How and when plate tectonics initiated remain uncertain. In part, this is because many signals that have been interpreted as diagnostic of plate tectonics can be alternatively explained via hot stagnant-lid tectonics. One such signal involves early Archean phaneritic ultramafic rocks. In the Eoarchean Isua supracrustal belt of southwestern Greenland, some ultramafic rocks have been interpreted as tectonically-exhumed mantle during Eoarchean subduction. To explore whether all Archean phaneritic ultramafic rocks originated as cumulate and/or komatiite – i.e., without requiring plate tectonics – we examined the petrology and geochemistry of such rocks in the Isua supracrustal belt and the Paleoarchean East Pilbara Terrane of northwestern Australia, with Pilbara ultramafic rocks interpreted as representative of rocks from non-plate tectonic settings. We found that Pilbara ultramafic samples have relict cumulate textures, relative enrichment of whole-rock Os, Ir, and Ru versus Pt and Pd, and spinel with variable TiO2, relatively consistent Cr#, and variable and low Mg#. Similar geochemical characteristics also occur in variably altered Isua ultramafic rocks. We show that Isua and Pilbara ultramafic rocks should have interacted with low Pt and Pd melts generated by sequestration of Pd and Pt into sulphide and/or alloy during magma generation or crystallization. Such melts cannot have interacted with a mantle wedge. Furthermore, altered mantle rocks and altered cumulates could have similar rock textures and whole-rock geochemistry such that they may not distinguish mantle from cumulate. Our findings suggest that depleted mantle interpretations are not consistent with geochemistry and/or rock textures obtained from Isua and Pilbara ultramafic rocks. Instead, cumulate textures of Pilbara samples, whole-rock Pt and Pd concentrations, and spinel geochemistry of Isua and Pilbara ultramafic rocks support cumulate origins and metasomatism involving co-genetic melts that formed in hot stagnant-lid settings. Collectively, these findings permit ≤ 3.2 Ga initiation of plate tectonics on Earth.

The Xuefengshan tectonic belt in South China is generally regarded as the collisional zone between the Yangtze and Cathaysia blocks in the Neoproterozoic. Recent studies suggest a buried Paleoproterozoic orogenic belt in the Xuefengshan belt, but its distribution and deep structure are still not clear. The 2.4-km-deep Xuefengshan pilot borehole in the Yuanma Basin penetrates the unconformity between the Cretaceous and Cambrian strata and reveals the Neoproterozoic Nanhua rifting formations. By employing the Kirchhoff prestack depth migration, we obtain the SE-trending depth migrated sections of two 30-km-long seismic profiles at the drilling site and a 550-km-long SinoProbe seismic profile across the Xuefengshan belt. The folded strong reflectors beneath the seismically transparent Neoproterozoic strata confirm the existence of the Paleoproterozoic orogen in the chevron-type syncline zone and the Xuefengshan belt. The Xuefengshan belt was then reactivated by collision between the Yangtze Block and the Wuyi terrane in 860–830 Ma, the intracontinental rifting in 820–690 Ma, and Early Paleozoic and Mesozoic intracontinental orogeny. Subduction of the Izanagi Plate beneath South China triggered crustal shortening in the eastern Yangtze thrust-fold belt in the Middle Jurassic, which was accommodated by a SE-dipping décollement to the Moho and a crustal root to a depth of 44 km beneath the Baimashan pluton in the eastern Xuefengshan belt. The Cretaceous crustal extension was limited in the upper crust in the Yuanma Basin. Therefore, the Xuefengshan belt provides a classic example how an ancient orogenic belt was repeatedly reworked in response to multiphase thermal-tectonic events.