Page 1
Geological and geochemical studies with Dolphin 3K
on North Hiyoshi seamount, Izu-Bonin-Mariana arc
Chih-Hsien SUN81 Robert J. STERN*1 Jh-o NAKA*2
Izumi SAKAMOTO*2 Makoto ARIMA*3
Shoshoniles (SiOs ≪ 55 to GO wi%; 100 = 3.0 lo 5.2 wi%) we recovered by R0V "Dolphin 3K'
on North Hiyoshi seamouiu (Dives 350 and 351) in the northernmost Mariana Aj-c. The major
mineial phase? in ihcse rocks are: olivine. clmopyroxene. and plagioclase with subordinate
sanidtne, magnetite, and biotile. These samples have elevated concentrations of lai-ge ion
lilhophile elements with respect lo high field strength elements. Light rare earth elements are
highly enriched, with <U/Yl ≫v = 13.0-K9. Their chemical characteristics and radiogenic (Sr-.
Nd. and Pb) is<*topic compositions are similar to other (avas from the Alkalic Volcano Province.
Glass inclusions in olivine and clino|≫yroxene are more felsic (SiOs = 61-74 wify and potassic
(KsO ・ 4 5-6.S wi%) 1 han the host lavas. Systematic valuations >≫ the compositions of glass
inclusions suggest that magma mixing controlled the petrogenesis of North HiyoShi magmas.
Keywords: fzu^Bomn-Manana, subducrion. shosliomie. Alkalic Volcano Province, glass inclusion
* I CenU'i foi Ulhosplu'rio Slmlios. Utiivorsily of 'IVxasal Pallas, I ISA
* 2 I)ocp S<*a Research itmom.Jai>an Marine*Soieiw -md 'IVclmoloyvCviuci.Jajwm
*'S (icolosical IhsUimU'. Yokohama National I'ninTsiiv.Japan
JAMSTEC J. Oeep Se* Res . J4 U!KW>
Page 2
1. Introduction
Convergent margins arc The most complex regions In
plate tectonics. Petrogenetically, the complexity is
induced by interactions between encountering plates and
subducting material, and between mantle wedge and fluid
derived from dehydration of the subducted plate. Each
has important influences on magma compositions. To
better understand the importance of each factor, it is best
to study a simple arc system. "Andean-type" arcs, though
easily accessible and widely studied, do not satisfy the
requirement, because contamination of mantle-derived
magma by continental crust may obscure the effects of
other controlling factors. For this reason, studies of
convergent margin pedogenesis focus increasingly on
intraoceanic arcs, Intraoceanic arcs provide the simplest
setting due to their entirely oceanic nature. Studies of
intraoceanic arcs have mostly focused on subaerial
edifices, which represent only a minor portion of the arc
system. Submarine edifices, the less mature portion of
intraoceanic arc systems, are poorly understood. Detailed
studies on the submerged portions of intraoceanic arcs
should provide a better understanding of convergent
margin magmatic systems. Modern marine technology,
such as the ROV "Dolphin 31? provides a valuable too) to
study submarine volcanoes. This report concerns
application of ROV "Dolphin 3fC' to studying one such
submarine volcano.
The hu-Bonin-Mariana (IBM) arc in the western Pacific
is an outstanding example of an intJ*aoceanic arc (Figure
1A). The two plates controlling the arc system are
oceanic, so that possible continental crustal contamination
can be neglected North Hiyoshi seamount lies in the
northernmost part of Mariana arc (Figure IB), about 1.100
kin south of Tokyo. Geochemical and penological studies
of North Hiyoshi and other volcanoes of the Alkalic
Volcano Province (AVP) m the centra! IBM arc. including*
Iwo Jima, Fukutoku-oka-nobu. Fukutoku seamount.
Central Hiyoshi. South Hiyoshi, and Ko-Hiyoshi show
chemical characteristics different from lavas of most IBM
volcanoes, which are tholehtic or calc-alkalic (Bloomer et
al.. 19R9b; Jin et al, 1989). The most striking difference of
AVP lav≪s arc their elevated Mi. and U*KK contents,
which show shoshonitie affinities.
The origin of shoshoniiic magmas is controversial.
MO
Attempts to interpret the origin of these melts in the IBM
Arc include, for instance, source mixing with an OlB'like
enriched mantle ([.in et al.. 1989), and involvement of
subduction components in the mantle wedge (Lin. 1992).
Stem et al. (1988) argued that propagating rifting of the
Mariana Trough backarc basin unroofed an enriched
region at the top of asthenosphere and triggered
shoshonitic volcanism. In this contribution we present
new geochemical and penological data for lavas collected
with ROV "Dolphin 3K' and use these to interpret the
roagmatic evolution of North Hiyoshi volcano.
2, Sampling and Analytical Methods
Observation and sampling was conducted on Nov.l,
1997 by the unmanned ROV "Dolphin 3K \ and mother
ship R/V "Natsushimo'. lis capability of underwater
operation noi only has yielded valuable samples, bui also
allows us to investigate the morphology and volcanology
of the submarine volcano. Two dives, Dives 350 and 351.
were performed before rough sea associated with a
typhoon halted operations.
Lava samples were sawed to remove Mn-coated and
altered outer rinds and then slabbed
Each slab was
cleaned in IN HCI for one hour and in de-ionized
water for
a half hour in ultrasonic bath. Sample powders were
prepared for major element analyses in a tungsten carbide
canister; for trace element and isotope compositions in an
alumina canister to avoid Ta and Nb contamination. All
microprobe sections wei*e doubly polished.
Major and trace element (including KEE) analyses were
carried out at the University of Kansas by ICP-ARS
(lnductively≪coupled plasma atomic emission speetrometiy)
and ICP-MS (Inductivcly≫coupled plasma mass spectro*
metry). respectively. Sample preparation for major
elemeni chemical analysis used 30 nig of powder and a
standard procedure, fusion with Uihium inelaborate
(UBOs) in a graphite ciuciblc, quenching and dissolution
in 0.1N HNO.i, and dilution to a total volume of 50 ml by
0.)N HNO:<. For trace element analysis, 100 mg of powder
was decomposed in HNO* and Hl \ evaporated, re-
dissolved completely in 4N HNOs. and diluted to a total
mass of 150g
Approximately 100 mg of powder was dissolved for
radiogenic isotopie compositions. Sr. Nd. and Pb were
JAMSTEC J. Ofiep Sea fies. 14{l<K)8)
Page 3
rig. I (A) Tectonic setting of the l≫i≪Honin-M.tri≪oia arc in ihc western Pacific. Water depth less than 3kni is outlined by shaded partem
PjuIhhI ivctnnglc shows Hip location of Flpuro IB. (B) Bathymc-tnc map of Otc Alkslic Volcano Province g^VPl- Only on? volcano.
Minami Iwo Jii≫a, dor-> not show alkalic affinity. 'Itlis i"≪y> also shows (wo drwljfp tracks. D53 & W. on Noith Myosin by R/V Thomas
VtoMpsoH in tf>85 during cru*e Tr-192 (Bitumen <! a].. 19S&0. I5iv<? siles in this stu<ly I* wiiliiii die d!*Jie<l rectanylv d≪aik<t in
Figure 2 Maud names arc given in italics; oilier AVP edifices also sliown
separated by conventional ion*exehange techniques. Pb
was further purified by the single-bead technique
(Maiaon. 1986). S>\ Nd-. and Pb isotopes were determined
using the Fitinigan MAT261 mullicotleetor mass
spectrometer at the University of Texas at Dallas (UTD).
Sr data ace fraction at ioii<o riveted to *RSr/ShSr = 0.1194
and noi inalixcd lo "'Sr/^Sr - 0.70800 for K&A SrCOs
standard. Five analyses of K&A Si CO * dining this study
yielded a mean *"Sr/ShSr of 0.70801? ± 0.000017 (total
range). M'*Nd/,1JNd are fractiomitioivcoiiecU'd to
M'vNcf/,,|*1N<l- 0.7^l<)0. Th<* mean ,nN<l/ ≪*≫Nd is
0.511857 s 0.000009 foi four measurements of ihel* Julia
JAMSTEC j. Deep Sea Ret., 14 (ISWRl
standard during this study- (Nd was calculated by using
theNd<i ≪otopicdataotPid'^al (19R9) forth?UJollaNd
standard and bulk earth to compute the bulk earth
'・^Nd/'^N'd value appropriate for the UTD lab. Pb
isotopes were analyzed
at 1.350 V and the ratios were
corrected for a 0.15% / amu thermal fraction at ion. Six
analyses of NBS 931 during ilus study gave a mean
ao≪p|,/20ipb. 16.930 a 0.004 (total range). 207Pb/-04Pb-
1 f>m* OM. ^Pb/^'Pb- 36.728 ± 0.020.
Minora) and glass inclusion compositions were
deu ≫i mined by using (he wavelength-dispersive sj^tcni of
(he JROl. .IX
A SfiOO eledron micro probe a l U'lIX An
i-n
Page 4
accelerating voltage of 15 )<V and a beam curreni o/10 nA
was used The beam diameter varied with deferent targets.
Olivine and pyroxene were analyzed using a beam diameier
of 1 nm; feldspar and glass inclusions were analyzed using a
beam diameter of5ym to minimize Na≫loss.
3. ResuJts
3,1 General Geology
North Hiyoshi seamount is a dormant edifice composed
of one major peak and five subordinate peaks, with
summit depths ranging from less than 200 to almost 600m
(Figure 2), These volcanic peaks extend approximately
10km E-W from 141 °39'E to about 141'45'E. Such a
cross-chain arrangement is a notable feature in portions of
the Mariana arc (Bloomer et aL, 1989a), Several neighboring
volcanoes (e.g. Iwo Jima and Central Hiyoshi Knoll, around
North Hiyoshi Seamount) show similar structures.
The major peak (14JR43'E) and two subordinate peaks
tl2
were surveyed by Dives 350 and 351. The two track* were
separated by a distance of about 4 km and both traversed
NW-SE. 350 started at a depth of 522m on a small
depression between the major peak and a subordinate
peak to the north, and climbed to 216m along a distance of
about 2 km. The region surveyed is mostly covered by
lava flow or pyroclastic deposits {Figure 3A-C). The
un sorted grain size and irregular shapes o/ clasts imply
that they are near the source. Sediment cover is thin and
only found in some gentle-sloping regions or depressions.
Nearly all exposed rocks have a thin coating of Mn-oxide.
Five samples were obtained from 4 localities in this (live
(Table i).
Two subordinate peaks wem investigated by Dive 351.
The dive began at 641m and traversed across the
northwestern peak, passed across a 674m deep saddle,
and ended on the slope of the western subordinate peak.
The traversed area has a rugged surface covcrcd most
JAMSTCC J Peep Sea Res., n
Page 5
Table. 1 Han(lspecim剛dcs criplionof sajllplesrefi)veredby “Z)alμxis 3K~ 350 ajld 351divcs.
Sample Mn-coating Comments3K350-1 film Porphyritic with plagioclase phenocrysts up to 1 cm long.
3K350-2 film
This rock may have been slightly altered. White interstitial
material can be seen between grains.
3K350-3A film
Vesicular rock; The dominant minerals, pyroxene and
plagioclase, can be seen easily on fresh surfaces. The outer
portion of this rock has been intensely altered
3K350-3B film More vesicular than 3k350-3A.
3K350-4 110 Vesicular and outer portion is altered. Subhedral pyroxenes
can be observed on fresh surfaces.
3K351-1 thick Volcanic tuff and biological remains cemented by Mn-oxide.3K351-2 no Andesitic lava. The rock is highly altered, but a few minerals
(biotite and clinopyroxene) remain unaltered.
3K351-3A3K.351-3B3K351-3C
thick Same as 3K351-1
JAM STEC J.Deep Sea Res..1411! 〕S吋M?
Page 6
likely by thin Mn-oxicle (Figure 3D). Five samples were
recovered, which mostly consist of Mnoxide-cemeuted
volcanic tuff and biological remains. The only lava sample
(3K3S1-2) was collected on the northwestern cone and is
heavily altered, 'Ihe slabby appearance of some outcrops
(Figure 3£~F) reflects the original surface of the
pyrodaslic deports.
Bathymetric data obtained by surveying before the dives
was compared with US NAVY SASS batliymetiy reported
by Bloomer ct al. (1989a). The earlier survey reported a
summit depth of 349m, whereas the R/V "Natsiishima"
survey indicates a summit depth of <200iri. Given that the
volcano has not einptcd in the recent past, the difference
indicates either a technical problem with the earlier sujvey
or a remarkable amount (>150 m) of tectonic uplift. We
prefer the fu*st explanation. A few narrow fissures were
observed during the dives (Figure 3E-P). No
hydrotherma) vents were found along the fissures
144JAMSTEC J Dw> Sea Res. 14 (UWK)
Page 7
3.2 Petrography
Igneous rocks from the two dives have mostly
porphyritcc and vesicular textures, <ind
outer portions
exhibiting different degrees of alteration (Figure 4).
Two phenocryst assemblages were noted; plagioclasc *
clinopyroxene + K≫ feldspar (sanidine) in 3K350-1, and
plagioclase + clinopyroxene + olivine * biotite + magnetite
* K-fcldspar (sanidine) in 3K350-2, 3a. 3b. and 4.
Typical textures of these lavas are shown In Figure 5.
Plagioclase commonly occurs as glomerocrysts with
clinopyroxene. Grain stees range up to 1 cm. Lamellar
twinning is common in plagioclase: oscillatory zoning is
observed in a few coarser grains Due to alteration, some
plagiocla&e grains arc partly replaced by secondly phases.
Samdine occurs as phenocrysts in The ground mass
(Figure 5A). Single or simple twinning are the common
habits. Sanidine has not been reported previously in North
Hlyoslu lavas.
Biotite generally forms euhedral phenocrysts or
inclusions in plagioclase (Figure 5B). Olivine is homo-
geneous without optical zoning, and occurs as euhedral to
subhedral grains (figure 5C~D). Groundmass typically
consists of |>lagioclase microlites and glass.
Sample 3K351-2. the only lava sample recovered on
Dive 351, is intensely altered, with residual biotite and
clinopyroxene.
Brownish glass inclusions are found in a few olivine and
clinopyroxene phenocrysts These inclusions show
I'ig. 4 Sample plic>l<M. (A) I'rosli iwji i>h>rrilk amluwiv wiiJi IcMsimi ptieuoci y>>K ojh ≪i lotn itmj* (RMC) M<ki samples h.nv alioird
omo jwmloiah aiul Mn tkxklo cuais on tJioir surface. <1>) Typical Mn oxidf cmxi oji Nonli Hiyoslii iaimuuni. While iwrnons
ai'P biological li.tce*.
JAMSTEC ≪)・ Deep Sea Res , 14 O'isA)
Page 8
Ktg. 5 Photomicrograph# of typical North Hiyoshi lavas collccietl clunng Dives 360 & 361. (A) Sample 3K- 350-1. (13)-(D) Sample
3K3S0 2. Flag ・ |i(agioclasc, Cyx ・ clinopyroxonc. 01" olivine.
variable textures even within an individual host mineral
Glass inclusions in olivine are spherical
to ellipiical
(Figure 6A & C), but those in pyroxene range
in shape
fi'Cin #i)hencHJ tr> veiy httgvteir fi^giirc £>33 Sr I)), 'Jln'ir
dimensions range* from 10 jjiti (o lOOym. Gas bubbles are
common in the inclusions. Th* diatncleis of these bubbles
are iclatlvely constant al about 113-20 \ufi-
3.3 Bulk rock compositions
Chemical (lata arc limbed lo igneous samples from Dive
'*550 l)jw 350 lavas ai*o iiuei'mediale in com pnsj lion (55 lo
60 StO; Table 'D and are diaiaUCi'izod by high
cmUtfiils. so lhal Ihe rocks He m Uic slioshoniU* field on a
SiO; fO<) diagram (HKiiro 7). All samples have
K<0/Nay()*) and aie approximmdy >≫imnled In silica,
willi up lu A ) molfof noim alive email/. Hie low Nj and
Mb
Cr concern rniions. atxl Mg# indicaic these samples are
highly fracJionatcd.
Previous studies subdivide lavas in ihc IBM arc inio iwo
majoi categories: cnlculkaline-lholeiilic and shosbonitic
series (Meijer. 197C; Siern. 1979; Dixon and Batiza. 19?9:
Meijer and Keagan, 1981. 1983: Dixon and Stem. 15)83.
Hole et <?/.. 1984; Slern ft at.. 1984: Slein el a I.. 198S:
Hloonicr et al 1989b; l,in ft oi, 1989). The alkalic affinities
of Dive 350 iavsis arc similai lo bill sligblly more felsic
ih ≪n i Ik* samples JVom TJ' U>2 1)53 reported by Bloomer
<tt af (1989b), The mafic samples recovered by 'IT 192
1)54 aiv nol ropresonled in ihc J?350 suiie.
Ugh I rare enrlb clement (l.RKK) enrichment 5k one ol
th<* lypkiil ch.ii'aclcriMic* of ,nv lavas Chondrh*--
uomiaUx^d Rl\K gjiiUn'us iu 'be suing les display <c ≪i\r
l,RKk mm hmi'iil with a slitfhUy nrgalive (> anomaly
JAMST6C J Oeci) Sea Res, 14
Page 9
Pig (3 Scanning election photoflraplis of ^!as≪ inclusion*. .mil
lioa olivine six) woxeue. OlkVuic-hoftie<l gUs* inclusions (pliotos A And
R) commonly have sj>]i<?rical io subsphciiesil ≫hajjcs: while glass inclusions in clinopyroxeno liavc sulrtphertcal to irregular
forms. (A)&(C) Sample 3K -3S0 2. <I3K*(1>> Sample 3K 350 4 01 - Olivine: Cpx ≪ clino-pyroxwe: Mt ・ niagneiitc.
(Cc/Ce**0.9) (Figure SA). The (U/Yb)N values range
from 13 9 lo 14.9, comparable to previously documented
values of A VP lavas (8.5-16.7, Liu et al.. 1989). HRKK
patterns are nearly flat. which is a common chaiticleri sties
of all IBM arc lavas. Another important geochenueal
feature of North Hiyoshi lavas is their enrichment in lai*ge
ion lithophile (MI.) elements. As shown in N-MORU-
noDualized "Spidergram" (figure SB) > Norili Hiyoshi
lavas display strong enrichments in I.II, (sticli as Kb. Ha.
and Th) and Pb with respect to high field strength
elements (HFSE) They are also depleted in Nb> Ta. ami
Ti relative to adjacent elemenls on the spidergram.
"Sr/ ≫Sr of the sample* cluster wilhln a narrow nntge.
from 0.7c>3?^ to rt.7u3#f>. again overlapping the AVP field
(Figure J>A). Nd rsoloptc oomiiosiiion arc less radiogenic
JflMSTEC J. Oaep Sea Res.. 14 mm
than Those of the Choleiitic to ealc-alkahue scries and
mostly lie wlihiii the lower pari of the AVP field. Pb
isotopir compositions are similar lo other
AVP lavas
(Figure W)
3,4 Mineral and glass inclusion compositions
Chemical variations in Norlli Hiyoshi plienociysts ait
modest. Cfinopyroxene phenuoysls range in composition
from !・ W 043 Fsn to KmaWo.isFsw. Com position a)
zoning was not obsewed Our clinopyioxenc analyses aa-
similar to previously reported dala lor Norlh Hiyoshi
(Bloomer ct at.. 1989b). Olivines also show a fairly
uniform compositions of Fo;a, somewhai less magncston
Ih≪m reported for IT 192 l)f>4 basnlis. I'lagioclaso
aiv commmilv zoned. The variations of compos)lion
147
Page 10
MH
Table. 2 Chemical anil ≫soiopic analyses of "Dolphin 2fC 3S0 dive samples.
TiOj
Al,0,
FejOs*
MnO
CaO
MgO
Na?0
K30
P,0
Sum
UO.I
Mg*
Rb (ppm)
Sr
Ba
Zr
Y
La
Ce
Pr
Nd
Sm
Eu
Gcf
Tb
Dy
Ho
£r
Yb
Lu
Hf
Nb
Ta
Pb
Sc
V
Cr
Co
Ni
Cu
7n
Ca
Cs
Th
U
Cc/Ce*
MSr/"$r
lfl,Nd/M,N<3
£Nd
2MPt/ °*Pb
≫≫Pb/"4Pb
0.5017.355.010.144.471.894.1752!0.22
99.320.91
0.45
149
79478124126.755.995.010,636.16.271.835.090.854.570.852.422.620.435.4015.90.8514.61091145 }41096.25J7 A61.317,12.53252
7 6414.4092
0.703740.512762.38
38 SOI15.62518.973
0.7017.427.390.166.Z 83.253.644.060.40
100.150.33
0.49
1241021$0425529 357.110212 I43.97.782.076.32j.OO5.030.952.612.590.425.2915.70.7311.017.41957.9420 313.384.071.917.01.9717.85.5514.90.91
0 703830.512803 14
38 82715.64819.022
0.6817.48
7.01O.IS7.24
3.4$3.55
3.950.3599.020.50
0 52
1211064
78224528.054.497,411.441.47,442.056.180.9?4,900.932.522.57OA I5.2515.40.7910.917.5
21316.7121.714459972 5P.O
\ 9017.95.3514.30.92
0 703850.512803.16
38 80915.62218.996
0.6817.557.180.157.0*
3.421& \4,03
0.3799.780.43
0.51
1131056
77424128.454.497.911.541.57.632.056.500.995.090.962.5C2.620.425.2214 80.7210.418.222412.5822 814069075 217.51.82
17.65.1714.00.92
0.703840.512813.28
38,83515.62919.004
0.6717667.170.157.12
3 243.573.02
0 3299 590.61
0 50
1151057
76324726.95J.291.110.558.06.981.975.950.924.710.892.422.490.405.2514.90.7310.517.91899.4520 012.166.170.116.61.8817.65.151390 92
0 703820.512803.06
32.87715645IV 01 ?
JAMSTEC J Oeen Sea Res. 14 cJWW
Page 11
Fir. 7 SiOa-KzO plot for whole rock (open clrdcs). ddjustctl olivme≪hostc>d (shaded squares) an<l clinopyroxene-hosted glass
inclusions (sj>o(te<l diamonds). compared 10 other IBM air lavas, lavas from TT 152 D53 and D54 are encuded by stippled
lines Data Tor AJkalic Volcano Province (including Pukuioku-okA-no-ba) lavas are taken from Sun and Stem (1998) and Suri el al
(190S): <la!,i for other IBM arc lavas ai*e from IJIovhkt el til. (1980b) and Un el ≪/. (19S9). Adjusted compositions of olivine-
hosted glass inclusions is discussed in text. All gl.iss inclusion analyse* are nui tnalized lo 100
between core and rink in a single grain can be as much as
10 mole % An. The plagioclase analyses in Table 3 range
from An73 10 An??, in the middle of (he broad range
defined by TT 192 D53 and D54 (Answn).
Microprobe analyses of glass inclusions are presented
in Table 3. Hie results show ihal the inclusions are more
fclsic (Si02 = 61―74 w(% ) titan (he whole rocks and are
more felsic than any AVP lava. These are classified as
shoshonites and banakites on the SiCb K2O diagram
(Figure 7). These inclusions are characterized by high
K>0 content and fo()/Na*0 ratio but low Mj*0 and CaO
contents. Normative compositions indicate these glasses
arc strongly oversalu rated in silica (5 1-42.3 mole%
quartz). All filusS inclusions aie highly (Yactionntcd. with
(Vfg# < 51 Pyroxene-hosted glass inclusions, as a whok*.
JAMSTEC J. Oeeo $*a R≪s . 14 (\<m)
have higher SiOs than olivine-hosted ones. Glass
inclusions napped in olivine have higher CaO and MgO,
*nd generally lower T10* and AJsO*.
4. Discussion
One of the main contributions of this research is the
recognition of evolved alkalic glass inclusions within
olivines and pyroxenes in North Hiyoshi lavas.
Glass inclusions {01) are known lo provide useful
records of iritigina evolution (Anderson. 1Q76; Donaldson
find Brown. 1977). Glass inclusions in pbcnoerysis show
vaiions compositions, which are observed not only in
different host minerals but also m the same host pliases
from dilleivm samples. The variability has two possible
explanations: (I) Gl coniposilions weie modified by (>t≫st≪
)■!!)
Page 12
Fig. 8 (A) N-MORB-normalized rare earth element (REE) patterns North Iliyoshi lavas have highly enriched LRF.fi and flal UREE
patterns similar to those of Alkalic Volcano Province lavas. The dominant IBM cale-alkalic anit ihtoleiitic lavas are much less
enriched in LR£B, Data for Alkalic Volcano Province arc from Sun anil Stern (unpublished dala); for Average IBM arc lavas are
from LSn et ai (1989), Elltnu el ui (1997), and Pc-ate and Pearcc (1998). (B) Flemcni comimtiWliiy diagiam. Data for Alkalic
Volcano Province and Average IBM arc lavas are compiled /rom Lin et at. (1989), Btoomer et al. (1989W. Elliott et ai (1997), Sun
H cl. (199&). Peate and Pearce (1998), Sun d ai (1998) and Sun and Stem (unpublished dauh
entrapment crystallization; or (2) GI reflect melt
compositions captured during different stages of magma
evolution. A handful of articles discuss (he chemical
modification of glass inclusions by post-entrapment
crystallisation (e.g. Dnngan and Rhodes, 1978; Kamenctsky
et al., 1995; Lee and Stern. 1998). The possibility of post-
entrapment crystallisation for Gf in olivine can be
assessed by calculating the Fe- Mg Kj> (Kp - (Fed/Mgoi)
/(Feii<i/Mgiiri) | of host mineral-glass inclusion pairs
(apparent Kjj) and comparing this with the value of 0.30
expected at equilibrium (Roeder and Emslie, 1970).
Olivine hosted glass inclusions in caloalkaline and
th ol eii tic lavas of (he IBM arc have apparen i Kus of 0 14 to
0.20 (Ice and Stern, 1998). Apparent K)>s of North Hiyoshi
glass Inclusions i tinge from 0.10 to 0.21, indicating significant
<H≫)nciH u'ysiaUisalion. The- compositions of olivjnc-
hosied glass inclusions in this study werf adjusted to the
ISO
equilibrium Kn value of 0.30 (Roeder and Emslie, 1970) by
adding 2 - '3% of olivine. This indicates thai post-
entrapment crystallization is a minor effect. The corrected
results are tabulated along w[(h the raw data in Table 3.
Connection for post-entrapment ciystullization in pyroxene-
hosted glass inclusions is currently not feasible clue to a
lack of procedure for evaluating equilibrium between
melt-clinopyroxene pair*. However, the effect of post*
entrapment crystallization in clinopyroxcne is not
negligible. 'Hie decrease of CaO/AlsOs in dinopyroxene-
hosted inclusions may be due to post ≪entra)>ment
crystallization of di no pyroxene Notably. Che significant
differences in Na and total oxides between olivine-hosted
and pyroxene-hosted melt inclusions probably resulted
from Na-loss during micro pi obt> analyses, especially for
(hose analyses with low total (<9? wt%)
JAMSTEC J Oe≪f> Sea Res., 14
Page 13
I;kfi. 0 (A) fNd versus 8?Sr/*6Sr diagram showing lhr> North Hiyoshi data of this study relative to other IBM arc lavas. The data arc
from: Alkalic Volcano Province (Sun et aL 1998 ; Sun and Stem. uniHiblishcd (lata):
Oiher IBM arc lavas (Di≪ni and Sicni, 1083:
Stem and Bibec. 1981. Woodhead and Prater. 19S5: Uo and Stern, im Woodhead. 19S9; Un ct al. 1990). (B) Pb isotopic
compositions of Nonh Hiyoshi lavas compared to the oilier IBM arc bvas. The field for AJkalic Volcano Province (Sun et oi, in
press; Sun and Sscrn, unpublished data) is shown, along with Uto field for other IBM arc lavns (Mcijer. 197& Woodhead el al,
1987: Slern el al. 1993: Elliou et qL 199?).
Although glass inclusions have up to 14% more Si02
than the host lavas, they define a relatively continuous
trend with the whole rock data on a Si02-K*0 diagram
(Figure 7). A subset of the most fclsie glass inclusions fall
off the trend of increasing KiO and SrC? and fie in tfie
h'gh-K calcalkaline field It is noteworthy that these
anomalous Gl all are clinopyroxene-hosted G \ in sample
3K350-1.
The continuous trend among host rocks and glass
inclusions might result fi-oni factional crystalljnation. To
evaluate this, major elements and CaO/AhOs mtios of CU
<11* plotted in Harker dtagianis along with the whole rocks
(figure 10). Al'Ox C.tO. MgO, and FeO* inversely
correlate wah SiO:'. sn^^siine magmatic evolution was
controlled by fractional cryatafHstaiion. Olivinc-liostcd
compositions plot between bulk
rocks and clmopyroxeno-
hosted inclusions. As crystallisation proceeded. j| depleted
JAMST6C J Deep Sea ≪・≫.. H <!≪*>
such elements as Mg, Fe, Ca by crystallizing olivine and
clioopyroxene, Na was lately consumed by plagioclase
accumulation.
Felsic melts In the IBM arc are not very common. Host
lavas and glass Inclusions define a cafc-afkaline trend
toward high total alkali in AFM diagram (Fig 11),
Illustrating magma evolution. Lee et al
(1995) and l-ee
and Stern (1998) argued that felsic tephra and glass
inclusion hi Mariana Arc may be a result of li-actionation
from mafic melts. Meen ct al, (1998) argued that North
Hiyoshi lavas evolved by fractional crystallisation along
the olivineplagioclase-augite thermal divide. Norm
compositions of lavas and glass inclusions in Dive 350 are
piojected from Di onto a 01 Plag Qz/Ne Quaiternajy
diagram (Ilgiuv 12), Host lavas are quana-saturated and
basielly fall on ihe thermal divide, indicating fractional
crystallization of phcuocrysi phases of o!iv|ne≫nugito
tSl
Page 14
152
Table. 3 Representative mineral and glass inclusion compositions
(,) Raw analytical data.
{2> For calculation of Mg#, Fe2V(Fe2* + Fe3* ) - 0.9.
<3)01ivine-hosled glass inclusion compositions adjusted to Kd = 0.3,
where Kj> ■ [(Fc<n/Mgoi)/(Fenq/Mgnq)] (Roederand Emslie, 1970).,li} Raw analytical data.
JAMSTEC J. Oeep SeR Res., 14 (1D98)
Page 15
Fig, 10 Marker diagrams showing relationships ber\v≪en host rocks and glass inclusions. Whote rocks. adjusted oJivicie-hosted and
clmopyroxe/w-hosuxl glass inclusions are 1 epresenled by oi>et> circlcs, shaded squares. ajid sported diamonds, respectively.
plagioclase. Glass Inclusions, on the other hand, form a
trend from near (he Qz apex, through the Plag+Aug+low-
Ca Pyx. to the OI*Plag*Aug field. This trend is not
compatible with fractional crystallization of either
Plag+Aug+low-Ca Pyx or 01 -Plag-Aug alone. A plausible
explanation is that magma mixing, instead of fractional
ciystallization, create the liquid trend. An alkaline magma
mixed with low-K silicic magma may possibly generate a
serial fieri vant magma as shown in these North Hiyoshi
glass inclusions. Tlie alkaline endmember could be
potassic and sNiea-undersaturated, in order to pixxince
these alkalic Norih Hiyoshi lavas: while the felsic
endmember is most likely a subalkaline rhyolite Such a
process had been proposed by Meen et ai (in press) lo
interpret the* pedogenesis of Central and South Hiyoshi
lavjis. Fractional crystallization may be ihe dominant
process among the host lavas in North Hiyoshi. But.
evidence from (he glass inclusions implies that magma
mixing lias controlled th<* i*vokil5ou of Ihe whole magmsi
system beneath North Hiyoshi. as well ihe other
Hiyoshi volcanoes.
5. Conclusions
Lavas collected from North Hiyoshi seamount in (his
study are characterized by highly enriched R£E and L1L
contents. Their relatively rariwgerirc
Pb end $r isotapic
compositions and non-radiogenic
Nd aj* like those of
other AVP lavas.
Glass Inclusions within olivine and clinopvroxene are
more felsic than hosi lavas but illustrate a continuous
evolutionary trend The feature may be interpreted by
magma mixing between a fractional crysiiiUination of
olmne-pyroxenc-plagioclase-iriagiietile. This implies thai
magmaftc evolution ol N011I1 Hiyoshi lavas may bp mainly
con (rolled by lintiional crystallization
Acknowl ed ge ments
Wc are grateful to Uie captain and crew of the R/V
"NavsushinuT (01 iheir veiy professional effort? dining the
com*sc of this invcsiication. We ihank T. Plank, smd C Dan
Page 16
Fig. 11 AFM diagram show* a calc-aJkatine fractionation trend ol North Hiyoshi whole rocks and jilass inclusions, compared to whole
rock compositions of AVPand other IBM arc lavas Data far AJkalic Volcano Province are taken from Sun a≫<l Sicm (1906): Sun
ei for other IBM arc lauas arc compiled from Bloomer el ai (1989b) ami Lin at (19S9). Symbols are same as tlial in
Figure 10.
Fig 12 Ol-Plag-Qfc/Ne quaternary diagram projected from Di. Host rocks and glass inclusions arc plotted and compared with lavas
from IT-192 DS3 & DS4 (Bloomer ef aLl9$9b). The daia of lavas and tflass inclusions approximately form a trajectory of
fractional crystallization. Ol-olivine. Plaff-plagioclase. Qz-quartT, N'c≫)Cpl)≪ii≫≪. Di≪d ioesitlc. Aug-augite. low-Ca Pyx-low C ≫
pyroxene.
for the help in ICP-AES and ICP- MS. and D. Draper for
technical assistance with microprobc analyses. The
p-icdtipaAwn Saat> arid Stern rrttde possible by a gram
from NOAA through ihc US-Japanese Submersible Science
Progi-jiri This is UTD Geoscicnces conlnhulion A89I
References
Anderson, A.T. (1974) : Evidence for a picrUic, volatile-
rich magma beneath Mi. Shasta, California. Journal of
I'etrufogy, 15.243 267.
Bloomer. S.H., K.J. Siern and N.C Smoot, (iy8ya) :
Physical voleanoloKy of lti(- submaiine Mariana and
Sf>4
Volcano Arcs. Bulletin of Votcanology. 51.210- 224,
Bloomer, S.H.. RJ- Stern, K. Flsk and C H Geschwind
(1989b) : Shoshonilic volcanism m the Mariana Arc. 1.
Mineralogic and major and trace element character-
istics. journal of Geophysical Research, 94. 4469
4496.
Dixon, T.H. and R. Batiza (1079) : Petrology and chemistry
of recent lavas in the Northern Marianas' implications
for (lie origin of island arc basal I* Contributions to
Mineralogy nd Petrology. 70.167 181.
Dixon, T.H. and R,1 Sl<*rn (1<)83) : 1 Virology, ch^mishy
and isotonic composition of submarine volcanoes in
JAMSTECJ Deep Sea *es . J<s (nm)
Page 17
the southern Mariana arc, Geological Society of
America Bulletin, 94,1159-1172.
Donalson, C.H. and R.W. Brown (1977) : Refractory
megacrysts and magnesium-rich melt inclusions
within spinel in oceanic tholeiites: indicators of magma
mixing and parental magma composition. Earth and
Planetary Science Letters, 37,81 -89.
Dungan, MA. and J.M. Rhodes (1978) : Residual glasses
and melt inclusions in basalts from DSDP Legs 45 and
46: evidence for magma mixing. Contributions to
Mineralogy and Petrology. 67,417-431.
Klliott, T., T. Plank, A. Zindler, W. White and B Bourdon.
(1997) : Element transport from slab to volcanic front
at the Mariana arc. Journal of Geophysical Research,
102,7, 14991-15019.
Hoie, H.J., A.D. Saunders, G.F. Mairiner and j. Tarney
(1984) : Subduction of pelagic sediments: implications
for the origin of Ce-anomalous basalts fi*oin the Mariana
Islands. Journal of Geological Society of Loudon, 141,453
-472.
Ito, J! and RJ. Stern (1986) ; Oxygen* and Strontium-
isotopic investigations of subduction zone volcanism:
the case of the Volcano Arc and the Mariana Island
Arc Earth and Planetary Science Letters. 76,312-320.
Kamenetsky. V.. N. M(trich and R. Cioni (1995). Potassic
primary melts of Vulsini (Roman Province): evidence
from mineralogy and melt inclusions. Contributions to
Mineralogy and Petrology, 120, 186- 196.
Lee J. and RJ. Stern (1995): Forty million years ofmagmatic
evolution in the Mariana arc: The tepbra glass record.
Journal of Geophysical Research, 100,9.17671-17687.
l-ee, J. and RJ. Stern (1998) : Glass inclusions in Mariana
Arc phenocrysts: a new perspective on mag ma tic
evolution in a typical intra≫oceamc arc. The Journal of
Geology. 106.19 33.
I.in, R.N., RJ. Stem and S.H Hjoomer (1989) : Shoshonitic
volcanism in the northern Mar iana Arc: 2. Urge-ion
lithophile and rai? wulli element abundances: evidence
for the souixe of incompatible clement enrichments in
inliaoceanic arcs. Journal of Geophysical Research. 94.4.
4497 -4514.
Lin. R.N.. RJ Stern. J Morris and S.H. BIchhiici. (1990) :
Nd- and Sr-i so topic compositions of lavas from tlu*
northern Mariana and southern Volcano arcs.
JAMSTEC J. Oeep Sea fies . M iliMS)
implications for origin of island arc melts. Contributions
to Mineralogy and Petrology, 105.381-392.
Lin, P.N. (1992): Trace element and isotopic characteristics
of western Pacific pelagic sediments: implications for
the peirogenesis of Mariana Arc magmas. Geochemica
et Cosmochitnica Acta, 66,1641-1654
Manton, W.I. (1988): Separation of Pb from young zircons
by single ≫be<id ion exchange. Chemical Geology
(Isotope Geoscience Section). 73,14?-152.
Meen, J.K., RJ. Stern and S.H. Bloomer (1998). Evidence
for magma mixing in the Mariana arc system. The
Island An, 7.443-459.
Meijer, A. (1976) : Pb and Sr isotopic data bearing on the
oiipin of volcanic rocks from the Mariana island-arc
system. Geological Society of America Bulletin, 87,
1356-1369
Meijer, A. and M. Reagan (1981) : Petrology and
geochemistry of the island of Sarigan in the Mariana
Arc: calc-alkaline volcanism in an oceanic setting.
Contributions to Mineralogy and Petrology, 77,337-354.
Meijer, A. and M. Reagan (1983): Origin of KsO-SSO? trends
involamoesofthejVarirtMJ arc. Geology, 11,67-71.
Peate. D.W. and J.A. Pearce (1998) : Causes of spatial
compositional variations in Mariana arc lavas: trace
element evidence. The Island Arc, 7.479-495.
Pier, J.G., P. A. Podosek, J,P, Luhr, J .C. Brannon and J J.
Aran da* Gomez (1989) : Spinel-lherzolitc-bearing
Quaternary volcanic centers in San Luis Potosi.
Mexico.2 Sr and Nd isotopic systcmatics. Joumal of
Geophysical Research, 95,7941-7951.
Roeder. P.L. and R.F. Emslie (1970) : Olivine*tiquid
equilibrium. Contributions to Mineralogy and Petrology.
29,275- 289.
Stem, RJ. (1979). On the origin ofandesite in the northern
Mariana Island arc: implications from Agrigan.
Contributions to Mineralogy and Petrology. 68.207-219.
Stern, P.J. and L. Bibee (1984) : Esmeralda Bank:
Geochemistry of an active submarine volcano in the
Mariana Island aiv Contributions to Mineralogy cud
Penology. 86.159 -169.
Siem, RJ.. N.C. Smoot and M. Rubin (1984). Unzipping of
the Volcano Aiv. Japan. Tcctnnoi>hy$ic$. 102. 153 175,
Stem. IU-. S.H. Bloomer, I'.N. Lin. K. Ho and J Monis
(1M8) : Sbo^honitic magmas in unsccnt arcs: new
Page 18
evidence from submarine volcanoes in {he northern
Marianas. Geology. 16.426 - 430.
Stern, RJ.. M.C. Jackson, P. Fryer and £. 1(0 (1993) : 0,
Sr, Nd, and Pb isotopic composition of (he Kasuga
Cross-Chain in the Mariana Arc: a new perspective on
(he K-k relationship. Earth and Planetary Science
Letters1119,459-475.
Sun, C.H and R J. Stern (1998): The source controversy for
the origin of shoshonitic lavas in (he IBM arc system,
western Pacific. GSA Abstracts with Programs, 30,130.
Sun C.H., RJ Stern., T. Yosluda. and J.-I. Kimura (1998) :
Fuku(oku-oka ≪no*ba volcano: A new perspective on (he
Alkatlc Volcano Province in the IBM arc. The Island Arc.
7,432-442
/5ft
Wood head, J.D. and D.G. Fraser (1985): Pb, Sr and lOBe
isotopic studies of volcanic rocks from the Northern
Mariana Island, Implications foe magma genesis and
crustal recycling in the Western Pacific. Geochemica
et Cosmochimica Acta. 49, 1925- 19X0.
Woorthead, J.D., R.S. Harmon and D.G. Fraser (1987): 0,
S, Sr, and Pb isolope variations in Volcanic rocks from
the Northern Mariana islands: implications for crustal
recycling in intra-ocennic arcs. Earth and Planetary
Science Letters, 83,39-52.
Woodhead, J.D.(19S9) Geochemistry of the Mariana arc
(Western Pacific): source composition and processes.
Chemical Geology. 76, 1-24.
(Manuscript received 21, July 1998)
JAMST6C J Deep Sea Res, 14 <)')<≫)