et al(2000) and Blichert-Toft et al (2003) Leached powders ofUSGS reference material BHVO-2 yielded Pb isotopicratios of 206Pb204Pbfrac14186454 8 207Pb204Pbfrac14 1549105 and 208Pb204Pbfrac14 38222514 and BCR-2yielded 206Pb204Pbfrac141880466 207Pb204Pbfrac14 1562518 and 208Pb204Pbfrac14 3883496 (2 error isreported as times 104) These values are in agreement withthose for leached residues of BHVO-2 and BCR-2 fromWeis et al (2006)Hf isotopic compositions were analyzed following the
procedures detailed byWeis et al (2007) The configurationfor Hf analyses monitored Lu mass 175 andYb mass 172 toallow for interference correction to masses 174 and 176 Hfisotopic ratios were normalized internally for mass fractio-nation to a 179Hf177Hf ratio of 07325 using an exponentialcorrection Standards were run after every two samplesand sample results were normalized to the ratio of the in-run daily average and a 176Hf177Hf ratio for JMC-475 of0282160 During the course of analyses the Hf standardJMC-475 gave an average value 02821679 (nfrac14130)USGS reference materials BCR-2 and BHVO-2 were pro-cessed with the samples and yielded Hf isotopic ratios of02828675 and 02831005 respectively Publishedvalues for BCR-2 and BHVO-2 are 02828717 and0283104 8 respectively (Weis et al 2007)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
39
in a single eruptive phase (c 231^225 Ma) as an emergentbasalt sequence during a deeper-water shallow-water andsubaerial stage and subsided below sea level relativelysoon after its formation (Fig 16) The deeper-water stagecontains pillowed and massive submarine flows (3000mthick) that built up the submarine volcanic edifice from theseafloor The lowest pillow basalts were emplaced onunconsolidated and lithified fine-grained siliciclastic andcarbonaceous sediments containing Daonella beds(c 235^232 Ma) that are intruded by abundantKarmutsen mafic sills (Fig 16a) The different pillowedand massive submarine flows resulted primarily from dif-ferent effusive rates and local topography Some of themassive submarine flows commonly recognized by radiat-ing columnar jointing represent master tubes for the deliv-ery of lava to distal parts of submarine flow fields Thepillowed flows which commonly have uneven topography
in modern sequences may have served as levees to themassive submarine flows rather than allowing unob-structed spreading and inflation typical of the massive sub-aerial flows that form on shallow slopes in continentalflood basalt provinces (eg Self et al 1997) The rapid suc-cession of flows is evident by the near absence of sedimentbetween submarine flows Minor volumes of picritic pillowlavas occur in the upper part of the pillowed flow sequencein areas of the plateau exposed on northern VancouverIsland (Fig 16 Nixon et al 2008)The growth of the plateau in shallow water led to
emplacement of an increasing proportion of volcaniclastics(pillow breccias and hyaloclastites) that accumulated up to1500m thick in some areas (Fig 16b Nixon et al 2008)Horizons of pillowed flows occur locally within the volca-niclastic stratigraphy Sedimentary structures (eg gradedbedding fluidization structures) and petrographic textures
Late Devonian-Mississippian arc
Daonella beds
Middle Triassic
deeper water
Late Devonian-Mississippian arc
pillowed flows massive submarine
flows
pillow brecciahyaloclastiteshallow water
~4 km
~6 km
Late Triassic limestoneinterflow limestonelenses
Karmutsen Formation
mafic intrusions sediment-sill complex
Karmutsen Formation
(a) (b)
(c) (d)
subaerial flows
Late Devonian-Mississippian arc Late Devonian-Mississippian arc
picritic pillow basalts
Miss to Permian marine sediments
sills
Fig 16 Growth of the Wrangellia oceanic plateau onVancouver Island (a) Deep-water eruptive phase of pillowed and massive submarineflows with almost no intervening sediments onto MiddleTriassic shale containing Daonella beds (c 235^232 Ma) The basement of the plateauis composed of Devonian to Mississippian arc sequences overlain by Mississippian to Permian carbonates and siliciclastic sediments(b) Shallow-water flow emplacement with increasing proportion of volcaniclastics resulted primarily from cooling^contraction granulationPicritic pillow basalts occur primarily near the transition between pillowed submarine flows and volcaniclastic units Pillowed and massiveflows occur locally within the volcaniclastic unit (c) Emergent stage of growth consists of overlapping subaerial flow fields similar to continen-tal flood basalts with pahoehoe flow structures Multi-tiered columnar jointing occurrs locally in response to ponding of flows (eg Bondre et al2004) Interflow carbonate deposits and plagioclase megacrystic flows developed locally during the waning stage of volcanism (d) Post-volcanicsubsidence and deposition of Late Carnian to Norian (c 212^223 Ma) platform carbonates on top of the plateau Units are labeled in the dia-grams Thicknesses of the underlying units are not to scale
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
33
within the volcaniclastic rocks indicate formation primar-ily via cooling^contraction granulation magma^water^steam interaction autobrecciation resedimentation andmass-wasting rather than pyroclastic eruption (Nixonet al 2008)A broad subaerial platform was constructed by emplace-
ment of subaerial lava flows during an emergent subaerialstage (Fig 16c) The subaerial unit predominantly com-prises well-layered sheet flows that formed in a similarway to continental flood basalts as hundred(s) of pahoehoeflow fields erupted on shallow slopes (Self et al 1997) how-ever volcaniclastic and pillowed flow units are locally pre-sent mostly in the upper parts of the stratigraphyLocalized ponding of flows led to the formation of multi-tiered columnar jointing in some subaerial flows (egBondre et al 2004) As volcanism waned local interflowcarbonate deposits developed and plagioclase megacrysticflows were erupted Subsidence of the plateau is recordedby the interflow lenses and deposits of platform carbonates(40^1500m thick) on top of the Karmutsen basalts thatcontain Late Carnian to Early Norian fossils (Fig 16d c216^227 Ma Nixon et al 2008)The evolution of the internal architecture of the
Wrangellia oceanic plateau was primarily related to theconditions of emplacement (ie water depth and eruptionenvironment) There does not appear to have been a sys-tematic coincident evolution or shift of lava compositionsonVancouver Island as has been recognized in most con-tinental flood basalts and parts of the Wrangellia oceanicplateau in Alaska and Yukon (Greene et al 2008 2009)In general with the exception of the picritic pillow lavasin the submarine stratigraphy and local variations betweenhigh-MgO and tholeiitic lavas there are no significantcompositional differences between lavas erupted duringthe different stages of growth of the plateau (ie submarineand subaerial flows) nor are substantial compositional dif-ferences generally found between massive submarine flowsin contact with pillowed flows tholeiitic basalt flows nearthe base of the Karmutsen Formation are compositionallysimilar to flows sampled near the top of the volcanicsequence Where it was possible to sample continuous sec-tions of volcanic stratigraphy greater than several hun-dreds of meters (eg Keogh^Maynard Lake area west ofthe Karmutsen Range Fig 2) there is no clear relationshipbetween stratigraphic position and compositionalvariation
CONCLUSIONSThe Karmutsen Formation covers extensive areas ofVancouver Island (20 000 km2) and forms part of a largeoceanic plateau erupted over a geologically short interval(c 225^231 Ma) The tripartite volcanic stratigraphyexposed on Vancouver Island is at least 6 km thick andcomprises a succession of submarine flows volcaniclastic
deposits and massive subaerial flowsThese volcanologicaldifferences are primarily related to the eruption environ-ment (deep-water shallow-water or subaerial) The rapidgrowth of the plateau prevented significant accumulationsof intervening sediment except in the uppermost stratigra-phy where isolated limestone lenses commonly associatedwith pillow basalt and hyaloclastite deposits preserve arecord of the final stages of subsidence of the plateauThe Wrangellia plateau on Vancouver Island was con-
structed dominantly of tholeiitic basalt with restrictedmajor- and trace-element and isotopic compositional var-iations Minor volumes of picritic pillow basalts eruptedlate in the submarine phase The Karmutsen basalts wereemplaced onto older Paleozoic arc volcanic sequences andlow NbLa for the picrites may indicate a contributionfrom the sub-arc lithospheric mantle Modeling of the pet-rogenesis of the Karmutsen picrites indicates melting ofanomalously hot mantle (15008C) and extensive degreesof partial melting (23^27) and is consistent with aplume initiation model Combined Sr^Nd^Hf^Pb isotopicsystematics indicate that the picritic and tholeiitic lavaswere derived from a depleted plume-type mantle sourcedistinct from the source of MORB that has compositionalsimilarities to the source of the lavas forming theCaribbean Plateau Picrites were generated by melting ofspinel lherzolite whereas the tholeiitic basalts formed asaggregate melts of both garnet and spinel lherzolite Thetholeiitic basalts underwent extensive low-pressure fractio-nation (52^3 kbar) and seismic studies indicate that someof theWrangellia crust beneathVancouver Island may cor-respond to the plutonic residues of the Karmutsen floodbasalts
ACKNOWLEDGEMENTSWe are grateful to Nick Arndt for helping us get this pro-ject started and Nick Massey for insights into VancouverIsland geology We greatly appreciate the assistance withPRIMELT modeling from Claude Herzberg We thankMikkel Schau for his insight and enthusiasm during field-work Jane Barling assisted with analyses by MC-ICP-MSThoughtful and thorough reviews of the manuscript byJohn Mahoney Fred Frey Wendy Bohrson DougalJerram and Teal Riley are greatly appreciated Fundingwas generously provided by the Rocks to Riches Programadministered by the BC amp Yukon Chamber of Mines in2004 by the BC Geological Survey and by NSERCDiscovery Grants to JSS and DW ARG was supportedby a University Graduate Fellowship at UBC
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
34
REFERENCESBarker F Brown A S Budahn J R amp Plafker G (1989) Back-arc
with frontal-arc component origin of Triassic Karmutsen basaltBritish Columbia Canada Chemical Geology 75 81^102
Blichert-Toft JWeis D Maerschalk C Agranier A amp Albare de F(2003) Hawaiian hot spot dynamics as inferred from the Hf and Pbisotope evolution of Mauna Kea volcano Geochemistry GeophysicsGeosystems 4(2) 1^27 doi1010292002GC000340
Bondre N R Duraiswami R A amp Dole G (2004) Morphologyand emplacement of flows from the Deccan Volcanic ProvinceIndia Bulletin of Volcanology 66 29^45
Carlisle D (1963) Pillow breccias and their aquagene tuffs QuadraIsland British Columbia Journal of Geology 71 48^71
Carlisle D (1972) Late Paleozoic to Mid-Triassic sedimentary-volcanic sequence on Northeastern Vancouver Island Geological
Society of Canada Paper Report of Activities 72-1 Part B 24^30Carlisle D amp SuzukiT (1974) Emergent basalt and submergent car-
bonate^clastic sequences including the Upper Triassic Dilleri andWelleri zones on Vancouver Island Canadian Journal of Earth
Sciences 11 254^279Chauvel C amp Blichert-Toft J (2001) A hafnium isotope and trace
element perspective on melting of the depleted mantle Earth and
Planetary Science Letters 190 137^151Chayes F (1954)The theory of thin-section analysis Journal of Geology
62 92^101Cho M Liou J G amp Maruyama S (1987) Transition from the zeo-
lite to prehnite^pumpellyite facies in the Karmutsen metabasitesVancouver Island British Columbia Journal of Petrology 27(2)467^494
Clowes R M Zelt C A Amor J R amp Ellis R M (1995)Lithospheric structure in the southern Canadian Cordillera froma network of seismic refraction lines Canadian Journal of Earth
Sciences 32(10) 1485^1513Coffin M F amp Eldholm O (1994) Large igneous provinces Crustal
structure dimensions and external consequences Reviews of
Geophysics 32(1) 1^36Cox K G (1980) A model for flood basalt volcanism Journal of
Petrology 21 629^650Eldholm O amp Coffin M F (2000) Large igneous provinces
and plate tectonics In Richards M A Gordon R G amp vander Hilst R D (eds) The History and Dynamics of Global
Plate Motions American Geophysical Union Geophysical Monograph 121309^326
Elliott T R Hawkesworth C J amp Grolaquo nvold K (1991) Dynamicmelting of the Iceland plume Nature 351 201^206
Escuder-Viruete J Pecurren rez-Estaucurren n A Contreras F Joubert MWeis D Ullrich T D amp Spadea P (2007) Plume mantle sourceheterogeneity through time Insights from the Duarte ComplexHispaniola northeastern Caribbean Journal of Geophysical Research112(B04203) doi1010292006JB004323
Farnetani C G Richards M A amp Ghiorso M S (1996)Petrological models of magma evolution and deep crustal structurebeneath hotspots and flood basalt provinces Earth and Planetary
Science Letters 143 81^94Fitton J G amp Godard M (2004) Origin and evolution of magmas
on the Ontong Java Plateau In Fitton J G Mahoney J JWallace P J amp Saunders A D (eds) Origin and Evolution of the
Ontong Java Plateau Geological Society London Special Publications 229151^178
Fitton J G Saunders A D Norry M J Hardarson B S ampTaylor R N (1997) Thermal and chemical structure of theIceland plume Earth and Planetary Science Letters 153 197^208
Fitton J G Saunders A D Kempton P D amp Hardarson B S(2003) Does depleted mantle form an intrinsic part of the Icelandplume Geochemistry Geophysics Geosystems 4(3) 1032 doi1010292002GC000424
Frey F A Huang S Blichert-Toft J Regelous M amp Boyet M(2005) Origin of depleted components in basalt related to theHawaiian hotspot Evidence from isotopic and incompatible ele-ment ratios Geochemistry Geophysics Geosystems 6(1) 23 doi1010292004GC000757
Galer S J G amp Abouchami W (1998) Practical application of leadtriple spiking for correction of instrumental mass discriminationMineralogical Magazine 62A 491^492
Ghiorso M S amp Sack R O (1995) Chemical mass transfer in mag-matic processes IV A revised and internally consistent thermody-namic model for the interpolation and extrapolation of liquid^solidequilibria in magmatic systems at elevated temperatures and pres-sures Contributions to Mineralogy and Petrology 119 197^212
Greene A R Scoates J S amp Weis D (2005)WrangelliaTerrane onVancouver Island Distribution of flood basalts with implicationsfor potential Ni^Cu^PGE mineralization In Grant B (ed)Geological Fieldwork 2004 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2005-1 209^220Greene A R Scoates J S Nixon G T amp Weis D (2006) Picritic
lavas and basal sills in the Karmutsen flood basalt provinceWrangellia northern Vancouver Island In Grant B (ed)Geological Fieldwork 2005 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2006-1 39^52Greene A R Scoates J S amp Weis D (2008) Wrangellia flood
basalts in Alaska A record of plume^lithosphere interaction in aLate Triassic accreted oceanic plateau Geochemistry Geophysics
Geosystems 9(Q12004) doi1010292008GC002092Greene A R Scoates J S Weis D amp Israel S (2009)
Geochemistry of triassic flood basalts from the Yukon (Canada)segment of the accreted Wrangellia oceanic plateau Lithosdoi101016jlithos200811010
Hauff F Hoernle K Tilton G Graham D W amp Kerr A C(2000) Large volume recycling of oceanic lithosphere over shorttime scales geochemical constraints from the Caribbean LargeIgneous Province Earth and Planetary Science Letters 174 247^263
Hauff F Hoernle K amp Schmidt A (2003) Sr^Nd^Pb compositionof Mesozoic Pacific oceanic crust (Site 1149 and 801 ODP Leg185)Implications for alteration of ocean crust and the input into theIzu^Bonin^Mariana subduction system Geochemistry Geophysics
Geosystems 4(8) doi1010292002GC000421Herzberg C (2004) Partial melting below the Ontong Java Plateau
In Fitton J G Mahoney J J Wallace P J amp Saunders A D(eds) Origin and Evolution of the Ontong Java Plateau Geological SocietyLondon Special Publications 229 179^183
Herzberg C (2006) Petrology and thermal structure of the Hawaiianplume from Mauna Kea volcano Nature 444(7119) 605
Herzberg C amp Asimow P D (2008) Petrology of some oceanicisland basalts PRIMELT2XLS software for primary magma cal-culation Geochemistry Geophysics Geosystems 9(Q09001) doi1010292008GC002057
Herzberg C amp OrsquoHara M J (2002) Plume-associated ultramaficmagmas of Phanerozoic age Journal of Petrology 43(10) 1857^1883
Herzberg C Asimow P D Arndt N Niu Y Lesher C MFitton J G Cheadle M J amp Saunders A D (2007)Temperatures in ambient mantle and plumes Constraints frombasalts picrites and komatiites Geochemistry Geophysics Geosystems8(Q02006) doi1010292006GC001390
Huang S amp Frey F A (2005) Trace element abundances of MaunaKea basalt from phase 2 of the Hawaii Scientific Drilling Project
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
35
Petrogenetic implications of correlations with major element con-tent and isotopic ratios Geochemistry Geophysics Geosystems 4(6)doi1010292002GC000322
Hussong D M Wipperman L K amp Kroenke L W (1979) Thecrustal structure of the Ontong Java and Manihiki OceanicPlateaus Journal of Geophysical Research 84(B11) 6003^6010
Jerram D A amp Widdowson M (2005) The anatomy of ContinentalFlood Basalt Provinces geological constraints on the processes andproducts of flood volcanism Lithos 79(3^4) 385^405
Jones D L Silberling N J amp Hillhouse J (1977)Wrangellia a dis-placed terrane in northwestern North America CanadianJournal ofEarth Sciences 14(11) 2565^2577
Kerr A C (2003) Oceanic Plateaus In Rudnick R (ed) The Crust
Treatise on GeochemistryVol 3 Oxford Elsevier Science pp 537^565Kerr A C amp Mahoney J J (2007) Oceanic plateaus Problematic
plumes potential paradigms Chemical Geology 241 332^353Kerr A CTarney J Marriner G F Klaver GT Saunders A D
amp Thirlwall M F (1996) The geochemistry and petrogenesis ofthe Late-Cretaceous picrites and basalts of Curacao NetherlandsAntilles a remnant of an oceanic plateau Contributions to
Mineralogy and Petrology 124(1) 29^43Kerr A C Marriner G F Tarney J Nivia A Saunders A D
Thirlwall M F amp Sinton A C (1997) Cretaceous basaltic ter-ranes in Western Colombia Elemental chronological and Sr-Ndisotopic constraints on petrogenesis Journal of Petrology 38(6)677^702
Kinzler R J (1997) Melting of mantle peridotite at pressuresapproaching the spinel to garnet transition Application to mid-ocean ridge basalt petrogenesis Journal of Geophysical Research
102(B1) 853^874Kinzler R J amp Grove T L (1992) Primary magmas of mid-ocean
ridge basalts 1 Experiments and methods Journal of Geophysical
Research 97(B5) 6885^6906Korenaga J (2005) Why did not the Ontong Java Plateau form sub-
aerially Earth and Planetary Science Letters 234 385^399Kuniyoshi S (1972) Petrology of the Karmutsen (Volcanic) Group
northeastern Vancouver Island British Columbia PhD disserta-tion University of California Los Angeles 242 pp
Lassiter J C DePaolo D J amp Mahoney J J (1995) Geochemistry ofthe Wrangellia flood basalt province Implications for the role ofcontinental and oceanic lithosphere in flood basalt genesis Journalof Petrology 36(4) 983^1009
LeBas M J (2000) IUGS reclassification of the high-Mg and picriticvolcanic rocks Journal of Petrology 41(10) 1467^1470
Longhi J (2002) Some phase equilibrium systematics of lherzolitemelting I Geochemistry Geophysics Geosystems 3(3) doi1010292001GC000204
MacDonald G A amp Katsura T (1964) Chemical composition ofHawaiian lavas Journal of Petrology 5 82^133
Mahoney J J (1987) An isotopic survey of Pacific oceanic plateausimplications for their nature and origin In Keating B HFryer P Batiza R amp Boehlert GW (eds) Seamounts Islands andAtolls American Geophysical Union Geophysical Monograph 43 207^220
Mahoney J LeRoex A P Peng Z Fisher R L amp Natland J H(1992) Southwestern limits of Indian Ocean Ridge mantle and theorigin of low 206Pb204Pb mid-ocean ridge basalt isotope systema-tics of the central Southwest Indian Ridge (178^508E) Journal ofGeophysical Research 97(B13) 19771^19790
Mahoney J J Sinton J M Kurz M D MacDougall J DSpencer K J amp Lugmair GW (1994) Isotope and trace elementcharacteristics of a super-fast spreading ridge East Pacific rise13^238S Earth and Planetary Science Letters 121 173^193
Massey N W D (1995a) Geology and mineral resources of theAlberni^Nanaimo Lakes sheet Vancouver Island 92F1W 92F2Eand part of 92F7E British Columbia Ministry of Energy Mines and
Petroleum Resources Paper 1992-2 132 ppMassey N W D (1995b) Geology and mineral resources of the
Cowichan Lake sheet Vancouver Island 92C16 British Columbia
Ministry of Energy Mines and Petroleum Resources Paper 1992-3 112 ppMassey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005a) Digital Geology Map of British Columbia Tile NM9 MidCoast BC British Columbia Ministry of Energy and Mines Geofile
2005-2Massey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005b) Digital Geology Map of British Columbia Tile NM10Southwest BC British Columbia Ministry of Energy and Mines Geofile
2005-3McDonough W F amp Sun S (1995) The composition of the Earth
Chemical Geology 120 223^253Monger JW H amp Journeay J M (1994) Basement geology and tec-
tonic evolution of theVancouver region In Monger JW H (ed)Geology and Geological Hazards of the Vancouver Region Southwestern
British Columbia Geological Survey of Canada Bulletin 481 3^25Muller J E (1977) Geology of Vancouver Island Geological Survey of
Canada Open File Map 463Muller J E (1980)The Paleozoic Sicker Group of Vancouver Island British
Columbia Geological Survey of Canada Paper 79-30 22 ppMuller J E Northcote K E amp Carlisle D (1974) Geology and
mineral deposits of Alert Bay^Cape Scott map area VancouverIsland British Columbia Geological Survey of Canada Paper 74-8 77
Neal C R Mahoney J J Kroenke L W Duncan R A ampPetterson M G (1997) The Ontong^Java Plateau InMahoney J J amp Coffin M F (eds) Large Igneous Provinces
Continental Oceanic and Planetary FloodVolcanism American Geophysical
Union Geophysical Monograph 100 183^216Niu Y Collerson K D Batiza R Wendt J I amp Regelous M
(1999) Origin of enriched-typed mid-ocean ridge basalt at ridgesfar from mantle plumes The East Pacific Rise at 11820rsquoN Journalof Geophysical Research 104 7067^7088
Nixon G T amp Orr A J (2007) Recent revisions to the EarlyMesozoic stratigraphy of Northern Vancouver Island (NTS 102I092L) and metallogenic implicatinos British Columbia InGrant B (ed) Geological Fieldwork 2006 British Columbia Ministry of
Energy Mines and Petroleum Resources Paper 2007-1 163^177Nixon GT Hammack J L KoyanagiV M Payie G J Haggart
JW Orchard M JTozerT Friedman R M Archibald D APalfy J amp Cordey F (2006a) Geology of the Quatsino^PortMcNeill area northernVancouver Island British Columbia Ministry
of Energy Mines and Petroleum Resources Geoscience Map 2006-2 scale150 000
Nixon G T Kelman M C Stevenson D Stokes L A ampJohnston K A (2006b) Preliminary geology of the Nimpkishmap area (NTS 092L07) northern Vancouver Island BritishColumbia In Grant B amp Newell J M (eds) Geological Fieldwork2005 British Columbia Ministry of Energy Mines and Petroleum Resources
Paper 2006-1 135^152Nixon G T Laroque J Pals A Styan J Greene A R amp
Scoates J S (2008) High-Mg lavas in the Karmutsen flood basaltsnorthernVancouver Island (NTS 092L) Stratigraphic setting andmetallogenic significance In Grant B (ed) Geological Fieldwork
2007 British Columbia Ministry of Energy Mines and Petroleum
Resources Paper 2008-1 175^190Nowell G M Kempton P D Noble S R Fitton J G
Saunders A Mahoney J J amp Taylor R N (1998) High precisionHf isotope measurements of MORB and OIB by thermal ionisation
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
36
mass spectrometry insights into the depleted mantle Chemical
Geology 149 211^233Parrish R R amp McNicoll V J (1992) U^Pb age determinations
from the southern Vancouver Island area British Columbia InRadiogenic Age and Isotopic Studies Report 5 Geological Survey of
Canada Paper 91-2 79^86Peate DW (1997) The Parana^Etendeka Province In Coffin M F
amp Mahoney J (eds) Large Igneous Provinces Continental Oceanic andPlanetary Flood Volcanism American Geophysical Union Geophysical
Monograph 100 217^245Perfit M R Fornari D J Ridley W I Kirk P D Casey J
Kastens K A Reynolds J R Edwards M Desonie DShuster R amp Paradis S (1996) Recent volcanism in the Siqueirostransform fault picritic basalts and implications for MORBmagma genesis Earth and Planetary Science Letters 141(1^4) 91^108
Pretorius W Weis D Williams G Hanano D Kieffer B ampScoates J S (2006) Complete trace elemental characterization ofgranitoid (USGSG-2GSP-2) reference materials by high resolutioninductively coupled plasma-mass spectrometry Geostandards and
Geoanalytical Research 30(1) 39^54Regelous M Niu Y Wendt J I Batiza R Greig A amp
Collerson K D (1999) Variations in the geochemistry of magma-tism on the East Pacific Rise at 10830rsquoN since 800 ka Earth and
Planetary Science Letters 168 45^63Recurren villon S Chauvel C Arndt N T Pik R Martineau F
Fourcade S amp Marty B (2002) Heterogeneity of the Caribbeanplateau mantle source Sr O and He isotopic compositions of oli-vine and clinopyroxene from Gorgona Island Earth and Planetary
Science Letters 205(1^2) 91Rhodes J M (1996) Geochemical stratigraphy of lava flows samples
by the Hawaii Scientific Drilling Project Journal of Geophysical
Research 101(B5) 11729^11746Rhodes J M amp Vollinger M J (2004) Composition of basaltic lavas
sampled by phase-2 of the Hawaii Scientific Drilling ProjectGeochemical stratigraphy and magma types Geochemistry
Geophysics Geosystems 5(3) doi1010292002GC000434Richards M A Jones D L Duncan R A amp DePaolo D J (1991)
A mantle plume initiation model for the Wrangellia flood basaltand other oceanic plateaus Science 254 263^267
Salters V J M amp Hart S R (1991) The mantle sources of oceanridges islands and arcs the Hf-isotope connection Earth and
Planetary Science Letters 104 364^380Salters V J M amp Stracke A (2004) Composition of the depleted
SaltersV J amp WhiteW (1998) Hf isotope constraints on mantle evo-lution Chemical Geology 145 447^460
SanoT amp Yamashita S (2004) Experimental petrology of basementlavas from Ocean Drilling Program Leg 192 implications for dif-ferentiation processes in Ontong Java Plateau magmas InFitton J G Mahoney J JWallace P J amp Saunders A D (eds)Origin and Evolution of the Ontong Java Plateau Geological Society
London Special Publications 229 185^218Saunders A D (2005) Large igneous provinces Origin and environ-
mental consequences Elements 1 259^263Self S ThordarsonT amp Keszthely L (1997) Emplacement of conti-
nental flood basalt lava flows In Mahoney J J amp Coffin M F(eds) Large Igneous Provinces Continental Oceanic and Planetary Flood
Volcanism American Geophysical Union Geophysical Monograph 100381^410
Shaw D M (2000) Continuous (dynamic) melting theory revisitedCanadian Mineralogist 38(5) 1041^1063
Sluggett C L (2003) Uranium^lead age and geochemical constraintson Paleozoic and Early Mesozoic magmatism in WrangelliaTerrane Saltspring Island British Columbia BSc thesisUniversity of British ColumbiaVancouver 84 pp
Storey M Mahoney J J amp Saunders A D (1997) Cretaceousbasalts in Madagascar and the transition between plume and con-tinental lithospheric mantle sources In Mahoney J J ampCoffin M F (eds) Large Igneous Provinces Continental Oceanic and
Planetary Flood Volcanism Aamerican Geophysical Union Geophysical
Monograph 100 95^122Surdam R C (1967) Low-grade metamorphism of the Karmutsen
Group PhD dissertation University of California Los Angeles288 pp
Sutherland-Brown A Yorath C J Anderson R G amp Dom K(1986) Geological maps of southern Vancouver IslandLITHOPROBE I 92C10 11 14 16 92F1 2 7 8 Geological Survey ofCanada Open File 1272
Tejada M L G Mahoney J J Castillo P R Ingle S P Sheth HC amp Weis D (2004) Pin-pricking the elephant evidence on theorigin of the Ontong Java Plateau from Pb^Sr^Hf^Nd isotopiccharacteristics of ODP Leg 192 basalts In Fitton J GMahoney J J Wallace P J amp Saunders A D (eds) Origin and
Evolution of the Ontong Java Plateau Geological Society London Special
Publications 229 133^150Thompson P M E Kempton P D amp Kerr A C (2008) Evaluation
of the effects of alteration and leaching on Sm^Nd and Lu^Hf sys-tematics in submarine mafic rocks Lithos 104(1^4) 164^176
Thompson P M E Kempton P D White R V Kerr A CTarney J Saunders A D Fitton J G amp McBirney A (2003)Hf-Nd isotope constraints on the origin of the CretaceousCaribbean plateau and its relationship to the Galapagos plumeEarth and Planetary Science Letters 217(1^2) 59^75
Vervoort J D Patchett P Blichert-Toft J amp Albare de F (1999)Relationships between Lu^Hf and Sm^Nd isotopic systems in theglobal sedimentary system Earth and Planetary Science Letters 16879^99
Walter M J (1998) Melting of garnet peridotite and the origin ofkomatiite and depleted lithosphere Journal of Petrology 39(1) 29^60
Weis D Kieffer B Maerschalk C Barling J de Jong JWilliams G A Hanano D Mattielli N Scoates J SGoolaerts A Friedman R A amp Mahoney J B (2006) High-pre-cision isotopic characterization of USGS reference materials byTIMS and MC-ICP-MS Geochemistry Geophysics Geosystems
7(Q08006) doi1010292006GC001283Weis D Kieffer B Hanano D Silva I N Barling J
Pretorius W Maerschalk C amp Mattielli N (2007) Hf isotopecompositions of US Geological Survey reference materialsGeochemistry Geophysics Geosystems 8(Q06006) doi1010292006GC001473
Wheeler J O amp McFeely P (1991) Tectonic assemblage map of theCanadian Cordillera and adjacent part of the United States ofAmerica Geological Survey of Canada Map 1712A
WhiteW M Albare de F amp Tecurren louk P (2000) High-precision analy-sis of Pb isotope ratios by multi-collector ICP-MS Chemical Geology167 257^270
Yorath C J Sutherland Brown A amp Massey N W D (1999)LITHOPROBE southern Vancouver Island British Columbia Geological
Survey of Canada Bulletin 498 145 ppZou H (1998) Trace element fractionation during modal and
non-model dynamic melting and open-system melting Amathematical treatment Geochimica et Cosmochimica Acta 62(11)1937^1945
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
37
Zou H amp Reid M R (2001) Quantitative modeling of trace elementfractionation during incongruent dynamic melting Geochimica et
Cosmochimica Acta 65(1) 153^162
APPENDIX SAMPLEPREPARAT ION ANDANALYT ICAL METHODSOnly the freshest rocks in the field were sampled and onlythe least altered samples were selected for chemical andisotopic analysis based on thorough petrographic inspec-tion Fifty-nine of the 129 collected samples were crushed(400 g) into small pieces 52mm in diameter in aRocklabs hydraulic piston crusher between WC plates tominimize contamination The coarse-crush was mixedand 100 g was powdered in a planetary mill using agatejars and balls cleaned with quartz sand between samples
ActLabs analytical methodsThe major- and trace-element compositions of the whole-rock powders were determined at Activation LaboratoriesLtd (Actlabs) in Ancaster Ontario Analytical techniquesand detection limits are also available from Actlabs (httpwwwactlabscommethsub_code4erehtm) The analyticalmethod for each of the elements analyzed is indicated inTable 2 For the major elements a 02 g sample was mixedwith a lithium metaborate^lithium tetraborate mixtureand fused in a graphite crucible The molten mixture waspoured into a 5 HNO3 solution and shaken until dis-solved (30min) The samples were analyzed for majoroxides and selected trace elements by inductively coupledplasma optical emission spectrometry (ICP-OES) on acombination simultaneous^sequential Thermo Jarrell-AshEnviro II system Internal calibration was achieved usinga variety of international reference materials (eg W-2BIR-1 DNC-1) and independent control samplesAdditional trace elements were analyzed by both instru-mental neutron activation analysis (INAA) and induc-tively couple plasma mass spectrometry (ICP-MS) ForINAA15^25 g of sample was weighed into small polyeth-ylene vials and irradiated with a control international ref-erence material CANMET WMS-1 and NiCr flux wires ata thermal neutron flux of 71012 neutronscm2 per s in theMcMaster Nuclear Reactor Following a 7 day waitingperiod the samples were measured on an Ortec high-purity Ge detector linked to a Canberra Series 95 multi-channel analyzer Activities for each element weredecay- and weight-corrected and compared with a detectorcalibration developed from multiple international certifiedreference materials For ICP-MS 025 g of sample wasdigested in HF followed by HNO3^HClO4 heated andtaken to dryness The samples were brought back into solu-tion with HCl Samples were analyzed using a PerkinElmer Optima 3000 ICP In-lab standards or certified ref-erence materials (eg W-2 BIR-1 DNC-1) were used for
quality control A total of 15 blind duplicates were ana-lyzed from the same sample powders to assessreproducibility
University of Massachusetts analyticalmethodsFourteen sample duplicate powders taken from the same100 g sample aliquot (high-MgO basalts and picrites)were also analyzed at the Ronald B Gilmore X-RayFluorescence (XRF) Laboratory at the University ofMassachusetts Major elements were measured on a fusedLa-bearing lithium borate glass disc using a SiemensMRS-400 spectrometer Trace element concentrations(Rb Sr Ba Ce Nb Zr Y Pb Zn Ga Ni Cr V) weremeasured on a separate powder pellet using a PhilipsPW2400 sequential spectrometer using a Rh tube LOIand ferrous iron measurements were made as describedby Rhodes amp Vollinger (2004) Precision and accuracy esti-mates for the data have been described by Rhodes (1996)and Rhodes amp Vollinger (2004) Results for each sampleare the average of two separate analyses and are shown inElectronic Appendix 2 for a comparison of analyses ofduplicate powders using XRFand ICP-MS
PCIGR analytical methodsA subset of 19 samples was selected for high-precisiontrace-element analysis and Sr Nd Pb and Hf isotopicanalysis at the Pacific Centre for Isotopic andGeochemical Research (PCIGR) at the University ofBritish Columbia (UBC) Samples were selected from the59 samples analyzed for whole-rock chemistry at ActLabsbased on major- and trace-element chemistry alteration(low LOI and petrographic alteration index) and samplelocation Samples were prepared for trace-element analysisat the PCIGR by the technique described by Pretorius et al(2006) on unleached rock powders Sample powders(100mg) were weighed in 7ml screw-top Savillex bea-kers and dissolved in 1ml 14N HNO3 and 5ml 48 HFon a hotplate for 48 h at 1308C with periodic ultrasonica-tion Samples were dried and redissolved in 6ml 6N HClon a hotplate for 24 h and then dried and redissolved in1ml concentrated HNO3 for 24 h before final dryingTrace element abundances were measured by high-resolution inductively coupled plasma-mass spectrometry(HR-ICP-MS) with aThermo Finnigan Element2 systemfollowing the procedures described by Pretorius et al(2006) within 24 h of redissolution HFSE and LILE weremeasured in medium-resolution mode at 2000 dilutionusing a PFATeflon spray chamber washed with aqua regiafor 3min between samples REE were measured in high-resolution mode and U and Pb in low-resolution modeat 2000 dilution using a glass spray chamber washedwith 2 HNO3 between samples Total procedural blanksand reference materials (BCR-2 BHVO-2) were analyzedwith the batch of samples Indium was used as an internal
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
38
standard in all samples and standard solutionsBackground and standard solutions were analyzed afterevery five samples to detect memory effects and massdrift Results for PCIGR trace-element analyses areshown in Electronic Appendix 3 for the full set of analyzedtrace elementsSample digestion for purification of Sr Nd Hf and Pb
for column chemistry began by weighing each samplepowder (400^500mg) prior to leaching All samples wereleached with 6N HCl and placed in an ultrasonic bath for15min Samples were rinsed twice with 18 M-cm H2Obetween each leaching step (15 total) until the supernatantliquid was clear [following the method of Mahoney (1987)]Samples were then dried on a hotplate for 24 h andweighed again Sample solutions were then prepared bydissolving 100^250mg of the leached powder dissolvedin 1ml 14N HNO3 and 10ml 48 HF on a hotplate for48 h at 1308C with periodic ultrasonication Samples weredried and redissolved in 6ml 6N HCl on a hotplate for24 h and then dried Pb was separated using anion exhangecolumns and the discard was used for Sr REE and Hfseparation Nd was separated from the other REE and Hfrequired two additional purification steps Detailed proce-dures for column chemistry for separating Sr Nd and Pbat the PCIGR have been described byWeis et al (2006) andHf purification has been described byWeis et al (2007) Srand Nd isotope ratios were measured by thermal ioniza-tion mass spectrometry (TIMS) on a Thermo FinniganTriton system in static mode with relay matrix rotation ona single Ta and double Re^Ta filament respectively Fourto five filaments per barrel of 21 were occupied by stan-dards (NBS 987 for Sr and LaJolla for Nd) for each barrelwhere samples were run Sample Sr and Nd isotopic com-positions were corrected for mass fractionation using86Sr88Srfrac14 01194 and 146Nd144Ndfrac14 07219 Each samplewas then normalized using the barrel average of the refer-ence material relative to the values of143Nd144Ndfrac14 0511858 and 87Sr86Srfrac14 0710248 (Weiset al 2006) During the course of the Vancouver Islandanalyses the La Jolla Nd standard gave an average valueof 05118566 (nfrac14 7) and NBS987 standard gave an aver-age of 07102408 (nfrac1411) (2 error is reported as times106) 147Sm144Nd errors are 15 or 0006 USGeological Survey (USGS) reference material BHVO-2was processed with the samples and yielded Sr and Nd iso-topic ratios of 07034607 and 05129786 respectivelyThese are in agreement with the published values of070347920 and 051298411 respectively (Weis et al2006)
Pb and Hf isotopic compositions were analyzed by mul-tiple collector-inductively coupled plasma-mass spectrome-try (MC-ICP-MS) on a static multi-collection on a NuPlasma (Nu Instruments) system The detailed analyticalprocedure for Pb isotopic analyses on the Nu Plasma atthe PCIGR has been described by Weis et al (2006) Theconfiguration for Pb analyses allows for collection of PbTl and Hg togetherTl and Hg are used to monitor instru-mental mass discrimination and isobaric overlap repec-tively All sample solutions were analyzed withapproximately the same PbTl ratio (4) as the referencematerial NIST SRM 981 To accomplish this a small ali-quot of each sample solution from the Pb columns was ana-lyzed on the Element2 to determine the precise amount ofPb available for analysis on the Nu Plasma The SRM 981standard was run after every two samples on the NuPlasma During the time samples were run analyses ofthe SRM 981 Pb reference material gave values of206Pb204Pbfrac1416940322 207Pb204Pbfrac1415495823and 208Pb204Pbfrac14 36713164 (nfrac14 61 2 error is reportedas times 104) these values are in excellent agreement withreported TIMS triple-spike values of Galer amp Abouchami(1998) Fractionation-corrected Pb isotopic ratios were fur-ther corrected by the sample-standard bracketing methodor the ln^ln correction method described by White et al(2000) and Blichert-Toft et al (2003) Leached powders ofUSGS reference material BHVO-2 yielded Pb isotopicratios of 206Pb204Pbfrac14186454 8 207Pb204Pbfrac14 1549105 and 208Pb204Pbfrac14 38222514 and BCR-2yielded 206Pb204Pbfrac141880466 207Pb204Pbfrac14 1562518 and 208Pb204Pbfrac14 3883496 (2 error isreported as times 104) These values are in agreement withthose for leached residues of BHVO-2 and BCR-2 fromWeis et al (2006)Hf isotopic compositions were analyzed following the
procedures detailed byWeis et al (2007) The configurationfor Hf analyses monitored Lu mass 175 andYb mass 172 toallow for interference correction to masses 174 and 176 Hfisotopic ratios were normalized internally for mass fractio-nation to a 179Hf177Hf ratio of 07325 using an exponentialcorrection Standards were run after every two samplesand sample results were normalized to the ratio of the in-run daily average and a 176Hf177Hf ratio for JMC-475 of0282160 During the course of analyses the Hf standardJMC-475 gave an average value 02821679 (nfrac14130)USGS reference materials BCR-2 and BHVO-2 were pro-cessed with the samples and yielded Hf isotopic ratios of02828675 and 02831005 respectively Publishedvalues for BCR-2 and BHVO-2 are 02828717 and0283104 8 respectively (Weis et al 2007)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
39
within the volcaniclastic rocks indicate formation primar-ily via cooling^contraction granulation magma^water^steam interaction autobrecciation resedimentation andmass-wasting rather than pyroclastic eruption (Nixonet al 2008)A broad subaerial platform was constructed by emplace-
ment of subaerial lava flows during an emergent subaerialstage (Fig 16c) The subaerial unit predominantly com-prises well-layered sheet flows that formed in a similarway to continental flood basalts as hundred(s) of pahoehoeflow fields erupted on shallow slopes (Self et al 1997) how-ever volcaniclastic and pillowed flow units are locally pre-sent mostly in the upper parts of the stratigraphyLocalized ponding of flows led to the formation of multi-tiered columnar jointing in some subaerial flows (egBondre et al 2004) As volcanism waned local interflowcarbonate deposits developed and plagioclase megacrysticflows were erupted Subsidence of the plateau is recordedby the interflow lenses and deposits of platform carbonates(40^1500m thick) on top of the Karmutsen basalts thatcontain Late Carnian to Early Norian fossils (Fig 16d c216^227 Ma Nixon et al 2008)The evolution of the internal architecture of the
Wrangellia oceanic plateau was primarily related to theconditions of emplacement (ie water depth and eruptionenvironment) There does not appear to have been a sys-tematic coincident evolution or shift of lava compositionsonVancouver Island as has been recognized in most con-tinental flood basalts and parts of the Wrangellia oceanicplateau in Alaska and Yukon (Greene et al 2008 2009)In general with the exception of the picritic pillow lavasin the submarine stratigraphy and local variations betweenhigh-MgO and tholeiitic lavas there are no significantcompositional differences between lavas erupted duringthe different stages of growth of the plateau (ie submarineand subaerial flows) nor are substantial compositional dif-ferences generally found between massive submarine flowsin contact with pillowed flows tholeiitic basalt flows nearthe base of the Karmutsen Formation are compositionallysimilar to flows sampled near the top of the volcanicsequence Where it was possible to sample continuous sec-tions of volcanic stratigraphy greater than several hun-dreds of meters (eg Keogh^Maynard Lake area west ofthe Karmutsen Range Fig 2) there is no clear relationshipbetween stratigraphic position and compositionalvariation
CONCLUSIONSThe Karmutsen Formation covers extensive areas ofVancouver Island (20 000 km2) and forms part of a largeoceanic plateau erupted over a geologically short interval(c 225^231 Ma) The tripartite volcanic stratigraphyexposed on Vancouver Island is at least 6 km thick andcomprises a succession of submarine flows volcaniclastic
deposits and massive subaerial flowsThese volcanologicaldifferences are primarily related to the eruption environ-ment (deep-water shallow-water or subaerial) The rapidgrowth of the plateau prevented significant accumulationsof intervening sediment except in the uppermost stratigra-phy where isolated limestone lenses commonly associatedwith pillow basalt and hyaloclastite deposits preserve arecord of the final stages of subsidence of the plateauThe Wrangellia plateau on Vancouver Island was con-
structed dominantly of tholeiitic basalt with restrictedmajor- and trace-element and isotopic compositional var-iations Minor volumes of picritic pillow basalts eruptedlate in the submarine phase The Karmutsen basalts wereemplaced onto older Paleozoic arc volcanic sequences andlow NbLa for the picrites may indicate a contributionfrom the sub-arc lithospheric mantle Modeling of the pet-rogenesis of the Karmutsen picrites indicates melting ofanomalously hot mantle (15008C) and extensive degreesof partial melting (23^27) and is consistent with aplume initiation model Combined Sr^Nd^Hf^Pb isotopicsystematics indicate that the picritic and tholeiitic lavaswere derived from a depleted plume-type mantle sourcedistinct from the source of MORB that has compositionalsimilarities to the source of the lavas forming theCaribbean Plateau Picrites were generated by melting ofspinel lherzolite whereas the tholeiitic basalts formed asaggregate melts of both garnet and spinel lherzolite Thetholeiitic basalts underwent extensive low-pressure fractio-nation (52^3 kbar) and seismic studies indicate that someof theWrangellia crust beneathVancouver Island may cor-respond to the plutonic residues of the Karmutsen floodbasalts
ACKNOWLEDGEMENTSWe are grateful to Nick Arndt for helping us get this pro-ject started and Nick Massey for insights into VancouverIsland geology We greatly appreciate the assistance withPRIMELT modeling from Claude Herzberg We thankMikkel Schau for his insight and enthusiasm during field-work Jane Barling assisted with analyses by MC-ICP-MSThoughtful and thorough reviews of the manuscript byJohn Mahoney Fred Frey Wendy Bohrson DougalJerram and Teal Riley are greatly appreciated Fundingwas generously provided by the Rocks to Riches Programadministered by the BC amp Yukon Chamber of Mines in2004 by the BC Geological Survey and by NSERCDiscovery Grants to JSS and DW ARG was supportedby a University Graduate Fellowship at UBC
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
34
REFERENCESBarker F Brown A S Budahn J R amp Plafker G (1989) Back-arc
with frontal-arc component origin of Triassic Karmutsen basaltBritish Columbia Canada Chemical Geology 75 81^102
Blichert-Toft JWeis D Maerschalk C Agranier A amp Albare de F(2003) Hawaiian hot spot dynamics as inferred from the Hf and Pbisotope evolution of Mauna Kea volcano Geochemistry GeophysicsGeosystems 4(2) 1^27 doi1010292002GC000340
Bondre N R Duraiswami R A amp Dole G (2004) Morphologyand emplacement of flows from the Deccan Volcanic ProvinceIndia Bulletin of Volcanology 66 29^45
Carlisle D (1963) Pillow breccias and their aquagene tuffs QuadraIsland British Columbia Journal of Geology 71 48^71
Carlisle D (1972) Late Paleozoic to Mid-Triassic sedimentary-volcanic sequence on Northeastern Vancouver Island Geological
Society of Canada Paper Report of Activities 72-1 Part B 24^30Carlisle D amp SuzukiT (1974) Emergent basalt and submergent car-
bonate^clastic sequences including the Upper Triassic Dilleri andWelleri zones on Vancouver Island Canadian Journal of Earth
Sciences 11 254^279Chauvel C amp Blichert-Toft J (2001) A hafnium isotope and trace
element perspective on melting of the depleted mantle Earth and
Planetary Science Letters 190 137^151Chayes F (1954)The theory of thin-section analysis Journal of Geology
62 92^101Cho M Liou J G amp Maruyama S (1987) Transition from the zeo-
lite to prehnite^pumpellyite facies in the Karmutsen metabasitesVancouver Island British Columbia Journal of Petrology 27(2)467^494
Clowes R M Zelt C A Amor J R amp Ellis R M (1995)Lithospheric structure in the southern Canadian Cordillera froma network of seismic refraction lines Canadian Journal of Earth
Sciences 32(10) 1485^1513Coffin M F amp Eldholm O (1994) Large igneous provinces Crustal
structure dimensions and external consequences Reviews of
Geophysics 32(1) 1^36Cox K G (1980) A model for flood basalt volcanism Journal of
Petrology 21 629^650Eldholm O amp Coffin M F (2000) Large igneous provinces
and plate tectonics In Richards M A Gordon R G amp vander Hilst R D (eds) The History and Dynamics of Global
Plate Motions American Geophysical Union Geophysical Monograph 121309^326
Elliott T R Hawkesworth C J amp Grolaquo nvold K (1991) Dynamicmelting of the Iceland plume Nature 351 201^206
Escuder-Viruete J Pecurren rez-Estaucurren n A Contreras F Joubert MWeis D Ullrich T D amp Spadea P (2007) Plume mantle sourceheterogeneity through time Insights from the Duarte ComplexHispaniola northeastern Caribbean Journal of Geophysical Research112(B04203) doi1010292006JB004323
Farnetani C G Richards M A amp Ghiorso M S (1996)Petrological models of magma evolution and deep crustal structurebeneath hotspots and flood basalt provinces Earth and Planetary
Science Letters 143 81^94Fitton J G amp Godard M (2004) Origin and evolution of magmas
on the Ontong Java Plateau In Fitton J G Mahoney J JWallace P J amp Saunders A D (eds) Origin and Evolution of the
Ontong Java Plateau Geological Society London Special Publications 229151^178
Fitton J G Saunders A D Norry M J Hardarson B S ampTaylor R N (1997) Thermal and chemical structure of theIceland plume Earth and Planetary Science Letters 153 197^208
Fitton J G Saunders A D Kempton P D amp Hardarson B S(2003) Does depleted mantle form an intrinsic part of the Icelandplume Geochemistry Geophysics Geosystems 4(3) 1032 doi1010292002GC000424
Frey F A Huang S Blichert-Toft J Regelous M amp Boyet M(2005) Origin of depleted components in basalt related to theHawaiian hotspot Evidence from isotopic and incompatible ele-ment ratios Geochemistry Geophysics Geosystems 6(1) 23 doi1010292004GC000757
Galer S J G amp Abouchami W (1998) Practical application of leadtriple spiking for correction of instrumental mass discriminationMineralogical Magazine 62A 491^492
Ghiorso M S amp Sack R O (1995) Chemical mass transfer in mag-matic processes IV A revised and internally consistent thermody-namic model for the interpolation and extrapolation of liquid^solidequilibria in magmatic systems at elevated temperatures and pres-sures Contributions to Mineralogy and Petrology 119 197^212
Greene A R Scoates J S amp Weis D (2005)WrangelliaTerrane onVancouver Island Distribution of flood basalts with implicationsfor potential Ni^Cu^PGE mineralization In Grant B (ed)Geological Fieldwork 2004 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2005-1 209^220Greene A R Scoates J S Nixon G T amp Weis D (2006) Picritic
lavas and basal sills in the Karmutsen flood basalt provinceWrangellia northern Vancouver Island In Grant B (ed)Geological Fieldwork 2005 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2006-1 39^52Greene A R Scoates J S amp Weis D (2008) Wrangellia flood
basalts in Alaska A record of plume^lithosphere interaction in aLate Triassic accreted oceanic plateau Geochemistry Geophysics
Geosystems 9(Q12004) doi1010292008GC002092Greene A R Scoates J S Weis D amp Israel S (2009)
Geochemistry of triassic flood basalts from the Yukon (Canada)segment of the accreted Wrangellia oceanic plateau Lithosdoi101016jlithos200811010
Hauff F Hoernle K Tilton G Graham D W amp Kerr A C(2000) Large volume recycling of oceanic lithosphere over shorttime scales geochemical constraints from the Caribbean LargeIgneous Province Earth and Planetary Science Letters 174 247^263
Hauff F Hoernle K amp Schmidt A (2003) Sr^Nd^Pb compositionof Mesozoic Pacific oceanic crust (Site 1149 and 801 ODP Leg185)Implications for alteration of ocean crust and the input into theIzu^Bonin^Mariana subduction system Geochemistry Geophysics
Geosystems 4(8) doi1010292002GC000421Herzberg C (2004) Partial melting below the Ontong Java Plateau
In Fitton J G Mahoney J J Wallace P J amp Saunders A D(eds) Origin and Evolution of the Ontong Java Plateau Geological SocietyLondon Special Publications 229 179^183
Herzberg C (2006) Petrology and thermal structure of the Hawaiianplume from Mauna Kea volcano Nature 444(7119) 605
Herzberg C amp Asimow P D (2008) Petrology of some oceanicisland basalts PRIMELT2XLS software for primary magma cal-culation Geochemistry Geophysics Geosystems 9(Q09001) doi1010292008GC002057
Herzberg C amp OrsquoHara M J (2002) Plume-associated ultramaficmagmas of Phanerozoic age Journal of Petrology 43(10) 1857^1883
Herzberg C Asimow P D Arndt N Niu Y Lesher C MFitton J G Cheadle M J amp Saunders A D (2007)Temperatures in ambient mantle and plumes Constraints frombasalts picrites and komatiites Geochemistry Geophysics Geosystems8(Q02006) doi1010292006GC001390
Huang S amp Frey F A (2005) Trace element abundances of MaunaKea basalt from phase 2 of the Hawaii Scientific Drilling Project
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
35
Petrogenetic implications of correlations with major element con-tent and isotopic ratios Geochemistry Geophysics Geosystems 4(6)doi1010292002GC000322
Hussong D M Wipperman L K amp Kroenke L W (1979) Thecrustal structure of the Ontong Java and Manihiki OceanicPlateaus Journal of Geophysical Research 84(B11) 6003^6010
Jerram D A amp Widdowson M (2005) The anatomy of ContinentalFlood Basalt Provinces geological constraints on the processes andproducts of flood volcanism Lithos 79(3^4) 385^405
Jones D L Silberling N J amp Hillhouse J (1977)Wrangellia a dis-placed terrane in northwestern North America CanadianJournal ofEarth Sciences 14(11) 2565^2577
Kerr A C (2003) Oceanic Plateaus In Rudnick R (ed) The Crust
Treatise on GeochemistryVol 3 Oxford Elsevier Science pp 537^565Kerr A C amp Mahoney J J (2007) Oceanic plateaus Problematic
plumes potential paradigms Chemical Geology 241 332^353Kerr A CTarney J Marriner G F Klaver GT Saunders A D
amp Thirlwall M F (1996) The geochemistry and petrogenesis ofthe Late-Cretaceous picrites and basalts of Curacao NetherlandsAntilles a remnant of an oceanic plateau Contributions to
Mineralogy and Petrology 124(1) 29^43Kerr A C Marriner G F Tarney J Nivia A Saunders A D
Thirlwall M F amp Sinton A C (1997) Cretaceous basaltic ter-ranes in Western Colombia Elemental chronological and Sr-Ndisotopic constraints on petrogenesis Journal of Petrology 38(6)677^702
Kinzler R J (1997) Melting of mantle peridotite at pressuresapproaching the spinel to garnet transition Application to mid-ocean ridge basalt petrogenesis Journal of Geophysical Research
102(B1) 853^874Kinzler R J amp Grove T L (1992) Primary magmas of mid-ocean
ridge basalts 1 Experiments and methods Journal of Geophysical
Research 97(B5) 6885^6906Korenaga J (2005) Why did not the Ontong Java Plateau form sub-
aerially Earth and Planetary Science Letters 234 385^399Kuniyoshi S (1972) Petrology of the Karmutsen (Volcanic) Group
northeastern Vancouver Island British Columbia PhD disserta-tion University of California Los Angeles 242 pp
Lassiter J C DePaolo D J amp Mahoney J J (1995) Geochemistry ofthe Wrangellia flood basalt province Implications for the role ofcontinental and oceanic lithosphere in flood basalt genesis Journalof Petrology 36(4) 983^1009
LeBas M J (2000) IUGS reclassification of the high-Mg and picriticvolcanic rocks Journal of Petrology 41(10) 1467^1470
Longhi J (2002) Some phase equilibrium systematics of lherzolitemelting I Geochemistry Geophysics Geosystems 3(3) doi1010292001GC000204
MacDonald G A amp Katsura T (1964) Chemical composition ofHawaiian lavas Journal of Petrology 5 82^133
Mahoney J J (1987) An isotopic survey of Pacific oceanic plateausimplications for their nature and origin In Keating B HFryer P Batiza R amp Boehlert GW (eds) Seamounts Islands andAtolls American Geophysical Union Geophysical Monograph 43 207^220
Mahoney J LeRoex A P Peng Z Fisher R L amp Natland J H(1992) Southwestern limits of Indian Ocean Ridge mantle and theorigin of low 206Pb204Pb mid-ocean ridge basalt isotope systema-tics of the central Southwest Indian Ridge (178^508E) Journal ofGeophysical Research 97(B13) 19771^19790
Mahoney J J Sinton J M Kurz M D MacDougall J DSpencer K J amp Lugmair GW (1994) Isotope and trace elementcharacteristics of a super-fast spreading ridge East Pacific rise13^238S Earth and Planetary Science Letters 121 173^193
Massey N W D (1995a) Geology and mineral resources of theAlberni^Nanaimo Lakes sheet Vancouver Island 92F1W 92F2Eand part of 92F7E British Columbia Ministry of Energy Mines and
Petroleum Resources Paper 1992-2 132 ppMassey N W D (1995b) Geology and mineral resources of the
Cowichan Lake sheet Vancouver Island 92C16 British Columbia
Ministry of Energy Mines and Petroleum Resources Paper 1992-3 112 ppMassey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005a) Digital Geology Map of British Columbia Tile NM9 MidCoast BC British Columbia Ministry of Energy and Mines Geofile
2005-2Massey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005b) Digital Geology Map of British Columbia Tile NM10Southwest BC British Columbia Ministry of Energy and Mines Geofile
2005-3McDonough W F amp Sun S (1995) The composition of the Earth
Chemical Geology 120 223^253Monger JW H amp Journeay J M (1994) Basement geology and tec-
tonic evolution of theVancouver region In Monger JW H (ed)Geology and Geological Hazards of the Vancouver Region Southwestern
British Columbia Geological Survey of Canada Bulletin 481 3^25Muller J E (1977) Geology of Vancouver Island Geological Survey of
Canada Open File Map 463Muller J E (1980)The Paleozoic Sicker Group of Vancouver Island British
Columbia Geological Survey of Canada Paper 79-30 22 ppMuller J E Northcote K E amp Carlisle D (1974) Geology and
mineral deposits of Alert Bay^Cape Scott map area VancouverIsland British Columbia Geological Survey of Canada Paper 74-8 77
Neal C R Mahoney J J Kroenke L W Duncan R A ampPetterson M G (1997) The Ontong^Java Plateau InMahoney J J amp Coffin M F (eds) Large Igneous Provinces
Continental Oceanic and Planetary FloodVolcanism American Geophysical
Union Geophysical Monograph 100 183^216Niu Y Collerson K D Batiza R Wendt J I amp Regelous M
(1999) Origin of enriched-typed mid-ocean ridge basalt at ridgesfar from mantle plumes The East Pacific Rise at 11820rsquoN Journalof Geophysical Research 104 7067^7088
Nixon G T amp Orr A J (2007) Recent revisions to the EarlyMesozoic stratigraphy of Northern Vancouver Island (NTS 102I092L) and metallogenic implicatinos British Columbia InGrant B (ed) Geological Fieldwork 2006 British Columbia Ministry of
Energy Mines and Petroleum Resources Paper 2007-1 163^177Nixon GT Hammack J L KoyanagiV M Payie G J Haggart
JW Orchard M JTozerT Friedman R M Archibald D APalfy J amp Cordey F (2006a) Geology of the Quatsino^PortMcNeill area northernVancouver Island British Columbia Ministry
of Energy Mines and Petroleum Resources Geoscience Map 2006-2 scale150 000
Nixon G T Kelman M C Stevenson D Stokes L A ampJohnston K A (2006b) Preliminary geology of the Nimpkishmap area (NTS 092L07) northern Vancouver Island BritishColumbia In Grant B amp Newell J M (eds) Geological Fieldwork2005 British Columbia Ministry of Energy Mines and Petroleum Resources
Paper 2006-1 135^152Nixon G T Laroque J Pals A Styan J Greene A R amp
Scoates J S (2008) High-Mg lavas in the Karmutsen flood basaltsnorthernVancouver Island (NTS 092L) Stratigraphic setting andmetallogenic significance In Grant B (ed) Geological Fieldwork
2007 British Columbia Ministry of Energy Mines and Petroleum
Resources Paper 2008-1 175^190Nowell G M Kempton P D Noble S R Fitton J G
Saunders A Mahoney J J amp Taylor R N (1998) High precisionHf isotope measurements of MORB and OIB by thermal ionisation
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
36
mass spectrometry insights into the depleted mantle Chemical
Geology 149 211^233Parrish R R amp McNicoll V J (1992) U^Pb age determinations
from the southern Vancouver Island area British Columbia InRadiogenic Age and Isotopic Studies Report 5 Geological Survey of
Canada Paper 91-2 79^86Peate DW (1997) The Parana^Etendeka Province In Coffin M F
amp Mahoney J (eds) Large Igneous Provinces Continental Oceanic andPlanetary Flood Volcanism American Geophysical Union Geophysical
Monograph 100 217^245Perfit M R Fornari D J Ridley W I Kirk P D Casey J
Kastens K A Reynolds J R Edwards M Desonie DShuster R amp Paradis S (1996) Recent volcanism in the Siqueirostransform fault picritic basalts and implications for MORBmagma genesis Earth and Planetary Science Letters 141(1^4) 91^108
Pretorius W Weis D Williams G Hanano D Kieffer B ampScoates J S (2006) Complete trace elemental characterization ofgranitoid (USGSG-2GSP-2) reference materials by high resolutioninductively coupled plasma-mass spectrometry Geostandards and
Geoanalytical Research 30(1) 39^54Regelous M Niu Y Wendt J I Batiza R Greig A amp
Collerson K D (1999) Variations in the geochemistry of magma-tism on the East Pacific Rise at 10830rsquoN since 800 ka Earth and
Planetary Science Letters 168 45^63Recurren villon S Chauvel C Arndt N T Pik R Martineau F
Fourcade S amp Marty B (2002) Heterogeneity of the Caribbeanplateau mantle source Sr O and He isotopic compositions of oli-vine and clinopyroxene from Gorgona Island Earth and Planetary
Science Letters 205(1^2) 91Rhodes J M (1996) Geochemical stratigraphy of lava flows samples
by the Hawaii Scientific Drilling Project Journal of Geophysical
Research 101(B5) 11729^11746Rhodes J M amp Vollinger M J (2004) Composition of basaltic lavas
sampled by phase-2 of the Hawaii Scientific Drilling ProjectGeochemical stratigraphy and magma types Geochemistry
Geophysics Geosystems 5(3) doi1010292002GC000434Richards M A Jones D L Duncan R A amp DePaolo D J (1991)
A mantle plume initiation model for the Wrangellia flood basaltand other oceanic plateaus Science 254 263^267
Salters V J M amp Hart S R (1991) The mantle sources of oceanridges islands and arcs the Hf-isotope connection Earth and
Planetary Science Letters 104 364^380Salters V J M amp Stracke A (2004) Composition of the depleted
SaltersV J amp WhiteW (1998) Hf isotope constraints on mantle evo-lution Chemical Geology 145 447^460
SanoT amp Yamashita S (2004) Experimental petrology of basementlavas from Ocean Drilling Program Leg 192 implications for dif-ferentiation processes in Ontong Java Plateau magmas InFitton J G Mahoney J JWallace P J amp Saunders A D (eds)Origin and Evolution of the Ontong Java Plateau Geological Society
London Special Publications 229 185^218Saunders A D (2005) Large igneous provinces Origin and environ-
mental consequences Elements 1 259^263Self S ThordarsonT amp Keszthely L (1997) Emplacement of conti-
nental flood basalt lava flows In Mahoney J J amp Coffin M F(eds) Large Igneous Provinces Continental Oceanic and Planetary Flood
Volcanism American Geophysical Union Geophysical Monograph 100381^410
Shaw D M (2000) Continuous (dynamic) melting theory revisitedCanadian Mineralogist 38(5) 1041^1063
Sluggett C L (2003) Uranium^lead age and geochemical constraintson Paleozoic and Early Mesozoic magmatism in WrangelliaTerrane Saltspring Island British Columbia BSc thesisUniversity of British ColumbiaVancouver 84 pp
Storey M Mahoney J J amp Saunders A D (1997) Cretaceousbasalts in Madagascar and the transition between plume and con-tinental lithospheric mantle sources In Mahoney J J ampCoffin M F (eds) Large Igneous Provinces Continental Oceanic and
Planetary Flood Volcanism Aamerican Geophysical Union Geophysical
Monograph 100 95^122Surdam R C (1967) Low-grade metamorphism of the Karmutsen
Group PhD dissertation University of California Los Angeles288 pp
Sutherland-Brown A Yorath C J Anderson R G amp Dom K(1986) Geological maps of southern Vancouver IslandLITHOPROBE I 92C10 11 14 16 92F1 2 7 8 Geological Survey ofCanada Open File 1272
Tejada M L G Mahoney J J Castillo P R Ingle S P Sheth HC amp Weis D (2004) Pin-pricking the elephant evidence on theorigin of the Ontong Java Plateau from Pb^Sr^Hf^Nd isotopiccharacteristics of ODP Leg 192 basalts In Fitton J GMahoney J J Wallace P J amp Saunders A D (eds) Origin and
Evolution of the Ontong Java Plateau Geological Society London Special
Publications 229 133^150Thompson P M E Kempton P D amp Kerr A C (2008) Evaluation
of the effects of alteration and leaching on Sm^Nd and Lu^Hf sys-tematics in submarine mafic rocks Lithos 104(1^4) 164^176
Thompson P M E Kempton P D White R V Kerr A CTarney J Saunders A D Fitton J G amp McBirney A (2003)Hf-Nd isotope constraints on the origin of the CretaceousCaribbean plateau and its relationship to the Galapagos plumeEarth and Planetary Science Letters 217(1^2) 59^75
Vervoort J D Patchett P Blichert-Toft J amp Albare de F (1999)Relationships between Lu^Hf and Sm^Nd isotopic systems in theglobal sedimentary system Earth and Planetary Science Letters 16879^99
Walter M J (1998) Melting of garnet peridotite and the origin ofkomatiite and depleted lithosphere Journal of Petrology 39(1) 29^60
Weis D Kieffer B Maerschalk C Barling J de Jong JWilliams G A Hanano D Mattielli N Scoates J SGoolaerts A Friedman R A amp Mahoney J B (2006) High-pre-cision isotopic characterization of USGS reference materials byTIMS and MC-ICP-MS Geochemistry Geophysics Geosystems
7(Q08006) doi1010292006GC001283Weis D Kieffer B Hanano D Silva I N Barling J
Pretorius W Maerschalk C amp Mattielli N (2007) Hf isotopecompositions of US Geological Survey reference materialsGeochemistry Geophysics Geosystems 8(Q06006) doi1010292006GC001473
Wheeler J O amp McFeely P (1991) Tectonic assemblage map of theCanadian Cordillera and adjacent part of the United States ofAmerica Geological Survey of Canada Map 1712A
WhiteW M Albare de F amp Tecurren louk P (2000) High-precision analy-sis of Pb isotope ratios by multi-collector ICP-MS Chemical Geology167 257^270
Yorath C J Sutherland Brown A amp Massey N W D (1999)LITHOPROBE southern Vancouver Island British Columbia Geological
Survey of Canada Bulletin 498 145 ppZou H (1998) Trace element fractionation during modal and
non-model dynamic melting and open-system melting Amathematical treatment Geochimica et Cosmochimica Acta 62(11)1937^1945
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
37
Zou H amp Reid M R (2001) Quantitative modeling of trace elementfractionation during incongruent dynamic melting Geochimica et
Cosmochimica Acta 65(1) 153^162
APPENDIX SAMPLEPREPARAT ION ANDANALYT ICAL METHODSOnly the freshest rocks in the field were sampled and onlythe least altered samples were selected for chemical andisotopic analysis based on thorough petrographic inspec-tion Fifty-nine of the 129 collected samples were crushed(400 g) into small pieces 52mm in diameter in aRocklabs hydraulic piston crusher between WC plates tominimize contamination The coarse-crush was mixedand 100 g was powdered in a planetary mill using agatejars and balls cleaned with quartz sand between samples
ActLabs analytical methodsThe major- and trace-element compositions of the whole-rock powders were determined at Activation LaboratoriesLtd (Actlabs) in Ancaster Ontario Analytical techniquesand detection limits are also available from Actlabs (httpwwwactlabscommethsub_code4erehtm) The analyticalmethod for each of the elements analyzed is indicated inTable 2 For the major elements a 02 g sample was mixedwith a lithium metaborate^lithium tetraborate mixtureand fused in a graphite crucible The molten mixture waspoured into a 5 HNO3 solution and shaken until dis-solved (30min) The samples were analyzed for majoroxides and selected trace elements by inductively coupledplasma optical emission spectrometry (ICP-OES) on acombination simultaneous^sequential Thermo Jarrell-AshEnviro II system Internal calibration was achieved usinga variety of international reference materials (eg W-2BIR-1 DNC-1) and independent control samplesAdditional trace elements were analyzed by both instru-mental neutron activation analysis (INAA) and induc-tively couple plasma mass spectrometry (ICP-MS) ForINAA15^25 g of sample was weighed into small polyeth-ylene vials and irradiated with a control international ref-erence material CANMET WMS-1 and NiCr flux wires ata thermal neutron flux of 71012 neutronscm2 per s in theMcMaster Nuclear Reactor Following a 7 day waitingperiod the samples were measured on an Ortec high-purity Ge detector linked to a Canberra Series 95 multi-channel analyzer Activities for each element weredecay- and weight-corrected and compared with a detectorcalibration developed from multiple international certifiedreference materials For ICP-MS 025 g of sample wasdigested in HF followed by HNO3^HClO4 heated andtaken to dryness The samples were brought back into solu-tion with HCl Samples were analyzed using a PerkinElmer Optima 3000 ICP In-lab standards or certified ref-erence materials (eg W-2 BIR-1 DNC-1) were used for
quality control A total of 15 blind duplicates were ana-lyzed from the same sample powders to assessreproducibility
University of Massachusetts analyticalmethodsFourteen sample duplicate powders taken from the same100 g sample aliquot (high-MgO basalts and picrites)were also analyzed at the Ronald B Gilmore X-RayFluorescence (XRF) Laboratory at the University ofMassachusetts Major elements were measured on a fusedLa-bearing lithium borate glass disc using a SiemensMRS-400 spectrometer Trace element concentrations(Rb Sr Ba Ce Nb Zr Y Pb Zn Ga Ni Cr V) weremeasured on a separate powder pellet using a PhilipsPW2400 sequential spectrometer using a Rh tube LOIand ferrous iron measurements were made as describedby Rhodes amp Vollinger (2004) Precision and accuracy esti-mates for the data have been described by Rhodes (1996)and Rhodes amp Vollinger (2004) Results for each sampleare the average of two separate analyses and are shown inElectronic Appendix 2 for a comparison of analyses ofduplicate powders using XRFand ICP-MS
PCIGR analytical methodsA subset of 19 samples was selected for high-precisiontrace-element analysis and Sr Nd Pb and Hf isotopicanalysis at the Pacific Centre for Isotopic andGeochemical Research (PCIGR) at the University ofBritish Columbia (UBC) Samples were selected from the59 samples analyzed for whole-rock chemistry at ActLabsbased on major- and trace-element chemistry alteration(low LOI and petrographic alteration index) and samplelocation Samples were prepared for trace-element analysisat the PCIGR by the technique described by Pretorius et al(2006) on unleached rock powders Sample powders(100mg) were weighed in 7ml screw-top Savillex bea-kers and dissolved in 1ml 14N HNO3 and 5ml 48 HFon a hotplate for 48 h at 1308C with periodic ultrasonica-tion Samples were dried and redissolved in 6ml 6N HClon a hotplate for 24 h and then dried and redissolved in1ml concentrated HNO3 for 24 h before final dryingTrace element abundances were measured by high-resolution inductively coupled plasma-mass spectrometry(HR-ICP-MS) with aThermo Finnigan Element2 systemfollowing the procedures described by Pretorius et al(2006) within 24 h of redissolution HFSE and LILE weremeasured in medium-resolution mode at 2000 dilutionusing a PFATeflon spray chamber washed with aqua regiafor 3min between samples REE were measured in high-resolution mode and U and Pb in low-resolution modeat 2000 dilution using a glass spray chamber washedwith 2 HNO3 between samples Total procedural blanksand reference materials (BCR-2 BHVO-2) were analyzedwith the batch of samples Indium was used as an internal
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
38
standard in all samples and standard solutionsBackground and standard solutions were analyzed afterevery five samples to detect memory effects and massdrift Results for PCIGR trace-element analyses areshown in Electronic Appendix 3 for the full set of analyzedtrace elementsSample digestion for purification of Sr Nd Hf and Pb
for column chemistry began by weighing each samplepowder (400^500mg) prior to leaching All samples wereleached with 6N HCl and placed in an ultrasonic bath for15min Samples were rinsed twice with 18 M-cm H2Obetween each leaching step (15 total) until the supernatantliquid was clear [following the method of Mahoney (1987)]Samples were then dried on a hotplate for 24 h andweighed again Sample solutions were then prepared bydissolving 100^250mg of the leached powder dissolvedin 1ml 14N HNO3 and 10ml 48 HF on a hotplate for48 h at 1308C with periodic ultrasonication Samples weredried and redissolved in 6ml 6N HCl on a hotplate for24 h and then dried Pb was separated using anion exhangecolumns and the discard was used for Sr REE and Hfseparation Nd was separated from the other REE and Hfrequired two additional purification steps Detailed proce-dures for column chemistry for separating Sr Nd and Pbat the PCIGR have been described byWeis et al (2006) andHf purification has been described byWeis et al (2007) Srand Nd isotope ratios were measured by thermal ioniza-tion mass spectrometry (TIMS) on a Thermo FinniganTriton system in static mode with relay matrix rotation ona single Ta and double Re^Ta filament respectively Fourto five filaments per barrel of 21 were occupied by stan-dards (NBS 987 for Sr and LaJolla for Nd) for each barrelwhere samples were run Sample Sr and Nd isotopic com-positions were corrected for mass fractionation using86Sr88Srfrac14 01194 and 146Nd144Ndfrac14 07219 Each samplewas then normalized using the barrel average of the refer-ence material relative to the values of143Nd144Ndfrac14 0511858 and 87Sr86Srfrac14 0710248 (Weiset al 2006) During the course of the Vancouver Islandanalyses the La Jolla Nd standard gave an average valueof 05118566 (nfrac14 7) and NBS987 standard gave an aver-age of 07102408 (nfrac1411) (2 error is reported as times106) 147Sm144Nd errors are 15 or 0006 USGeological Survey (USGS) reference material BHVO-2was processed with the samples and yielded Sr and Nd iso-topic ratios of 07034607 and 05129786 respectivelyThese are in agreement with the published values of070347920 and 051298411 respectively (Weis et al2006)
Pb and Hf isotopic compositions were analyzed by mul-tiple collector-inductively coupled plasma-mass spectrome-try (MC-ICP-MS) on a static multi-collection on a NuPlasma (Nu Instruments) system The detailed analyticalprocedure for Pb isotopic analyses on the Nu Plasma atthe PCIGR has been described by Weis et al (2006) Theconfiguration for Pb analyses allows for collection of PbTl and Hg togetherTl and Hg are used to monitor instru-mental mass discrimination and isobaric overlap repec-tively All sample solutions were analyzed withapproximately the same PbTl ratio (4) as the referencematerial NIST SRM 981 To accomplish this a small ali-quot of each sample solution from the Pb columns was ana-lyzed on the Element2 to determine the precise amount ofPb available for analysis on the Nu Plasma The SRM 981standard was run after every two samples on the NuPlasma During the time samples were run analyses ofthe SRM 981 Pb reference material gave values of206Pb204Pbfrac1416940322 207Pb204Pbfrac1415495823and 208Pb204Pbfrac14 36713164 (nfrac14 61 2 error is reportedas times 104) these values are in excellent agreement withreported TIMS triple-spike values of Galer amp Abouchami(1998) Fractionation-corrected Pb isotopic ratios were fur-ther corrected by the sample-standard bracketing methodor the ln^ln correction method described by White et al(2000) and Blichert-Toft et al (2003) Leached powders ofUSGS reference material BHVO-2 yielded Pb isotopicratios of 206Pb204Pbfrac14186454 8 207Pb204Pbfrac14 1549105 and 208Pb204Pbfrac14 38222514 and BCR-2yielded 206Pb204Pbfrac141880466 207Pb204Pbfrac14 1562518 and 208Pb204Pbfrac14 3883496 (2 error isreported as times 104) These values are in agreement withthose for leached residues of BHVO-2 and BCR-2 fromWeis et al (2006)Hf isotopic compositions were analyzed following the
procedures detailed byWeis et al (2007) The configurationfor Hf analyses monitored Lu mass 175 andYb mass 172 toallow for interference correction to masses 174 and 176 Hfisotopic ratios were normalized internally for mass fractio-nation to a 179Hf177Hf ratio of 07325 using an exponentialcorrection Standards were run after every two samplesand sample results were normalized to the ratio of the in-run daily average and a 176Hf177Hf ratio for JMC-475 of0282160 During the course of analyses the Hf standardJMC-475 gave an average value 02821679 (nfrac14130)USGS reference materials BCR-2 and BHVO-2 were pro-cessed with the samples and yielded Hf isotopic ratios of02828675 and 02831005 respectively Publishedvalues for BCR-2 and BHVO-2 are 02828717 and0283104 8 respectively (Weis et al 2007)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
39
REFERENCESBarker F Brown A S Budahn J R amp Plafker G (1989) Back-arc
with frontal-arc component origin of Triassic Karmutsen basaltBritish Columbia Canada Chemical Geology 75 81^102
Blichert-Toft JWeis D Maerschalk C Agranier A amp Albare de F(2003) Hawaiian hot spot dynamics as inferred from the Hf and Pbisotope evolution of Mauna Kea volcano Geochemistry GeophysicsGeosystems 4(2) 1^27 doi1010292002GC000340
Bondre N R Duraiswami R A amp Dole G (2004) Morphologyand emplacement of flows from the Deccan Volcanic ProvinceIndia Bulletin of Volcanology 66 29^45
Carlisle D (1963) Pillow breccias and their aquagene tuffs QuadraIsland British Columbia Journal of Geology 71 48^71
Carlisle D (1972) Late Paleozoic to Mid-Triassic sedimentary-volcanic sequence on Northeastern Vancouver Island Geological
Society of Canada Paper Report of Activities 72-1 Part B 24^30Carlisle D amp SuzukiT (1974) Emergent basalt and submergent car-
bonate^clastic sequences including the Upper Triassic Dilleri andWelleri zones on Vancouver Island Canadian Journal of Earth
Sciences 11 254^279Chauvel C amp Blichert-Toft J (2001) A hafnium isotope and trace
element perspective on melting of the depleted mantle Earth and
Planetary Science Letters 190 137^151Chayes F (1954)The theory of thin-section analysis Journal of Geology
62 92^101Cho M Liou J G amp Maruyama S (1987) Transition from the zeo-
lite to prehnite^pumpellyite facies in the Karmutsen metabasitesVancouver Island British Columbia Journal of Petrology 27(2)467^494
Clowes R M Zelt C A Amor J R amp Ellis R M (1995)Lithospheric structure in the southern Canadian Cordillera froma network of seismic refraction lines Canadian Journal of Earth
Sciences 32(10) 1485^1513Coffin M F amp Eldholm O (1994) Large igneous provinces Crustal
structure dimensions and external consequences Reviews of
Geophysics 32(1) 1^36Cox K G (1980) A model for flood basalt volcanism Journal of
Petrology 21 629^650Eldholm O amp Coffin M F (2000) Large igneous provinces
and plate tectonics In Richards M A Gordon R G amp vander Hilst R D (eds) The History and Dynamics of Global
Plate Motions American Geophysical Union Geophysical Monograph 121309^326
Elliott T R Hawkesworth C J amp Grolaquo nvold K (1991) Dynamicmelting of the Iceland plume Nature 351 201^206
Escuder-Viruete J Pecurren rez-Estaucurren n A Contreras F Joubert MWeis D Ullrich T D amp Spadea P (2007) Plume mantle sourceheterogeneity through time Insights from the Duarte ComplexHispaniola northeastern Caribbean Journal of Geophysical Research112(B04203) doi1010292006JB004323
Farnetani C G Richards M A amp Ghiorso M S (1996)Petrological models of magma evolution and deep crustal structurebeneath hotspots and flood basalt provinces Earth and Planetary
Science Letters 143 81^94Fitton J G amp Godard M (2004) Origin and evolution of magmas
on the Ontong Java Plateau In Fitton J G Mahoney J JWallace P J amp Saunders A D (eds) Origin and Evolution of the
Ontong Java Plateau Geological Society London Special Publications 229151^178
Fitton J G Saunders A D Norry M J Hardarson B S ampTaylor R N (1997) Thermal and chemical structure of theIceland plume Earth and Planetary Science Letters 153 197^208
Fitton J G Saunders A D Kempton P D amp Hardarson B S(2003) Does depleted mantle form an intrinsic part of the Icelandplume Geochemistry Geophysics Geosystems 4(3) 1032 doi1010292002GC000424
Frey F A Huang S Blichert-Toft J Regelous M amp Boyet M(2005) Origin of depleted components in basalt related to theHawaiian hotspot Evidence from isotopic and incompatible ele-ment ratios Geochemistry Geophysics Geosystems 6(1) 23 doi1010292004GC000757
Galer S J G amp Abouchami W (1998) Practical application of leadtriple spiking for correction of instrumental mass discriminationMineralogical Magazine 62A 491^492
Ghiorso M S amp Sack R O (1995) Chemical mass transfer in mag-matic processes IV A revised and internally consistent thermody-namic model for the interpolation and extrapolation of liquid^solidequilibria in magmatic systems at elevated temperatures and pres-sures Contributions to Mineralogy and Petrology 119 197^212
Greene A R Scoates J S amp Weis D (2005)WrangelliaTerrane onVancouver Island Distribution of flood basalts with implicationsfor potential Ni^Cu^PGE mineralization In Grant B (ed)Geological Fieldwork 2004 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2005-1 209^220Greene A R Scoates J S Nixon G T amp Weis D (2006) Picritic
lavas and basal sills in the Karmutsen flood basalt provinceWrangellia northern Vancouver Island In Grant B (ed)Geological Fieldwork 2005 British Columbia Ministry of Energy Mines
and Petroleum Resources Paper 2006-1 39^52Greene A R Scoates J S amp Weis D (2008) Wrangellia flood
basalts in Alaska A record of plume^lithosphere interaction in aLate Triassic accreted oceanic plateau Geochemistry Geophysics
Geosystems 9(Q12004) doi1010292008GC002092Greene A R Scoates J S Weis D amp Israel S (2009)
Geochemistry of triassic flood basalts from the Yukon (Canada)segment of the accreted Wrangellia oceanic plateau Lithosdoi101016jlithos200811010
Hauff F Hoernle K Tilton G Graham D W amp Kerr A C(2000) Large volume recycling of oceanic lithosphere over shorttime scales geochemical constraints from the Caribbean LargeIgneous Province Earth and Planetary Science Letters 174 247^263
Hauff F Hoernle K amp Schmidt A (2003) Sr^Nd^Pb compositionof Mesozoic Pacific oceanic crust (Site 1149 and 801 ODP Leg185)Implications for alteration of ocean crust and the input into theIzu^Bonin^Mariana subduction system Geochemistry Geophysics
Geosystems 4(8) doi1010292002GC000421Herzberg C (2004) Partial melting below the Ontong Java Plateau
In Fitton J G Mahoney J J Wallace P J amp Saunders A D(eds) Origin and Evolution of the Ontong Java Plateau Geological SocietyLondon Special Publications 229 179^183
Herzberg C (2006) Petrology and thermal structure of the Hawaiianplume from Mauna Kea volcano Nature 444(7119) 605
Herzberg C amp Asimow P D (2008) Petrology of some oceanicisland basalts PRIMELT2XLS software for primary magma cal-culation Geochemistry Geophysics Geosystems 9(Q09001) doi1010292008GC002057
Herzberg C amp OrsquoHara M J (2002) Plume-associated ultramaficmagmas of Phanerozoic age Journal of Petrology 43(10) 1857^1883
Herzberg C Asimow P D Arndt N Niu Y Lesher C MFitton J G Cheadle M J amp Saunders A D (2007)Temperatures in ambient mantle and plumes Constraints frombasalts picrites and komatiites Geochemistry Geophysics Geosystems8(Q02006) doi1010292006GC001390
Huang S amp Frey F A (2005) Trace element abundances of MaunaKea basalt from phase 2 of the Hawaii Scientific Drilling Project
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
35
Petrogenetic implications of correlations with major element con-tent and isotopic ratios Geochemistry Geophysics Geosystems 4(6)doi1010292002GC000322
Hussong D M Wipperman L K amp Kroenke L W (1979) Thecrustal structure of the Ontong Java and Manihiki OceanicPlateaus Journal of Geophysical Research 84(B11) 6003^6010
Jerram D A amp Widdowson M (2005) The anatomy of ContinentalFlood Basalt Provinces geological constraints on the processes andproducts of flood volcanism Lithos 79(3^4) 385^405
Jones D L Silberling N J amp Hillhouse J (1977)Wrangellia a dis-placed terrane in northwestern North America CanadianJournal ofEarth Sciences 14(11) 2565^2577
Kerr A C (2003) Oceanic Plateaus In Rudnick R (ed) The Crust
Treatise on GeochemistryVol 3 Oxford Elsevier Science pp 537^565Kerr A C amp Mahoney J J (2007) Oceanic plateaus Problematic
plumes potential paradigms Chemical Geology 241 332^353Kerr A CTarney J Marriner G F Klaver GT Saunders A D
amp Thirlwall M F (1996) The geochemistry and petrogenesis ofthe Late-Cretaceous picrites and basalts of Curacao NetherlandsAntilles a remnant of an oceanic plateau Contributions to
Mineralogy and Petrology 124(1) 29^43Kerr A C Marriner G F Tarney J Nivia A Saunders A D
Thirlwall M F amp Sinton A C (1997) Cretaceous basaltic ter-ranes in Western Colombia Elemental chronological and Sr-Ndisotopic constraints on petrogenesis Journal of Petrology 38(6)677^702
Kinzler R J (1997) Melting of mantle peridotite at pressuresapproaching the spinel to garnet transition Application to mid-ocean ridge basalt petrogenesis Journal of Geophysical Research
102(B1) 853^874Kinzler R J amp Grove T L (1992) Primary magmas of mid-ocean
ridge basalts 1 Experiments and methods Journal of Geophysical
Research 97(B5) 6885^6906Korenaga J (2005) Why did not the Ontong Java Plateau form sub-
aerially Earth and Planetary Science Letters 234 385^399Kuniyoshi S (1972) Petrology of the Karmutsen (Volcanic) Group
northeastern Vancouver Island British Columbia PhD disserta-tion University of California Los Angeles 242 pp
Lassiter J C DePaolo D J amp Mahoney J J (1995) Geochemistry ofthe Wrangellia flood basalt province Implications for the role ofcontinental and oceanic lithosphere in flood basalt genesis Journalof Petrology 36(4) 983^1009
LeBas M J (2000) IUGS reclassification of the high-Mg and picriticvolcanic rocks Journal of Petrology 41(10) 1467^1470
Longhi J (2002) Some phase equilibrium systematics of lherzolitemelting I Geochemistry Geophysics Geosystems 3(3) doi1010292001GC000204
MacDonald G A amp Katsura T (1964) Chemical composition ofHawaiian lavas Journal of Petrology 5 82^133
Mahoney J J (1987) An isotopic survey of Pacific oceanic plateausimplications for their nature and origin In Keating B HFryer P Batiza R amp Boehlert GW (eds) Seamounts Islands andAtolls American Geophysical Union Geophysical Monograph 43 207^220
Mahoney J LeRoex A P Peng Z Fisher R L amp Natland J H(1992) Southwestern limits of Indian Ocean Ridge mantle and theorigin of low 206Pb204Pb mid-ocean ridge basalt isotope systema-tics of the central Southwest Indian Ridge (178^508E) Journal ofGeophysical Research 97(B13) 19771^19790
Mahoney J J Sinton J M Kurz M D MacDougall J DSpencer K J amp Lugmair GW (1994) Isotope and trace elementcharacteristics of a super-fast spreading ridge East Pacific rise13^238S Earth and Planetary Science Letters 121 173^193
Massey N W D (1995a) Geology and mineral resources of theAlberni^Nanaimo Lakes sheet Vancouver Island 92F1W 92F2Eand part of 92F7E British Columbia Ministry of Energy Mines and
Petroleum Resources Paper 1992-2 132 ppMassey N W D (1995b) Geology and mineral resources of the
Cowichan Lake sheet Vancouver Island 92C16 British Columbia
Ministry of Energy Mines and Petroleum Resources Paper 1992-3 112 ppMassey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005a) Digital Geology Map of British Columbia Tile NM9 MidCoast BC British Columbia Ministry of Energy and Mines Geofile
2005-2Massey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005b) Digital Geology Map of British Columbia Tile NM10Southwest BC British Columbia Ministry of Energy and Mines Geofile
2005-3McDonough W F amp Sun S (1995) The composition of the Earth
Chemical Geology 120 223^253Monger JW H amp Journeay J M (1994) Basement geology and tec-
tonic evolution of theVancouver region In Monger JW H (ed)Geology and Geological Hazards of the Vancouver Region Southwestern
British Columbia Geological Survey of Canada Bulletin 481 3^25Muller J E (1977) Geology of Vancouver Island Geological Survey of
Canada Open File Map 463Muller J E (1980)The Paleozoic Sicker Group of Vancouver Island British
Columbia Geological Survey of Canada Paper 79-30 22 ppMuller J E Northcote K E amp Carlisle D (1974) Geology and
mineral deposits of Alert Bay^Cape Scott map area VancouverIsland British Columbia Geological Survey of Canada Paper 74-8 77
Neal C R Mahoney J J Kroenke L W Duncan R A ampPetterson M G (1997) The Ontong^Java Plateau InMahoney J J amp Coffin M F (eds) Large Igneous Provinces
Continental Oceanic and Planetary FloodVolcanism American Geophysical
Union Geophysical Monograph 100 183^216Niu Y Collerson K D Batiza R Wendt J I amp Regelous M
(1999) Origin of enriched-typed mid-ocean ridge basalt at ridgesfar from mantle plumes The East Pacific Rise at 11820rsquoN Journalof Geophysical Research 104 7067^7088
Nixon G T amp Orr A J (2007) Recent revisions to the EarlyMesozoic stratigraphy of Northern Vancouver Island (NTS 102I092L) and metallogenic implicatinos British Columbia InGrant B (ed) Geological Fieldwork 2006 British Columbia Ministry of
Energy Mines and Petroleum Resources Paper 2007-1 163^177Nixon GT Hammack J L KoyanagiV M Payie G J Haggart
JW Orchard M JTozerT Friedman R M Archibald D APalfy J amp Cordey F (2006a) Geology of the Quatsino^PortMcNeill area northernVancouver Island British Columbia Ministry
of Energy Mines and Petroleum Resources Geoscience Map 2006-2 scale150 000
Nixon G T Kelman M C Stevenson D Stokes L A ampJohnston K A (2006b) Preliminary geology of the Nimpkishmap area (NTS 092L07) northern Vancouver Island BritishColumbia In Grant B amp Newell J M (eds) Geological Fieldwork2005 British Columbia Ministry of Energy Mines and Petroleum Resources
Paper 2006-1 135^152Nixon G T Laroque J Pals A Styan J Greene A R amp
Scoates J S (2008) High-Mg lavas in the Karmutsen flood basaltsnorthernVancouver Island (NTS 092L) Stratigraphic setting andmetallogenic significance In Grant B (ed) Geological Fieldwork
2007 British Columbia Ministry of Energy Mines and Petroleum
Resources Paper 2008-1 175^190Nowell G M Kempton P D Noble S R Fitton J G
Saunders A Mahoney J J amp Taylor R N (1998) High precisionHf isotope measurements of MORB and OIB by thermal ionisation
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
36
mass spectrometry insights into the depleted mantle Chemical
Geology 149 211^233Parrish R R amp McNicoll V J (1992) U^Pb age determinations
from the southern Vancouver Island area British Columbia InRadiogenic Age and Isotopic Studies Report 5 Geological Survey of
Canada Paper 91-2 79^86Peate DW (1997) The Parana^Etendeka Province In Coffin M F
amp Mahoney J (eds) Large Igneous Provinces Continental Oceanic andPlanetary Flood Volcanism American Geophysical Union Geophysical
Monograph 100 217^245Perfit M R Fornari D J Ridley W I Kirk P D Casey J
Kastens K A Reynolds J R Edwards M Desonie DShuster R amp Paradis S (1996) Recent volcanism in the Siqueirostransform fault picritic basalts and implications for MORBmagma genesis Earth and Planetary Science Letters 141(1^4) 91^108
Pretorius W Weis D Williams G Hanano D Kieffer B ampScoates J S (2006) Complete trace elemental characterization ofgranitoid (USGSG-2GSP-2) reference materials by high resolutioninductively coupled plasma-mass spectrometry Geostandards and
Geoanalytical Research 30(1) 39^54Regelous M Niu Y Wendt J I Batiza R Greig A amp
Collerson K D (1999) Variations in the geochemistry of magma-tism on the East Pacific Rise at 10830rsquoN since 800 ka Earth and
Planetary Science Letters 168 45^63Recurren villon S Chauvel C Arndt N T Pik R Martineau F
Fourcade S amp Marty B (2002) Heterogeneity of the Caribbeanplateau mantle source Sr O and He isotopic compositions of oli-vine and clinopyroxene from Gorgona Island Earth and Planetary
Science Letters 205(1^2) 91Rhodes J M (1996) Geochemical stratigraphy of lava flows samples
by the Hawaii Scientific Drilling Project Journal of Geophysical
Research 101(B5) 11729^11746Rhodes J M amp Vollinger M J (2004) Composition of basaltic lavas
sampled by phase-2 of the Hawaii Scientific Drilling ProjectGeochemical stratigraphy and magma types Geochemistry
Geophysics Geosystems 5(3) doi1010292002GC000434Richards M A Jones D L Duncan R A amp DePaolo D J (1991)
A mantle plume initiation model for the Wrangellia flood basaltand other oceanic plateaus Science 254 263^267
Salters V J M amp Hart S R (1991) The mantle sources of oceanridges islands and arcs the Hf-isotope connection Earth and
Planetary Science Letters 104 364^380Salters V J M amp Stracke A (2004) Composition of the depleted
SaltersV J amp WhiteW (1998) Hf isotope constraints on mantle evo-lution Chemical Geology 145 447^460
SanoT amp Yamashita S (2004) Experimental petrology of basementlavas from Ocean Drilling Program Leg 192 implications for dif-ferentiation processes in Ontong Java Plateau magmas InFitton J G Mahoney J JWallace P J amp Saunders A D (eds)Origin and Evolution of the Ontong Java Plateau Geological Society
London Special Publications 229 185^218Saunders A D (2005) Large igneous provinces Origin and environ-
mental consequences Elements 1 259^263Self S ThordarsonT amp Keszthely L (1997) Emplacement of conti-
nental flood basalt lava flows In Mahoney J J amp Coffin M F(eds) Large Igneous Provinces Continental Oceanic and Planetary Flood
Volcanism American Geophysical Union Geophysical Monograph 100381^410
Shaw D M (2000) Continuous (dynamic) melting theory revisitedCanadian Mineralogist 38(5) 1041^1063
Sluggett C L (2003) Uranium^lead age and geochemical constraintson Paleozoic and Early Mesozoic magmatism in WrangelliaTerrane Saltspring Island British Columbia BSc thesisUniversity of British ColumbiaVancouver 84 pp
Storey M Mahoney J J amp Saunders A D (1997) Cretaceousbasalts in Madagascar and the transition between plume and con-tinental lithospheric mantle sources In Mahoney J J ampCoffin M F (eds) Large Igneous Provinces Continental Oceanic and
Planetary Flood Volcanism Aamerican Geophysical Union Geophysical
Monograph 100 95^122Surdam R C (1967) Low-grade metamorphism of the Karmutsen
Group PhD dissertation University of California Los Angeles288 pp
Sutherland-Brown A Yorath C J Anderson R G amp Dom K(1986) Geological maps of southern Vancouver IslandLITHOPROBE I 92C10 11 14 16 92F1 2 7 8 Geological Survey ofCanada Open File 1272
Tejada M L G Mahoney J J Castillo P R Ingle S P Sheth HC amp Weis D (2004) Pin-pricking the elephant evidence on theorigin of the Ontong Java Plateau from Pb^Sr^Hf^Nd isotopiccharacteristics of ODP Leg 192 basalts In Fitton J GMahoney J J Wallace P J amp Saunders A D (eds) Origin and
Evolution of the Ontong Java Plateau Geological Society London Special
Publications 229 133^150Thompson P M E Kempton P D amp Kerr A C (2008) Evaluation
of the effects of alteration and leaching on Sm^Nd and Lu^Hf sys-tematics in submarine mafic rocks Lithos 104(1^4) 164^176
Thompson P M E Kempton P D White R V Kerr A CTarney J Saunders A D Fitton J G amp McBirney A (2003)Hf-Nd isotope constraints on the origin of the CretaceousCaribbean plateau and its relationship to the Galapagos plumeEarth and Planetary Science Letters 217(1^2) 59^75
Vervoort J D Patchett P Blichert-Toft J amp Albare de F (1999)Relationships between Lu^Hf and Sm^Nd isotopic systems in theglobal sedimentary system Earth and Planetary Science Letters 16879^99
Walter M J (1998) Melting of garnet peridotite and the origin ofkomatiite and depleted lithosphere Journal of Petrology 39(1) 29^60
Weis D Kieffer B Maerschalk C Barling J de Jong JWilliams G A Hanano D Mattielli N Scoates J SGoolaerts A Friedman R A amp Mahoney J B (2006) High-pre-cision isotopic characterization of USGS reference materials byTIMS and MC-ICP-MS Geochemistry Geophysics Geosystems
7(Q08006) doi1010292006GC001283Weis D Kieffer B Hanano D Silva I N Barling J
Pretorius W Maerschalk C amp Mattielli N (2007) Hf isotopecompositions of US Geological Survey reference materialsGeochemistry Geophysics Geosystems 8(Q06006) doi1010292006GC001473
Wheeler J O amp McFeely P (1991) Tectonic assemblage map of theCanadian Cordillera and adjacent part of the United States ofAmerica Geological Survey of Canada Map 1712A
WhiteW M Albare de F amp Tecurren louk P (2000) High-precision analy-sis of Pb isotope ratios by multi-collector ICP-MS Chemical Geology167 257^270
Yorath C J Sutherland Brown A amp Massey N W D (1999)LITHOPROBE southern Vancouver Island British Columbia Geological
Survey of Canada Bulletin 498 145 ppZou H (1998) Trace element fractionation during modal and
non-model dynamic melting and open-system melting Amathematical treatment Geochimica et Cosmochimica Acta 62(11)1937^1945
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
37
Zou H amp Reid M R (2001) Quantitative modeling of trace elementfractionation during incongruent dynamic melting Geochimica et
Cosmochimica Acta 65(1) 153^162
APPENDIX SAMPLEPREPARAT ION ANDANALYT ICAL METHODSOnly the freshest rocks in the field were sampled and onlythe least altered samples were selected for chemical andisotopic analysis based on thorough petrographic inspec-tion Fifty-nine of the 129 collected samples were crushed(400 g) into small pieces 52mm in diameter in aRocklabs hydraulic piston crusher between WC plates tominimize contamination The coarse-crush was mixedand 100 g was powdered in a planetary mill using agatejars and balls cleaned with quartz sand between samples
ActLabs analytical methodsThe major- and trace-element compositions of the whole-rock powders were determined at Activation LaboratoriesLtd (Actlabs) in Ancaster Ontario Analytical techniquesand detection limits are also available from Actlabs (httpwwwactlabscommethsub_code4erehtm) The analyticalmethod for each of the elements analyzed is indicated inTable 2 For the major elements a 02 g sample was mixedwith a lithium metaborate^lithium tetraborate mixtureand fused in a graphite crucible The molten mixture waspoured into a 5 HNO3 solution and shaken until dis-solved (30min) The samples were analyzed for majoroxides and selected trace elements by inductively coupledplasma optical emission spectrometry (ICP-OES) on acombination simultaneous^sequential Thermo Jarrell-AshEnviro II system Internal calibration was achieved usinga variety of international reference materials (eg W-2BIR-1 DNC-1) and independent control samplesAdditional trace elements were analyzed by both instru-mental neutron activation analysis (INAA) and induc-tively couple plasma mass spectrometry (ICP-MS) ForINAA15^25 g of sample was weighed into small polyeth-ylene vials and irradiated with a control international ref-erence material CANMET WMS-1 and NiCr flux wires ata thermal neutron flux of 71012 neutronscm2 per s in theMcMaster Nuclear Reactor Following a 7 day waitingperiod the samples were measured on an Ortec high-purity Ge detector linked to a Canberra Series 95 multi-channel analyzer Activities for each element weredecay- and weight-corrected and compared with a detectorcalibration developed from multiple international certifiedreference materials For ICP-MS 025 g of sample wasdigested in HF followed by HNO3^HClO4 heated andtaken to dryness The samples were brought back into solu-tion with HCl Samples were analyzed using a PerkinElmer Optima 3000 ICP In-lab standards or certified ref-erence materials (eg W-2 BIR-1 DNC-1) were used for
quality control A total of 15 blind duplicates were ana-lyzed from the same sample powders to assessreproducibility
University of Massachusetts analyticalmethodsFourteen sample duplicate powders taken from the same100 g sample aliquot (high-MgO basalts and picrites)were also analyzed at the Ronald B Gilmore X-RayFluorescence (XRF) Laboratory at the University ofMassachusetts Major elements were measured on a fusedLa-bearing lithium borate glass disc using a SiemensMRS-400 spectrometer Trace element concentrations(Rb Sr Ba Ce Nb Zr Y Pb Zn Ga Ni Cr V) weremeasured on a separate powder pellet using a PhilipsPW2400 sequential spectrometer using a Rh tube LOIand ferrous iron measurements were made as describedby Rhodes amp Vollinger (2004) Precision and accuracy esti-mates for the data have been described by Rhodes (1996)and Rhodes amp Vollinger (2004) Results for each sampleare the average of two separate analyses and are shown inElectronic Appendix 2 for a comparison of analyses ofduplicate powders using XRFand ICP-MS
PCIGR analytical methodsA subset of 19 samples was selected for high-precisiontrace-element analysis and Sr Nd Pb and Hf isotopicanalysis at the Pacific Centre for Isotopic andGeochemical Research (PCIGR) at the University ofBritish Columbia (UBC) Samples were selected from the59 samples analyzed for whole-rock chemistry at ActLabsbased on major- and trace-element chemistry alteration(low LOI and petrographic alteration index) and samplelocation Samples were prepared for trace-element analysisat the PCIGR by the technique described by Pretorius et al(2006) on unleached rock powders Sample powders(100mg) were weighed in 7ml screw-top Savillex bea-kers and dissolved in 1ml 14N HNO3 and 5ml 48 HFon a hotplate for 48 h at 1308C with periodic ultrasonica-tion Samples were dried and redissolved in 6ml 6N HClon a hotplate for 24 h and then dried and redissolved in1ml concentrated HNO3 for 24 h before final dryingTrace element abundances were measured by high-resolution inductively coupled plasma-mass spectrometry(HR-ICP-MS) with aThermo Finnigan Element2 systemfollowing the procedures described by Pretorius et al(2006) within 24 h of redissolution HFSE and LILE weremeasured in medium-resolution mode at 2000 dilutionusing a PFATeflon spray chamber washed with aqua regiafor 3min between samples REE were measured in high-resolution mode and U and Pb in low-resolution modeat 2000 dilution using a glass spray chamber washedwith 2 HNO3 between samples Total procedural blanksand reference materials (BCR-2 BHVO-2) were analyzedwith the batch of samples Indium was used as an internal
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
38
standard in all samples and standard solutionsBackground and standard solutions were analyzed afterevery five samples to detect memory effects and massdrift Results for PCIGR trace-element analyses areshown in Electronic Appendix 3 for the full set of analyzedtrace elementsSample digestion for purification of Sr Nd Hf and Pb
for column chemistry began by weighing each samplepowder (400^500mg) prior to leaching All samples wereleached with 6N HCl and placed in an ultrasonic bath for15min Samples were rinsed twice with 18 M-cm H2Obetween each leaching step (15 total) until the supernatantliquid was clear [following the method of Mahoney (1987)]Samples were then dried on a hotplate for 24 h andweighed again Sample solutions were then prepared bydissolving 100^250mg of the leached powder dissolvedin 1ml 14N HNO3 and 10ml 48 HF on a hotplate for48 h at 1308C with periodic ultrasonication Samples weredried and redissolved in 6ml 6N HCl on a hotplate for24 h and then dried Pb was separated using anion exhangecolumns and the discard was used for Sr REE and Hfseparation Nd was separated from the other REE and Hfrequired two additional purification steps Detailed proce-dures for column chemistry for separating Sr Nd and Pbat the PCIGR have been described byWeis et al (2006) andHf purification has been described byWeis et al (2007) Srand Nd isotope ratios were measured by thermal ioniza-tion mass spectrometry (TIMS) on a Thermo FinniganTriton system in static mode with relay matrix rotation ona single Ta and double Re^Ta filament respectively Fourto five filaments per barrel of 21 were occupied by stan-dards (NBS 987 for Sr and LaJolla for Nd) for each barrelwhere samples were run Sample Sr and Nd isotopic com-positions were corrected for mass fractionation using86Sr88Srfrac14 01194 and 146Nd144Ndfrac14 07219 Each samplewas then normalized using the barrel average of the refer-ence material relative to the values of143Nd144Ndfrac14 0511858 and 87Sr86Srfrac14 0710248 (Weiset al 2006) During the course of the Vancouver Islandanalyses the La Jolla Nd standard gave an average valueof 05118566 (nfrac14 7) and NBS987 standard gave an aver-age of 07102408 (nfrac1411) (2 error is reported as times106) 147Sm144Nd errors are 15 or 0006 USGeological Survey (USGS) reference material BHVO-2was processed with the samples and yielded Sr and Nd iso-topic ratios of 07034607 and 05129786 respectivelyThese are in agreement with the published values of070347920 and 051298411 respectively (Weis et al2006)
Pb and Hf isotopic compositions were analyzed by mul-tiple collector-inductively coupled plasma-mass spectrome-try (MC-ICP-MS) on a static multi-collection on a NuPlasma (Nu Instruments) system The detailed analyticalprocedure for Pb isotopic analyses on the Nu Plasma atthe PCIGR has been described by Weis et al (2006) Theconfiguration for Pb analyses allows for collection of PbTl and Hg togetherTl and Hg are used to monitor instru-mental mass discrimination and isobaric overlap repec-tively All sample solutions were analyzed withapproximately the same PbTl ratio (4) as the referencematerial NIST SRM 981 To accomplish this a small ali-quot of each sample solution from the Pb columns was ana-lyzed on the Element2 to determine the precise amount ofPb available for analysis on the Nu Plasma The SRM 981standard was run after every two samples on the NuPlasma During the time samples were run analyses ofthe SRM 981 Pb reference material gave values of206Pb204Pbfrac1416940322 207Pb204Pbfrac1415495823and 208Pb204Pbfrac14 36713164 (nfrac14 61 2 error is reportedas times 104) these values are in excellent agreement withreported TIMS triple-spike values of Galer amp Abouchami(1998) Fractionation-corrected Pb isotopic ratios were fur-ther corrected by the sample-standard bracketing methodor the ln^ln correction method described by White et al(2000) and Blichert-Toft et al (2003) Leached powders ofUSGS reference material BHVO-2 yielded Pb isotopicratios of 206Pb204Pbfrac14186454 8 207Pb204Pbfrac14 1549105 and 208Pb204Pbfrac14 38222514 and BCR-2yielded 206Pb204Pbfrac141880466 207Pb204Pbfrac14 1562518 and 208Pb204Pbfrac14 3883496 (2 error isreported as times 104) These values are in agreement withthose for leached residues of BHVO-2 and BCR-2 fromWeis et al (2006)Hf isotopic compositions were analyzed following the
procedures detailed byWeis et al (2007) The configurationfor Hf analyses monitored Lu mass 175 andYb mass 172 toallow for interference correction to masses 174 and 176 Hfisotopic ratios were normalized internally for mass fractio-nation to a 179Hf177Hf ratio of 07325 using an exponentialcorrection Standards were run after every two samplesand sample results were normalized to the ratio of the in-run daily average and a 176Hf177Hf ratio for JMC-475 of0282160 During the course of analyses the Hf standardJMC-475 gave an average value 02821679 (nfrac14130)USGS reference materials BCR-2 and BHVO-2 were pro-cessed with the samples and yielded Hf isotopic ratios of02828675 and 02831005 respectively Publishedvalues for BCR-2 and BHVO-2 are 02828717 and0283104 8 respectively (Weis et al 2007)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
39
Petrogenetic implications of correlations with major element con-tent and isotopic ratios Geochemistry Geophysics Geosystems 4(6)doi1010292002GC000322
Hussong D M Wipperman L K amp Kroenke L W (1979) Thecrustal structure of the Ontong Java and Manihiki OceanicPlateaus Journal of Geophysical Research 84(B11) 6003^6010
Jerram D A amp Widdowson M (2005) The anatomy of ContinentalFlood Basalt Provinces geological constraints on the processes andproducts of flood volcanism Lithos 79(3^4) 385^405
Jones D L Silberling N J amp Hillhouse J (1977)Wrangellia a dis-placed terrane in northwestern North America CanadianJournal ofEarth Sciences 14(11) 2565^2577
Kerr A C (2003) Oceanic Plateaus In Rudnick R (ed) The Crust
Treatise on GeochemistryVol 3 Oxford Elsevier Science pp 537^565Kerr A C amp Mahoney J J (2007) Oceanic plateaus Problematic
plumes potential paradigms Chemical Geology 241 332^353Kerr A CTarney J Marriner G F Klaver GT Saunders A D
amp Thirlwall M F (1996) The geochemistry and petrogenesis ofthe Late-Cretaceous picrites and basalts of Curacao NetherlandsAntilles a remnant of an oceanic plateau Contributions to
Mineralogy and Petrology 124(1) 29^43Kerr A C Marriner G F Tarney J Nivia A Saunders A D
Thirlwall M F amp Sinton A C (1997) Cretaceous basaltic ter-ranes in Western Colombia Elemental chronological and Sr-Ndisotopic constraints on petrogenesis Journal of Petrology 38(6)677^702
Kinzler R J (1997) Melting of mantle peridotite at pressuresapproaching the spinel to garnet transition Application to mid-ocean ridge basalt petrogenesis Journal of Geophysical Research
102(B1) 853^874Kinzler R J amp Grove T L (1992) Primary magmas of mid-ocean
ridge basalts 1 Experiments and methods Journal of Geophysical
Research 97(B5) 6885^6906Korenaga J (2005) Why did not the Ontong Java Plateau form sub-
aerially Earth and Planetary Science Letters 234 385^399Kuniyoshi S (1972) Petrology of the Karmutsen (Volcanic) Group
northeastern Vancouver Island British Columbia PhD disserta-tion University of California Los Angeles 242 pp
Lassiter J C DePaolo D J amp Mahoney J J (1995) Geochemistry ofthe Wrangellia flood basalt province Implications for the role ofcontinental and oceanic lithosphere in flood basalt genesis Journalof Petrology 36(4) 983^1009
LeBas M J (2000) IUGS reclassification of the high-Mg and picriticvolcanic rocks Journal of Petrology 41(10) 1467^1470
Longhi J (2002) Some phase equilibrium systematics of lherzolitemelting I Geochemistry Geophysics Geosystems 3(3) doi1010292001GC000204
MacDonald G A amp Katsura T (1964) Chemical composition ofHawaiian lavas Journal of Petrology 5 82^133
Mahoney J J (1987) An isotopic survey of Pacific oceanic plateausimplications for their nature and origin In Keating B HFryer P Batiza R amp Boehlert GW (eds) Seamounts Islands andAtolls American Geophysical Union Geophysical Monograph 43 207^220
Mahoney J LeRoex A P Peng Z Fisher R L amp Natland J H(1992) Southwestern limits of Indian Ocean Ridge mantle and theorigin of low 206Pb204Pb mid-ocean ridge basalt isotope systema-tics of the central Southwest Indian Ridge (178^508E) Journal ofGeophysical Research 97(B13) 19771^19790
Mahoney J J Sinton J M Kurz M D MacDougall J DSpencer K J amp Lugmair GW (1994) Isotope and trace elementcharacteristics of a super-fast spreading ridge East Pacific rise13^238S Earth and Planetary Science Letters 121 173^193
Massey N W D (1995a) Geology and mineral resources of theAlberni^Nanaimo Lakes sheet Vancouver Island 92F1W 92F2Eand part of 92F7E British Columbia Ministry of Energy Mines and
Petroleum Resources Paper 1992-2 132 ppMassey N W D (1995b) Geology and mineral resources of the
Cowichan Lake sheet Vancouver Island 92C16 British Columbia
Ministry of Energy Mines and Petroleum Resources Paper 1992-3 112 ppMassey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005a) Digital Geology Map of British Columbia Tile NM9 MidCoast BC British Columbia Ministry of Energy and Mines Geofile
2005-2Massey NW D MacIntyre D G Desjardins P J amp Cooney RT
(2005b) Digital Geology Map of British Columbia Tile NM10Southwest BC British Columbia Ministry of Energy and Mines Geofile
2005-3McDonough W F amp Sun S (1995) The composition of the Earth
Chemical Geology 120 223^253Monger JW H amp Journeay J M (1994) Basement geology and tec-
tonic evolution of theVancouver region In Monger JW H (ed)Geology and Geological Hazards of the Vancouver Region Southwestern
British Columbia Geological Survey of Canada Bulletin 481 3^25Muller J E (1977) Geology of Vancouver Island Geological Survey of
Canada Open File Map 463Muller J E (1980)The Paleozoic Sicker Group of Vancouver Island British
Columbia Geological Survey of Canada Paper 79-30 22 ppMuller J E Northcote K E amp Carlisle D (1974) Geology and
mineral deposits of Alert Bay^Cape Scott map area VancouverIsland British Columbia Geological Survey of Canada Paper 74-8 77
Neal C R Mahoney J J Kroenke L W Duncan R A ampPetterson M G (1997) The Ontong^Java Plateau InMahoney J J amp Coffin M F (eds) Large Igneous Provinces
Continental Oceanic and Planetary FloodVolcanism American Geophysical
Union Geophysical Monograph 100 183^216Niu Y Collerson K D Batiza R Wendt J I amp Regelous M
(1999) Origin of enriched-typed mid-ocean ridge basalt at ridgesfar from mantle plumes The East Pacific Rise at 11820rsquoN Journalof Geophysical Research 104 7067^7088
Nixon G T amp Orr A J (2007) Recent revisions to the EarlyMesozoic stratigraphy of Northern Vancouver Island (NTS 102I092L) and metallogenic implicatinos British Columbia InGrant B (ed) Geological Fieldwork 2006 British Columbia Ministry of
Energy Mines and Petroleum Resources Paper 2007-1 163^177Nixon GT Hammack J L KoyanagiV M Payie G J Haggart
JW Orchard M JTozerT Friedman R M Archibald D APalfy J amp Cordey F (2006a) Geology of the Quatsino^PortMcNeill area northernVancouver Island British Columbia Ministry
of Energy Mines and Petroleum Resources Geoscience Map 2006-2 scale150 000
Nixon G T Kelman M C Stevenson D Stokes L A ampJohnston K A (2006b) Preliminary geology of the Nimpkishmap area (NTS 092L07) northern Vancouver Island BritishColumbia In Grant B amp Newell J M (eds) Geological Fieldwork2005 British Columbia Ministry of Energy Mines and Petroleum Resources
Paper 2006-1 135^152Nixon G T Laroque J Pals A Styan J Greene A R amp
Scoates J S (2008) High-Mg lavas in the Karmutsen flood basaltsnorthernVancouver Island (NTS 092L) Stratigraphic setting andmetallogenic significance In Grant B (ed) Geological Fieldwork
2007 British Columbia Ministry of Energy Mines and Petroleum
Resources Paper 2008-1 175^190Nowell G M Kempton P D Noble S R Fitton J G
Saunders A Mahoney J J amp Taylor R N (1998) High precisionHf isotope measurements of MORB and OIB by thermal ionisation
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
36
mass spectrometry insights into the depleted mantle Chemical
Geology 149 211^233Parrish R R amp McNicoll V J (1992) U^Pb age determinations
from the southern Vancouver Island area British Columbia InRadiogenic Age and Isotopic Studies Report 5 Geological Survey of
Canada Paper 91-2 79^86Peate DW (1997) The Parana^Etendeka Province In Coffin M F
amp Mahoney J (eds) Large Igneous Provinces Continental Oceanic andPlanetary Flood Volcanism American Geophysical Union Geophysical
Monograph 100 217^245Perfit M R Fornari D J Ridley W I Kirk P D Casey J
Kastens K A Reynolds J R Edwards M Desonie DShuster R amp Paradis S (1996) Recent volcanism in the Siqueirostransform fault picritic basalts and implications for MORBmagma genesis Earth and Planetary Science Letters 141(1^4) 91^108
Pretorius W Weis D Williams G Hanano D Kieffer B ampScoates J S (2006) Complete trace elemental characterization ofgranitoid (USGSG-2GSP-2) reference materials by high resolutioninductively coupled plasma-mass spectrometry Geostandards and
Geoanalytical Research 30(1) 39^54Regelous M Niu Y Wendt J I Batiza R Greig A amp
Collerson K D (1999) Variations in the geochemistry of magma-tism on the East Pacific Rise at 10830rsquoN since 800 ka Earth and
Planetary Science Letters 168 45^63Recurren villon S Chauvel C Arndt N T Pik R Martineau F
Fourcade S amp Marty B (2002) Heterogeneity of the Caribbeanplateau mantle source Sr O and He isotopic compositions of oli-vine and clinopyroxene from Gorgona Island Earth and Planetary
Science Letters 205(1^2) 91Rhodes J M (1996) Geochemical stratigraphy of lava flows samples
by the Hawaii Scientific Drilling Project Journal of Geophysical
Research 101(B5) 11729^11746Rhodes J M amp Vollinger M J (2004) Composition of basaltic lavas
sampled by phase-2 of the Hawaii Scientific Drilling ProjectGeochemical stratigraphy and magma types Geochemistry
Geophysics Geosystems 5(3) doi1010292002GC000434Richards M A Jones D L Duncan R A amp DePaolo D J (1991)
A mantle plume initiation model for the Wrangellia flood basaltand other oceanic plateaus Science 254 263^267
Salters V J M amp Hart S R (1991) The mantle sources of oceanridges islands and arcs the Hf-isotope connection Earth and
Planetary Science Letters 104 364^380Salters V J M amp Stracke A (2004) Composition of the depleted
SaltersV J amp WhiteW (1998) Hf isotope constraints on mantle evo-lution Chemical Geology 145 447^460
SanoT amp Yamashita S (2004) Experimental petrology of basementlavas from Ocean Drilling Program Leg 192 implications for dif-ferentiation processes in Ontong Java Plateau magmas InFitton J G Mahoney J JWallace P J amp Saunders A D (eds)Origin and Evolution of the Ontong Java Plateau Geological Society
London Special Publications 229 185^218Saunders A D (2005) Large igneous provinces Origin and environ-
mental consequences Elements 1 259^263Self S ThordarsonT amp Keszthely L (1997) Emplacement of conti-
nental flood basalt lava flows In Mahoney J J amp Coffin M F(eds) Large Igneous Provinces Continental Oceanic and Planetary Flood
Volcanism American Geophysical Union Geophysical Monograph 100381^410
Shaw D M (2000) Continuous (dynamic) melting theory revisitedCanadian Mineralogist 38(5) 1041^1063
Sluggett C L (2003) Uranium^lead age and geochemical constraintson Paleozoic and Early Mesozoic magmatism in WrangelliaTerrane Saltspring Island British Columbia BSc thesisUniversity of British ColumbiaVancouver 84 pp
Storey M Mahoney J J amp Saunders A D (1997) Cretaceousbasalts in Madagascar and the transition between plume and con-tinental lithospheric mantle sources In Mahoney J J ampCoffin M F (eds) Large Igneous Provinces Continental Oceanic and
Planetary Flood Volcanism Aamerican Geophysical Union Geophysical
Monograph 100 95^122Surdam R C (1967) Low-grade metamorphism of the Karmutsen
Group PhD dissertation University of California Los Angeles288 pp
Sutherland-Brown A Yorath C J Anderson R G amp Dom K(1986) Geological maps of southern Vancouver IslandLITHOPROBE I 92C10 11 14 16 92F1 2 7 8 Geological Survey ofCanada Open File 1272
Tejada M L G Mahoney J J Castillo P R Ingle S P Sheth HC amp Weis D (2004) Pin-pricking the elephant evidence on theorigin of the Ontong Java Plateau from Pb^Sr^Hf^Nd isotopiccharacteristics of ODP Leg 192 basalts In Fitton J GMahoney J J Wallace P J amp Saunders A D (eds) Origin and
Evolution of the Ontong Java Plateau Geological Society London Special
Publications 229 133^150Thompson P M E Kempton P D amp Kerr A C (2008) Evaluation
of the effects of alteration and leaching on Sm^Nd and Lu^Hf sys-tematics in submarine mafic rocks Lithos 104(1^4) 164^176
Thompson P M E Kempton P D White R V Kerr A CTarney J Saunders A D Fitton J G amp McBirney A (2003)Hf-Nd isotope constraints on the origin of the CretaceousCaribbean plateau and its relationship to the Galapagos plumeEarth and Planetary Science Letters 217(1^2) 59^75
Vervoort J D Patchett P Blichert-Toft J amp Albare de F (1999)Relationships between Lu^Hf and Sm^Nd isotopic systems in theglobal sedimentary system Earth and Planetary Science Letters 16879^99
Walter M J (1998) Melting of garnet peridotite and the origin ofkomatiite and depleted lithosphere Journal of Petrology 39(1) 29^60
Weis D Kieffer B Maerschalk C Barling J de Jong JWilliams G A Hanano D Mattielli N Scoates J SGoolaerts A Friedman R A amp Mahoney J B (2006) High-pre-cision isotopic characterization of USGS reference materials byTIMS and MC-ICP-MS Geochemistry Geophysics Geosystems
7(Q08006) doi1010292006GC001283Weis D Kieffer B Hanano D Silva I N Barling J
Pretorius W Maerschalk C amp Mattielli N (2007) Hf isotopecompositions of US Geological Survey reference materialsGeochemistry Geophysics Geosystems 8(Q06006) doi1010292006GC001473
Wheeler J O amp McFeely P (1991) Tectonic assemblage map of theCanadian Cordillera and adjacent part of the United States ofAmerica Geological Survey of Canada Map 1712A
WhiteW M Albare de F amp Tecurren louk P (2000) High-precision analy-sis of Pb isotope ratios by multi-collector ICP-MS Chemical Geology167 257^270
Yorath C J Sutherland Brown A amp Massey N W D (1999)LITHOPROBE southern Vancouver Island British Columbia Geological
Survey of Canada Bulletin 498 145 ppZou H (1998) Trace element fractionation during modal and
non-model dynamic melting and open-system melting Amathematical treatment Geochimica et Cosmochimica Acta 62(11)1937^1945
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
37
Zou H amp Reid M R (2001) Quantitative modeling of trace elementfractionation during incongruent dynamic melting Geochimica et
Cosmochimica Acta 65(1) 153^162
APPENDIX SAMPLEPREPARAT ION ANDANALYT ICAL METHODSOnly the freshest rocks in the field were sampled and onlythe least altered samples were selected for chemical andisotopic analysis based on thorough petrographic inspec-tion Fifty-nine of the 129 collected samples were crushed(400 g) into small pieces 52mm in diameter in aRocklabs hydraulic piston crusher between WC plates tominimize contamination The coarse-crush was mixedand 100 g was powdered in a planetary mill using agatejars and balls cleaned with quartz sand between samples
ActLabs analytical methodsThe major- and trace-element compositions of the whole-rock powders were determined at Activation LaboratoriesLtd (Actlabs) in Ancaster Ontario Analytical techniquesand detection limits are also available from Actlabs (httpwwwactlabscommethsub_code4erehtm) The analyticalmethod for each of the elements analyzed is indicated inTable 2 For the major elements a 02 g sample was mixedwith a lithium metaborate^lithium tetraborate mixtureand fused in a graphite crucible The molten mixture waspoured into a 5 HNO3 solution and shaken until dis-solved (30min) The samples were analyzed for majoroxides and selected trace elements by inductively coupledplasma optical emission spectrometry (ICP-OES) on acombination simultaneous^sequential Thermo Jarrell-AshEnviro II system Internal calibration was achieved usinga variety of international reference materials (eg W-2BIR-1 DNC-1) and independent control samplesAdditional trace elements were analyzed by both instru-mental neutron activation analysis (INAA) and induc-tively couple plasma mass spectrometry (ICP-MS) ForINAA15^25 g of sample was weighed into small polyeth-ylene vials and irradiated with a control international ref-erence material CANMET WMS-1 and NiCr flux wires ata thermal neutron flux of 71012 neutronscm2 per s in theMcMaster Nuclear Reactor Following a 7 day waitingperiod the samples were measured on an Ortec high-purity Ge detector linked to a Canberra Series 95 multi-channel analyzer Activities for each element weredecay- and weight-corrected and compared with a detectorcalibration developed from multiple international certifiedreference materials For ICP-MS 025 g of sample wasdigested in HF followed by HNO3^HClO4 heated andtaken to dryness The samples were brought back into solu-tion with HCl Samples were analyzed using a PerkinElmer Optima 3000 ICP In-lab standards or certified ref-erence materials (eg W-2 BIR-1 DNC-1) were used for
quality control A total of 15 blind duplicates were ana-lyzed from the same sample powders to assessreproducibility
University of Massachusetts analyticalmethodsFourteen sample duplicate powders taken from the same100 g sample aliquot (high-MgO basalts and picrites)were also analyzed at the Ronald B Gilmore X-RayFluorescence (XRF) Laboratory at the University ofMassachusetts Major elements were measured on a fusedLa-bearing lithium borate glass disc using a SiemensMRS-400 spectrometer Trace element concentrations(Rb Sr Ba Ce Nb Zr Y Pb Zn Ga Ni Cr V) weremeasured on a separate powder pellet using a PhilipsPW2400 sequential spectrometer using a Rh tube LOIand ferrous iron measurements were made as describedby Rhodes amp Vollinger (2004) Precision and accuracy esti-mates for the data have been described by Rhodes (1996)and Rhodes amp Vollinger (2004) Results for each sampleare the average of two separate analyses and are shown inElectronic Appendix 2 for a comparison of analyses ofduplicate powders using XRFand ICP-MS
PCIGR analytical methodsA subset of 19 samples was selected for high-precisiontrace-element analysis and Sr Nd Pb and Hf isotopicanalysis at the Pacific Centre for Isotopic andGeochemical Research (PCIGR) at the University ofBritish Columbia (UBC) Samples were selected from the59 samples analyzed for whole-rock chemistry at ActLabsbased on major- and trace-element chemistry alteration(low LOI and petrographic alteration index) and samplelocation Samples were prepared for trace-element analysisat the PCIGR by the technique described by Pretorius et al(2006) on unleached rock powders Sample powders(100mg) were weighed in 7ml screw-top Savillex bea-kers and dissolved in 1ml 14N HNO3 and 5ml 48 HFon a hotplate for 48 h at 1308C with periodic ultrasonica-tion Samples were dried and redissolved in 6ml 6N HClon a hotplate for 24 h and then dried and redissolved in1ml concentrated HNO3 for 24 h before final dryingTrace element abundances were measured by high-resolution inductively coupled plasma-mass spectrometry(HR-ICP-MS) with aThermo Finnigan Element2 systemfollowing the procedures described by Pretorius et al(2006) within 24 h of redissolution HFSE and LILE weremeasured in medium-resolution mode at 2000 dilutionusing a PFATeflon spray chamber washed with aqua regiafor 3min between samples REE were measured in high-resolution mode and U and Pb in low-resolution modeat 2000 dilution using a glass spray chamber washedwith 2 HNO3 between samples Total procedural blanksand reference materials (BCR-2 BHVO-2) were analyzedwith the batch of samples Indium was used as an internal
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
38
standard in all samples and standard solutionsBackground and standard solutions were analyzed afterevery five samples to detect memory effects and massdrift Results for PCIGR trace-element analyses areshown in Electronic Appendix 3 for the full set of analyzedtrace elementsSample digestion for purification of Sr Nd Hf and Pb
for column chemistry began by weighing each samplepowder (400^500mg) prior to leaching All samples wereleached with 6N HCl and placed in an ultrasonic bath for15min Samples were rinsed twice with 18 M-cm H2Obetween each leaching step (15 total) until the supernatantliquid was clear [following the method of Mahoney (1987)]Samples were then dried on a hotplate for 24 h andweighed again Sample solutions were then prepared bydissolving 100^250mg of the leached powder dissolvedin 1ml 14N HNO3 and 10ml 48 HF on a hotplate for48 h at 1308C with periodic ultrasonication Samples weredried and redissolved in 6ml 6N HCl on a hotplate for24 h and then dried Pb was separated using anion exhangecolumns and the discard was used for Sr REE and Hfseparation Nd was separated from the other REE and Hfrequired two additional purification steps Detailed proce-dures for column chemistry for separating Sr Nd and Pbat the PCIGR have been described byWeis et al (2006) andHf purification has been described byWeis et al (2007) Srand Nd isotope ratios were measured by thermal ioniza-tion mass spectrometry (TIMS) on a Thermo FinniganTriton system in static mode with relay matrix rotation ona single Ta and double Re^Ta filament respectively Fourto five filaments per barrel of 21 were occupied by stan-dards (NBS 987 for Sr and LaJolla for Nd) for each barrelwhere samples were run Sample Sr and Nd isotopic com-positions were corrected for mass fractionation using86Sr88Srfrac14 01194 and 146Nd144Ndfrac14 07219 Each samplewas then normalized using the barrel average of the refer-ence material relative to the values of143Nd144Ndfrac14 0511858 and 87Sr86Srfrac14 0710248 (Weiset al 2006) During the course of the Vancouver Islandanalyses the La Jolla Nd standard gave an average valueof 05118566 (nfrac14 7) and NBS987 standard gave an aver-age of 07102408 (nfrac1411) (2 error is reported as times106) 147Sm144Nd errors are 15 or 0006 USGeological Survey (USGS) reference material BHVO-2was processed with the samples and yielded Sr and Nd iso-topic ratios of 07034607 and 05129786 respectivelyThese are in agreement with the published values of070347920 and 051298411 respectively (Weis et al2006)
Pb and Hf isotopic compositions were analyzed by mul-tiple collector-inductively coupled plasma-mass spectrome-try (MC-ICP-MS) on a static multi-collection on a NuPlasma (Nu Instruments) system The detailed analyticalprocedure for Pb isotopic analyses on the Nu Plasma atthe PCIGR has been described by Weis et al (2006) Theconfiguration for Pb analyses allows for collection of PbTl and Hg togetherTl and Hg are used to monitor instru-mental mass discrimination and isobaric overlap repec-tively All sample solutions were analyzed withapproximately the same PbTl ratio (4) as the referencematerial NIST SRM 981 To accomplish this a small ali-quot of each sample solution from the Pb columns was ana-lyzed on the Element2 to determine the precise amount ofPb available for analysis on the Nu Plasma The SRM 981standard was run after every two samples on the NuPlasma During the time samples were run analyses ofthe SRM 981 Pb reference material gave values of206Pb204Pbfrac1416940322 207Pb204Pbfrac1415495823and 208Pb204Pbfrac14 36713164 (nfrac14 61 2 error is reportedas times 104) these values are in excellent agreement withreported TIMS triple-spike values of Galer amp Abouchami(1998) Fractionation-corrected Pb isotopic ratios were fur-ther corrected by the sample-standard bracketing methodor the ln^ln correction method described by White et al(2000) and Blichert-Toft et al (2003) Leached powders ofUSGS reference material BHVO-2 yielded Pb isotopicratios of 206Pb204Pbfrac14186454 8 207Pb204Pbfrac14 1549105 and 208Pb204Pbfrac14 38222514 and BCR-2yielded 206Pb204Pbfrac141880466 207Pb204Pbfrac14 1562518 and 208Pb204Pbfrac14 3883496 (2 error isreported as times 104) These values are in agreement withthose for leached residues of BHVO-2 and BCR-2 fromWeis et al (2006)Hf isotopic compositions were analyzed following the
procedures detailed byWeis et al (2007) The configurationfor Hf analyses monitored Lu mass 175 andYb mass 172 toallow for interference correction to masses 174 and 176 Hfisotopic ratios were normalized internally for mass fractio-nation to a 179Hf177Hf ratio of 07325 using an exponentialcorrection Standards were run after every two samplesand sample results were normalized to the ratio of the in-run daily average and a 176Hf177Hf ratio for JMC-475 of0282160 During the course of analyses the Hf standardJMC-475 gave an average value 02821679 (nfrac14130)USGS reference materials BCR-2 and BHVO-2 were pro-cessed with the samples and yielded Hf isotopic ratios of02828675 and 02831005 respectively Publishedvalues for BCR-2 and BHVO-2 are 02828717 and0283104 8 respectively (Weis et al 2007)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
39
mass spectrometry insights into the depleted mantle Chemical
Geology 149 211^233Parrish R R amp McNicoll V J (1992) U^Pb age determinations
from the southern Vancouver Island area British Columbia InRadiogenic Age and Isotopic Studies Report 5 Geological Survey of
Canada Paper 91-2 79^86Peate DW (1997) The Parana^Etendeka Province In Coffin M F
amp Mahoney J (eds) Large Igneous Provinces Continental Oceanic andPlanetary Flood Volcanism American Geophysical Union Geophysical
Monograph 100 217^245Perfit M R Fornari D J Ridley W I Kirk P D Casey J
Kastens K A Reynolds J R Edwards M Desonie DShuster R amp Paradis S (1996) Recent volcanism in the Siqueirostransform fault picritic basalts and implications for MORBmagma genesis Earth and Planetary Science Letters 141(1^4) 91^108
Pretorius W Weis D Williams G Hanano D Kieffer B ampScoates J S (2006) Complete trace elemental characterization ofgranitoid (USGSG-2GSP-2) reference materials by high resolutioninductively coupled plasma-mass spectrometry Geostandards and
Geoanalytical Research 30(1) 39^54Regelous M Niu Y Wendt J I Batiza R Greig A amp
Collerson K D (1999) Variations in the geochemistry of magma-tism on the East Pacific Rise at 10830rsquoN since 800 ka Earth and
Planetary Science Letters 168 45^63Recurren villon S Chauvel C Arndt N T Pik R Martineau F
Fourcade S amp Marty B (2002) Heterogeneity of the Caribbeanplateau mantle source Sr O and He isotopic compositions of oli-vine and clinopyroxene from Gorgona Island Earth and Planetary
Science Letters 205(1^2) 91Rhodes J M (1996) Geochemical stratigraphy of lava flows samples
by the Hawaii Scientific Drilling Project Journal of Geophysical
Research 101(B5) 11729^11746Rhodes J M amp Vollinger M J (2004) Composition of basaltic lavas
sampled by phase-2 of the Hawaii Scientific Drilling ProjectGeochemical stratigraphy and magma types Geochemistry
Geophysics Geosystems 5(3) doi1010292002GC000434Richards M A Jones D L Duncan R A amp DePaolo D J (1991)
A mantle plume initiation model for the Wrangellia flood basaltand other oceanic plateaus Science 254 263^267
Salters V J M amp Hart S R (1991) The mantle sources of oceanridges islands and arcs the Hf-isotope connection Earth and
Planetary Science Letters 104 364^380Salters V J M amp Stracke A (2004) Composition of the depleted
SaltersV J amp WhiteW (1998) Hf isotope constraints on mantle evo-lution Chemical Geology 145 447^460
SanoT amp Yamashita S (2004) Experimental petrology of basementlavas from Ocean Drilling Program Leg 192 implications for dif-ferentiation processes in Ontong Java Plateau magmas InFitton J G Mahoney J JWallace P J amp Saunders A D (eds)Origin and Evolution of the Ontong Java Plateau Geological Society
London Special Publications 229 185^218Saunders A D (2005) Large igneous provinces Origin and environ-
mental consequences Elements 1 259^263Self S ThordarsonT amp Keszthely L (1997) Emplacement of conti-
nental flood basalt lava flows In Mahoney J J amp Coffin M F(eds) Large Igneous Provinces Continental Oceanic and Planetary Flood
Volcanism American Geophysical Union Geophysical Monograph 100381^410
Shaw D M (2000) Continuous (dynamic) melting theory revisitedCanadian Mineralogist 38(5) 1041^1063
Sluggett C L (2003) Uranium^lead age and geochemical constraintson Paleozoic and Early Mesozoic magmatism in WrangelliaTerrane Saltspring Island British Columbia BSc thesisUniversity of British ColumbiaVancouver 84 pp
Storey M Mahoney J J amp Saunders A D (1997) Cretaceousbasalts in Madagascar and the transition between plume and con-tinental lithospheric mantle sources In Mahoney J J ampCoffin M F (eds) Large Igneous Provinces Continental Oceanic and
Planetary Flood Volcanism Aamerican Geophysical Union Geophysical
Monograph 100 95^122Surdam R C (1967) Low-grade metamorphism of the Karmutsen
Group PhD dissertation University of California Los Angeles288 pp
Sutherland-Brown A Yorath C J Anderson R G amp Dom K(1986) Geological maps of southern Vancouver IslandLITHOPROBE I 92C10 11 14 16 92F1 2 7 8 Geological Survey ofCanada Open File 1272
Tejada M L G Mahoney J J Castillo P R Ingle S P Sheth HC amp Weis D (2004) Pin-pricking the elephant evidence on theorigin of the Ontong Java Plateau from Pb^Sr^Hf^Nd isotopiccharacteristics of ODP Leg 192 basalts In Fitton J GMahoney J J Wallace P J amp Saunders A D (eds) Origin and
Evolution of the Ontong Java Plateau Geological Society London Special
Publications 229 133^150Thompson P M E Kempton P D amp Kerr A C (2008) Evaluation
of the effects of alteration and leaching on Sm^Nd and Lu^Hf sys-tematics in submarine mafic rocks Lithos 104(1^4) 164^176
Thompson P M E Kempton P D White R V Kerr A CTarney J Saunders A D Fitton J G amp McBirney A (2003)Hf-Nd isotope constraints on the origin of the CretaceousCaribbean plateau and its relationship to the Galapagos plumeEarth and Planetary Science Letters 217(1^2) 59^75
Vervoort J D Patchett P Blichert-Toft J amp Albare de F (1999)Relationships between Lu^Hf and Sm^Nd isotopic systems in theglobal sedimentary system Earth and Planetary Science Letters 16879^99
Walter M J (1998) Melting of garnet peridotite and the origin ofkomatiite and depleted lithosphere Journal of Petrology 39(1) 29^60
Weis D Kieffer B Maerschalk C Barling J de Jong JWilliams G A Hanano D Mattielli N Scoates J SGoolaerts A Friedman R A amp Mahoney J B (2006) High-pre-cision isotopic characterization of USGS reference materials byTIMS and MC-ICP-MS Geochemistry Geophysics Geosystems
7(Q08006) doi1010292006GC001283Weis D Kieffer B Hanano D Silva I N Barling J
Pretorius W Maerschalk C amp Mattielli N (2007) Hf isotopecompositions of US Geological Survey reference materialsGeochemistry Geophysics Geosystems 8(Q06006) doi1010292006GC001473
Wheeler J O amp McFeely P (1991) Tectonic assemblage map of theCanadian Cordillera and adjacent part of the United States ofAmerica Geological Survey of Canada Map 1712A
WhiteW M Albare de F amp Tecurren louk P (2000) High-precision analy-sis of Pb isotope ratios by multi-collector ICP-MS Chemical Geology167 257^270
Yorath C J Sutherland Brown A amp Massey N W D (1999)LITHOPROBE southern Vancouver Island British Columbia Geological
Survey of Canada Bulletin 498 145 ppZou H (1998) Trace element fractionation during modal and
non-model dynamic melting and open-system melting Amathematical treatment Geochimica et Cosmochimica Acta 62(11)1937^1945
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
37
Zou H amp Reid M R (2001) Quantitative modeling of trace elementfractionation during incongruent dynamic melting Geochimica et
Cosmochimica Acta 65(1) 153^162
APPENDIX SAMPLEPREPARAT ION ANDANALYT ICAL METHODSOnly the freshest rocks in the field were sampled and onlythe least altered samples were selected for chemical andisotopic analysis based on thorough petrographic inspec-tion Fifty-nine of the 129 collected samples were crushed(400 g) into small pieces 52mm in diameter in aRocklabs hydraulic piston crusher between WC plates tominimize contamination The coarse-crush was mixedand 100 g was powdered in a planetary mill using agatejars and balls cleaned with quartz sand between samples
ActLabs analytical methodsThe major- and trace-element compositions of the whole-rock powders were determined at Activation LaboratoriesLtd (Actlabs) in Ancaster Ontario Analytical techniquesand detection limits are also available from Actlabs (httpwwwactlabscommethsub_code4erehtm) The analyticalmethod for each of the elements analyzed is indicated inTable 2 For the major elements a 02 g sample was mixedwith a lithium metaborate^lithium tetraborate mixtureand fused in a graphite crucible The molten mixture waspoured into a 5 HNO3 solution and shaken until dis-solved (30min) The samples were analyzed for majoroxides and selected trace elements by inductively coupledplasma optical emission spectrometry (ICP-OES) on acombination simultaneous^sequential Thermo Jarrell-AshEnviro II system Internal calibration was achieved usinga variety of international reference materials (eg W-2BIR-1 DNC-1) and independent control samplesAdditional trace elements were analyzed by both instru-mental neutron activation analysis (INAA) and induc-tively couple plasma mass spectrometry (ICP-MS) ForINAA15^25 g of sample was weighed into small polyeth-ylene vials and irradiated with a control international ref-erence material CANMET WMS-1 and NiCr flux wires ata thermal neutron flux of 71012 neutronscm2 per s in theMcMaster Nuclear Reactor Following a 7 day waitingperiod the samples were measured on an Ortec high-purity Ge detector linked to a Canberra Series 95 multi-channel analyzer Activities for each element weredecay- and weight-corrected and compared with a detectorcalibration developed from multiple international certifiedreference materials For ICP-MS 025 g of sample wasdigested in HF followed by HNO3^HClO4 heated andtaken to dryness The samples were brought back into solu-tion with HCl Samples were analyzed using a PerkinElmer Optima 3000 ICP In-lab standards or certified ref-erence materials (eg W-2 BIR-1 DNC-1) were used for
quality control A total of 15 blind duplicates were ana-lyzed from the same sample powders to assessreproducibility
University of Massachusetts analyticalmethodsFourteen sample duplicate powders taken from the same100 g sample aliquot (high-MgO basalts and picrites)were also analyzed at the Ronald B Gilmore X-RayFluorescence (XRF) Laboratory at the University ofMassachusetts Major elements were measured on a fusedLa-bearing lithium borate glass disc using a SiemensMRS-400 spectrometer Trace element concentrations(Rb Sr Ba Ce Nb Zr Y Pb Zn Ga Ni Cr V) weremeasured on a separate powder pellet using a PhilipsPW2400 sequential spectrometer using a Rh tube LOIand ferrous iron measurements were made as describedby Rhodes amp Vollinger (2004) Precision and accuracy esti-mates for the data have been described by Rhodes (1996)and Rhodes amp Vollinger (2004) Results for each sampleare the average of two separate analyses and are shown inElectronic Appendix 2 for a comparison of analyses ofduplicate powders using XRFand ICP-MS
PCIGR analytical methodsA subset of 19 samples was selected for high-precisiontrace-element analysis and Sr Nd Pb and Hf isotopicanalysis at the Pacific Centre for Isotopic andGeochemical Research (PCIGR) at the University ofBritish Columbia (UBC) Samples were selected from the59 samples analyzed for whole-rock chemistry at ActLabsbased on major- and trace-element chemistry alteration(low LOI and petrographic alteration index) and samplelocation Samples were prepared for trace-element analysisat the PCIGR by the technique described by Pretorius et al(2006) on unleached rock powders Sample powders(100mg) were weighed in 7ml screw-top Savillex bea-kers and dissolved in 1ml 14N HNO3 and 5ml 48 HFon a hotplate for 48 h at 1308C with periodic ultrasonica-tion Samples were dried and redissolved in 6ml 6N HClon a hotplate for 24 h and then dried and redissolved in1ml concentrated HNO3 for 24 h before final dryingTrace element abundances were measured by high-resolution inductively coupled plasma-mass spectrometry(HR-ICP-MS) with aThermo Finnigan Element2 systemfollowing the procedures described by Pretorius et al(2006) within 24 h of redissolution HFSE and LILE weremeasured in medium-resolution mode at 2000 dilutionusing a PFATeflon spray chamber washed with aqua regiafor 3min between samples REE were measured in high-resolution mode and U and Pb in low-resolution modeat 2000 dilution using a glass spray chamber washedwith 2 HNO3 between samples Total procedural blanksand reference materials (BCR-2 BHVO-2) were analyzedwith the batch of samples Indium was used as an internal
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
38
standard in all samples and standard solutionsBackground and standard solutions were analyzed afterevery five samples to detect memory effects and massdrift Results for PCIGR trace-element analyses areshown in Electronic Appendix 3 for the full set of analyzedtrace elementsSample digestion for purification of Sr Nd Hf and Pb
for column chemistry began by weighing each samplepowder (400^500mg) prior to leaching All samples wereleached with 6N HCl and placed in an ultrasonic bath for15min Samples were rinsed twice with 18 M-cm H2Obetween each leaching step (15 total) until the supernatantliquid was clear [following the method of Mahoney (1987)]Samples were then dried on a hotplate for 24 h andweighed again Sample solutions were then prepared bydissolving 100^250mg of the leached powder dissolvedin 1ml 14N HNO3 and 10ml 48 HF on a hotplate for48 h at 1308C with periodic ultrasonication Samples weredried and redissolved in 6ml 6N HCl on a hotplate for24 h and then dried Pb was separated using anion exhangecolumns and the discard was used for Sr REE and Hfseparation Nd was separated from the other REE and Hfrequired two additional purification steps Detailed proce-dures for column chemistry for separating Sr Nd and Pbat the PCIGR have been described byWeis et al (2006) andHf purification has been described byWeis et al (2007) Srand Nd isotope ratios were measured by thermal ioniza-tion mass spectrometry (TIMS) on a Thermo FinniganTriton system in static mode with relay matrix rotation ona single Ta and double Re^Ta filament respectively Fourto five filaments per barrel of 21 were occupied by stan-dards (NBS 987 for Sr and LaJolla for Nd) for each barrelwhere samples were run Sample Sr and Nd isotopic com-positions were corrected for mass fractionation using86Sr88Srfrac14 01194 and 146Nd144Ndfrac14 07219 Each samplewas then normalized using the barrel average of the refer-ence material relative to the values of143Nd144Ndfrac14 0511858 and 87Sr86Srfrac14 0710248 (Weiset al 2006) During the course of the Vancouver Islandanalyses the La Jolla Nd standard gave an average valueof 05118566 (nfrac14 7) and NBS987 standard gave an aver-age of 07102408 (nfrac1411) (2 error is reported as times106) 147Sm144Nd errors are 15 or 0006 USGeological Survey (USGS) reference material BHVO-2was processed with the samples and yielded Sr and Nd iso-topic ratios of 07034607 and 05129786 respectivelyThese are in agreement with the published values of070347920 and 051298411 respectively (Weis et al2006)
Pb and Hf isotopic compositions were analyzed by mul-tiple collector-inductively coupled plasma-mass spectrome-try (MC-ICP-MS) on a static multi-collection on a NuPlasma (Nu Instruments) system The detailed analyticalprocedure for Pb isotopic analyses on the Nu Plasma atthe PCIGR has been described by Weis et al (2006) Theconfiguration for Pb analyses allows for collection of PbTl and Hg togetherTl and Hg are used to monitor instru-mental mass discrimination and isobaric overlap repec-tively All sample solutions were analyzed withapproximately the same PbTl ratio (4) as the referencematerial NIST SRM 981 To accomplish this a small ali-quot of each sample solution from the Pb columns was ana-lyzed on the Element2 to determine the precise amount ofPb available for analysis on the Nu Plasma The SRM 981standard was run after every two samples on the NuPlasma During the time samples were run analyses ofthe SRM 981 Pb reference material gave values of206Pb204Pbfrac1416940322 207Pb204Pbfrac1415495823and 208Pb204Pbfrac14 36713164 (nfrac14 61 2 error is reportedas times 104) these values are in excellent agreement withreported TIMS triple-spike values of Galer amp Abouchami(1998) Fractionation-corrected Pb isotopic ratios were fur-ther corrected by the sample-standard bracketing methodor the ln^ln correction method described by White et al(2000) and Blichert-Toft et al (2003) Leached powders ofUSGS reference material BHVO-2 yielded Pb isotopicratios of 206Pb204Pbfrac14186454 8 207Pb204Pbfrac14 1549105 and 208Pb204Pbfrac14 38222514 and BCR-2yielded 206Pb204Pbfrac141880466 207Pb204Pbfrac14 1562518 and 208Pb204Pbfrac14 3883496 (2 error isreported as times 104) These values are in agreement withthose for leached residues of BHVO-2 and BCR-2 fromWeis et al (2006)Hf isotopic compositions were analyzed following the
procedures detailed byWeis et al (2007) The configurationfor Hf analyses monitored Lu mass 175 andYb mass 172 toallow for interference correction to masses 174 and 176 Hfisotopic ratios were normalized internally for mass fractio-nation to a 179Hf177Hf ratio of 07325 using an exponentialcorrection Standards were run after every two samplesand sample results were normalized to the ratio of the in-run daily average and a 176Hf177Hf ratio for JMC-475 of0282160 During the course of analyses the Hf standardJMC-475 gave an average value 02821679 (nfrac14130)USGS reference materials BCR-2 and BHVO-2 were pro-cessed with the samples and yielded Hf isotopic ratios of02828675 and 02831005 respectively Publishedvalues for BCR-2 and BHVO-2 are 02828717 and0283104 8 respectively (Weis et al 2007)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
39
Zou H amp Reid M R (2001) Quantitative modeling of trace elementfractionation during incongruent dynamic melting Geochimica et
Cosmochimica Acta 65(1) 153^162
APPENDIX SAMPLEPREPARAT ION ANDANALYT ICAL METHODSOnly the freshest rocks in the field were sampled and onlythe least altered samples were selected for chemical andisotopic analysis based on thorough petrographic inspec-tion Fifty-nine of the 129 collected samples were crushed(400 g) into small pieces 52mm in diameter in aRocklabs hydraulic piston crusher between WC plates tominimize contamination The coarse-crush was mixedand 100 g was powdered in a planetary mill using agatejars and balls cleaned with quartz sand between samples
ActLabs analytical methodsThe major- and trace-element compositions of the whole-rock powders were determined at Activation LaboratoriesLtd (Actlabs) in Ancaster Ontario Analytical techniquesand detection limits are also available from Actlabs (httpwwwactlabscommethsub_code4erehtm) The analyticalmethod for each of the elements analyzed is indicated inTable 2 For the major elements a 02 g sample was mixedwith a lithium metaborate^lithium tetraborate mixtureand fused in a graphite crucible The molten mixture waspoured into a 5 HNO3 solution and shaken until dis-solved (30min) The samples were analyzed for majoroxides and selected trace elements by inductively coupledplasma optical emission spectrometry (ICP-OES) on acombination simultaneous^sequential Thermo Jarrell-AshEnviro II system Internal calibration was achieved usinga variety of international reference materials (eg W-2BIR-1 DNC-1) and independent control samplesAdditional trace elements were analyzed by both instru-mental neutron activation analysis (INAA) and induc-tively couple plasma mass spectrometry (ICP-MS) ForINAA15^25 g of sample was weighed into small polyeth-ylene vials and irradiated with a control international ref-erence material CANMET WMS-1 and NiCr flux wires ata thermal neutron flux of 71012 neutronscm2 per s in theMcMaster Nuclear Reactor Following a 7 day waitingperiod the samples were measured on an Ortec high-purity Ge detector linked to a Canberra Series 95 multi-channel analyzer Activities for each element weredecay- and weight-corrected and compared with a detectorcalibration developed from multiple international certifiedreference materials For ICP-MS 025 g of sample wasdigested in HF followed by HNO3^HClO4 heated andtaken to dryness The samples were brought back into solu-tion with HCl Samples were analyzed using a PerkinElmer Optima 3000 ICP In-lab standards or certified ref-erence materials (eg W-2 BIR-1 DNC-1) were used for
quality control A total of 15 blind duplicates were ana-lyzed from the same sample powders to assessreproducibility
University of Massachusetts analyticalmethodsFourteen sample duplicate powders taken from the same100 g sample aliquot (high-MgO basalts and picrites)were also analyzed at the Ronald B Gilmore X-RayFluorescence (XRF) Laboratory at the University ofMassachusetts Major elements were measured on a fusedLa-bearing lithium borate glass disc using a SiemensMRS-400 spectrometer Trace element concentrations(Rb Sr Ba Ce Nb Zr Y Pb Zn Ga Ni Cr V) weremeasured on a separate powder pellet using a PhilipsPW2400 sequential spectrometer using a Rh tube LOIand ferrous iron measurements were made as describedby Rhodes amp Vollinger (2004) Precision and accuracy esti-mates for the data have been described by Rhodes (1996)and Rhodes amp Vollinger (2004) Results for each sampleare the average of two separate analyses and are shown inElectronic Appendix 2 for a comparison of analyses ofduplicate powders using XRFand ICP-MS
PCIGR analytical methodsA subset of 19 samples was selected for high-precisiontrace-element analysis and Sr Nd Pb and Hf isotopicanalysis at the Pacific Centre for Isotopic andGeochemical Research (PCIGR) at the University ofBritish Columbia (UBC) Samples were selected from the59 samples analyzed for whole-rock chemistry at ActLabsbased on major- and trace-element chemistry alteration(low LOI and petrographic alteration index) and samplelocation Samples were prepared for trace-element analysisat the PCIGR by the technique described by Pretorius et al(2006) on unleached rock powders Sample powders(100mg) were weighed in 7ml screw-top Savillex bea-kers and dissolved in 1ml 14N HNO3 and 5ml 48 HFon a hotplate for 48 h at 1308C with periodic ultrasonica-tion Samples were dried and redissolved in 6ml 6N HClon a hotplate for 24 h and then dried and redissolved in1ml concentrated HNO3 for 24 h before final dryingTrace element abundances were measured by high-resolution inductively coupled plasma-mass spectrometry(HR-ICP-MS) with aThermo Finnigan Element2 systemfollowing the procedures described by Pretorius et al(2006) within 24 h of redissolution HFSE and LILE weremeasured in medium-resolution mode at 2000 dilutionusing a PFATeflon spray chamber washed with aqua regiafor 3min between samples REE were measured in high-resolution mode and U and Pb in low-resolution modeat 2000 dilution using a glass spray chamber washedwith 2 HNO3 between samples Total procedural blanksand reference materials (BCR-2 BHVO-2) were analyzedwith the batch of samples Indium was used as an internal
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
38
standard in all samples and standard solutionsBackground and standard solutions were analyzed afterevery five samples to detect memory effects and massdrift Results for PCIGR trace-element analyses areshown in Electronic Appendix 3 for the full set of analyzedtrace elementsSample digestion for purification of Sr Nd Hf and Pb
for column chemistry began by weighing each samplepowder (400^500mg) prior to leaching All samples wereleached with 6N HCl and placed in an ultrasonic bath for15min Samples were rinsed twice with 18 M-cm H2Obetween each leaching step (15 total) until the supernatantliquid was clear [following the method of Mahoney (1987)]Samples were then dried on a hotplate for 24 h andweighed again Sample solutions were then prepared bydissolving 100^250mg of the leached powder dissolvedin 1ml 14N HNO3 and 10ml 48 HF on a hotplate for48 h at 1308C with periodic ultrasonication Samples weredried and redissolved in 6ml 6N HCl on a hotplate for24 h and then dried Pb was separated using anion exhangecolumns and the discard was used for Sr REE and Hfseparation Nd was separated from the other REE and Hfrequired two additional purification steps Detailed proce-dures for column chemistry for separating Sr Nd and Pbat the PCIGR have been described byWeis et al (2006) andHf purification has been described byWeis et al (2007) Srand Nd isotope ratios were measured by thermal ioniza-tion mass spectrometry (TIMS) on a Thermo FinniganTriton system in static mode with relay matrix rotation ona single Ta and double Re^Ta filament respectively Fourto five filaments per barrel of 21 were occupied by stan-dards (NBS 987 for Sr and LaJolla for Nd) for each barrelwhere samples were run Sample Sr and Nd isotopic com-positions were corrected for mass fractionation using86Sr88Srfrac14 01194 and 146Nd144Ndfrac14 07219 Each samplewas then normalized using the barrel average of the refer-ence material relative to the values of143Nd144Ndfrac14 0511858 and 87Sr86Srfrac14 0710248 (Weiset al 2006) During the course of the Vancouver Islandanalyses the La Jolla Nd standard gave an average valueof 05118566 (nfrac14 7) and NBS987 standard gave an aver-age of 07102408 (nfrac1411) (2 error is reported as times106) 147Sm144Nd errors are 15 or 0006 USGeological Survey (USGS) reference material BHVO-2was processed with the samples and yielded Sr and Nd iso-topic ratios of 07034607 and 05129786 respectivelyThese are in agreement with the published values of070347920 and 051298411 respectively (Weis et al2006)
Pb and Hf isotopic compositions were analyzed by mul-tiple collector-inductively coupled plasma-mass spectrome-try (MC-ICP-MS) on a static multi-collection on a NuPlasma (Nu Instruments) system The detailed analyticalprocedure for Pb isotopic analyses on the Nu Plasma atthe PCIGR has been described by Weis et al (2006) Theconfiguration for Pb analyses allows for collection of PbTl and Hg togetherTl and Hg are used to monitor instru-mental mass discrimination and isobaric overlap repec-tively All sample solutions were analyzed withapproximately the same PbTl ratio (4) as the referencematerial NIST SRM 981 To accomplish this a small ali-quot of each sample solution from the Pb columns was ana-lyzed on the Element2 to determine the precise amount ofPb available for analysis on the Nu Plasma The SRM 981standard was run after every two samples on the NuPlasma During the time samples were run analyses ofthe SRM 981 Pb reference material gave values of206Pb204Pbfrac1416940322 207Pb204Pbfrac1415495823and 208Pb204Pbfrac14 36713164 (nfrac14 61 2 error is reportedas times 104) these values are in excellent agreement withreported TIMS triple-spike values of Galer amp Abouchami(1998) Fractionation-corrected Pb isotopic ratios were fur-ther corrected by the sample-standard bracketing methodor the ln^ln correction method described by White et al(2000) and Blichert-Toft et al (2003) Leached powders ofUSGS reference material BHVO-2 yielded Pb isotopicratios of 206Pb204Pbfrac14186454 8 207Pb204Pbfrac14 1549105 and 208Pb204Pbfrac14 38222514 and BCR-2yielded 206Pb204Pbfrac141880466 207Pb204Pbfrac14 1562518 and 208Pb204Pbfrac14 3883496 (2 error isreported as times 104) These values are in agreement withthose for leached residues of BHVO-2 and BCR-2 fromWeis et al (2006)Hf isotopic compositions were analyzed following the
procedures detailed byWeis et al (2007) The configurationfor Hf analyses monitored Lu mass 175 andYb mass 172 toallow for interference correction to masses 174 and 176 Hfisotopic ratios were normalized internally for mass fractio-nation to a 179Hf177Hf ratio of 07325 using an exponentialcorrection Standards were run after every two samplesand sample results were normalized to the ratio of the in-run daily average and a 176Hf177Hf ratio for JMC-475 of0282160 During the course of analyses the Hf standardJMC-475 gave an average value 02821679 (nfrac14130)USGS reference materials BCR-2 and BHVO-2 were pro-cessed with the samples and yielded Hf isotopic ratios of02828675 and 02831005 respectively Publishedvalues for BCR-2 and BHVO-2 are 02828717 and0283104 8 respectively (Weis et al 2007)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND
39
standard in all samples and standard solutionsBackground and standard solutions were analyzed afterevery five samples to detect memory effects and massdrift Results for PCIGR trace-element analyses areshown in Electronic Appendix 3 for the full set of analyzedtrace elementsSample digestion for purification of Sr Nd Hf and Pb
for column chemistry began by weighing each samplepowder (400^500mg) prior to leaching All samples wereleached with 6N HCl and placed in an ultrasonic bath for15min Samples were rinsed twice with 18 M-cm H2Obetween each leaching step (15 total) until the supernatantliquid was clear [following the method of Mahoney (1987)]Samples were then dried on a hotplate for 24 h andweighed again Sample solutions were then prepared bydissolving 100^250mg of the leached powder dissolvedin 1ml 14N HNO3 and 10ml 48 HF on a hotplate for48 h at 1308C with periodic ultrasonication Samples weredried and redissolved in 6ml 6N HCl on a hotplate for24 h and then dried Pb was separated using anion exhangecolumns and the discard was used for Sr REE and Hfseparation Nd was separated from the other REE and Hfrequired two additional purification steps Detailed proce-dures for column chemistry for separating Sr Nd and Pbat the PCIGR have been described byWeis et al (2006) andHf purification has been described byWeis et al (2007) Srand Nd isotope ratios were measured by thermal ioniza-tion mass spectrometry (TIMS) on a Thermo FinniganTriton system in static mode with relay matrix rotation ona single Ta and double Re^Ta filament respectively Fourto five filaments per barrel of 21 were occupied by stan-dards (NBS 987 for Sr and LaJolla for Nd) for each barrelwhere samples were run Sample Sr and Nd isotopic com-positions were corrected for mass fractionation using86Sr88Srfrac14 01194 and 146Nd144Ndfrac14 07219 Each samplewas then normalized using the barrel average of the refer-ence material relative to the values of143Nd144Ndfrac14 0511858 and 87Sr86Srfrac14 0710248 (Weiset al 2006) During the course of the Vancouver Islandanalyses the La Jolla Nd standard gave an average valueof 05118566 (nfrac14 7) and NBS987 standard gave an aver-age of 07102408 (nfrac1411) (2 error is reported as times106) 147Sm144Nd errors are 15 or 0006 USGeological Survey (USGS) reference material BHVO-2was processed with the samples and yielded Sr and Nd iso-topic ratios of 07034607 and 05129786 respectivelyThese are in agreement with the published values of070347920 and 051298411 respectively (Weis et al2006)
Pb and Hf isotopic compositions were analyzed by mul-tiple collector-inductively coupled plasma-mass spectrome-try (MC-ICP-MS) on a static multi-collection on a NuPlasma (Nu Instruments) system The detailed analyticalprocedure for Pb isotopic analyses on the Nu Plasma atthe PCIGR has been described by Weis et al (2006) Theconfiguration for Pb analyses allows for collection of PbTl and Hg togetherTl and Hg are used to monitor instru-mental mass discrimination and isobaric overlap repec-tively All sample solutions were analyzed withapproximately the same PbTl ratio (4) as the referencematerial NIST SRM 981 To accomplish this a small ali-quot of each sample solution from the Pb columns was ana-lyzed on the Element2 to determine the precise amount ofPb available for analysis on the Nu Plasma The SRM 981standard was run after every two samples on the NuPlasma During the time samples were run analyses ofthe SRM 981 Pb reference material gave values of206Pb204Pbfrac1416940322 207Pb204Pbfrac1415495823and 208Pb204Pbfrac14 36713164 (nfrac14 61 2 error is reportedas times 104) these values are in excellent agreement withreported TIMS triple-spike values of Galer amp Abouchami(1998) Fractionation-corrected Pb isotopic ratios were fur-ther corrected by the sample-standard bracketing methodor the ln^ln correction method described by White et al(2000) and Blichert-Toft et al (2003) Leached powders ofUSGS reference material BHVO-2 yielded Pb isotopicratios of 206Pb204Pbfrac14186454 8 207Pb204Pbfrac14 1549105 and 208Pb204Pbfrac14 38222514 and BCR-2yielded 206Pb204Pbfrac141880466 207Pb204Pbfrac14 1562518 and 208Pb204Pbfrac14 3883496 (2 error isreported as times 104) These values are in agreement withthose for leached residues of BHVO-2 and BCR-2 fromWeis et al (2006)Hf isotopic compositions were analyzed following the
procedures detailed byWeis et al (2007) The configurationfor Hf analyses monitored Lu mass 175 andYb mass 172 toallow for interference correction to masses 174 and 176 Hfisotopic ratios were normalized internally for mass fractio-nation to a 179Hf177Hf ratio of 07325 using an exponentialcorrection Standards were run after every two samplesand sample results were normalized to the ratio of the in-run daily average and a 176Hf177Hf ratio for JMC-475 of0282160 During the course of analyses the Hf standardJMC-475 gave an average value 02821679 (nfrac14130)USGS reference materials BCR-2 and BHVO-2 were pro-cessed with the samples and yielded Hf isotopic ratios of02828675 and 02831005 respectively Publishedvalues for BCR-2 and BHVO-2 are 02828717 and0283104 8 respectively (Weis et al 2007)
GREENE et al WRANGELLIA FLOOD BASALTS ON VANCOUVER ISLAND