The primary driving force behind present-day structural, magmatic, sedimentary and metamorphic processes is plate tectonics, resulting from the flow of matter and energy between the lithosphere and mantle along divergent, convergent, and transform plate boundaries. Operation of plate tectonics and eruption of hot spot lavas from mantle plumes, stemming from the core-mantle boundary, appears to be coupled in that subducting oceanic plates pile up at the core-mantle boundary and then rise as buoyant plumes to feed hot spot volcanoes (see Hofmann, 1997; Burke, 2011). How far back in Earth history were these geological processes driven by plate tectonics? Were Archean geological processes dominated by density-driven, vertical crustal overturns and diapirs, without modern analogs? How did Archean continents grow? How did Archean oceanic crust form? Did Archean oceanic crust recycle into the mantle at subduction zones? These questions remain controversial. Excellent exposures and a prolonged geological record spanning 3.85–2.5 Ga in the Archean craton of southwestern Greenland provide a unique opportunity to test hypotheses proposed for the early evolution of Earth.

This craton consists mainly of Eoarchean to Neoarchean (ca. 3.8–2.7 Ga) metamorphosed tonalite-trondhjemite-granodiorite suites (TTGs), amphibolite-dominated greenstone belts (also known as supracrustal belts), and layered anorthosite complexes (Friend and Nutman, 2005; Windley and Garde, 2009; Polat et al., 2011). These all underwent several phases of deformation and greenschist to granulite facies metamorphism. All geological data are consistent with formation through accretion of oceanic island arcs and continental blocks (Friend and Nutman, 2005; Polat et al., 2011). The remnants of oceans closed during accretion are marked by abundant pillow lavas and ultramafic rocks in the Isua, Ivisaartoq-Ujarassuit and Tartoq greenstone belts (Polat et al., 2011; Kisters et al., 2012), which occur as fault-bounded lithotectonic assemblages, intruded by contemporaneous TTGs. The structural and lithological characteristics of these belts are comparable to those of Phanerozoic accretionary complexes (see Şengör et al., 1993).

Despite some lithological differences between Archean and Phanerozoic terranes (i.e., higher abundance of komatiites, TTG intrusions, layered anorthosites, banded iron formations in Archean terranes), signatures of plate tectonic processes are well preserved in the former (de Wit, 1998; Kerrich and Polat, 2006; Percival et al., 2006; Burke, 2011). The lithological differences likely resulted from higher mantle temperatures and an oxygen-poor atmosphere in the Archean. Plate tectonics appear to have shaped the evolution of Earth since the formation of the oldest known rocks (Burke, 2011).

Modern basalts from different tectonic settings, including mid-oceanic ridge basalt (MORB), oceanic island arc basalt (IAB), ocean island basalt (OIB), and oceanic plateau basalt (OPB), have different trace element signatures (see Pearce and Peate, 1995; Hofmann, 1997; Kerr, 2003). This geochemical behavior likely also prevailed in the Archean, given that certain groups of elements behave consistently in specific petrogenetic settings (Polat and Kerrich, 2006). The chemical composition of basalts in Archean greenstone belts carries the fingerprints of their geodynamic settings (Puchtel et al., 1999; Polat and Kerrich, 2006), suggesting that they are remnants of Archean oceanic crust. These fragments of oceanic crust were assembled at convergent plate boundaries and intruded predominantly by TTGs, forming Archean greenstone belt-granitoid terranes (Kusky and Polat, 1999; Burke, 2011).

Many geochemical studies on Archean volcanic rocks used primitive- or MORB-normalized trace element diagrams to interpret their geodynamic setting (see Puchtel et al., 1999; O’Neil et al., 2011). On N-MORB-normalized diagrams, the Eoarchean Isua, and the Mesoarchean Ivisaartoq-Ujarassuit and Fiskenæsset basaltic volcanic rocks in southwestern Greenland are characterized by large negative Nb anomalies relative to Th and La (Polat et al., 2011), indicating that hydrous fluids and/or melts originating from the subducted oceanic slabs metasomatized their mantle sources. These volcanic rocks plot in the arc field on an Nb/Yb–Th/Yb proxy diagram, as do Phanerozoic supra-subduction zone ophiolites and oceanic island arc basalts (Pearce, 2008; Polat et al., 2011).

Archean oceanic island basalts could have not been generated without subduction of oceanic crust, but Archean basalts with unambiguous MORB-like signatures are extremely rare (Polat and Kerrich, 2006). Basalts with OIB-like trace element characteristics have so far been recognized only in Neoarchean Wawa greenstone belts (Polat et al., 1999). The scarcity of MORB, OIB, and mélanges in Archean terranes, together with the lack of intact ophiolites and blueschists, have been used to argue against the interpretation of Nb-depleted Archean rocks as subduction zone products, and the formation of Archean continental crust at convergent plate margins.

In this issue of Geology, Jenner et al. (2013, p. 327–330) report OIB-like trace element patterns for Eoarchean (ca. 3.75 Ga) basaltic amphibolites from Innersuartuut Island, southwestern Greenland. The authors convincingly show that crustal contamination and metamorphic alteration can be ruled out, providing geochemical evidence for the existence of hot spots as early as 3.75 Ga. These are the oldest known, plume-derived intra-plate volcanic rocks and appear to have been accreted to a convergent plate margin and intruded by TTGs. In contrast to other Archean basaltic rocks in southwestern Greenland, which plot in the IAB field on a log-transformed La/Th–Nb/Th–Sm/Th–Yb/Th discrimination diagram, the Innersuartuut Island basaltic amphibolites plot predominantly in the OIB field, supporting the interpretation of these rocks as OIB (Figs. 1A and 1B). These rocks probably represent only a tiny relic of a large ocean island fragment that appears to have been mostly destroyed by deformation, erosion and TTG intrusion following its accretion.

The presence of Archean IAB and OIB in southwestern Greenland implies the generation and destruction of oceanic crust at spreading centers and subduction zones, respectively, in those ancient times. The question then arises: Why has Archean MORB not been extensively preserved in greenstone belts? Given its negative buoyancy, almost all Archean oceanic crust generated at spreading centers probably was recycled into the mantle by subduction, similar to its modern counterpart. In contrast, thicker, buoyant oceanic plateaus, island arcs and fore-arcs would have accreted to convergent plate boundaries (Kerrich and Polat, 2006). Following their accretion, oceanic islands arcs and plateaus were multiply deformed, metamorphosed, and intruded by TTGs, becoming part of Archean continental crust. Oceanic island arcs and plateaus thus have been preferentially preserved.

Despite the lack of direct evidence for Archean MORB in most greenstone belts, the geochemical fingerprints of these rocks are found in subduction-derived Eoarchean to Mesoarchean boninites, picrites, and tholeiitic basalts in southwestern Greenland (Polat et al., 2011). On the log-transformed La/Th–Nb/Th–Sm/Th–Yb/Th diagram of Agrawal et al. (2008), these mafic to ultramafic volcanic rocks display a trend from MORB to IAB (Figs. 1C and 1D), providing indirect evidence for the involvement of MORB mantle source in their petrogenesis. This trend is attributed to the conversion of the Archean depleted upper mantle, the source of MORB, to sub-arc mantle wedges by the development of intra-oceanic subduction systems (Polat et al., 2011). Accordingly, subduction zone trace element compositions in southwestern Greenland Archean volcanic rocks are interpreted as reflecting the enrichment of depleted upper mantle by subduction-derived melts and fluids. Collectively, the presence of IAB and OIB, along with the traces of MORB, in Eoarchean to Mesoarchean greenstone belts in southwestern Greenland is consistent with the existence of diverse geodynamic settings (e.g., arcs, back-arcs, mid-ocean ridges, and ocean islands) in the early Earth.

Future studies on mafic to ultramafic volcanic rocks in Archean cratons should focus on the recognition of petrogenetic links between trace element systematics and geodynamics. Geochemical data and petrological constraints will allow students of Archean geology to determine the remnants of diverse geodynamic settings in Archean greenstone belts and to provide new criteria to test the competing uniformitarian and non-uniformitarian hypotheses for the early Earth.