Introduction:
When, how, and why plate tectonics began on Earth remain among the most important unresolved questions in plate tectonic theory (e.g., Bauer et al., 2020; Beall et al., 2018; Brown and Johnson, 2018; Condie and Puetz, 2019; Hansen, 2007; Harrison, 2009; Korenaga, 2011; Nutman et al., 2020; Stern, 2008; Tang et al., 2020). Investigations of plate tectonic initiation have significant implications for questions associated with the evolution of early terrestrial planets, including (1) whether early Earth experienced any pre-plate tectonic global geodynamics/cooling after the magma ocean stage (e.g., Bédard, 2018; Collins et al., 1998; Lenardic, 2018; Moore and Webb, 2013; O’Neill and Debaille, 2014); and (2) why other terrestrial planets in the solar system appear to lack plate tectonic records (e.g., Moore et al., 2017; Stern et al., 2017; cf. Yin, 2012a; Yin, 2012b).
Many proposed signals for the initiation or early operation of plate tectonics on Earth are controversial due to the issue of non-uniqueness. For instance, the origin of Hadean zircons from the Jack Hills of western Australia have been contrastingly interpreted as (1) detrital crystals from felsic magmas generated by ~4.3 Ga plate subduction (Harrison, 2009; Hopkins et al., 2008); (2) zircons crystallized via impact heating and ejecta sheet burial (Marchi et al., 2014) or (3) low pressure melting of Hadean mafic crustal materials (Reimink et al., 2020). Similarly, researchers continue to debate whether the presence of Archean low-Ti mafic lava (also termed as boninite or boninitic basalts) must indicate subduction initiation as early as ~3.7 Ga (cf. Pearce and Reagan, 2019; Polat and Hofmann, 2003). Another example is how a ~3.2 Ga shift in zircon Hf-isotope signatures has been variably interpreted to indicate the onset of plate tectonics (Næraa et al., 2012) or enhanced mantle melting during a proposed Earth’s thermal peak (Kirkland et al., 2021). Due to these equivocal interpretations, the initiation of plate tectonics has been suggested to be ≤3.2 Ga using geological records that are generally considered unique to plate tectonics (e.g., paired metamorphic belts, ultra-high pressure terranes, and passive margins) (e.g., Brown and Johnson, 2018; Cawood et al., 2018; Stern, 2008; cf. Bauer et al., 2020; Foley et al., 2014; Harrison, 2009; Korenaga, 2011; Nutman et al., 2020). The ≤3.2 Ga onset of plate tectonics requires early Earth tectonic evolution to be non-uniformitarian, involving some form of single-plate stagnant-lid tectonics (e.g., Bédard, 2018; Collins et al., 1998; Moore and Webb, 2013).
One proposed signal of early plate tectonics is the preservation of phaneritic ultramafic rocks in Eo- and Paleoarchean terranes. However, the issue of non-uniqueness also extends to their interpretations. In the Eoarchean Isua supracrustal belt and adjacent meta-tonalite bodies exposed in southwestern Greenland (Fig. 1a ), some dunites and harzburgites have been interpreted to represent melt-depleted mantle rocks that were emplaced on top of crustal rocks during the Eoarchean plate tectonic subduction (e.g., Friend and Nutman, 2011; Nutman et al., 2020; Van de Löcht et al., 2018), similar to how modern ophiolitic ultramafic rocks are preserved in collisional massifs (e.g., Boudier et al., 1988; Lundeen, 1978; Wal and Vissers, 1993). In contrast, Szilas et al. (2015) argue that dunites and harzburgites in the Isua supracrustal belt can be interpreted as crustal cumulates based on their geochemical signatures. Crustal cumulates are not exclusive to plate tectonics, and have been used to explain the emplacement of other phaneritic ultramafic rocks in Eo- and Paleoarchean terranes. Examples include phaneritic ultramafic rocks in the Eoarchean Tussapp ultramafic complex of southwestern Greenland (McIntyre et al., 2019), the Paleoarchean East Pilbara Terrane of northwestern Australia (e.g., Smithies et al., 2007), and the Paleoarchean Barberton Greenstone Belt of South Africa (e.g., Byerly et al., 2019). Therefore, ultramafic rocks in the Isua supracrustal belt potentially formed in a different tectonic setting compared to those of other early Archean terranes. Because all early Archean terranes preserve voluminous tonalite-trondhjemite-granodiorite (TTG) suites surrounded by deformed, dominantly mafic supracrustal belts (e.g.,Fig. 1 ; also Condie, 2019), a different origin for the Isua supracrustal belt may be an artifact of interpretive non-uniqueness. If the phaneritic ultramafic rocks of the Isua supracrustal belt can be similarly interpreted to have cumulate or volcanic origins (which is questioned by many recent studies, see section 2 for a review), then these rocks cannot be used as unequivocal indicators of plate tectonics.
This contribution explores the origins of Isua ultramafic rocks via analysis of new and published geochemical and petrological findings, including comparative analysis of the key Isua rocks and similar rocks from settings considered representative of hot stagnant-lid tectonics [In this study, we follow tectonic taxonomy from Lenardic (2018)]. The Paleoarchean geology of the East Pilbara Terrane is widely accepted as representing hot stagnant-lid tectonics (Hickman, 2021; Johnson et al., 2014; Smithies et al., 2007, 2021; Van Kranendonk et al., 2004, 2007); Pilbara ultramafic samples are also investigated in this study (Fig. 1b ) as examples of ultramafic rocks from non-plate tectonic regimes. We also compare the petrology and geochemistry of Isua ultramafic rocks with compiled (1) ultramafic cumulate rocks; (2) modelled ultramafic cumulate rocks; (3) melt-depleted mantle rocks from plate tectonic settings; and (4) modelled melt-depleted mantle rocks. We then examine whether the generation of Isua and Pilbara ultramafic rocks is compatible with the predictions of hot stagnant-lid tectonics. Our findings help to evaluate whether plate tectonics is indeed required to explain the Eoarchean assembly of the Isua supracrustal belt.