Onset of plate tectonics by the Eoarchean
Graphical abstract
Introduction
There have been two types of orogens, accretionary and collisional, throughout Earth history (Windley and Condie, 1992, Windley, 1995, Şengör and Natal’in, 1996a) typically expressed as low-grade greenstone-granite belts and high-grade granulite-gneiss belts as defined by Windley and Bridgwater (1971), and most recently by hafnium isotope data in terms of external and internal orogens (Collins et al., 2011).
Accretionary orogens (Cawood et al., 2009) form by the accretion of diverse components from the ocean floor, such as oceanic crust formed at mid-ocean ridges and associated hydrothermal vent deposits, epidosites, ophiolites, ocean plate stratigraphic sections, seamounts, and oceanic plateaus, as well as accretionary prisms, and island arcs with their forearcs and back-arcs. However, we only see today minor relicts of these components in accretionary orogens, particularly because of preferential subduction erosion during off-scraping/underplating/subduction type accretion (Stern, 2011, Scholl and von Huene, 2007, Scholl and von Huene, 2009) as in the subducting margins of eastern Asia and other parts of the circum-Pacific (Kimura and Ludden, 1995, Kusky and Young, 1999, Wakita et al., 2013) where ocean-derived ultramafic rocks, gabbros, sheeted dykes and most of the basalts are typically subducted, leaving only a small portion of the upper basalts and the overlying sediments to be accreted. This is the main reason why the largest subduction-accretion orogen in Japan contains only one ocean-derived, fore-arc ophiolite (the 580 Ma O-eyama ophiolite), but over 1000 small accreted fragments of seamounts, oceanic crust, oceanic plateaus, and arcs (Isozaki, 1996) appropriately termed ‘ophirags’ by Şengör and Natal’in, 1996b) and Şengör and Natal'in (2004). These fragmentary relations in Phanerozoic accretionary orogens bear heavily on the occurrence and preservation of comparable relicts of oceanic lithosphere in Archean accretionary orogens/belts (e.g., Kusky, 1989, Kusky, 2020, Kusky and Polat, 1999, Kusky et al., 2001, Kusky et al., 2018).
In contrast, continent–continent collisional orogens typified by the Alpine-Himalayan Chain form by the collision of two large continental blocks like India/Afro-Arabia and Eurasia (Treloar and Searle, 2019). Such collisional orogens can be traced back in Earth history with high degrees of confidence only to the Paleoproterozoic, when there is substantial evidence for the first existence of major continents (Wise et al., 1974, Windley, 1977, Hynes, 2001, Kasting et al., 2006, Flament et al., 2008, Pope et al., 2012, Johnson and Wing, 2020) and perhaps even the first supercontinent (Pehrsson et al., 2013). The Wilson Cycle, as traditionally viewed (Wilson et al., 2019), could only have started to operate when there were continental blocks large enough to rift, enabling the separation and drift of individual continents, which were then susceptible to subduction on their margins leading to the formation of Andean-type or active magmatic, granite-dominated orogens, and eventually to collision with another continent (Şengör, 1990, Wilson et al., 2019).
The essential point arising out of the above differences in understanding of orogens in Earth history is that we accept the geological evidence for the presence of accretionary orogens throughout the geological record into the Eoarchean (Condie, 2007, Cawood et al., 2009, Kusky et al., 2018), and therefore for the existence of plate tectonics back to at least 3.9–4.0 Ga or beyond (Harrison, 2009, Harrison, 2020, Turner et al., 2020), because accretionary orogens necessarily form by subduction and accretion via horizontal motion of plates with associated sea-floor spreading, transform motions between plates, and convergence and subduction. In contrast, since no large continental blocks have been demonstrated to have formed until perhaps the latest Archean (Pehrsson et al., 2013), the Wilson Cycle of rifting and re-assembling of large continents was only able to start near the Archean-Proterozoic boundary and therefore large continent–continent collisional orogens were only able to initially form to a significant degree in the Paleoproterozoic. These relations were all described and explained at some length in The Evolving Continents (Windley, 1995).
Today it is well recognized that subduction-accretion-erosion has played a major role in the differential preservation of the geological record (Kimura and Ludden, 1995, Isozaki, 1996, Scholl and von Huene, 2007, Scholl and von Huene, 2009, Stern, 2011). However, the possibility that there has been so much removal of the geological record that it would make it impossible for us to make a viable assessment of the differences seen today across the Archean-Proterozoic boundary palls into almost insignificance in comparison with the vast differences between the geological record of the late Archean and the early Proterozoic (Windley, 1995, Condie and O’Neill, 2010, and hundreds of other publications); as Goodwin (1996) put it “The early Proterozoic era represents a major watershed in Earth’s crustal history.” Accordingly, models that do not take account of the details and implications of the structural characteristics, field relationships, lithotectonic associations, and derivative geochemical-isotopic character of accretionary orogens, compared with Wilson Cycle orogens, only provide a partial analysis of the available data from the preserved record, and must be abjured. These relations and their consequences will become apparent in the detailed documentation below.
While applying a uniformitarian view in the interpretation of early Archean rock sequences, or applying the null hypothesis in which it is assumed that plate tectonics operated until it can be “proven guilty through negligence” and shown to not have operated (e.g., Kusky et al., 2018, Harrison, 2020), one must also be astutely aware of the effects of secular cooling of the planet's interior and changes in atmospheric chemistry on the constituents and composition of rocks from ancient terranes (Wang et al., 2020). While some models (e.g., Dhuime et al., 2015, Tang et al., 2016, Chen et al., 2019) have assumed a significantly more mafic early continental crust, and others have assumed from that, that plate tectonic processes were not operating early on, Keller and Harrison (2020) recently demonstrated that incompatible and compatible elements that are commonly used as proxies to estimate crustal compositions have significantly different concentrations and relationships, at constant silica values, for progressively higher mantle temperatures with time. Thus, studies that use trace element ratios as proxies to infer drastic changes in composition of continental crust with time (e.g., Dhuime et al., 2015) are “significantly undermined” (Keller and Harrison, 2020) by not accounting for secular changes in mantle temperature, because the same results can be produced at higher temperatures but with constant Si contents. At the same time, there remains a significant body of evidence that the early crust was slightly more enriched in mafic elements (Mg, Fe) and Cr, and that these have gradually evolved through time to their current values, whereas K2O has gradually increased with time (Condie, 1993). Other secular changes on the planet including atmospheric oxidation and biological activity also have significant feedback effects that can affect estimates of early crustal composition and volume (Harrison, 2020, Guo and Korenaga, 2020).
The Earth’s continents contain the fragmentary record of its tectonic evolution and crustal growth (not the same) from 4.0 Ga to the Present, the first third of which constitutes the Archean Eon that extends from ca. 4.0 Ga to 2.5 Ga, preceded by the Hadean, for which there is no preserved rock record (Harrison, 2020). Multi-disciplinary estimates suggest that ca. 85% of the present-day volume of continental crust had formed by the end of the Archean at an average growth rate of 11.7 Pg/a (Dewey and Windley, 1981) or 65% by 3.0 Ga (Cawood et al., 2013), which is comparable to that of the present-day crust (Scholl and von Huene, 2009). When considering models of crustal growth and composition however, we must consider the possibility of selective preservation of terranes that have compositions that would resist deformation, whereas those prone to deformation such as having high concentrations of heat-producing elements or being located along margins that could be subducted, or on the upper plates of collisions, may be preferentially under-represented (Morgan, 1985, Morgan, 1989, Harrison, 2009, Kusky et al., 2018, Hildebrand et al., 2018).
With the above in mind, we note that the Archean continental crust preserves segments of 1: upper crust represented by volcanic-dominated, low-grade, greenstone-granite belts (Williams et al., 1991, Polat and Kerrich, 2001a, Kusky and Polat, 1999, Percival et al., 2012), which show evidence of deposition at or near the ambient sea-level in the Neoarchean of the Superior and Slave Provinces in Canada (Percival and Helmstaedt, 2004), but at abyssal ocean depths (2.5–3.8 km) in others such as the Paleoarchean of the Eastern Pilbara (Kitajima et al., 2001a), and 1.7–4 km in the Paleoarchean (3.56–3.33 Ga) Barberton belt of South Africa (de Wit et al., 2011), and 2: lower crust represented by TTG-dominated granulite-gneiss belts that were metamorphosed in the high amphibolite-granulite facies, as in West Greenland (Nutman et al., 2004a, Windley and Garde, 2009), the North China Craton (Kusky et al., 2016), and the Barberton Greenstone Belt (Furnes et al., 2013, de Wit et al., 2018). Sedimentary rocks are present in the Eoarchean belts but form a subordinate component, including generally chert, banded iron formation (BIF), sandstones, conglomerates, and mafic shales (Eriksson et al., 2007). In the 4.0–3.2 Ga greenstone belts, the volcanic units greatly predominate in abundance, but occur with associated generally mafic volcaniclastic sediments, along with BIF's, cherts, pelites, syn-orogenic turbidites, conglomerates, sandstones, and even less-common carbonate remnants, and stromatolitic evaporites (Fedo et al., 2001). Mesoarchean-Neoarchean upper crustal greenstone belts show a greater preponderance of distal to proximal graywacke-turbidite deposits such as comprise much of the Slave Province (Kusky, 1989, Kusky, 1991, Kusky and De Paor, 1991), along with the appearance of high-gradient alluvial fan deposits, low-sinuosity river deposits, and shallow marine deposits formed near storm base (Windley, 1995, Eriksson et al., 2007, Fedo et al., 2001) with the oldest well-preserved tidalites from the 3.1 Ga Mozaan Group of the Pongola rift on the Kaavvaal Craton (Burke et al., 1985, Luskin et al., 2019). The first widespread appearance of passive margin type deposits are preserved from the Neoarchean of the Superior, Slave, Kaapvaal, and Pilbara cratons (Bradley, 2008). Some generally low-grade cratons such as the Superior Province also contain belts of high-grade gneisses as in the Pikwitonei belt and Kapuskasing structural zone (Bell, 1978, Percival et al., 2012), and conversely some high-grade Archean cratons as in West Greenland (Windley and Garde, 2009) contain strips of amphibolite facies, volcanic-dominated, upper crustal rocks such as the Eoarchean Isua belt (Komiya et al., 2004, Nutman and Friend, 2009), and the Mesoarchean Ivisaartoq belt (Polat et al., 2007), the Qussuk, Bjorneøen and Storø belts (Szilas et al., 2016), and the Tartoq belt (Kisters et al., 2012, Polat et al., 2016; Szilas et al., 2013a). These dual relations are important because they show us how the upper and lower crust were intercalated by sub-horizontal thrusting, and also how both crustal levels contain the same diagnostic lithologies, although in vastly different quantities.
A distinctive lithology in many Archean granulite-gneiss belts, as in West Greenland (e.g. Windley and Garde, 2009), South India (Subramaniam, 1956), South Canada (Bell, 1978), West Australia (Myers, 1988), and the Superior Province (Fig. 1), occurs in strips up to 2 km wide of layered and cumulate megacrystic calcic anorthosite-leucogabbro-gabbro-ultramafic complexes, which are not considered in most recent Archean crustal-tectonic syntheses (e.g. Cawood et al., 2013, Cawood et al., 2018); and yet they are considerable in volume and importance, amounting for example to at least 500 km strike-length in West Greenland. Eoarchean to Neoarchean anorthosites share the geochemical and petrological characteristics of Tethyan ophiolite-hosted anorthosites, suggesting that Archean anorthosite-bearing layered intrusions largely formed at intra-oceanic convergent plate margins (Sotiriou and Polat, 2020).
Section snippets
Models of Archean tectonics
There are many models to explain the formation of Early Archean tectonic belts (Fig. 1), which we will discuss later, but here we outline the two most important trends in thought:
Nulliak, Labrador, Canada
Following early work organised by Ken Collerson on the early Archean rocks in the Saglek area of northern Labrador (Bridgwater et al., 1975, Collerson et al., 1976, Komiya et al., 2015; Komiya et al., 2017) made a very detailed and comprehensive study, in five different areas, of the Nulliak volcano-sedimentary belt (Fig. 1, Fig. 2). Protoliths of the Uivak orthogneisses, which have U-Pb zircon ages of >3.95 Ga (Shimojo et al., 2013, Shimojo et al., 2016), were intruded into and
Components and rock associations
Here we describe the most important components that make-up Eoarchean orogens in terms of understanding their origin, and we discuss whether the components and structures are consistent with a plate tectonic origin, or require the operation of some alternative mechanism of planetary heat loss in the Eoarchean.
Interpretation of the trace element compositions in Archean igneous rocks
The geochemical characteristics of Eoarchean to Neoarchean volcanic and intrusive rocks and their geodynamic implications were discussed in detail in many recent studies (Dilek and Furnes, 2011, Polat et al., 2011a, Arndt, 2013, Turner et al., 2014, Furnes et al., 2014, Blichert-Toft et al., 2015, Wang et al., 2015, Grosch and Slama, 2017, Hildebrand et al., 2018; Hastie and Fitton, 2019; Nakamura et al., 2020, Van de Löcht et al., 2020, Sobolev et al., 2019, Lowrey et al., 2019, Gamal El Dien
Archean orogens and modern analogues
If one is going to consider tectonic development in terms of ancient and modern analogues, it is essential to have a clear view of the variability of modern tectonic settings, and which are most applicable to serve as analogues of possible Archean tectonic environments. This of course must be done with an appreciation of any changes caused by the secular cooling of the planet, and other changes in rock types influenced by changes in biology, oceanic and atmospheric chemistry, proximity of the
Conclusions
The period of transition in the geochemical record of geodynamic growth of continental crust and mantle at ca. 3.2–3.0 Ga (Fig. 18) was the result of the change from the dominant melting of the mantle source of juvenile oceanic and arc magmas from ca. >4.0 to 3.3 Ga to the rise of melting of enriched mantle wedges and of increasing active continental margins from ca. 3.2 to 3.0 Ga, which continued to well beyond the end of the Archean. As Halla et al. (2017) pointed out, these changes were
Funding
This work was supported by the National Natural Science Foundation of China (Grant Numbers: 41890834, 91755213, 41961144020, and 41602234), Chinese Ministry of Education (BP0719022), the Chinese Academy of Sciences (QYZDY-SSWDQC017), the MOST Special Fund (MSF-GPMR02-3), and the Open Fund (GPMR201704) of the State Key Laboratory of Geological Processes and Mineral Resources, China University of Geosciences (Wuhan), and the Fundamental Research Fund (CUGL 180406) from the China University of
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
We wish to express our thanks to our many colleagues with whom, over the years, we have spent many months in the field mapping different Archean terranes, some of which are discussed in this manuscript. This paper is dedicated to the memory of Alfred Kröner, whose seminal contributions to the field of Archean geology and geochronology paved the way for many advances, debates, and theories on the development of the early Earth. Without his high-precision zircon geochronology from numerous
References (468)
- et al.
Hydrothermal alteration of upper oceanic crust formed at a fast-spreading ridge: mineral, chemical, and isotopic evidence from ODP Site 801
Chem. Geol.
(2003) - et al.
Remnants of an early Archean (>3.75 Ga) sea-floor, hydrothermal system in the Isua greenstone belt
Precambr. Res.
(2001) - et al.
The K/U ratio of the silicate Earth: Insights into mantle composition, structure and thermal evolution
Earth Planet. Sci. Lett.
(2009) - et al.
Crustal evolution and the temporality of anorthosites
Earth Sci. Rev.
(2017) The Messina Layered Intrusion, Limpopo Belt, South Africa: an example of in-situ contamination of an Archean anorthositic complex by continental crust
Precambr. Res.
(1996)A catalytic delamination-driven model for coupled genesis of Archaean crust and sub-continental lithospheric mantle
Geochim. Cosmochim. Acta
(2006)How many arcs can dance on the head of a plume? A ‘Comment’on: a critical assessment of Neoarchen ‘plume only’ geodynamics: evidence from the Superior Province, by Derek Wyman
Precambr. Res.
(2013)Stagnant lids and mantle overturns: Implications for Archaean tectonics, magmagenesis, crustal growth, mantle evolution, and the start of plate tectonics
Geosci. Front.
(2018)From the LIPS of a serial killer: endogenic retardation of biological evolution on unstable stagnant-lid planets
Planet. Space Sci.
(2020)- et al.
The hunting of the snArc
Precambr. Res.
(2013)