AN OVERVIEW OF THE YILGARN CRATON AND ITS CRUSTAL EVOLUTION D.C. Champion & K.F. Cassidy Geoscience Australia, GPO Box 378, Canberra ACT 2601, Australia Introduction The Archean Yilgarn Craton of Western Australia, is not only one of the largest extant fragments of Archean crust in the world, but is also one of the most richly-mineralised regions in the world. Accordingly, understanding the four-dimensional evolution of the craton is important for reasons other than just studies of the geodynamics of early Earth crustal growth. In particular, better understanding the space-time evolution of the craton and its components provides important constraints on the geodynamics and architecture of mineral systems. It also has the potential to explain the heterogeneous distribution of mineralisation within the craton; that is, is there a geodynamic reason why the Kalgoorlie Terrane, for example, is so strongly enriched in nickel and gold? One approach to constraining the four-dimensional evolution of any region is by using the geochemical, isotopic and geochronological characteristics of granites and other felsic rocks. The integrated information these characteristics convey provide constraints on timing of crustal growth, nature of crustal growth (lateral versus vertical), and the degree of crustal reworking, as well as delineating crustal domains. This approach is particularly effective for the Yilgarn Craton where granites comprise over 70% of the rock types by surface area and at least similar amounts by volume, based on seismic refraction studies (Drummond, 1988). This paper presents the results of granite studies undertaken within the Yilgarn Craton by Geoscience Australia and collaborating partners in the last ten years. In particular, it draws on the results of two AMIRA projects, P482 (Cassidy et al., 2002) and P624 (Barley et al., 2003). The data are used to show that the Yilgarn Craton comprises distinct crustal terranes with differing crustal histories. More importantly, pre-existing continental crust has played a major role in younger orogenic events and appears to have exerted a significant, but poorly understood, spatial control on the distribution of mineral systems, such as gold, komatiite-associated nickel sulphide and volcanic-hosted massive sulphide (VHMS) base metal systems. Yilgarn geology - new subdivisions The Yilgarn Craton dominantly comprises typical Archean granite-greenstone terranes that formed largely between ca. 3.10 and 2.60 Ga (Fig. 1). Minor older rocks (back to 3.7 Ga) occur within the Narryer Terrane. Greenstone sequences, typically dominated by mafic volcanism, but including ultramafic to felsic metavolcanic rocks and metasedimentary rocks, form less than 30% by area of the craton. Greenstone sequences formed episodically from ca. 3.0-2.9 Ga in the western Yilgarn to 2.72-2.655 Ga in the east (Fig. 1). Episodic felsic magmatism occupies a similar time frame, and was largely synchronous with greenstone formation. The only exception to this is the late (ca. 2.655 to 2.63 Ga) craton-wide granite event, that effectively marks cratonization (Fig. 1). The great majority of the felsic magmatic record in the Yilgarn Craton is ca. 2.76 to 2.63 Ga in age. Extant felsic plutonism older than this is poorly preserved, largely as isolated, often gneissic, remnants; it is best recorded, paradoxically, within the older greenstone belts. Previous subdivisions of the Yilgarn Craton have been based on greenstone morphology (Gee et al., 1981), fault-bounded tectonostratigraphy (Myers, 1995), or both (e.g. Swager, 1997; Barley et al., 1998, 2002, 2003). These subdivisions, however, have not taken into account the felsic magmatic rocks that form over 70% of the craton. Geochemistry, geochronology and Sm-Nd isotopic data from the Yilgarn granites (Figs 1, 2; Cassidy et al., 2006; Champion et al., 2006), in conjunction with detailed greenstone studies in the eastern Yilgarn (Barley et al., 2002, 2003), indicate that the bulk of the craton can be subdivided into two major provinces (see Cassidy et al., 2006) – an older craton nucleus, the Youanmi Terrane (YT; formerly Southern Cross and Murchison Provinces), and a collage of terranes in the east – the Eastern Goldfields Superterrane (EGST; formerly Eastern Goldfields Province). Granite groups The bulk of granites of the Yilgarn Craton are mineralogically very similar, with >80% being biotite-bearing monzogranites or granodiorites (Table 1); accordingly geochemistry and geochronology is required to help characterize and discriminate specific granite groups. Champion and Sheraton (1997), studying granites in the Leonora-Laverton region, recognized five main groups largely on the basis of petrological and geochemical characteristics. More recent work has confirmed and extended this general subdivision for granites across the Yilgarn Craton (Table 1). Granites can be divided into two major geochemical groups: High-Ca and Low-Ca, comprising over 60% and 20%, respectively, of the total granites, and three minor (High-HFSE, Syenitic and Mafic) geochemical groups (Table 1; Figure 1). All groups, except the syenites, occur within both the YT and EGST; the syenites are localised within the latter. The three minor groups appear to be strongly localized within or marginal to greenstone belts. The High-Ca granites (2.72-2.66 Ga) are felsic (68-77% SiO2), dominantly sodic rocks (Table 1), that form a LILE-rich end-member to the ubiquitous tonalite-trondhjemite-granodiorite (TTG) group of granites that occur in all Archean cratons. The High-Ca granites have chemical compositions consistent with derivation from a broadly basaltic precursor but with an additional crustal component. Like TTGs elsewhere, the HighCa granites are (mostly) characterized by Sr-undepleted, Y-depleted signatures, indicating they were largely produced at high pressures; either within thickened mafic crust or by melting of subducting oceanic crust. The younger Low-Ca group is chemically distinct from the High-Ca granites. The former are characterised by high-LILE, strong enrichments in LREE and some of the HFSE, and have compositions consistent with crystal fractionation, and were derived by partial melting of crust of broadly tonalitic composition; that is, they represent reworking of older (High-Ca) felsic crust. Other granite types in the EGP are volumetrically minor and include: a) the high SiO2 (>74% SiO2) HighHFSE group (2.685-2.66 Ga) with distinctive A-type characteristics but low LILE contents, especially Rb and Pb; b) a geochemically diverse but isotopically similar group of more mafic (<60 to >70% SiO2) granites (2.69-2.65 Ga) that exhibit large between-suite variations in LILE and LREE; and c) younger syenites (2.652.64 Ga). Age distribution U-Pb zircon geochronology of granites (Nelson, 1997; Fletcher et al., 2001; Cassidy et al., 2002; Dunphy et al., 2003; Geological Survey of Western Australia, 2006; Sircombe et al., 2007) indicates a range in the age of granitoid magmatism from ca. 3.0 to 2.63 Ga, with an apparent continuum in magmatism between 2.72 and 2.63 Ga across the craton. Oldest rocks belong to the High-Ca, High-HFSE and Mafic groups. Importantly, the geochronological database clearly confirms distinct episodes of specific granite magmatism (Fig. 1). This is best illustrated by the High-Ca granites, which display a peak between 2.72 and 2.68 Ga in the YT and a younger pronounced peak between 2.68 and 2.655 Ga in the EGST. Notably, while voluminous High-Ca magmatism was occurring in the EGST (2.68-2.655 Ga), at this time the YT was the site of minimal magmatism of any type. The period ca. 2.68 Ga is important, therefore, as it not only marks the cessation of significant plutonism in the YT, but also the switch from localised Mafic and High-HFSE granites in the EGST to widespread High-Ca granites. The other significant feature confirmed by the geochronology is the switch from High-Ca to Low-Ca magmatism at ca. 2.655 Ga (Fig. 1). Low-Ca granites are late in the magmatic sequence with ages between 2.655 and 2.63 Ga. These ages are found across the Yilgarn Craton and record what must have been a voluminous craton-wide event of some significance. Sm-Nd isotopes and crustal development Regional Sm-Nd isotopic and zircon inheritance data from felsic magmatic rocks confirm that there is a significant crustal component in both the High-Ca and Low-Ca granites and, as such, the isotopic data for these granites can be used to provide strong constraints on the crustal prehistory of the Yilgarn Craton. Champion and Cassidy (in prep.) used Nd depleted-mantle model ages (TDM), calculated from Sm-Nd isotopic data for the granites, to produce a gridded map showing relative crustal ages (not absolute age) for the Yilgarn Craton (Fig. 2). Results of this work, from both the High-Ca and Low-Ca granites, clearly identify a major isotopic break between the YT and the EGST (Fig. 2). The simplest interpretation for such an isotopic break is that it represents a major change in average crustal age. The fact that the isotopic break, for the most part, closely corresponds with the ground position of the Ida Fault, supports this idea. The isotopic data, therefore, effectively subdivide the craton into a larger older crustal block (YT) that is bounded to the east by the significantly younger, economically-important, EGST. The isotopic data also delineate two isotopically younger domains that broadly correspond to geological domains; one in the central part of the YT, co-incident with layered mafic intrusives, and one within the EGST (Fig. 2). Importantly, these domains are either characterized by primitive isotopic signatures (i.e. Nd TDM ages close to crystallisation ages), and/or contain evidence for arc-related magmatism (e.g., calc-alkaline andesites). Constraints from felsic magmatism on tectonic models for the EGST Champion and Cassidy (in Barley et al., 2003), proposed preferred tectonic environments for the formation for each granite group. Although tectonic models based on granite data alone are equivocal, they can be used, in conjunction with the points raised above and the Sm-Nd isotopic data, to provide important constraints for tectonic models of the Yilgarn Craton, particularly the EGST. Favoured tectonic models indicate a variety of arc environments (largely pre-existing continental crust - not island arc-like) with or without various rifting regimes, pre ca. 2.655 Ga. Granite magmatism pre-2.68 Ga appears to be largely syn-volcanic, with clear match-ups between intrusive and extrusive units. It is not surprising, therefore, that tectonic models interpreted from these granites (e.g. Cassidy et al., 2002), closely match those suggested for the volcanic and sedimentary sequences of this age (Barley et al., 2002, 2003). Overall, there is a dominance of continental margin signatures as well as evidence of magmatic recycling of older arc-related crust. The two most significant changes in magmatic style in the EGST occur at ca. 2.675 and 2.655 Ga. The earlier changes in style of magmatism are interpreted to represent variations within an overall subduction-related environment. The change at 2.675 Ga, from localised Mafic and High-HFSE granites to widespread dominantly High-Ca granites (within all terranes of the EGST), coupled with the first significant appearance of LILE-enriched magmatism around this time (high-LILE subgroups of the High-Ca and Mafic groups), suggests that significant recycling of crust commenced during this period, possibly either as a consequence of a subducted crustal component, and/or, more likely, crustal thickening (e.g. Hildreth and Moorbath, 1988), and/or, speculatively, perhaps a significant change (shallowing) in the angle of slab subduction, with commensurate associated crustal shortening and deformation (e.g. Gutscher, 2002). If correct, then the appearance of High-Ca magmatism at this time across all terranes within the EGST, suggests that crustal thickening (or some other process) was similarly occurring across all terranes. Importantly, this period also corresponds to the timing of cessation of High-Ca magmatism in the YT. That is, widespread High-Ca magmatism ceased in the YT, and commenced in the EGST at much the same time. These two events are most likely related and together with the crustal thickening hypothesis, may indicate some form of terrane accretion occurred at this time. The 2.655 Ga change from High-Ca dominated to Low-Ca dominated magmatism reflects a significant change in switch in magmatic style, and presumably represents some fundamental re-ordering of the tectonic environment at that time, interpreted by Champion and Cassidy (in Barley et al., 2003) to be the cessation of subduction. From this time on, but certainly post-2.648 Ga, there is no preserved (or recognised) record of subduction-related Archean magmatism within the Yilgarn Craton. In this regard, the cessation of Low-Ca magmatism (ca. 2.62-2.63 Ga) marks the timing of effective cratonisation of the Yilgarn Craton as a whole. Crustal isotopic domains and mineral systems Consideration of the relationship between crustal age and the distribution of mineral deposits shows that there is a close spatial association between komatiite-associated nickel sulphide, VHMS base-metal and gold mineralisation with specific crustal domains, with this best expressed for the EGST. These associations are: 1) komatiite-associated nickel-sulphide deposits are concentrated in terranes with pre-existing crust; that is, Youanmi, Kalgoorlie, and an ‘old’ domain of the Kurnalpi Terrane. 2) VHMS base-metal systems appear spatially associated with terranes with juvenile crust, for example, the ‘young’ domains in the central Youanmi and Kurnalpi Terranes (Fig. 2). The Golden Grove deposits in the central YT and the Teutonic Bore and Jaguar deposits in the Kurnalpi Terrane are all associated with juvenile crust. 3) although gold mineralisation occurs in all terranes across the Yilgarn Craton, the majority of the gold endowment is spatially restricted to areas underlain by crust of intermediate age (Nd TDM ages of 2.95-3.1 Ga in Fig. 2); specifically the Kalgoorlie Terrane and an ‘old’ domain of the Kurnalpi Terrane. In the EGST, at least, there is little significant gold mineralization in domains with more juvenile signatures. Huston et al. (2005) suggested that the high heat flow and extensional structures that are characteristics of juvenile crustal domains encourage formation of VHMS deposits in such an environment. This relationship of metal endowment and specific crustal domains effectively explains why the Abitibi Subprovince in Canada is more endowed in VHMS mineralisation than the Yilgarn Craton; that is, the crust is not juvenile enough in the Yilgarn. Why these apparent relationships hold for komatiite-associated nickel-sulphide and gold mineral systems is not entirely clear, and it is noted that the relationship for gold does not appear to hold for the extremely well (gold) endowed Abitibi Subprovince. Using empirical relationships it is possible that, unlike VHMS systems, gold endowment in a specific terrane is not related to the nature of the underlying lithosphere, but related to the superposition of plume-related and subsequent subduction-derived magmatism on a pre-existing lithospheric architecture. The development of specific mineral systems therefore possibly reflects the interaction of specific geotectonic processes, such as later lithospheric-scale orogeny, with pre-existing crustal and mantle reservoirs. Gold potential is likely enhanced where accretionary fluid events using lithospheric-scale plumbing systems are able to interact with gold-rich lithospheric source rocks. This association is best developed in the Kalgoorlie and eastern Kurnalpi terranes in the EGST and the young crustal domain in the YT where there is evidence for both plume-related and subduction-derived magmatism. It is possible that the absence of a critical feature (e.g. plume-related komatiites in the ‘young’ crustal belt, Kurnalpi Terrane) decreases the metal inventory and/or the ability to extract fertile material during orogeny, making these crustal domains less prospective. Acknowledgements The authors wish to acknowledge that many of the results presented here were made possible through a number of industry-funded AMIRA projects, especially P482 and P624. The authors would also like to acknowledge co-workers from the Centre for Global Metallogeny, UWA, Mark Barley, Stuart Brown, Ian Fletcher, Stephen Gardoll, David Groves, Bryan Krapez, Neal McNaughton, Jan Dunphy; Geoscience Australia: Richard Blewett, Paul Henson, Tanya Fomin, Alan Whitaker; and, the Geological Survey of Western Australia: Bruce Groenewald, Steve Wyche. References Barley M.E., Krapež B., Brown S.J.A., Hand J. & Cas R.A.F. 1998, Mineralised volcanic and sedimentary successions in the Eastern Goldfields Province, Western Australia, Aust. Miner. Ind. Res. Assoc., Project P437 Final Report, 282p. Barley M.E., Brown S.J.A., Krapež B. & Cas R.A.F. 2002, Tectonostratigraphic analysis of the Eastern Yilgarn Craton: an improved geological framework for exploration in Archaean Terranes, Aust. Miner. Ind. Res. Assoc., Project P437A, Final Report, 257p. Barley M.E., Brown S.J.A., Cas R.A.F., Cassidy K.F., Champion D.C., Gardoll S.J. & Krapež B. 2003, An integrated geological and metallogenic framework for the eastern Yilgarn Craton: developing geodynamic models of highly mineralised Archaean granite-greenstone terranes, Aust. Miner. Ind. Res. Assoc., Project P624, Final Report, 192 188p. Cassidy K.F., Champion D.C., McNaughton N.J., Fletcher I.R., Whitaker A.J., Bastrakova I.V. & Budd A.R. 2002, Characterization and metallogenic significance of Archaean granitoids of the Yilgarn Craton, Western Australia: Aust. Miner. Ind. Res. Assoc. Project P482/MERIWA Project M281, Final Report, 514 p. Cassidy K.F., Champion D.C., Krapež B., Barley M.E., Brown S.J.A., Blewett R.S., Groenewald P.B. & Tyler I.M. 2006, A revised geological framework for the Yilgarn Craton, Western Australia, Western Australia Geological Survey, Record 2006/8, pp8. Champion D.C. & Sheraton J.W. 1997, Geochemistry and Sm-Nd isotopic systematics of Archaean granitoids of the Eastern Goldfields Province, Yilgarn Craton, Australia: constraints on crustal growth. Precambrian Research, 83, 109-132. Champion D.C., Cassidy K.F. & Smithies R.H. 2006, Sm-Nd isotope characteristics of the Pilbara and Yilgarn Cratons, Western Australia: implications for crustal growth in the Archean, Geochim. Cosmochim. Acta, 70, Issue 18, Supplement 1, A95 Drummond B.J. 1988, A review of crust/upper mantle structure in the Precambrian areas of Australia and implications for Precambrian crustal evolution. Precambrian Research, 40/41, 101–116. Dunphy J.M., Fletcher I.R., Cassidy K.F. & Champion D.C. 2003, Compilation of SHRIMP U-Pb geochronology data, Yilgarn Craton, Western Australia, 2001-02. Geoscience Australia Record 2003/15, 139p. Fletcher I.R., Dunphy J.M., Cassidy K.F. & Champion D.C. 2001, Compilation of SHRIMP U-Pb geochronology data, Yilgarn Craton, Western Australia, 2000-01. Geoscience Australia Record 2001/47, 110p. Gee R.D., Baxter J.L., Wilde S.A. & Williams I.R. 1981, Crustal development in the Archaean Yilgarn Block, Western Australia. In: Glover J.E. & Groves D.I. (eds), Archaean Geology: Second International Symposium, Perth, 1980, Geol. Soc. Aust. Spec. Publ., 7, 43-56. Geological Survey of Western Australia 2006, Compilation of geochronology data, June 2006 update. Western Australia Geological Survey, Digital Data, ANZWA1220000726. Gutscher M-.A. 2002, Andean subduction styles and their effect on thermal structure and interplate coupling. Journal of South American Earth Sciences, 15, 3-10. Hildreth E.W. & Moorbath S. 1988, Crustal contributions to arc magmatism in the Andes of Central Chile. Contributions to Mineralogy & Petrology, 98, 455-489. Huston D.L., Champion D.C. & Cassidy K.F. 2005, Tectonic controls on the endowment of Archean cratons in VHMS deposits: evidence from Pb and Nd isotopes. In: Mao J. & Bierlein, F.P. (eds), Mineral Deposit Research: Meeting the Global Challenge, Berlin/Heidelberg, Springer, 15-18. Myers J. 1995, The generation and assembly of an Archaean supercontinent: evidence from the Yilgarn craton, Western Australia. In: Coward M.P. & Ries A.C. (eds), Early Precambrian Processes, Geological Society of London Special Publication, 95, 143-154. Nelson D.R. 1997, Evolution of the Archaean granite-greenstone terrane of the Eastern Goldfields, Western Australia: SHRIMP zircon constraints. Precambrian Research, 83, 57-81. Sircombe K.N., Cassidy K.F., Champion D.C. & Tripp G. 2007, Compilation of SHRIMP U-Pb geochronological data, Yilgarn Craton, Western Australia, 2004-2006, Geoscience Australia Record 2007/01, 182p. Swager C.P. 1997, Tectono-stratigraphy of late Archaean greenstone terrains in the southern Eastern Goldfields, Western Australia. Precambrian Research, 83, 11-41. Figure 1. Distribution of granites groups within the Youanmi Terrane and Eastern Goldfields Superterrane of the Yilgarn Craton. Granite groups are as outlined in Table 1. Figure from Cassidy et al. (2002). Inset shows the relative age distributions of greenstone rocks and granites. Figure 2. Nd depleted-mantle model age map of the Yilgarn Craton. Image produced by gridding Nd depleted-mantle model ages calculated from Sm-Nd point data. (from Champion and Cassidy, in prep.) Group Area % Lithologies Field Characteristics Geochemistry Youanmi Terrane Eastern Goldfields Superterrane Comments High-Ca granodiorite, granite, trondhjemite. strongly deformed to gneissic to mildly deformed; typically ovoid plutons (elongate parallel to structural grain); minor dykes/sills and small bodies; some allanite or sphene mildly deformed to undeformed, locally strongly deformed; large intrusions (sheet-like) to small pods and dykes; biotite-dominant; allanite, sphene and fluorite-bearing variably deformed, highlevel intrusives, commonly spatially associated with volcanic rocks and volcanic complexes; presence of amphibole in a very felsic rock is diagnostic variably deformed, distinctive dark-looking granites; form moderate to small sized plutons and common dikes and sills; common amphibole ± biotite ± pyroxene commonly undeformed; distinctive red granites with green pyroxene (or amphibole); K-feldsparrich, little or no quartz high Na2O, Na2O/K2O, low Th, LREE, Zr; mostly Y-depleted, Srundepleted; range of LILE, LREE and Th contents; younger rocks extend to more LILEenriched compositions high K2O, low Na2O, high Rb, Th, LREE, Zr; moderately fractionated end-members >3.0 Ga-2.9? Ga, ca. 2.81 Ga, ca. 2.76 to 2.68 Ga; mostly 2.73-2.68 Ga ca. 2.8 Ga (minor remnants); 2.74-2.65 Ga; majority 2.685 to 2.655 Ga; youngest members appear to occur within Kalgoorlie Terrane high pressure partial melting of basaltic protoliths 2.65-2.6? Ga; mostly 2.65 to 2.63 Ga; possibly also 2.685 Ga 2.655 to 2.63 Ga partial melting of High-Ca type source rocks distinctive combination of high FeO*, MgO, TiO2, Y, Zr with low Rb, Pb, Sr, Al2O3 3.01 to 2.92 Ga?, ca. 2.81 Ga, 2.76 Ga to 2.45 & younger?; also 2655-2620 Ma spatial association with VHMS mineral systems low SiO2 (55-70+%), moderate to high Ni, Cr, MgO; range of LILE, LREE and Th; subdivided into high- and low-LILE 3.01 to 2.92 Ga?, ca. 2.81 Ga, 2.76 Ga to 2.10 Ga >2.72 Ga to 2.665 Ga; 2.7 to 2.68 Ga most common; mostly geographically restricted to Kurnalpi Terrane & north-east Kalgoorlie Terrane >2.72 Ga to 2.65 Ga; possibly younger? LILE-enriched members tend to be ca. 2.665 and younger high total alkalis (Na2O + K2O) 10-12%; commonly low MgO, FeO*, TiO2 none ca. 2.65 Ga, and 2.655-2.645 Ga some spatial association with gold mineralisation >60% distributed both within and external to greenstone belts Low-Ca granodiorite, granite >20% mostly external to greenstone belts HighHFSE granite, minor granodiorite 5-10% mostly internal or marginal to greenstone belts Mafic diorite, granodiorite, granite, tonalite & trondhjemite 5-10% Syenitic syenite, quartz syenite <5% internal or marginal to greenstone belt Table 1. Characteristics of the five granite groups in the Yilgarn Craton. common spatial association with gold mineralisation, especially highLILE members