geochemistry of paleoproterozoic metasedimentary rocks - terrapub

215

Geochemical Journal, Vol. 38, pp. 215 to 228, 2004

*Corresponding author (e-mail: [email protected])

Copyright © 2004 by The Geochemical Society of Japan.

Geochemistry of Paleoproterozoic metasedimentary rocks fromthe Birim diamondiferous field, southern Ghana:

Implications for provenance and crustal evolution atthe Archean-Proterozoic boundary

D. K. ASIEDU,1* S. B. DAMPARE,2 P. ASAMOAH SAKYI,3 B. BANOENG-YAKUBO,1 S. OSAE,2

B. J. B. NYARKO2 and J. MANU1

1Department of Geology, University of Ghana, P.O. Box LG 58, Legon, Ghana2National Nuclear Research Institute, Ghana Atomic Energy Commission, P.O. Box LG 80, Legon, Ghana

3Institute of Environment and Resources, Technical University of Denmark, DK 2800 Lyngby, Denmark

(Received February 17, 2003; Accepted October 23, 2003)

Metagraywackes and metapelites from the Paleoproterozoic Birimian Supergroup in the Birim diamondiferous field,southern Ghana, were analyzed for their major and trace element contents. Compared to early Proterozoic crust, themetasedimentary rocks are enriched in ferromagnesian elements but depleted in rare earth elements (REE), high fieldstrength elements (HFSE) (with exception of Zr), and Th. They show REE patterns similar to their Archean counterparts.The chemical data indicate that the sediments were derived from a local source of mixed felsic—mafic composition, withthe latter dominating. The source rocks were the basaltic to dacitic volcanic rocks and granitoids within the Birimiangreenstone belts. The chemical data further suggest their deposition in a tectonic setting comparable to modern islandarcs, and that minimal old upper crust (i.e., pre-Birimian sources) was involved in their formation. The analyzedmetasedimentary rocks have Eu-anomalies and GdN/YbN, Sm/Nd, Th/Sc, Cr/Sc and Cr/Th ratios that closely resemblethose of their Archean counterparts, and therefore inconsistent with models suggesting abrupt compositional changes inupper crust at the Archean-Proterozoic boundary.

Keywords: Ghana, Birimian, geochemistry, Archean-Proterozoic boundary, provenance

various generations of granitoids during the Eburneanevent at 2.1 Ga (Leube et al., 1990; Hirdes et al., 1992).There is controversy regarding the lithostratigraphicsuccessions of the main units (i.e., the volcano-sedimen-tary and bimodal volcanics). Some workers (e.g., Junner,1940; Milési et al., 1992) have proposed that the volcano-sedimentary unit is older whereas others (e.g., Tagini,1971; Hottin and Quedraogo, 1975) hold a contrary opin-ion, that the volcanic unit is older. However, Leube et al.(1990) have proposed that the two units formed quasi-contemporaneously as lateral facies equivalent.

Various alternative models have been proposed withrespect to the tectonic setting of the Birimian rocks inWest Africa: intraplate oceanic plateaus (Abouchami etal., 1990); from intacratonic-rift to oceanic-spreading andfinally to an accretion-related setting (Leube et al., 1990);immature island arcs built on oceanic crust (Sylvester andAttoh, 1992); and back-arc basin (e.g., Vidal and Alric,1994). Most geochemical studies on the Birimian haveconcentrated on the metavolcanic rocks within thegreenstone belts (e.g., Abouchami et al., 1990; Leube etal., 1990; Sylvester and Attoh, 1992; Béziat et al., 2000)

INTRODUCTION

The Paleoproterozoic Birimian Supergroup (Junner,1940) forms the northern and eastern portions of the Leo-Man Shield, which occupies the southern segment of theWest African Craton (Fig. 1A). It consists of two majorunits: one unit is composed dominantly of volcano—de-trital rocks and the other mostly represented by bimodalalthough largely tholeiitic volcanics (Abouchami et al.,1990). Geochronological studies on the Birimian rockssuggest they were formed during the time interval ~2.3to 2.0 Ga (Abouchami et al., 1990; Liégeois et al., 1991;Boher et al., 1992; Taylor et al., 1992). The rocks weredeposited on deformed and metamorphosed basementrocks of no more than 50 m.y. older (Abouchami et al.,1990), as sediments and volcanics in an extensivegeosyncline. They were later folded, metamorphosed un-der mostly greenschist-facies conditions and invaded by

216 D. K. Asiedu et al.

and to a lesser degree, the granitoids (e.g., Leube et al.,1990; Doumbia et al., 1998; Loh and Hirdes, 1999). Incomparison to the volcanic rocks, the Birimian sedimen-tary rocks have received very li t t le attention ingeochemical studies even though sedimentary rocks con-tain a wealth of information about provenance and crustalevolution (e.g., McLennan et al., 1990, 1995).

It is now well established that part of a record of geo-logic history is retained in detrital sediments. Petrographicexamination has traditionally been an important methodin extracting this information (e.g., Dickinson and Suczek,1979; Dickinson et al., 1983), but has generally not provedvery useful in the examination of mudstones and meta-morphosed sediments. Geochemical examination, particu-larly REE has no such restrictions however, and can be

used effectively for all types of clastic sediments to evalu-ate the nature and evolution of the provenance and of sedi-mentary history (McLennan et al . , 1995). Thegeochemical approach is, therefore, more appropriate tothis study of metamorphosed sedimentary rocks.

In this paper, we examine the geochemistry of theBirimian metasedimentary rocks (i.e., metagraywackesand metapelites) in the Birim diamondiferous field, situ-ated about 110 km northwest of Accra, Ghana (Figs. 1and 2). The major purpose of this study is to constrainthe sources of the sediments and the tectonic setting ofthe sedimentary basin. The implications of these data forthe crustal evolution at the Archean-Proterozoic bound-ary will also be discussed. The Birimian rocks have un-dergone up to amphibolite-facies metamorphism and otheralterations that may significantly affect the mobility ofsome elements such as Si, Na, K, Ca, Mg, Rb, and Sr(Abouchami et al., 1990). Accordingly, we put muchemphasis on relatively immobile elements such as theREE, HFSE, Cr, Co, Th and Sc (Taylor and McLennan,1985; Bhatia and Crook, 1986). These elements are gen-erally thought to be transferred nearly quantitatively andunfractionated into the sedimentary record during sedi-mentary processes (Taylor and McLennan, 1985;McDaniel et al., 1994), and thus may reflect the signa-ture of the parent material (Taylor and McLennan, 1985;McLennan, 1989). However, in some cases a possiblefractionation of REE during weathering and diagenesishas also been observed (e.g., Nesbitt, 1979; Anwiller andMack, 1991).

GEOLOGIC SETTING

In Ghana, the Birimian is characterized by sedimen-tary basins, which separate a series of sub-parallel,roughly equally spaced, north-easterly-trending volcanicbelts (Fig. 1B). The sedimentary basins are composed ofvolcaniclastics, wackes and argillites and are intruded byaluminous granitic plutons and metamorphosed toamphibolite facies (Leube et al., 1990). The volcanic beltsconsist predominantly of metamorphosed tholeiitic lavas,minor volcaniclastics and “belt-type” granitoids, andmetamorphosed to greenschist facies (Leube et al., 1990;Sylvester and Attoh, 1992). The basement (i .e. ,“Dabakalian”) over which these supracrustal rocks arelaid is unknown in Ghana, though fragments of this base-ment have been reported in the Bole-Navrongo belt (Kutu,per. comm.).

The study area lies in the Birim diamondiferous field(Fig. 2), which is located within the Cape Coast Basin(Fig. 1B). Junner (1943) has extensively described thegeology of the diamondiferous field. The BirimianSupergroup in the study area is mainly composed ofmetasedimentary rocks comprising tuffaceous

Fig. 1. (A) Simplified geological map of West Africa (modifiedafter Abouchami et al., 1990). (B) The geology of the south-western part of Ghana (modified from Leube et al., 1990).

Metasedimentary rocks from southern Ghana 217

metagraywackes with subordinate quartzites and interbed-ded gray and black phyllites and schists. Mafic toultramafic lavas and sub-volcanic rocks are common. Inthe southern portion of the study area, themetasedimentary rocks are intruded by basin-typegranitoids. Other post-Birimian intrusive rocks includegranite, aplite, porphyry and pegmatite. The Birimianrocks in the study area generally exhibit up to greenschist-facies regional metamorphism (Junner, 1943). Both themetasedimentary rocks and the ultramafic intrusives closeto granite batholith display various degrees of contactmetamorphism (Junner, 1943). The rocks have been af-fected by low-grade metasomatic alterations, involvingsilicification and widespread formation of secondarychlorite and sericite. Overlying the Birimian rocks arePliocene to Recent alluvial deposits that host most of thediamonds recovered from the diamondiferous field.

The Akwatia metagraywackes are characterized bypoor sorting and angular to sub-rounded grains, with somegrains showing elongation along the direction of folia-

tion. They consist of about 60 vol.% sand- to silt-sizedclasts made up of quartz, plagioclase, orthoclase, biotite,calcite, muscovite, sericite and chlorite dispersed in a fine-grained matrix of quartz, clay minerals and opaque min-erals (~40 vol.%). They also contain accessory mineralssuch as pyrite, staurolite, garnet, zircon and iron oxideminerals. The metapelites in general have sericite andchlorite mineralogy with angular to subangular grains ofquartz, feldspar and calcite embedded in the matrix. Thinsections of 50 metasedimentary rocks were examinedpetrographically. Twenty-four least altered samples, com-prising 19 metagraywackes and 5 metapelites were care-fully selected for major and trace element analysis byInstrumental Neutron Activation Analysis (INAA) at theGhana Research Reactor-1 (GHARR-1) facility of theNational Nuclear Research Institute, Ghana Atomic En-ergy Commission, following the method of Koeberl(1993). The estimated mineralogical compositions and theexact locations of the analyzed samples are given in Ap-pendix A.

Fig. 2. Geological map of the study area (modified after Junner, 1943).


Table 1. Geochemical data for metasedimentary rocks from the Birim diamondiferous field

ANALYTICAL PROCEDURE

Rock samples were crushed with a steel jaw crusherand then pulverized with an agate mortar. About 100–200mg of each of the powdered samples were sealed intopolyethene, packed into a plastic rabbit capsule or vial

and heat-sealed. Irradiation of the samples was then car-ried out using the GHARR-1 reactor operating at 15 kWwith a corresponding thermal neutron flux of 5.0 × 1011

n cm–2s–1. Irradiation times were chosen to account forshort-, medium- and long-live radionuclides, and thusranged from 30 s to 6 hrs. After cooling times of 10 min,


Eu/Eu* = EuN/(SmN × GdN)0.5; Ce/Ce* = CeN/(LaN × NdN)0.5.nd = not determined; Subscript N denotes chondrite-normalized values. Normalization factors are from Taylor and McLennan (1985).

Table 1. (continued)

3–5 days and 1–2 weeks after irradiation, samples werecounted for 10 min, 1 hr, 3 hrs and 12 to 24 hrs. Measure-ments of spectral intensities were carried out using a PC-based gamma-ray system consisting of a N-type high pu-

rity germanium (HPGe) detector Model GR 2518; a HVpower supply Model 3105 and a spectroscopy amplifierModel 2020; an 8K ACCUSPEC multi-channel analyzer(MCA) emulation card and a 486 microcomputer for spec-


trum data evaluation and analysis. The detector operatedon a bias voltage of (-ve) 3000 V with relative efficiencyof 25% and had a resolution of 8 keV (FWHM) for Co-60gamma-ray energy of 1332 keV. The irradiation condi-tions as well as the accuracy of synthetic standards werechecked with geological standards. The validation of theanalytical procedure was undertaken by irradiating astandard reference material IAEA SOIL 7 and countingunder identical experimental conditions (Appendix B).By using the methods described by Nyarko (1999), theconcentrations of identified elements in the samples weredetermined.

ANALYTICAL RESULTS

The major and trace element compositions of theanalyzed samples are given in Table 1. Comparison ofaverage compositions reveal that there are fewgeochemical differences between the metagraywackes andthe metapelites. However, SiO2 content is relatively lower,and Fe2O3 (total Fe as Fe2O3) and K2O relatively higherin the metapelites, evidently due to a decrease in quartzcontent and an increase in clay fraction in the metapelites.The MgO contents in the metapelite are also significantlyhigher than those of the metagraywackes.

MetagraywackesThe metagraywacke samples generally show wide

variations in their major and trace element compositions.They belong to the sodic and ferromagnesian potassicsandstones in the nomenclature of Blatt et al. (1980). Mostof the metagraywackes display andesitic composition intheir major-element composition. The SiO2/Al2O3 ratios(6.3–1.3) are low, indicative of their immaturity, and K2O/Na2O ratios (1.5–0.5) are variable but mostly less than 1(mean = 0.8). Compared to average early Proterozoicgraywackes (Condie, 1993), the analyzed metagraywackesare enriched in Al2O3, Zr, and the ferromagnesian ele-ments (MgO, Cr, Co, and V) and depleted in total REEabundances, Th, and Ta. However, the HFSE (i.e., Zr, Hf,and Ta) and Th concentrations are significantly higherthan those reported in the volcanic rocks of the Birimiangreenstone belts (Sylvester and Attoh, 1992).

The metagraywackes show typical Archean chondrite-normalized REE distribution patterns (Fig. 3A); light-REE(LREE) enrichment (LaN/SmN = 2.17–4.19), slightlynegative to slightly positive Eu-anomalies (Eu/Eu* =0.64–1.21; average = 0.93) and heavy-REE (HREE) de-pletion (GdN/YbN = 1.87–3.36). HREE-depletion is alsocommon in dacitic to andesitic volcanic rocks of theBirimian greenstone belts (Sylvester and Attoh, 1992).Several of the analyzed samples (particularly AB 7, AKW24B, GCD 25, and AKW 24A) show significant negativeCe-anomalies (Table 1). Negative Ce-anomalies are a

common feature in the tholeiitic basalts from the Birimiangreenstone belts (Abouchami et al., 1990; Leube et al.,1990; Sylvester and Attoh, 1992). Negative Ce-anoma-lies in sedimentary rocks can result either by prolongedsub-aquatic weathering (Küster and Liégeois, 2001) orby post-depositional alterations such as pervasive circu-lation of hydrothermal fluids (Abouchami et al., 1990).However, Ce-anomalies in sediments can also be inher-ited from their source rocks, in that case from magmaticrocks where they are common (Shimizu et al., 1992).

MetapelitesLike the metagraywackes, the metapelites also show

variable major and trace element compositions. One sam-ple, AJ 1, has exceptionally low Al2O3 and high MgOand Cr content (Table 1). Another sample, FOL 21, havelow SiO2 content but high Al2O3 and K2O contents, evi-

Fig. 3. Chondrite-normalized REE diagrams for (A) selectedmetagraywacke samples, and (B) metapeli te samples.Chondrite-normalized values after Taylor and McLennan(1985). Plotted for comparison are averages of early Archeangraywacke, early Proterozoic crust, Archean shale, andProterozoic shale (data from Condie, 1993).


dently due to high clay mineral content. Like themetagraywackes, the metapelites have low but variableSiO2/Al2O3 ratios (7.6–0.8). K2O contents and K2O/Na2Oratios are variable, but generally higher than that ofmetagraywackes. Compared to average Proterozoic shale(Condie, 1993) the metapelites are enriched in MgO,Fe2O3, Sc, V, Cr, and Co, and depleted in REE, Th, Hf,and Ta. The Zr contents, however, are comparable to thoseof average Proterozoic shale.

The chondrite-normalized REE distribution patternsfor the five analyzed samples (Fig. 3B) are also similarto their Archean counterparts (Condie, 1993), and arecharacterized by LREE enrichment (LaN/SmN = 2.14–3.20), slightly negative Eu anomalies (Eu/Eu* = 0.77–0.94; average = 0.88) and flat HREE (GdN/YbN = 1.55–1.78). No negative Ce anomalies are present in theanalyzed metapelites. The generally greater depletions inEu (i.e., lower Eu/Eu*) in the metapelite relative to theassociated metagraywackes may be related to concentra-tion of plagioclase in sands during sorting (McLennan etal., 1990).

DISCUSSION

Influence of heavy mineral accumulation and metamor-phism

A number of heavy minerals (particularly zircon,allanite and monazite) are dominated by trace elements,and thus their accumulation in high concentration maysignificantly influence trace element concentrations insedimentary rocks (McLennan et al., 1993). Of these threeminerals, only zircon is visible in the studied samples.Zircon enrichment in sediments can be reflected by rela-tionships between Th/Sc and Zr/Sc (McLennan et al.,1993). On this diagram the analyzed samples follow thegeneral provenance-dependent compositional variationtrend, and there is no sample falling in the high Zr/Scrange typical of zircon accummulation associated withsediment recycling and sorting (Fig. 4A). In addition, zir-con preferentially incorporates HREE relative to LREE,and its accumulation would lead to HREE enrichment anda decrease in LaN/YbN ratio with increasing Zr content.However, no such direct relationship exist between Zr andLaN/YbN ratios of the analyzed samples (Fig. 4B).Monazite, a REE-enriched heavy mineral, has a very steepchondrite-normalized REE pattern; even small amounts(<0.01%) can result in significant increases in the GdN/YbN ratio (McLennan et al., 1993). The HREE-depletionpattern observed in the metagraywackes could be the re-sult of monazite enrichment though this is highly unlikelygiven that no monazite grains were observed in thin-section.

As mentioned earlier, the Birimian metasedimentaryrocks has undergone greenschist to amphibolite facies

metamorphism which may affect the mobility of majorand trace elements. Great care was taken in selecting theleast altered samples for the analysis. Samples contain-ing visible quartz veins, which may representremobilization, were avoided for the analysis. Severallines of evidence argue against large-scale elementremobilization of some particular trace elements in theanalyzed samples. The analyzed samples, particularly forthe metapelites, show uniform and fairly smooth REEpatterns, which would not be expected duringremobilization (Yang et al., 1998). In addition thecovariance between Th/Sc and Zr/Sc ratios (Fig. 4A) andbetween Th/Sc and Cr/Th ratios (Fig. 5) argue against

Fig. 4. (A) Th/Sc versus Zr/Sc graph for the metasedimentaryrocks. Trend lines are defined by McLennan et al. (1993).(B) LaN/YbN versus Zr graph for the metasedimentary rocks.


large-scale remobilization, at least for these elements, inthe analyzed samples. Although some elements, particu-larly the large-ion lithophile elements, were probablyremobilized, large-scale remobilization of the REEs, Th,Zr, Sc, Cr, and Co is unlikely, therefore making these el-ements useful in provenance and tectonic setting discrimi-nation.

ProvenanceThe analyzed metasedimentary rocks have concentra-

tions of ferromagnesian elements (i.e., Mg, Cr, V, Co)several times higher than that of average early Proterozoicupper crust (EPC), indicating the presence of mafic rocksin their source area (Fig. 6). This mafic component in theanalyzed samples can be found in the matrix which isinvariably chloritic, and which formed largely from de-cay of labile constituents such as feldspar and rock frag-ment. Because of this decay, mafic rock fragments arescarce. The concentrations of HFSE such as Zr, Hf andTa, which are generally enriched in felsic relative to maficrocks, are generally depleted in the Akwatiametasediments relative to those of EPC. An exception isZr and Hf concentrations that are comparable to EPC and,therefore, may suggest a minor provenance componentof felsic rocks. The elemental ratios, La/Sc, La/Cr, La/Co, Th/Sc, Th/Cr and Th/Co are particularly critical ofprovenance (Cullers and Podkovyrov, 2000). Comparedto early Proterozoic volcanic rocks of different composi-tions, the analyzed metasedimentary rocks show La/Scand Th/Sc ratios intermediate between andesites and felsic

rocks, and La/Cr, La/Co, Th/Cr, Th/Co ratios betweenbasalts and andesites (Table 2). Cr/Th and Th/Sc ratiosare particularly sensitive to the composition of sedimentsources (Taylor and McLennan, 1985). On a plot of thesetwo ratios, the analyzed metasedimentary rocks define astraight line suggestive of mixing a felsic and a maficcomponent (Fig. 5).

REE patterns have been used widely in geochemicalstudies of metasedimentary rocks. The degree of differ-entiation of LREE from HREE is a measure of the pro-portion of felsic to mafic components in the source re-gion, whereas Eu anomalies may provide informationabout the nature of the processes affecting the source area,such as whether plagioclase has been removed from the

Fig. 5. Cr/Th - Th/Sc graph showing the distribution ofmetasedimentary rocks. Averages of volcanic rocks of earlyProterozoic age are plotted for comparison. FVO = felsicvolcanics; AND = andesites; BAS = basalts (after Condie,1993). Averages of volcanic rocks from the Kibi-Winneba andAshanti belts of the Birimian greenstone belts are also plottedfor comparison. BFV = Birimian rhyolites; BAV = Birimianandesites; BBV = Birimian basalts. The averages of theBirimian data were compiled from those reported by Sylvesterand Attoh (1992).

Fig. 6. The average multi-element pattern (with range repre-sented by bars) of the analyzed samples normalized to averageearly Proterozoic crust (EPC). The elements are arranged insuch a way that those mainly enriched in felsic rocks are plot-ted on the left-hand side and those enriched in mafic rocks areplotted on the right-hand side. EPC values are those reportedby Condie (1993).


ultimate igneous source areas of the sediments (Taylorand McLennan, 1985). The relatively less LREE enrich-ment of the analyzed samples (LaN/YbN mostly between2.1 and 3.5) compared to that of PAAS and EPC (LaN/YbN = 4.3 and 3.9 respectively; Condie, 1993) suggestthe dominance of mafic rocks over felsic rocks in thesource areas. In addition, the lack of significant negativeEu-anomalies for most of the metagraywackes and themetapelites suggest the dominance of andesitic and/orbasaltic rocks in the source region, and that K-rich gra-nitic rocks were not present in significant proportion(Taylor and McLennan, 1985).

The general lack of Eu anomalies, low SiO2/Al2O3,Th/Sc and Zr/Sc ratios and the generally variable majorand trace element compositions (Table 1) suggest localprovenance for the metasedimentary rocks (McLennan etal., 1993). The associated Birimian metavolcanics, whichis composed of basaltic and andesitic to dacitic rocks,appears the most likely source rocks. This interpretationsupport geochronological studies on the Birimianmetasedimentary rocks in Ghana, which indicate that thedetrital materials were mostly derived from the associ-ated Birimian volcanic belts (Taylor et al., 1992; Daviset al . , 1994). Even though compared to EPCmetasedimentary rocks show depletion in Th and HFSE(except Zr), these elements are significantly higher thanthose reported from the volcanic rocks of the Birimiangreenstone belts (Sylvester and Attoh, 1992). The levelsof HFSE and Th concentrations observed in the analyzedsamples are, however, comparable to those of Birimianbelt-type granitoids (Loh and Hirdes, 1999). Single-zircon-grain dating of a volcaniclastic wacke from theKumasi Basin (Fig. 1B) yielded ages of ~2185–2155 Ma(Davis et al., 1994). This age range encompasses U-Pbzircon ages of 2179 and 2172 Ma obtained for belt-typegranitoids of southwest Ghana (Hirdes et al., 1992; Daviset al., 1994). It is, therefore, possible that a portion of thedetrital sediments were supplied by Birimian belt-typegranitoids.

Comparison with sedimentary rocks from known tectonicsettings

In the last two decades, sedimentologists have endeav-oured to distinguish the tectonic conditions prevalentduring deposition of sediments on the basis ofgeochemistry (e.g., Bhatia and Crook, 1986; McLennanet al., 1993). Even though tectonic fields identified bythese studies are originally intended for Phanerozoic clas-tic sedimentary rock, they have gained wide applicationin Precambrian sedimentary rocks (e.g., McLennan et al.,1995; Kalsbeek et al., 1998; Yang et al., 1998; Toulkeridiset al., 1999; Bhat and Ghosh, 2001).

McLennan et al. (1993, 1995) described five majorprovenance types on the basis of geochemistry, the char-acteristics of which are summarized in Table 3. Severallines of evidence suggest that the analyzedmetasedimentary rocks are likely dominated by sourcesrepresentative of Young Undifferentiated Arc terranes.The most important evidence is (1) the unevolved majorand trace element composition (e.g., low SiO2/Al2O3, lowTh/Sc, low Zr/Sc), (2) the variable major and trace ele-ment compositions, (3) lack of significant negative Euanomalies, and (4) lower REE abundances, and variablebut less LREE enrichment when compared with those ofEPC and PAAS. This interpretation is supported by Sm-Nd isotopic study on the Birimian metasedimentary(Taylor et al., 1992) that gives neodymium model agesof within 0.2 Ga of the ~2.1 Ga sedimentation age (andformation age of associated metavolcanics). In the La-Th-Sc ternary diagram, the analyzed samples plot withinthe field defined by modern sediments deposited in mag-matic arc-related basins along active plate margins (Fig.7). Together, the geochemical and isotopic data indicateonly minimal contribution of old upper crust, suggestingthat not much significant volumes of pre-Birimian conti-nental sources was involved in their formation. This study,therefore, agrees with other petrographic studies (Leubeet al., 1990) and isotopic analyses (Taylor et al., 1992;Davis et al., 1994) on the Birimian sedimentary rocks that

Table 2. A comparison of provenance elemental ratios of the Birimian metasedimentary rocksand early Proterozoic volcanic rocks (Condie, 1993)

Birimian metasedimentary rocks Felsic volcanic rocks Andesites Basalts

Metagraywackes Metapelites

Range Mean Range Mean Mean Mean Mean

La/Sc 2.71–1.40 1.40 1.11–0.38 0.74 1.88 1.05 0.32La/Cr 0.36–0.72 0.17 0.16–0.01 0.09 3.00 0.31 0.08La/Co 0.64–0.24 0.38 0.39–0.12 0.31 5.00 0.74 0.29Th/Sc 0.47–0.09 0.29 0.30–0.10 0.22 0.50 0.17 0.07Th/Cr 0.11–0.01 0.04 0.04–0.002 0.03 0.80 0.05 0.02Th/Co 0.12–0.03 0.08 0.11–0.03 0.09 1.33 0.12 0.06


indicate their derivation from local sources within thegreenstone belts. However, significant contribution frombasement rocks cannot be entirely ruled out since the base-ment of the Birimian (i.e., “Dabakalian”) is no more than0.5 Ga older than the sedimentation age (Abouchami etal., 1990).

Crustal evolution at the Archean-Proterozoic boundarySedimentary rocks have been used to constrain the

average composition of the terrains exposed at the timeof deposition (e.g., Condie, 1993; McLennan et al., 1995).The Archean-Proterozoic boundary (A/P boundary) isrecognized as a fundamental benchmark in the chemicalevolution of the upper continental crust. Extensive stud-ies of surface samples from Precambrian shields in widegeographic areas such as North America, western Europe,Australia, India and southern Africa have shown that theArchean upper crust is generally different in chemicalcomposition from post-Archean upper crust (Condie,1993; Taylor and McLennan, 1985). In order to determinewhether changes in composition of the continental crustat the A/P boundary was worldwide, we test the proposedmodel of compositional change at the A/P boundary onthe chemical data of the Birimian metasediments, whichis located in one of the world’s less studied shields(Sylvester and Attoh, 1992).

Tabl

e 3.

Su

mm

ary

of g

eoch

emic

al c

hara

cter

isti

cs o

f pr

oven

ance

typ

es (

afte

r M

cLen

nan

et a

l.,

1993

, 19

95)

Exo

tic

com

pone

nts:

Che

mic

al s

igna

tur e

dep

ends

on

the

natu

r e o

f th

e co

mpo

nent

. F

or e

xam

ple,

ver

y hi

gh M

g, C

r, N

i, V

and

Cr/

V w

ould

be

dist

inct

ive

of o

phio

liti

c so

urce

s.

Fig. 7. La - Th - Sc ternary plot with fields defined by Girtyand Barber (1993). Source rock compositions (volcanics) areof early Proterozoic age (Condie, 1993). BAS = basalts;AND = andesites; FVO = felsic volcanic rocks; TTG =Proterozoic tonalite-trondhjemite-granodiorite; EPC = earlyProterozoic crust; EPG = early Proterozoc graywackes;PSS = Proterozoic sandstones; GRA = Proterozoic granites(Condie, 1993).


Fig. 8. Plot of Eu/Eu* versus GdN/YbN for the metasedimentaryrocks. Fields of Archean greenstone sediments and post-ArcheanAustralian Shale (PAAS) after McLennan et al. (1995).

The chemical differences between Archean and post-Archean upper crust are recorded in trace elements ofsedimentary rocks (Taylor and McLennan, 1985;McLennan and Hemming, 1992; Condie, 1993). Theseinclude: (1) a decrease in Eu-anomalies, (2) a decrease inGdN/YbN ratio from > 2.0 to 1.0–2.0; Condie (1993) has,however, indicated that unlike Archean graywackes,Archean shales lack HREE depletion, (3) a decrease inSm/Nd ratio from about 0.21 to 0.19, (4) a decrease inthe Cr/Th ratio from about 20 to 5.7, (5) a decrease in Cr/Sc ratio from about 13 to 4.1, and (6) an increase in theTh/Sc ratio from about 0.5 to 1.0 (possibly only in conti-nental sediments).

The analyzed metasedimentary rocks show typicalArchean signatures; low Th/Sc ratios (<0.5, with excep-tion of one sample), high Sm/Nd ratios (>2.1, with ex-ception of 3 samples), high Th/Sc ratios (>20, with ex-ception of 3 samples), and high Cr/Sc ratios (>6, withexception of 2 samples). In addition, the Eu/Eu* and GdN/YbN ratio compares very favourably with Archean sedi-mentary rocks (Fig. 8). It is also important to note thatlike Archean shales, the Akwatia metapelites lack HREEdepletion.

Sedimentary rocks of any age, derived primarily fromArchean terranes, especially from relatively localizedsources, may show Archean geochemical signatures.However, it is unlikely the Akwatia metasedimentaryrocks inherited these trace element features from theirpossible Archean source given the inferred juvenile arc

provenance. It is, rather most likely that the sedimentswere derived from rocks within the Birimian greenstonebelts, and that the composition of the continental crust ofthe study area (and possibly for West Africa) during theearly Proterozoic was broadly similar to that of Archeancrust. We therefore conclude that the secularcompositional trend at the A/P boundary is not applica-ble to the metasedimentary rocks in the Birimdiamondiferous field. This interpretation is supported bystudies on volcanic rocks from the Birimian greenstonebelts, which indicate that the Birimian has rock associa-tion and trace-element signatures similar to those ofArchean greenstone belts (Abouchami et al., 1990; Leubeet al., 1990; Sylvester and Attoh, 1992; Taylor et al.,1992). Whether these trace element features observed inthe Birimian metasedimentary rocks of the Birimdiamondiferous field are only unique to the Akwatia area(and maybe the Cape Coast Basin) but not a common fea-ture of the Birimian metasediments as a whole needs fur-ther geochemical investigations of other Birimian sedi-mentary basins.

CONCLUSIONS

The Birimian metasedimentary rocks of the Birimdiamondiferous field show REE and trace element char-acteristics that closely resemble those of their Archeancounterparts but differ significantly from post-Archeansediments, suggesting that the continental crust of thestudy area during the early Proterozoic had chemical com-positions similar to those of the Archean crust. This in-terpretation supports the suggestion by Sylvester andAttoh (1992) that the A/P boundary may not coincide witha worldwide, fundamental change in crustal evolution.The trace element data further suggest that themetasedimentary rocks were mainly derived from a juve-nile arc source of mixed felsic and mafic composition,and support geochronological data (Taylor et al., 1992;Davis et al., 1994) that indicate their derivation frommainly the Birimian volcanic belts.

Acknowledgments—We are grateful to the Ghana Atomic En-ergy Commission as well as the entire professional staff of itsGHARR-1 Centre, for making available to us their NAA facili-ties for the chemical analysis. We also acknowledge with thanksthe logistic assistance and reception given to us by J. Manfuland the entire staff of the Prospecting Department of the GhanaConsolidated Diamonds Ltd., Akwatia, during our fieldwork.The fieldwork was financially supported by the Ghana Miner-als Commission’s Mineral Development Fund. We are verygrateful to K. Sugitani and C. Koeberl for their critical reviewsand suggestions, which have substantially improved the manu-script.


REFERENCES

Abouchami, W., Boher, M., Michard, A. and Albarede, F. (1990)A major 2.1 Ga old event of mafic magmatism in WestAfrica: An early stage of crustal accretion. J. Geophys. Res.95, 17605–17629.

Anwiller, D. N. and Mack, L. E. (1991) Diagenetic modifica-tion of Sm-Nd model ages in Tertiary sandstones and shales,Texas Gulf Coast. Geology 57, 101–115.

Béziat, D., Bourges, F., Debat, P., Lompo, M., Martin, F. andTollon, F. (2000) A Paleoproterozoic ultramafic-mafic as-semblage and associated volcanic rocks of the Boromogreenstone belt: fractionates originating from island-arcvolcanic activity in the West African Craton. PrecambrianRes. 101, 25–47.

Bhat, M. I. and Ghosh, S. K. (2001) Geochemistry of the 2.51Ga old Rampur group pelite, western Himalayas: implica-tions for their provenance and weathering. Precambrian Res.108, 1–16.

Bhatia, M. R. and Crook, A. W. (1986) Trace element charac-teristics of greywackes and tectonic setting discriminationof sedimentary basins. Contrib. Mineral. Petrol. 92, 181–193.

Blatt, H. G., Middleton, G. V. and Murray, R. C. (1980) Originof Sedimentary Rocks. 2nd ed., Prentice-Hall, New Jersey.

Boher, M., Abouchami, W., Michard, A., Albarede, F. and Arndt,N. T. (1992) Crustal growth in West Africa at 2.1 Ga. J.Geophys. Res. 95, 345–369.

Condie, K. C. (1993) Chemical composition and evolution ofthe upper continental crust: Contrasting results from sur-face samples and shales. Chem. Geol. 104, 1–37.

Cullers, R. L. and Podkovyrov, V. N. (2000) Geochemistry ofthe Mesoproterozoic Lakhanda shales in southeasternYakutia, Russia: implications for mineralogical and prov-enance control, and recycling. Precambrian Res. 104, 77–93.

Davis, D. W., Hirdes, W., Schaltegger, U. and Nunoo, E. A.(1994) U-Pb age constraints on deposition and provenanceof Birimian and gold-bearing Tarkwaian sediments inGhana, West Africa. Precambrian Res. 67, 89–107.

Dickinson, W. R. and Suczek, C. A. (1979) Plate tectonics andsandstone compositions. Bull. Am. Assoc. Pet. Geol. 63,2164–2182.

Dickinson, W. R., Beard, L. S., Brakenridge, G. R., Erjavec, J.L., Ferguson, R. C., Inman, K. F., Knepp, R. A., Lindberg,F. A. and Ryberg, P. T. (1983) Provenance of North AmericaPhanerozoic sandstone in relation to tectonic setting. Bull.Geol. Soc. Am. 94, 222–235.

Doumbia, S., Pouclet, A., Kouamelan, A., Peucat, J. J., Vidal,M. and Delor, C. (1998) Petrogenesis of juvenile-typeBirimian (Paleoproterozoic) granitoids in Central Côte-d’Ivore, West Africa: geochemical and geochronology.Precambrian Res. 87, 33–63.

Girty, G. H. and Barber, R. W. (1993) REE, Th, and Sc evi-dence for the depositional setting and source rock charac-terist ics of the Quartz Hill chert , Sierra Nevada,California. Processes Controlling the Composition of ClasticSediments (Johnsson, M. J. and Basu, A., eds.), Geol. Soc.Am. Spec. Pap. 284, 109–119.

Hirdes, W., Davis, D. W. and Eisenlohr, B. N. (1992) Reassess-ment of Proterozoic granitoid ages in Ghana on the basis ofU/Pb zircon and monazite dating. Precambrian Res. 56, 89–96.

Hottin, G. and Quedraogo, O. F. (1975) Notice explicative dela carte géologoque a 1:1,000,000 de la République deHaute-Volta. Direction de la Géologie et des Mines,République de Haute-Volta, Ouagadougou, 58 pp.

Junner, N. R. (1940) Geology of the Gold Coast and WesternTogoland (with revised geological map). Bull. Gold CoastGeol. Surv. 11, 40 pp.

Junner, N. R. (1943) The diamond deposits of the Gold Coastwith notes on other diamond deposits in West Africa. Bull.Gold Coast Geol. Surv. 12, 55 pp.

Kalsbeek, F., Pulvertaft, T. C. R. and Nutman, A. P. (1998)Geochemistry, age and origin of metagreywackes from thePalaeoproterozoic Karrat Group, Rinkian Belt, West Green-land. Precambrian Res. 91, 383–399.

Koeberl, C. (1993) Instrumental neutron activation analysis ofgeochemical and cosmochemical samples: a fast and reli-able method for small sample analysis. J. Radioanal. Nucl.Chem. 168, 47–60.

Küster, D. and Liégeois, J.-P. (2001) Sr, Nd isotope andgeochemistry of the Bayuda Desert high-grade metamor-phic basement (Sudan): an early Pan-African oceanic con-vergent margin, not the edge of the East Sahara ghost craton?Precambrian Res. 109, 1–23.

Leube, A., Hirdes, W., Mauer, R. and Kesse, G. O. (1990) Theearly Proterozoic Birimian Supergroup of Ghana and someaspects of its associated gold mineralization. PrecambrianRes. 46, 139–165.

Liégeois, J. P., Claessens, W., Camara, D. and Klerkx, J. (1991)Short-lived Eburnean orogeny in southern Mali: geology,tectonics, U-Pb and Rb-Sr geochronology. Precambrian Res.50, 111–136.

Loh, G. and Hirdes, W. (1999) Explanatory notes for the geo-logical map of Southwest Ghana, 1: 100,000 sheets Sekondi(0402A) and Axim (0403B). Bull. Ghana Geol. Surv. 49,149 pp.

McDaniel, D. K., Hemming, S. R., McLennan, S. M. andHanson, G. N. (1994) Resetting of neodymium isotopes andredistribution of REEs during sedimentary processes: Theearly Proterozoic Chelmsford Formation, Sudbury Basin,Ontario, Canada. Geochim. Cosmochim. Acta 58, 931–941.

McLennan, S. M. (1989) Rare earth elements in sedimentaryrocks: Influence of provenance and sedimentary processes.Rev. Mineral. 21, 169–200.

McLennan, S. M. and Hemming, S. (1992) Samarium/neodym-ium elemental and isotopic systematics in sedimentaryrocks. Geochim. Cosmochim. Acta 56, 887–898.

McLennan, S. M., Taylor, S. R., McCulloch, M. T. and Maynard,J. B. (1990) Geochemical and Nd-Sr isotopic compositionof deep-sea turbidites: Crustal evolution and plate tectonicassociations. Geochim. Cosmochim. Acta 54, 2015–2050.

McLennan, S. M., Hemming, S. R., McDaniel, D. K. andHanson, G. N. (1993) Geochemical approaches to sedimen-tation, provenance and tectonics. Processes Controlling theComposition of Clastic Sediments (Johnsson, M. J. andBasu, A., eds.), Geol. Soc. Am. Spec. Pap. 284, 21–40.


McLennan, S. M., Hemming, S. R., Taylor, S. R. and Eriksson,K. A. (1995) Early Proterozoic crustal evolution:Geochemical and Nd-Pb isotopic evidence frommetasedimentary rocks, southern North America. Geochim.Cosmochim. Acta 59, 1153–1177.

Milési, J. P., Ledru, P., Feybesse, J. L., Dommanget, A. andMarcoux, E. (1992) Early Proterozoic ore deposits and tec-tonics of the Birimian orogenic belt , West Africa.Precambrian Res. 58, 305–344.

Nesbitt, H. W. (1979) Mobility and fractionation of rare earthelements during weathering of a granodiorite. Nature 299,715–717.

Nyarko, B. J. B. (1999) Calibration and validation of an Aptecgamma spectroscopy system at Gharr-1 Centre for instru-mental neutron activation analysis (unpublished M.PhilThesis). University of Ghana.

Shimizu, H., Sawatari, H., Kawata, Y., Dunkley, P. N. andMasuda, A. (1992) Ce and Nd isotope geochemistry on arcvolcanic rocks with negative Ce anomaly: existence ofsources with concave REE patterns in the mantle beneaththe Solomon and Bonin island arcs. Contrib. Mineral. Pet-rol. 110, 242–252.

Sylvester, P. J. and Attoh, K. (1992) Lithostratigraphy and com-position of 2.1 Ga greenstone belts of the West AfricanCraton and their bearing on crustal evolution and the

Archean-Proterozoic boundary. J. Geol. 100, 377–393.Tagini, B. (1971) Esquisse structurale de la Côte D’Ivoire. Essai

de géotectonique régionale. Thèse Univ. Lausanne. Sociétéd’Etat pour le Développement minière de la Côte D’Ivoire.

Taylor, P. N., Moorbath, S., Leube, A. and Hirdes, W. (1992)Early Proterozoic crustal evolution in the Birimian ofGhana: constraints from geochronology and isotope geol-ogy. Precambrian Res. 56, 97–111.

Taylor, S. R. and McLennan, S. M. (1985) The ContinentalCrust: Its Composition and Evolution. Blackwell Scientific,Oxford.

Toulkeridis, T., Clauer, N., Kröner, A., Reimer, T. and Todt, W.(1999) Characterization, provenance, and tectonic settingof Fig Tree greywackes from the Archaean BarbertonGreenstone Belt, South Africa. Sediment. Geol. 124, 113–129.

Vidal, M. and Alric, G. (1994) The Palaeoproterozoic (Birimian)of Haute-Comoé in the West African craton, Ivory Coast: atransitional back-arc basin. Precambrian Res. 65, 207–229.

Yang, H., Kyser, K. and Ansdell, K. (1998) Geochemical andNd isotopic compositions of the metasedimentary rocks inthe La Ronge Domain, Trans-Hudson Orogen, Canada: im-plications for the evolution of the domain. Precambrian Res.92, 37–64.

Appendix A. Locations and mineralogical assemblages of the analyzed samples

A (abundant), >20%; M (minor), 5–20%; T (trace), <5%. Mineral abbreviations: Q, quartz; P, plagioclase; K, K-feldspar; Bi, biotite;Ms, muscovite; Cal, calcite; Chl, chlorite; Cht, chert; Opq, pyrite and opaque minerals; Hm, heavy minerals; Sct, sericite.


Source of IAEA Soil 7: pages 35 and 36 of IAEA Analytical Quality Control Services (AQCS) 2002–2003, 112 pp.Precision of the counting system was calculated as a percentage relative standard deviation (% RSD) of six replicate samples analyzed.

Appendix B. Composition of IAEA Soil 7 reference material

Analyte Reported values This work % RSD

mg/kg 95% C.I.

Al 47000 44000–51000 46000 ± 500 1.1Ca 163000 157000–174000 168000 ± 10200 6.1Ce 61 50–63 59.5 ± 2.3 3.9Co 8.9 8.4–10.1 9.31 ± 0.51 5.4Cr 60 49–74 51.6 ± 2.5 4.8Dy 3.9 3.2–5.3 3.77 ± 0.15 3.9Eu 1 0.9–1.3 1.06 ± 0.06 5.4Fe 25700 25200–26300 25300 ± 460 1.8Hf 5.1 4.8–5.5 4.92 ± 0.29 5.8K 12100 11300–12700 12700 ± 700 5.5La 28 27–29 28.7 ± 1.1 3.8Lu 0.3 0.1–0.4 0.39 ± 0.02 5.5Mg 11300 11000–11800 11600 ± 450 3.9Mn 631 604–650 626 ± 29 4.2Na 2400 2300–2500 2300 ± 40 1.7Nd 30 22–34 32.5 ± 1.2 3.6Sc 8.3 6.9–9.0 8.21±0.23 2.8Si 180000 169000–201000 186000 ± 11300 6.1Sm 5.1 4.8–5.5 5.3 ± 0.25 4.8Ta 0.8 0.6–1.0 0.77 ± 0.04 4.6Tb 0.6 0.5–0.9 0.72 ± 0.03 4.5Th 8.2 6.5–8.7 6.62 ± 0.21 3.2Ti 3000 2600–3700 3300 ± 162 4.9V 66 59–73 63.5 ± 2.8 4.4Yb 2.4 1.9–2.6 2.2 ± 0.09 4.3Zr 185 180–201 189 ± 11.2 5.9

geochemistry of paleoproterozoic metasedimentary rocks - terrapub

Documents