Formation of Paleoarchean continental crust through infracrustal melting of enriched basalt

https://doi.org/10.1016/j.epsl.2009.03.003Get rights and content

Abstract

Paleoarchean rocks of the tonalite–trondhjemite–granodiorite (TTG) series require a basaltic source region more enriched in K, LILE, Th and LREE than the low-K tholeiites typical of Archean supracrustal sequences. Most TTG of the Pilbara Craton, in northwestern Australia, formed between 3.5 and 3.42 Ga through infracrustal melting of a source older than 3.5 Ga. Basaltic to andesitic rocks of the 3.51 Ga Coucal Formation, at the base of the Pilbara Supergroup, are amongst the only well-preserved remnants of pre-3.5 Ga supracrustal material on Earth, and may have formed a large proportion of pre-3.5 Ga Pilbara crust. These rocks are significantly enriched in K, LILE, Th and LREE compared to post-3.5 Ga Paleoarchean basalts and andesites, and form a compositionally suitable source for TTG. Enrichment in these basalts was not the result of crustal assimilation but was inherited from a mantle source that was less depleted than modern MORB-source and was enriched in recycled crustal components. We suggest that the formation of Paleoarchean TTG and of their voluminous mafic source regions reflects both a primitive stage in the thermal and compositional evolution of the mantle and a significant prehistory of crustal recycling.

Introduction

The processes that formed the Archean tonalite-trondhjemite-granodiorite (TTG) series that volumetrically dominates Archean crust (Martin, 1986, Martin, 1994) provide insights into how, when, and at what rate early continental crust formed. TTG suites comprise rocks with a high silica content (averaging > 68 wt.% SiO2) that are sodic (Na2O ~ 4.65; K2O/Na2O ~ 0.4) and form part of a group showing little or no increase in K2O/Na2O with increasing SiO2. These rocks typically have high La/Yb (~ 48) and low Yb (~ 0.6 ppm) (Martin et al., 2005), reflecting a garnet and/or amphibole residue, and high Sr, Eu and Ba reflecting the absence of plagioclase in the source. Their formation has been attributed to melting of subducted basaltic slabs or of overthickened mafic crust (e.g., Barker and Arth, Barker and Arth, 1976, Martin, 1986, Drummond and Defant, 1990, Rapp et al., 1991, Atherton and Petford, 1993).

However, while experimental studies and trace element modelling show that the source for true TTG was generally basaltic (e.g. Sen and Dunn, 1994, Rapp and Watson, 1995, Moyen and Stevens, 2006), these studies also impose several caveats. Most notable – but largely hidden in previous discussions of TTG petrogenesis – is that even TTG at the LILE-poor end of the compositional spectrum are too enriched in K, LILE, Th and LREE to be simply a result of melting of either MORB-like basaltic crust or of the slightly LILE- and LREE-enriched low-K tholeiites generally thought typical of Archean basaltic crust (e.g. Tarney and Jones, 1994, Kemp and Hawkesworth, 2003, Clemens et al., 2006, Champion and Smithies, 2007). These enrichments in TTG are not a result of fractional crystallisation, because there appears to be no systematic correlation between (for example) SiO2 or MgO, and LILE-, LREE-enrichments in most TTG suites (e.g., Martin, 1994, Clemens et al., 2006, Champion and Smithies, 2007, Moyen et al., 2007). More complex models involving magma mixing, assimilation or mingling are not only inconsistent with field evidence that shows that microgranular (or other) enclaves are very rare (Collins, 1993), but generally do not solve the requirement for LILE-enrichment.

These constraints on source composition are particularly important for Paleoarchean TTG that form a major component of the East Pilbara Terrane in the Pilbara Craton of northwestern Australia (Appendix Fig. 1), because these rocks appear to have resulted through melting of thickened mafic crust in a non-subduction environment (Smithies, 2000, Smithies et al., 2005, Champion and Smithies, 2007, Van Kranendonk et al., 2007a, Van Kranendonk et al., 2007b). This requires that source rocks with compositions uncharacteristic of most preserved Archean crust must once have formed a fundamental component of the supracrustal sequence.

In this paper, we describe a package of LILE-, Th- and LREE-enriched basalts and basaltic-andesites from the 3.53–3.51 Ga Coucal Formation in the stratigraphically lowest parts of the Warrawoona Group, Pilbara Craton. We show that > 3.5 Ga supracrustal packages like this were possibly the ultimate source of nearly all of the granites that volumetrically dominate this region. Implicit in this hypothesis is that such packages were more typical of the pre-3.5 Ga supracrustal sequences than the relatively LILE- and LREE- poor basalts that now dominate the Archean supracrustal record. We examine the implications that this hypothesis may have on models for early crustal growth of the Pilbara Craton. The existence of LREE-enriched basalts within the > 3.51 Ga Theespruit Formation, of the Barberton greenstone belt, South Africa (e.g. Hofmann and Harris, 2008) suggests that this hypothesis might also be relevant to other tracts of Paleoarchean crust.

Section snippets

Geological setting

The East Pilbara Terrane (Appendix Fig. 1) is one of the world's best-preserved Paleo- to Mesoarchean granite–greenstone terranes. The stratigraphically lowest Warrawoona Group is a 3.53–3.42 Ga volcanosedimentary succession, at least 12 km thick, dominated by pillowed tholeiitic and komatiitic basalt (see Van Kranendonk et al., 2007a, Van Kranendonk et al., 2007b). However, Pb- and Nd-isotopic data and inherited zircon age data indicate more than 200 m.y. of crustal pre-history (e.g. Van

Constraints on the genesis of Pilbara Paleoarchean TTG

Following Barker and Arth (1976), Champion and Smithies (2007) divided Paleoarchean TTG of the Pilbara Craton into two broad subgroups; a high Al2O3, Na2O, Sr, low Fe, Y, HREE (high-Al) subgroup, which forms the majority of rocks, and a low Al2O3, Na2O, Sr, high Fe, Y, HREE (low-Al) subgroup. Both are sodic with K2O/Na2O varying from < 0.25 to close to 1, with evidence for only limited K-enrichment. Granites of both groups range from slightly to strongly LREE-enriched when normalised to

Trace-element models of potential TTG source regions

Trace-element modelling was undertaken to test whether the Coonterunah F2 basalts and basaltic-andesites of the c. 3.51 Ga Coucal Formation formed a compositionally more viable source to the Pilbara TTG than typical Archean mafic crust (see also Champion and Smithies, 2007). As an example of typical Archean basalt, we have used the average compositions of tholeiitic basalts from the c. 3.46 Ga Apex Basalt (Table 1), which stratigraphically overlies the Coonterunah Subgroup. Melt compositions

Discussion

The voluminous TTG and TTG-like rocks of the East Pilbara Terrane mainly formed between 3.5 and 3.45 Ga, from a source older than 3.5 Ga (McCulloch, 1987, Champion and Smithies, 2001, Smithies et al., 2003) (Fig. 3). In the case of the East Pilbara Terrane, the fact that this TTG is neither mantle-derived nor subduction-related (i.e. are not partly or wholly mantle- or slab-melts) indicates that their enriched basaltic source must have formed a major part of the pre-3.5 Ga supracrustal sequence.

An infracrustal model for Paleoarchean crustal evolution

From studies of the O- and Hf-isotopic compositions of zircons, Kemp et al. (2006) concluded that the processes controlling the growth of ancient continental growth were intrinsically different from those that controlled crustal differentiation. The initial formation of Pilbara TTG and subsequent, protracted reworking (differentiation) of the resulting Pilbara crust to form high-K granites forms a spectacular example of these decoupled processes.

Our model for the generation of early Pilbara

Conclusions

  • The geochemistry of TTG requires that the source was more enriched than typical Archean basalt. In the case of the voluminous Paleoarchean (< 3.5 Ga) TTG of the Pilbara Craton, this source was older (> 3.5 Ga) than all but the oldest preserved fragment of the supracrustal sequence. These two important constraints are satisfied by the discovery that this oldest preserved fragment comprises 3.51 Ga basalts and andesites enriched in K, LILE, Th and LREE (Coonterunah F2-type rocks) to the extent that

Acknowledgements

We would like to thank Paul Morris and Chris Kirkland for comments on various drafts. Arthur Hickman is thanked for his support throughout our work in the Pilbara. Journal reviews by Tony Kemp and Chris Hawkesworth further significantly improved the manuscript. Suzanne Dowsett and Mike Prause are thanked for drafting the figures. Published with the permission of the Executive Director, Geological Survey of Western Australia and the Chief Executive Officer, Geoscience Australia.

References (55)

  • GreenM.G. et al.

    Growth and recycling of early Archaean continental crust: geochemical evidence from the Coonterunah and Warrawoona Groups, Pilbara Craton, Australia

    Tectonophysics

    (2000)
  • HawkesworthC.J. et al.

    The differentiation and rates of generation of the continental crust

    Chem. Geol.

    (2006)
  • HofmannA. et al.

    Silica alteration zones in the Barberton greenstone belt: a window into seafloor processes 3.5–3.3 Ga ago

    Chem. Geol.

    (2008)
  • JahnB.-M. et al.

    REE geochemistry and isotopic data of Archean silicic volcanics and granitoids from the Pilbara Block, Western Australia: implications for the early crustal evolution

    Geochim. et Cosmochim. Acta

    (1981)
  • KempA.I.S. et al.

    Granitic perspectives on the generation and secular evolution of the continental crust

  • KerrichR. et al.

    Trace element systematics of Mg- to Fe-tholeiitic basalt suites of the Superior Province: implications for Archean mantle reservoirs and greenstone belt genesis

    Lithos

    (1999)
  • KleinM. et al.

    Partitioning of high-field strength and rare-earth elements between amphibole and quartz-dioritic to tonalitic melts: and experimental study

    Chem. Geol.

    (1997)
  • KleinM. et al.

    Experimental partitioning of high field strength and rare earth elements between clinopyroxene and garnet in andesitic to tonalitic systems

    Geochim. et Cosmochim. Acta

    (2000)
  • KlemmeS. et al.

    Experimental constraints on major and trace element partitioning during partial melting of eclogite

    Geochim. et Cosmochim. Acta,

    (2002)
  • MartinH.

    The Archean grey gneiss and the genesis of continental crust

  • MartinH. et al.

    An overview of adakite, tonalite–trondhjemite–granodiorite (TTG), and sanukitoid: relationships and some implications for crustal evolution

    Lithos

    (2005)
  • OhtaH. et al.

    Field occurrence, geochemistry and petrogenesis of the Archaean mid-oceanic ridge basalts(A-MORBs) of the Cleaverville area, Pilbara craton, Western Australia

    Lithos

    (1996)
  • RappR.P. et al.

    Partial melting of amphibolite/eclogite and the origin of Archean trondhjemites and tonalites

    Precamb. Res.

    (1991)
  • SmithiesR.H.

    The Archaean tonalite–trondhjemite–granodiorite (TTG) series is not an analogue of Cenozoic adakite

    Earth Planet. Sci. Lett.

    (2000)
  • SmithiesR.H. et al.

    Formation of Earth's early Archaean continental crust

    Precamb. Res.

    (2003)
  • SmithiesR.H. et al.

    It started with a plume — early Archaean basaltic proto-continental crust

    Earth Planet. Sci. Lett.

    (2005)
  • AthertonM.P. et al.

    Generation of sodium-rich magmas from newly underplated basaltic crust

    Nature

    (1993)
  • Cited by (251)

    View all citing articles on Scopus
    View full text