Precambrian Research 164 (2008) 201213Contents lists available
at ScienceDirectPrecambrian Researchj our nal homepage: www. el
sevi er . com/ l ocat e/ pr ecamr esGeochemistry of ne-grained
clastic sedimentary rocks of the NeoproterozoicIkorongo Group, NE
Tanzania: Implications for provenance andsource rock
weatheringCharles Kasanzu, Makenya A.H. Maboko, Shukrani
ManyaDepartment of Geology, University of Dar es Salaam, P.O. Box
35052, Dar es Salaam, Tanzaniaarti cle i nfoArticle
history:Received 2 November 2007Received in revised form 23 April
2008Accepted 27 April 2008Keywords:Ikorongo
GroupMudrocksGeochemistryWeatheringProvenanceabstractThe
Neoproterozoic IkorongoGroup, whichlies unconformablyonthe late
ArchaeanNyanzianSupergroupof the Tanzania Craton, is comprised of
conglomerates, quartzites, shales, siltstones, red sandstones
withrare agstones and gritstones and is regionally subdivided into
four litho-stratigraphic units namely theMakobo, Kinenge, Sumuji
and Masati Formations.We report geochemical data for the mudrocks
(i.e., shales and siltstones) from the Ikorongo basin in anattempt
to constrain their provenance and source rock weathering. These
mudrocks are compositionallysimilar to PAAS and PS indicating
derivation from mixed macfelsic sources. However, the
siltstonesshow depletion in the transition elements (Cr, Ni, Cu, Sc
and V) and attest to a more felsic protolith thanthose for PAAS and
PS. The Chemical Index of Alteration (CIA: 5282) reveal a
moderately weatheredprotolith for the mudrocks. The consistent REE
patterns with LREE-enriched and HREE-depleted patterns((La/Yb)CN
=7.338.3) coupled with negative Eu anomalies (Eu/Eu* =0.71 on
average), which character-isticsaresimilartotheaveragePAASandPS,
illustratecratonicsourcesthatformedbyintra-crustaldifferentiation.GeochemicalconsiderationsandpalaeocurrentindicationssuggestthattheprovenanceoftheIko-rongo
Group include high-Mg basaltic-andesites, dacites, rhyolites and
granitoids from the NeoarchaeanMusoma-Mara Greenstone Belt to the
north of the Ikorongo basin. Mass balance calculations suggest
rel-ative contributions of 47%, 42% and 11% from granitoids,
high-magnesium basaltic-andesites and dacites,respectively to the
detritus that formed the shales. Corresponding contributions to the
siltstones detritusare 53%, 43% and 4%. 2008 Elsevier B.V. All
rights reserved.1. IntroductionInclasticsedimentaryrocks,
thetraceelementssuchasrareearthelements(REE)andTharesaidtoberelativelyinsolubleandasaresulttheiroriginalcompositionsarenotupsetduringweathering,
erosion and transportation from the parent rocks todepositional
environments (Taylor et al., 1986). Although, pro-cesses suchas
weathering, hydraulic sorting, and
post-depositionaldiagenesishavebeenreportedtodistort geochemical
informa-tionaboutthesourcearea(e.g., NesbittandYoung, 1982),
yet,ratiosoftheimmobiletraceelementsnormallyreectthoseofthe source
rocks rather than the effects of sedimentary processes(Taylor and
McLennan, 1985). It is on this basis, therefore, that thechemistry
of ne-grained clastic sedimentary rocks has been longutilized for
making inference on source rock compositions, palaeo-Corresponding
author. Tel.: +255 784 77 56 56; fax: +255 222 41 00 78.E-mail
address: [email protected] (C.
Kasanzu).climaticconditionsandtectonicsetting(TaylorandMcLennan,1985).The
geology of Tanzania (Fig. 1) can generally be subdivided intove
tectono-stratigraphic units: (1) Archaean cratonic rocks
(i.e.,theDodomanBelt, NyanzianandKavirondianSupergroups);(2)early
to late Proterozoic sedimentary covers and metamorphic ter-ranes
(e.g., the Karagwe-Ankolean; the Bukoban Supergroup,
theUbendian-Usagarani
Supergroups;andtheIkorongoGroup);(3)Pan-African metamorphic rocks
(i.e., the Mozambique Belt) locatedin the eastern margin of the
Tanzania Craton; (4) Phanerozoic sed-imentarybasins(e.g.,
TheKaroo);(5)NeogenevolcanicrocksofnorthernandsouthernTanzania(Cloutieretal.,
2005andrefer-ences therein).The Ikorongo Group of clastic
sedimentary rocks, which is thefocusof thisstudy,
hasbeencorrelatedtotheNeoproterozoicBukobanSupergroup(e.g.,
Shackleton,
1986)whichoverliestheTanzaniaCratoninwesternandnorthwesternTanzania(Fig.
1).Geochemical studies in Tanzania to date have mainly
concentratedontheArchaeancrustal rocks(e.g., Messo, 2004;Manya,
2005;0301-9268/$ see front matter 2008 Elsevier B.V. All rights
reserved.doi:10.1016/j.precamres.2008.04.007202 C. Kasanzu et al. /
Precambrian Research 164 (2008) 201213Fig. 1. Generalized
tectono-stratigraphic map of Tanzania (modied fromASGA/UNESCO,
1968). The regional setting of the study area shown in Fig. 2 is
indicated in the insetbox.Manyaetal., 2006, 2007a,b; Mtoro,
2007)andthePan-Africanorogeny(e.g., Maboko, 1995;Muhongoetal.,
2001)inthecon-text of deciphering the geological evolution of the
Tanzania Cratonand the neighboring mobile belts. On the other hand,
sedimentarycovershavereceivedlittleornoattention. Inthispaper,
there-fore, we present major and trace element geochemical data of
thene-grained sedimentary rocks (i.e., siltstones and shales)
fromtheIkorongo basin. The purpose of the study is to constrain
source(s)of the Ikorongo clastic sedimentary rocks and to
understand theweathering conditions in the source region(s).2.
Geological settingTheIkorongobasinliesinthenortheasternpartof
theTan-zania Craton (Fig. 1). It is comprised of clastic
sedimentaryrocks which unconformably overlie Archaean basement
rockscomposedofhighlydeformedmetamorphicrocks,
granitesandgreenstone sequences whichformpart of the
NyanzianSupergroup(Stockley, 1936). TheIkorongoGroupis
composedpredominantlyofconglomerates, quartzitic sandstones, brown
and green shales, silt-stones, red sandstones and subordinate
agstones and gritstones(Macfarlane, 1965; Kasanzu, 2007).The
general geology of the IkorongoGrouphas beendiscussedinPickering et
al. (1959) and recently summarized in Kasanzu (2007).Regionally,
the group is comprised of four litho-stratigraphic for-mations
namely, frombottomto top, the Makobo, Kinenge, Sumujiand Masati.
However, the Makobo Formation is not exposed in thestudyarea.
TheKinengeFormation, whichunconformablyover-lies basement rocks, is
comprised of quartzites and conglomeraticsandstones that are
overlain by shales and siltstones of the SumujiFormation(Fig. 2).
TheSumuji Formationisoverlain, throughagradual transition, by thick
bedded, ne- to medium-grainedcross-bedded red sandstones of the
Masati Formation. On the basis oftheirgeographicaldistribution,
therocksoftheIkorongoGroupwere most likely derived from weathering,
erosion and depositionof rocks that constituted the exposed upper
continental crust of theTanzania Craton during the late Proterozoic
(Kasanzu, 2007).3. Sampling and analytical methodologySamples
presentedhereincludeshales andsiltstones (heretermed as mudrocks)
collected fromthe Sumuji Formation in viewof the fact that
ne-grainedclastic sedimentary rocks are more use-ful in geochemical
studies than the coarser ones (e.g., Taylor andMcLennan, 1985).
Mudrocks are ne-grained siliciclastic rocks
richinclayminerals(Yong, 2002). Clayspreservesourcerockchem-ical
signatures due to the fact their mineralogy is rarely
affectedduringdiagenesisandmetamorphosis(Weaver, 1989).
Fifty-fourmudrock samples collected after careful geological
mapping weretrimmed to remove weathered surfaces and subsequently
crushedin a jig-saw crusher for size reduction. The particles were
washedand oven-dried at 70C overnight. The dried samples were
thenleft to cool for 24h. The samples were pulverized in an agate
plan-etary mill to a grain size of Na2O, it was assumed that the
concentration of CaOequals that of Na2O (Bock et al., 1998).
However, only one brownshale sample (IK 67) showed CaO contents
higher than Na2O.The calculated CIA values for the Ikorongo
mudrocks are pre-sented in Table 1. The brown and green shales have
comparable CIAvalues which range between 59 and 80 with a mean of
75 whereasinthesiltstonesthevaluesrangefrom57to73withameanof66. The
average CIA values in the shales are higher than that
forPAAS(CIA=69;noCIAdataforPS)andattesttoamoreweath-ered source. In
the shale samples, the elements Ca and Sr are moredepleted relative
to PAAS than Al, K, Cs and Ba. These discrepanciescan be explained
by two possible processes. Firstly, since Ca and Srare contained in
minerals that weather rapidly than those whichcontain K, Al, Cs and
Ba (e.g., Roddaz et al., 2006), their depletion(i.e., Ca and Sr) in
the shales, therefore, could be caused by weather-ing of parent
rocks. Secondly, suchdepletions canalso be attributedto source
rocks being poor in plagioclase since both Ca and Sr arecontained
in plagioclase (e.g., Roddaz et al.,
2006).ThefactthatsiltstonesarecharacterizedbylowerCIAvaluesthan
PAAS suggests that the Ikorongo siltstones were fromthe ero-sionof
a less weatheredsource thanthat for PAAS. This is supportedby the
lower contents in both Al and K relative to PAAS. Further-more, the
depletioninCaOinthe siltstone samples, relative toPAAS,is
suggestive of parent rocks being poor in plagioclase.
Therefore,besides weathering, it seems that these rocks were
derived fromthe erosion of plagioclase-poor rocks.In a plot of
Al/Na ratios versus CIA of Servaraj and Arthur (2006)all shales and
most siltstones plot, together with PAAS, in the inter-mediate
weathering eld except for a few samples which plot inthe low
weathering eld (Fig. 6). Of interest is the coexistence ofthe shale
samples with variable Al/Na ratios but displaying similarCIA values
(Fig. 6). This trend could be caused by diagenetic loss ofNa+(e.g.,
Nesbitt and Young, 1982).Most authors favor the use of Al2O3(CaO*
+Na2O)K2O(ACNK) ternary plot inevaluating the chemical
weatheringtrendsthansimplecomparisonofnumericalvalue(e.g.,
Nesbittand Young, 1984, 1989; Roddaz et al., 2006). The ACNK plot
isalso useful inindicating post-depositionmetasomatic
modicationofmajorelementcompositions(Fedoetal., 1995).
Insuchdia-grams, plots of mudrocks commonlyformaweatheringtrendthat
isalmost parallel to the ACNjoint. On the Al2O3(CaO*
+Na2O)K2O(ACNK) diagram, the siltstone samples plot sub-parallel to
theC. Kasanzu et al. / Precambrian Research 164 (2008) 201213
207Fig. 6. Scatter plot of Al/Na ratios versus Chemical Index of
Alteration (CIA) for theIkorongo mudrocks. Note the increase of
Al/Na ratios at nearly constant CIA valuessuggesting post
depositional loss of Na+. Fields are fromServaraj and Arthur
(2006).Symbols as in Fig. 3.ACNaxis scattering towards the area
between illite and muscoviteand, yet, dene an ideal trend for
weathering of a primary sourcewithagranodioriticcomposition(e.g.,
Fedoetal., 1995). Onthecontrary, shale samples cluster in a tight
group near the AK axis(Fig. 7), attesting to a uniformly weathered
source. The fact thatthe siltstone samples display variable CIA
values (Fig. 6), coupledwiththeirscatteringintheACNKdiagram,
demonstratethatthese samples might have been affected by other
processes thanweathering, mostlikelyK-metasomatism(Fig. 7).
Similartrendshavebeenreportedinancientweatheringproles(e.g.,
Nesbittand Young, 1989) and in Archaean shales (e.g., Fedo et al.,
1995),and are interpreted to be caused by diagenetic modication of
K bymetasomatism.5.2. Mineral sortingOf particular importance,
sedimentary processes suchas sortingmay modify the mineral
abundances and consequently the abun-danceofspecicelements.
McLennanetal. (1993)usedaTh/Scversus Zr/Sc plot to distinguish
between the contrasting effects
ofsourcecompositionandsedimentaryprocessesonthecomposi-tion of
clastic sedimentary rocks. An addition of zircon by mineralsorting
and/or recycling to samples would result in an increase inZr/Sc
ratios (McLennan et al., 1993). On the Th/Sc versus Zr/Sc dia-Fig.
7. Al2O3CaO* +Na2OK2O(NesbittandYoung,
1984)diagramforIkorongomudrocks. The scattering of the siltstone
samples is suggestive of K-metasomatism.Also included is the
position for an original granodioritic source (from Fedo et
al.,1995). Plg: Plagioclase; Ksp: K-Feldspar.Fig. 8. Th/Sc versus
Zr/Sc diagramfor the Ikorongo mudrocks (after McLennan et
al.,1993). Trend 1 represents sediments derived directly from
igneous rocks that havebeen least affected by sedimentary sorting
and recycling. Heavy mineral accumula-tion by sediment sorting and
recycling would result in Zr enrichment relative to Thas dened by
Trend 2. Symbols as in Fig. 3.gram, two trends, one showing direct
contribution from primarysource rocks (marked 1), and the other
showing the inuence ofsedimentaryprocesses (marked2) canbe
distinguished(Fig. 8). Thetrenddepicting sedimentary processes
reects the effect of mineralsorting and/or sediment recycling, the
effect of which is the prefer-ential Zr enrichment in the
sediments. On the Th/ScZr/Sc diagram(Fig. 8), all shale samples
cluster in the eld sub-parallel to Trend1 suggesting compositional
homogeneity and minimal inuence ofheavy mineral sorting. Onthe
other hand, the siltstones display twotrends with some samples
plotting along Trend 1 which is
indica-tiveofminimalinuenceofmineralsortingandothersplottingalong
Trend 2 which is indicative of heavy mineral accumulationby
sediment recycling and sorting (Fig.
8).Mineralsortingnormallytendstoincreasetheabundanceofnonclay
detrital minerals at the expense of clay minerals (NesbittandYoung,
1984). Therefore, thefactthatthesiltstonesplotonboth trends in Fig.
8 suggests that the relative differences in theabundances of some
trace elements might have been attributed tomineral sorting. On the
other hand, the lower contents in the ele-ments Cr, Ni, Sc and V
could be attributed to the lesser amounts ofAl2O3 which have been
reported to signicantly control trace ele-ments distribution (e.g.,
McLennan et al., 1983; Asiedu et al., 2000).Thisfact
issupportedbythepositivecorrelation(r2) betweenAl2O3 andmost of the
ferromagnesiantrace element abundances inthe siltstones (e.g.,
Al2O3Sc =0.9; Al2O3Cr =0.4; Al2O3Ni =0.6).We also suggest that the
lower abundances of the trace elements,particularly the REE and
ferromagnesian trace elements, in the silt-stones
couldbeattributedtoa dilutioneffect of quartz. For instance,a
siltstone sample IK 80 contains higher contents of SiO2
(92wt%)thanallsamples,
consequentlyitsREEcompositionsaresigni-cantly diluted (Fig.
5).Nonetheless, REEpatterns of all Ikorongosiltstonesamples(except
sample IK 80) are similar to those of PAAS and PS, suggest-ing that
they have not been intensively affected by factors, in thiscase
mineral sorting, that could disrupt source rock information.5.3.
ProvenanceAmongthefactorsthatcontrolthegeochemicalcompositionof
clastic sedimentary rocks include source rocks,
weather-ing/recycling, andpost-depositional diagenesis(e.g.,
Taylorand208 C. Kasanzu et al. / Precambrian Research 164 (2008)
201213Fig. 9. Plot of Th/U versus Th for the Ikorongo mudrocks. The
grey box shows thetypical range of upper crustal protoliths. The
arrowstands for an idealized weather-ing trend for sediments
derived from upper crust (McLennan et al., 1993). Symbolsas in Fig.
3.McLennan, 1985; McLennanet al., 1993). As shownabove,
nonethe-less, these processes can only be responsible for minor
variationsin major and trace element contents in the Ikorongo
mudrocks andrather the chemical characteristics reect the
composition of thesources.The major and trace element compositions
for the brown andgreen shales (Table 1; Figs. 4 and 5) are highly
comparable and aresuggestiveofasimilarprotolith.
HoweverthedepletioninTiO2,Al2O3 and the transition trace elements
particularly Co, Ni, Sc andV, which are normally enriched in mac
rocks (Rollinson, 1993), inthesiltstones,
couldprobablyindicaterelativelymorefelsic detritusthan that for the
shales.The Th/U ratios are very useful in determining the source
char-acteristics of clastic sedimentary rocks (Roddaz et al.,
2006). Thepresent day average crust has Th/Uratios of 4.254.30
whereas thevalues for upper andlower mantleare2.6and3.8,
respectively(Pauletal., 2003andreferencestherein).
AlthoughsometimeshigherTh/U ratios have been related to oxidative
weathering and removalof U, yet, clastic sedimentary rocks derived
from the upper crustarecharacterizedbyratiosequal
toorgreaterthan4whereasratioslowerthan4havebeenrelatedtoamantlecontribution(e.g.,
Roddaz et al., 2006). The Ikorongo siltstones, brown and
greenshales show mean Th/U ratios of 5.71, 6.90 and 5.86,
respectively,whichcharacteristicsaresuggestiveof uppercrustal
parentage.These ratios are, however, higher than the values for
PAAS (4.70)and PS (4.21) (Taylor and McLennan, 1985; Condie, 1993).
The ele-vatedTh/UratiosintheIkorongomudrockscouldbeattributedto
either increased weathering intensity or variation in
oxidationstate during deposition which would permit U mobility
(Roddaz etal., 2006 and references therein). On the Th/U versus Th
diagram,all mudrock samples from the Ikorongo basin follow the
idealizedweathering trend (McLennan et al., 1993) expected for
sedimentsderived from the upper crust (Fig. 9).The geochemical
variations between elements such as Th and La(indicative of a
felsic source) and Sc (indicative of a mac
source)havebeenusedtodistinguishbetweenfelsicandmacprove-nances by
various authors (e.g., McLennan et al., 1980). Th/Sc
ratiosareuseful indicators of sourcerocks processes
andareunaffectedbysedimentary processes (Taylor and McLennan,
1985). The Th versusSc plot (Fig. 10), adopted from McLennan et al.
(1993), reveals twodominant sourceareas, acontinental
sourcewithTh/Sc ratios near 1for siltstones and an almost 5050 mix
of continental and interme-diate component for both brown and green
shales (Fig. 10; see silt-Fig. 10. Thversus Sc diagramindicating
felsic andmac provenance for the Ikorongomudrocks.
Notethesiltstonesamples demarcatedbydottedline. Symbols as inFig.
3.stone samples demarcated by dotted line). This observation
furthersuggests that the siltstones were formed by more felsic
detritus.OnaLaThScternarydiagram(Fig. 11), whichis usedtoprecisely
discriminate felsic and mac provenance of clastic
sed-imentaryrocks(e.g., TaylorandMcLennan, 1985),
theshaleandsiltstone samples are almost indistinguishable and
cluster in
theeldformixedsourcesclosetoPAAS-andPS-likeprovenance.Morethanftysamplesgenerallyfall
betweentheTaylor andMcLennans (1985) approximation of the upper
continental crust(Fig. 11). Therefore, traceelement
datafortheIkorongoGroupmudrocks (Table 1; Fig. 11) strongly
demonstrate an upper crustprotolith, a feature also revealed in the
Th/U plot (Fig. 9).Palaeocurrent directions, alongside geochemical
data, can
alsohelptohighlightthepossibleprovenanceoftheIkorongorocks.Twosets
of palaeocurrent directions
weredocumentedinthebasin:north-north-west/south-south-east and
north/south. According tothegeologicalsettingoftheIkorongobasin,
possiblecandidatesfor provenance include older felsic and mac
igneous successionswhich are a major assemblage in the neighboring
Nyanzian Super-group of the Archaean Tanzania Craton. The
compositions of rocksin the terranes bordering the Ikorongo basin
have been well con-strained in Messo (2004), Manya (2005), Manya et
al. (2007a,b),Mtoro (2007) and Elisaimon (unpublished data).
Lithological unitsstudiedinthe area, whichare more likelytohave
fedsediments intoFig. 11. Ternary plot of LaThSc concentrations
(after Taylor and McLennan, 1985)for the Ikorongo mudrocks. Symbols
as in Fig. 3. UC=upper crust (data from Taylorand McLennan, 1985).
Symbols as in Fig. 3.C. Kasanzu et al. / Precambrian Research 164
(2008) 201213 209Fig. 12. Major greenstone domains of the Nyanzian
Supergroup showing the setting of possible provenance regions
(modied from Borg and Shackleton, 1997).the Ikorongo basin, include
basalts, dacites, rhyolites, rhyodacites,granites, TTGs,
basaltic-andesites and basaltic-trachyandesites.Therefore, in the
light of the two sets of palaeocurrent directions,possible
candidate sources include the felsic and mac volcanics ofTarime
and, possibly Suguti and Ikoma (Fig. 12). Both Tarime
andSugutiformpartoftheMusoma-MaraGreenstoneBelt(MMGB)whereas Ikoma
belongs to the Kilimafedha Greenstone Belts (KGB)of the Nyanzian
Supergroup (see Fig. 12 for locations).Intheir comprehensivestudies
onsedimentaryrocks, Taylor andMcLennan (1985) indicated the
signicance of relying on the ele-ments that are least mobile under
the expected range of geologicconditions in provenance
determination. This observation is basedonthe fact that during
weathering, the alkali and alkaline earthele-ments are quite
soluble whereas elements like Al, Zr, Hf, Sc, Y, Nb,Th and REE are
relatively immobile (Taylor and McLennan, 1985;NesbittandYoung,
1982). ThereforetheabundancesofREE, Th,and the transition trace
elements, especially Sc, and their respec-tive ratios are the best
proxies for provenance studies (Taylor andMcLennan, 1985).
Table2compares immobiletraceelements ratios,La/Sc, Co/Th, Cr/Th,
andTh/Sc for the Ikorongomudrocks withthoseofpossiblesourcerocks,
PAAS, PSandwell-establishedratiosofsands derived from mac and
felsic rocks. Although variations areevident,
La/ScandSc/ThratiosfortheIkorongorockspointtoafelsic
dominated-detritus (Table 2). When compared to PAAS andPS, La/Sc,
Sc/Th, Cr/Th and Co/Th ratios in the Ikorongo mudrocksindicate
derivation from a more felsic source than the PAAS and
PSsources.Thesiltstones showrelativelyhigher proportionof felsic
detritusthan shales do (see Table 2), an observation which is also
supportedby the multielement variation diagram in Fig. 4, which
shows thedepletion of compatible elements such as Cr, Ni, Sc and V
which areusually regarded as mac components (Asiedu et al.,
2000).The La/Sc and Co/Th ratios of the mudrocks are similar to
thoseof TTGs from the Tarime segment of the MMGB to the north of
theIkorongo basin suggesting that the TTG supplied bulk of the
felsiccomponent to the Ikorongo basin. In particular, the Sc/Th
ratios ofthe siltstones are very similar to those of the TTGs
suggesting thatthe siltstones were formed by detritus predominantly
derived fromthe weathering of the TTGs.
AcontributionfromMMGBrhyodacitesis also indicated by the close
similarity between the Cr/Th ratios ofthetwolithologies(siltstones
5.94;rhyodacites 5.54). Ontheother hand, dacites from the MMGB have
Cr/Th ratios which arecomparable to those of the shales suggesting
that the dacites mayhave also been source rocks for the shales.In
addition, the abundances of the REE have been used to
infersourcesofsedimentaryrocks(McLennanetal., 1993;Asieduetal.,
2000). For instance, mac rocks contain low LREE/HREE ratiosand no
Eu anomalies, whereas felsic rocks usually contain
higherLREE/HREEratios andnegativeEuanomalies (Taylor
andMcLennan,1985; Roddaz et al., 2006). The Eu anomaly in
sedimentary rocksis commonly regarded as inherited from the source
rocks. There-fore, the REE patterns obtained in sedimentary rocks
can help tomake inference on the nature of protolith (Taylor and
McLennan,1985). In this regard, further constraints on possible
source rocksfor the Ikorongo mudrocks can be made by using the REE
compo-sitions.Chondrite-normalizedabundances andpatterns (Fig. 5)
indicatethat, despite considerable variations in contents, most of
the Iko-210 C. Kasanzu et al. / Precambrian Research 164 (2008)
201213Table 2Range of some elemental ratios for the Ikorongo
mudstones in comparison with possible source rocks from the
Musoma-Mara Greenstone Belt and Kilimafedha GreenstoneBeltLithology
La/Sc Sc/Th Cr/Th Co/Th SourceRange Mean Range Mean Range Mean
Range MeanRhyolites 0.190.31 0.28 Manya (2005)Dacites 2.3321.78
8.63 Manya (2005)Gabbro 18.34665.40 140.99 Manya (2005)K-Granites
0.060.95 48.47 Manya (2005)Na-Granitoids 0.235.68 2.01 Manya
(2005)Rhyodacites 0.4029.43 5.54 Manya (2005)High-magnesium
basaltic-andesite 10.55166.52 48.47 Manya
(2005)Tonalite-Trondjemite-Granodiorites 2.9127.5 10.03 0.0412.05
0.67 01.80 0.12 0.202.94 0.88 Elisaimon (unpublished data)Biotite
granites 7.3128 23.68 0.030.32 0.11 0.020.17 0.08 Elisaimon
(unpublished data)Calcic granites 6.1230.68 13.68 0.070.54 0.23
0.040.72 0.24 Elisaimon (unpublished data)Basaltic-andesites
0.441.64 1.26 2.286.93 3.48 5.3091.61 19.4 5.4111.48 6.91 Messo
(2004)Basaltic-tranchyandesites 0.641.46 1.15 2.425.27 3.23 4.715
7.9 5.129 6.2 Messo (2004)Basalts 61.221850 608.5 110.2265 160.5
Mtoro (2007)Sands from felsic rocks 2.516 0.051.2 0.57.7 0.221.5
Taylor and McLennan (1985)Sands from basic rocks 0.441.1 2025 22100
7.18.3 Taylor and McLennan (1985)Post-Archaean Australian Shale
(PAAS) 2.81 1.27 7.94 1.59 Taylor and McLennan (1985)Average
Proterozoic Shale (PS) 2.23 1.18 8.04 1.25 Condie (1993)Brown
shales 2.025.20 3.25 0.741.58 1.08 5.0517.68 8.83 0.282.64 1.11
This studyGreen shales 2.097.06 2.92 0.751.22 1.01 6.4314.19 9.05
0.421.79 1.08 This studySiltstones 2.1013.84 4.39 0.091.11 0.66
1.3918.24 5.94 02.78 1.37 This studyAlso included are ratios for
PAAS, PS and sands derived from mac and felsic protoliths.rongo
mudrock samples studied have striking similarities in theirREE
patterns. All samples are characterized by an enrichment ofthe
LREE, negative Eu anomalies and relatively at HREE patterns.These
features, particularlythe negative Euanomalies suggest a
dif-ferentiated protolith, similar to granite (e.g., McLennan et
al., 1993;Asiedu et al., 2000). REE patterns for samples from the
IkorongoGroupshowcloserresemblancetothoseoftheTarimevolcanicand
plutonic rocks and to a lesser extent with rhyolites from theSuguti
(Fig. 13). On the other hand, magmatic rocks from Ikoma tothe
southeast of Ikorongo do not seemto have fed sediments to
theIkorongo basin since their REE patterns do not match with those
ofthe Ikorongo mudrocks (not shown).Based on the REE compositions
of the source candidates, mix-ing calculations of Albarede (2002)
were carried out to estimatethe relative contribution of source
materials required to generatethe Ikorongo mudrocks. Briey, for a
system o, containing severalelements (i =1, . . ., m) hosted in
phases (j =1, . . ., n), let Mj be themass of phase j and mijthe
mass of element (or species) i hostedin the phase j. Then, the
composition of species (or element) i inphase j can be
mathematically dened asCij =mijMjForthebulkmaterial,
massconservationrequiresthatM0 =
nj=1Mj.Therefore, for a given element, i, the proportion of fj
of the phasej is such that: fj=Mj/M0 and Ci0 = mi0/M0 =
nj=1mij/M0 (all equa-tions adopted from Albarede, 2002).The
high-magnesium-basaltic-andesites (HMBA), granitoidsand dacites
comprise major part of the exposed crust in the MMGB.Based on their
aerial distribution and geochemical afnity to theTable
3Representative REE compositions of possible source rocks located
in the MMGB and the Ikorongo mudrocksHMBA (N=13) Dacites (N=27)
Granitoids (N=33) Average shale Average siltstoneAverage STDEV
Average STDEV Average STDEV Average STDEV Average STDEVLa 28.79
7.29 48.10 19.22 52.34 21.38 57.42 57.74 36.58 13.68Ce 60.75 14.88
96.93 38.01 100.96 41.76 100.19 26.06 69.02 29.77Pr 7.22 1.73 10.94
4.14 11.04 4.54 11.97 23.61 7.59 3.15Nd 29.00 6.85 41.39 15.69
39.16 15.52 39.48 14.51 26.71 10.65Sm 5.31 1.12 6.53 2.43 6.12 2.41
7.22 7.26 5.59 1.96Eu 1.44 0.25 1.58 0.58 1.11 0.36 1.46 3.45 1.24
0.32Gd 4.55 0.75 5.17 1.71 4.78 2.01 5.45 3.77 5.18 1.40Tb 0.65
0.08 0.60 0.20 0.63 0.28 0.81 2.56 0.76 0.22Dy 3.61 0.33 2.93 0.99
3.28 1.47 4.35 2.02 4.03 1.20Ho 0.73 0.07 0.54 0.18 0.64 0.30 0.80
1.59 0.75 0.22Er 1.87 0.19 1.33 0.46 1.67 0.76 2.36 1.57 2.26
0.76Tm 0.29 0.03 0.20 0.07 0.26 0.12 0.35 1.45 0.34 0.13Yb 1.93
0.19 1.33 0.47 1.83 0.75 2.28 1.42 2.19 0.91Lu 0.28 0.03 0.19 0.07
0.27 0.11 0.33 1.34 0.34 0.16Eu/Eu* 0.90 0.05 0.83 0.04 0.64 0.02
0.71 0.03 0.71 0.05(La/Yb)CN10.15 2.73 24.31 3.80 20.38 5.60 15.20
2.06 11.14 2.7(La/Sm)CN3.41 0.25 4.63 0.19 5.18 0.10 5.02 0.42 4.11
0.86(Gd/Yb)CN1.91 0.13 3.13 0.05 2.07 0.02 1.95 0.02 1.91 0.45C.
Kasanzu et al. / Precambrian Research 164 (2008) 201213 211Fig. 13.
Comparison of the Chondrite-normalized REE patterns of the Ikorongo
mudrocks with those of possible source rocks from MMGB. Other
symbols as in Fig. 3.Ikorongo mudrocks (Figs. 12 and 13), 14 HMBA,
33 granitoids and27 dacite samples were used to model the detritus
that fed sedi-ments into the Ikorongo basin. The three rock types
were
treatedasdistinctcomponentsbecausetheyaremajorrocktypeswithdistinctive
geochemistry from which the Ikorongo mudrocks arethought to be
derived from.For this case, REE concentrations of shale samples
were aver-aged and the resulting composite was assigned average
shale. TheTable 4Results of mixing and mass balance calculations
and comparison between original and calculated parametersHMBA DCT
GRDProportions (%)Model shale 42 11 47Model siltstone 43 4 53Model
shale Average shale %Variation Model siltstone Average siltstone
%VariationLa 41.98 57.42 27 42.04 36.58 15Ce 83.63 100.19 17 83.51
69.02 21Pr 9.43 11.97 21 9.40 7.59 24Nd 35.14 39.48 11 34.88 26.71
31Sm 5.82 7.22 19 5.78 5.59 3Eu 1.30 1.46 11 1.27 1.24 3Gd 4.73
5.45 13 4.70 5.18 9Tb 0.64 0.81 21 0.64 0.76 16Dy 3.38 4.35 22 3.41
4.03 16Ho 0.67 0.80 17 0.67 0.75 10Er 1.72 2.36 27 1.74 2.26 23Tm
0.27 0.35 24 0.27 0.34 20Yb 1.81 2.28 20 1.85 2.19 16Lu 0.26 0.33
21 0.27 0.34 20(Eu/Eu*) 0.77 0.71 9 0.76 0.71 7(La/Yb)CN16.52 15.20
9 16.14 11.14 45(La/Sm)CN4.38 5.02 13 4.40 4.11 7(Gd/Yb)CN2.12 1.95
9 2.04 1.91 7DCT means dacites. GRD means granitoids.212 C. Kasanzu
et al. / Precambrian Research 164 (2008) 201213Fig. 14. Comparison
of Chondrite-normalized REE patterns between (a) average shale and
model shale, and (b) average siltstone and model
siltstone.sameprocedurewasperformedforsiltstonesamples(Table3).Conversely,
for the case of siltstones, three samples (IK 19, 79 and80) were
eliminated in the calculations since they are signicantlyenriched
in SiO2 (81.4692wt%). The HREE contents of these
sam-plesseemtobehighlydilutedbyquartzleadingtoinconsistentLREE/HREE
ratios.From the average data shown in Table 3, modeling for the
aver-age shale protolith was done using the ratios (La/Yb)CN,
(Gd/Yb)CNas well as theEu/Eu*, andthemixingcalculations was set
inamatrixform asEuEuLaYbGdYb=a b c0.9 0.83 0.6410.2 24.31 20.41.91
3.13 2.07abc=0.7115.21.95shaleameansHMBA; b=dacites; c =granitoids.
TheratiosLa/YbandGd/Yb are Chondrite-normalized.The results of the
mixing calculations (Table 4) showthat, gran-itoids and the HMBA
from the Tarime region were the main sourcerocks (granitoids 47%:
HMBA 42%), while the dacites suppliedlesser amounts of detritus
(11%). This observation, therefore, con-forms to an almost 1:1
mixture between felsic and mac protolithsimilar totheupper crust
composition(Taylor andMcLennan,1985). Similar
resultswerealsoobtainedfor siltstonesamplesexcept that granitoids
show a relatively higher proportion (53%)whereas the HMBA supplied
about 43% and dacites 4%.Fromthe calculations, optimal tting of
average REE concentra-tions of the Ikorongo Group with those of
source candidates wereachieved by mass balance
calculations.Toobtainthecalculatedvalues,
thefollowingequationfromAlbarede (2002) was used:WRmix=
C1+C2+C3WRmixrefers to the calculated whole rock compositions while
, andstand for the proportions of the HMBA, dacites, and
gran-itoids, respectively, obtained in mixing calculations. C1, C2
and C3stand for the respective species (elements) in the HMBA,
dacites,and granitoids used in the mixing calculations. The results
of thecalculations, whichwerebasedontheREEparameters, arepre-sented
in Table 4 and Fig. 14 for comparison.6. ConclusionsSource rock
weathering and provenance of the Ikorongo Grouphave
beenassessedusing geochemical studies. Major element com-positions
suggest that the Ikorongo mudrocks were derived frommoderately
weathered protoliths. Th/U ratios coupled with Th
ver-susScandLaThScplotssuggestanuppercrustalprotolithforthe
Ikorongo mudrocks similar to the PAAS and PS protolith.
ThefractionatedREEpatternsandthenegativeEu/Eu*anomaliesofthe
Ikorongo mudrocks further attest to an upper crust
provenancetypical of a craton interior. Based on palaeocurrent
measurements,thesourcerocksfortheIkorongoGroupliestothenorthofthebasin
suggesting that the MMGB, which comprises of older felsicand mac
igneous rocks, is a possible source terrane. The REE pat-terns and
elemental ratios such as La/Sc, Sc/Th, Cr/Th and Co/Thof the
studied mudrocks reveal that the source rocks include mag-matic
rocks from the Tarime and Suguti segments of the
MMGB.Basedonmixingandmassbalancecalculations,
theshaledetri-tuscanbemodeledbyamixtureof47%granitoids, 42%HMBAand
11% dacites. The siltstones, on the other hand, were derivedfrom a
rather more felsic protolith which corresponds to
mixingof53%granitoids,
43%HMBAand4%dacites(allproportionsbywt%).AcknowledgementsThis
research was nancially supported by Sida/SAREC
throughtheprojectGeologyandmineralizationofArchaeangreenstonebelts
in the Lake Victoria Goldelds of the Faculty of Science,
Uni-versityofDaresSalaam. TheReviewersRoserKorschandHughRollinson
are acknowledged for their constructive comments thatimproved the
output of the paper.ReferencesAlbarede, F., 2002.
IntroductiontoGeochemicalModeling. CambridgeUniversityPress, 543
pp.ASGA/UNESCO, 1968. Mineral Survey of the Lake Victoria Goldelds.
Final Report.New York Press, 58 pp.Asiedu, D.K., Suzuki, S.,
Nogami, K., Shibata, T., 2000. Geochemistry of lower creta-ceous
sediments, inner zone of southwest Japan: constraints on provenance
andtectonic environment. J. Geochem. 34, 155173.Bock, B., McLennan,
S.M., Hanson, G.N., 1998. Geochemistry and provenance of theMiddle
Ordovician Austin Glen Member (Normanskill Formation) and the
Taco-nian Orogeny in New England. J. Sediment. 45, 635655.Borg, G.,
Shackleton, R.M., 1997. The Tanzania and NE Zaire Cratons. In: de
Wit, M.J.,Ashwal, L.D. (Eds.), Greenstone Belts. Clarendon Press,
Oxford, pp. 608619.C. Kasanzu et al. / Precambrian Research 164
(2008) 201213 213Boynton, W.V., 1984. Geochemistry of the rare
earth elements: meteorite studies.In: Henderson, P. (Ed.), Rare
Earth Element Geochemistry. Elsevier, Amsterdam,pp. 63114.Button,
A., Tyler, N., 1979. Precambrian palaeoweathering and erosion
surfaces insouthernAfrica: reviewof their character andeconomic
signicance. Econ. Geol.Res. Unit, Univ. Witwatersrand,
Johannesburg, Info. Circ. 135, p. 37.Cloutier, J., Ross, K.S.,
Bardoux, M., 2005. Nd isotopic, petrogenetic and
geochemicalinvestigation of the Tulawaka East gold deposit,
Tanzania Craton. PrecambrianRes. 139, 147163.Condie, K.C., Noll
Jr., P.D., Conway, C.M., 1992. Geochemical and detrital mode
evi-dence for two sources of Early Proterozoic sedimentary rocks
from the TontoBasin Supergroup, central Arizona. Sediment. Geol.
77, 5176.Condie, K.C., 1993. Chemical composition and evolution of
the upper
continentalcrust:contrastingresultsfromsurfacesamplesandshales.
Chem. Geol. 104,137.Elisaimon, E., unpublished. A GIS-aided
aero-geophysical, geological and geochemi-cal investigations of the
late ArchaeanGranites of the Musoma-Mara GreenstoneBelt,
north-eastern Tanzania. M.Sc. Thesis. University of Dar es
Salaam.Fedo, C.M., Nesbitt, H.W., Young, G., 1995.
Unravelingtheeffectsof potassiummetasomatisminsedimentaryrocks
andpalaeosols, withimplications for palae-oweathering conditions
and provenance. J. Geol. 23, 921924.Herron, M.M., 1988.
Geochemicalclassicationofterrigeneoussandsandshalesfrom core or log
data. J. Sediment. Petrol. 58, 820829.Kasanzu, C., 2007. Geology
and geochemical investigations of clastic sedimentaryrocks of the
Neoproterozoic Ikorongo Group, NE Tanzania. M.Sc. Thesis.
Univer-sity of Dar es Salaam, 106 pp.Kimura, S., Shikazono, N.,
Kahiwagi, H., Nohara, M., 2003. MiddleMiocene-earlyPliocene
paleo-oceanic environment of Japan Sea deduced from
geochemicalfeatures of sedimentary rocks. J. Sediment. Res. 164,
105129.Maboko, M.A.H., 1995. Neodymiumisotopic constraints
ontheprotolithages of rocksinvolved in Pan-African tectonism in the
Mozambique Belt of Tanzania. J. Geol.Soc. London 152,
911916.Macfarlane,A.,1965.TheGeologyofQuarterDegreeSheet25.MineralResourceDivision,
Dodoma.Manya, S., 2005.
GeochemistryandGeochronologyofthevolcanicandplutonicrocks of the
Tarime segment of the Musoma-Mara Greenstone Belt,
NortheastTanzania. Ph.D. Thesis. University of Dar es Salaam, 159
pp.Manya, S., Kobayashi, K., Maboko, M.A.H., Nakamura, E., 2006.
Ion microprobe zirconUPb dating of the late Archaean metavolcanics
and associated granites of theMusoma-Mara Greenstone Belt,
North-eastern Tanzania: implications for thegeological evolution of
the Tanzania Craton. J. Afr. Earth Sci. 45, 355366.Manya, S.,
Makenya, A.H., Nakamura, E., 2007a. Geochemistry and Nd-isotopic
com-position of potassic magmatismin the Neoarchaean Musoma-Mara
GreenstoneBelt, northern Tanzania. Precambrian Res. 159,
231240.Manya, S., Makenya, A.H., Nakamura, E., 2007b.
Thegeochemistryof high-Mgandesite and associated adakitic rocks in
the Musoma-Mara Greenstone Belt,northern Tanzania: possible
evidence for Neoarchaean ridge subduction? Pre-cambrian Res. 159,
241259.McLennan, S.M., 1989. Rareearthelementsinsedimentaryrocks:
inuenceofprovenance and sedimentary processes. In: Lipin, B.R.,
MacKay, G.A. (Eds.), Geo-chemistry and Mineralogy of Rare Earth
Elements. Miner. Soc. Am., pp. 169200.McLennan, S.M., Nance, W.B.,
Taylor, S.R., 1980. Rare earth elementthorium
cor-relationinsedimentaryrocks,
andthecompositionofthecontinentalcrust.Geochim. Cosmochim. Acta 44,
18331839.McLennan, S.M., Taylor, S.R., Ericksson, K.A., 1983.
Geochemistry of Archaean shalesfrom the Pilbara Supergroup, Western
Australia. Geochim. Cosmochim. Acta 47,12111222.McLennan, S.M.,
Hemming, S., McDaniel, D.K., Hanson, G.N., 1993.
Geochemicalapproaches to sedimentation, provenance and tectonics.
Geol. Soc. Am. SpecialPaper 284, pp. 2140.Messo, W.C., 2004.
Geochemistry of Neo-Archaean volcanic rocks of the Ikoma areainthe
Kilimafedha Greenstone belt, NWTanzania. M.Sc. Thesis. University
of Dares Salaam, 101 pp.Mtoro, M., 2007. Geochemical investigation
of metavolcanic rocks of the Musoma-Mara Greenstone Belt in the
Suguti area, Tanzania. M.Sc. Thesis. University ofDar es Salaam,
105 pp.Muhongo, S., Kroner, A., Nemchin, A.A., 2001. Zirconages
fromgranulites facies rocksin the Mozambique belt of Tanzania and
implications for Gondwana assembly.J. Geol. 109, 171190.Nesbitt,
H.W.M., Young, G.M., 1982. Early Proterozoic climates and plate
motionsinferred from major element chemistry of lutites. Nature
299, 715717.Nesbitt, H.W., Young, G.M., 1984. Predictionof
someweatheringtrendsof plu-tonic and volcanic rocks based on
thermodynamic and kinetic considerations.Geochim. Cosmochim. Acta
48, 15231534.Nesbitt, H.W., Young, G.M., 1989. Formation and
diagenesis of weathering proles.J. Geol. 97, 129147.Paul, D.,
White, W.M., Turcotte, D.L., 2003.
Constraintsonthe232Th/238Uratio() of the continental crust.
Geochem. Geophys. Geosyst. 4 (12),
1102,doi:10.1029/2002GC000497.Pickering, R., Stockley, G.M.,
Macfarlane, A., 1959. Revised Stratigraphy of the Iko-rongo Group.
Mineral Resource Division, Dodoma, 56 pp.Piper, D.Z., 1974.
Seawater as the source of minor elements in black shales,
phos-phorites and other sedimentary rocks. Chem. Geol. 114,
95114.Roddaz, M., Viers, J., Brusset, S., Baby, P., Boucayrand, C.,
H erail, G., 2006. Controlson weathering and provenance in the
Amazonian foreland basin: insights frommajor andtraceelement
geochemistryof NeogeneAmazoniansediments. Chem.Geol. 226,
3165.Rollinson, H.R., 1993. Using Geochemical Data: Evaluation,
Presentation, Interpre-tation. Longman Group, UK, 1315.Servaraj,
K., Arthur, C.C., 2006. ModerateChemical weatheringof subtropical
Taiwan:constraints fromsolid-phase geochemistry of sediments
andsedimentary rocks.J. Geol. 114, 101116.Shackleton, R.M., 1986.
Precambrian collision tectonics in Africa. In: Coward, M.P.,Ries,
A.C. (Eds.), Collision Tectonics. Geological Society, Lond., Spec.
Pub. 19, pp.324349.Stockley, G.M., 1936. Geology of the
south-eastern regions of the Musoma District,Geol. Surv.
Tanganyika, Dar es Salaam, Short Paper 13, pp. 2023.Taylor, S.R.,
McLennan, S.M., 1985. The Continental Crust: Its Compositionand
Evolution. Blackwell Scientic Publications, Blackwell, Oxford,
UK,312 pp.Taylor, S.R., Rudnick, R.L., McLennan, S.M., Ericksson,
K.A., 1986. Rare earth elementpatterns in Archaean high-grade
metasediments and their tectonic signicance.Geochim. Cosmochim.
Acta 50, 22672279.Weaver, C.E., 1989. Clays, Muds, and Shales.
Elsevier, Amsterdam, 819 pp.Yong, I.L., 2002. Provenance derived
from the geochemistry of late Palaeozoic-earlyMesozoic mudrocks of
the Pyeongan Supergroup, Korea. Sediment. Geol. 149,219235.