ABSTRACT Recent studies have suggested various mechanisms to generate melt during exhumation of deeply subducted crust, including dehydroxylation of omphacite, fluid‐absent omphacite‐ and/or phengite‐breakdown melting and fluid‐present melting. However, the space–time relationships among these mechanisms remain uncertain. Here, we report microstructural evidence of the reactions responsible for, and compositions of, low‐volume melts frozen in situ (as leucosome pockets) within weakly deformed granitic veins that cross‐cut foliated UHP eclogite boudins in gneisses from the central Sulu belt, China. Phengite in the granitic veins records crystallization pressures of 3.4–2.7 GPa, and the granitic veins have whole‐rock Sr–Nd isotope compositions and trace element patterns consistent with derivation from the eclogite. The granitic veins likely crystallized from a solute‐rich supercritical fluid or hydrous melt generated by dehydroxylation of nominally anhydrous minerals during the early stage of exhumation (decompression). Subsequently, leucosome pockets and grain‐boundary films of melt formed in the granitic veins by the successive breakdown of omphacite and phengite. Based on the mineral modes and chemical compositions of 55 leucosome pockets (cf. microgranitoids), we distinguish those with (1) high Na/K ratios mainly composed of plagioclase and euhedral amphibole with skeletal omphacite and (2) low Na/K ratios predominantly composed of K‐feldspar and plagioclase and which contain phengite with corroded margins and fine‐grained biotite. High Na/K leucosome pockets (HLPs) are consistent with a melting reaction involving mostly omphacite‐breakdown, whereas low Na/K leucosome pockets (LLPs) are inferred to have formed by a reaction consuming variable proportions of phengite and omphacite. We argue that the reactions to form HHP then LLP initiated at approximately 1.5 and 1.1 GPa, respectively, and document two different mechanisms to generate melt during exhumation of deeply subducted crust in which breakdown of omphacite occurs before (deeper than) phengite. At higher temperatures, these melt‐producing reactions would generate a larger volume of melt that potentially could facilitate exhumation and increase crust–mantle interactions, thereby increasing the compositional heterogeneity of orogenic mantle.

Ultrahigh-temperature (UHT) metamorphism represents extremely hot crustal conditions, and is generally related to convergent plate margins during supercontinent assembly. However, the mechanisms to form Archean UHT metamorphism before supercontinent cycle are controversial. The Saglek-Hebron Complex (SHC) contains some of the oldest rocks on Earth that formed as early as ca 3.9 billion years ago and recorded polymetamorphism in Archean, although the exact metamorphic conditions and history are poorly understood. In this study, we investigate petrography, pressure-temperature (P-T) estimates, and zircon U-Pb geochronology of a metasedimentary rock from the SHC of northern Labrador, part of the North Atlantic Craton (NAC). The U-Pb age and trace element composition of zircon suggest zircon growth during melt crystallization and garnet resorption following an episode of ultrahigh-temperature (UHT) metamorphism at or before 2.72 Ga. Phase equilibrium modelling, two-felspar, and Zr-in-rutile thermometry together constrain UHT metamorphic P-T conditions to 7.2-9.2 kbar and 920-1000 degrees C. A rimward increase in the garnet grossular component, a change of garnet modes, and localized growth of matrix kyanite and rutile are consistent with a post-peak high-T retrograde stage within a P-T path associated with pressure increase and cooling (to similar to 11 kbar at similar to 850 degrees C). We propose that the UHT rocks formed previously in the lower crust of a hot and weak backarc region with elevated heat flow due to lithosphere extension and asthenosphere upwelling, which was closed by lateral thickening during subsequent multiple terrane convergence. Our findings are consistent with Hf isotopes in zircon from rocks of the NAC, the global record of metamorphism, and a period of cratonization of Earth's ancient crustal nuclei during the Neoarchean era that may be coincident with the operation of plate tectonics.

Understanding how incompatible elements are recycled into the mantle is essential for deciphering mantle heterogeneity and crustal evolution. However, a clear inventory of the mineral phases carrying these elements into the lower mantle remains debated. Here, we present a comprehensive mineralogical and geochemical investigation of magnesite (MgCO3) in subducted Neoproterozoic carbonatized serpentinites from the Arabian-Nubian Shield. Laser-ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) analyses reveal that the magnesite is enriched in incompatible fluid-mobile elements (e.g., B, Sb, As, Pb, and Mo), with concentrations ranging from several times to up to two orders of magnitude higher than those of the primitive mantle. Atom probe tomography (APT) shows that some moderately incompatible elements (e.g., Cu, Mn, and V) are homogeneously distributed within magnesite and, thereby, structurally bound. Given that the experimentally-determined stability of magnesite extends to lower mantle pressures, our findings suggest that magnesite is capable of transporting incompatible elements and carbon into Earth's deep interior, potentially contributing to the lower mantle source of some ocean island basalts (OIBs) and superdeep diamonds.

Eclogite-facies rocks record deep burial and exhumation of crust, and on the modern Earth are most commonly associated with subduction settings. The scarcity of blueschist- and eclogite-facies metamorphism from the Archean rock record indicates either that subduction was not operative or that the evidence has somehow been destroyed. We report an eclogite-facies mineral association (phengite + kyanite + rutile) preserved, with quartz, as inclusions within garnet porphyroblasts in a putative metasedimentary rock from the Lewisian Gneiss Complex in northwest Scotland, part of the North Atlantic craton. In situ Lu−Hf dating of garnet constrains eclogite-facies metamorphism to the Mesoarchean to Neoarchean transition at 2.81 ± 0.04 Ga. Phase equilibrium modeling, Zr-in-rutile thermometry, and phengite barometry constrain pressures (P) to 1.5−2.5 GPa and temperatures (T) to 580−660 °C, corresponding to thermobaric ratios (T/P) between 230 °C/GPa and 440 °C/GPa, markedly cooler than recorded by metamorphic rocks of a similar age elsewhere. The low T/P ratios may provide an early record of the transition from some pre-plate tectonic regime to one with an increasing role for stable subduction, and which ultimately became plate tectonics. This transition reflects thickening and strengthening of the lithosphere, the inevitable long-term consequence of secular cooling of the mantle.

Stromatolites are indicators of life on Earth since at least 3.5 Ga and possibly 3.7 Ga. From the mid-Archean onward, records of stromatolites display a range of morphologies and sizes, indicating that microorganisms were present in a variety of habitats. By comparison, the existing Paleoarchean stromatolite record is limited to small decimeter-scale domes, cones, and tufts that occur in evaporitic environments. Here, we document large domal stromatolites with similar to 1.5 m width and 0.8 m height from a chert-jasper-carbonate member in the Paleoarchean (ca. 3.47 Ga) Mount Ada Basalt, Pilbara craton, Western Australia. Sedimentological features and trace-element chemistry show that the stromatolites grew on seafloor pillow basalts in an anoxic environment with hydrothermal influence. Aqueous alteration of basaltic-komatiitic crust likely supplied key nutrients, extending benthic microbial habitats into the deeper Paleoarchean ocean and supporting a thriving biosphere independent of continental exposure.

On Earth, atmospheric oxygen is inferred to have risen over three major intervals before reaching modern levels, with each interval having a profound impact on the evolution of the biosphere. However, the principal driver behind these stepwise increases remains elusive. Here, we compile metamorphic thermobaric ratios (T/P) through time and use them as a first-order, probabilistic proxy for the likelihood of "cold" subduction (i.e., with T/P < 375 °C GPa-1) during secular cooling of Earth's mantle. Then, we couple this tectonic forcing to biogeochemical modeling to test whether more efficient cold subduction may have enhanced the net transfer of reduced organic carbon and pyrite to Earth's deep interior, thereby diminishing oxygen sinks and allowing surface oxygen levels to increase at geological timescales. Modeling results indicate that the progressive emergence of cold subduction could plausibly have contributed to the long-term oxygenation trajectory and associated secular trends in atmospheric carbon dioxide, seawater sulfate, sedimentary phosphorus, and marine redox conditions. Although the absolute magnitudes remain uncertain, the predicted trajectory of surface oxygenation is qualitatively consistent with the broad three-step pattern inferred from geochemical proxies. We propose that the progressive evolution of subduction may have been a key driver of long-term surface oxygenation, linking mantle cooling to the rise of conditions favorable for aerobic lifeforms.

The nature of Earth's crust during the Hadean eon [≥4.03 billion years ago (Ga)] is uncertain. Numerical models of early Earth geodynamics emphasize the control of mantle temperature but generally consider only internally derived heat, despite empirical evidence for an intense Hadean impact flux. Using a stochastic model of that flux, we show that the time-integrated heat due to impacts would have dwarfed that produced internally throughout the Hadean. Earth's Hadean crust would have been extensively molten at depths below a few kilometers, causing gravitational segregation of dense, iron- and magnesium-rich material and driving average crustal compositions to become increasingly silica rich. Globally, impact heating would have become much less important after 3.9 Ga, allowing the crust to thicken. That enduring continental crust appeared around this time is likely not a coincidence.

ABSTRACT Whether nominally anhydrous garnet pyroxenite can melt during exhumation from eclogite‐facies P–T conditions and, if so, what chemical composition such melts would have is largely unexplored. Here, we document petrographic, chemical, geochronological and thermodynamic modelling evidence for the initiation of in situ partial melting of nominally anhydrous eclogite‐facies garnet pyroxenite (garnet + omphacite/sodian augite + rutile) from the northern Sulu belt, China. The garnet pyroxenite forms the cores of amphibolitized boudins hosted in strongly deformed migmatitic felsic gneiss. The garnet pyroxenite is strongly foliated and comprises interlayered garnet‐rich and clinopyroxene‐rich layers. The latter contain abundant pockets (0.2–1 mm across) comprising volumetrically dominant plagioclase with euhedral grains of either clinopyroxene (Cpx + Pl leucosome) or amphibole (Amp + Pl leucosome), and locally grain‐boundary films of K‐feldspar (K‐bearing leucosome). These leucosome pockets are preferentially located at triple junctions between clinopyroxene grains and have microstructural features consistent with their crystallization from locally derived melt, including plagioclase seams extending along clinopyroxene grain boundaries in the garnet pyroxenite host. The calculated bulk composition of the leucosome pockets is consistent with melting through a process combining breakdown of clinopyroxene ± garnet, but requiring a contribution of mobile components (including H 2 O and K) from the host gneiss, yielding melts of gabbroic to monzonitic composition. U–Pb analysis of euhedral zircon cores yield Neoproterozoic ages of c. 740 Ma, whereas zircon rims yield Triassic metamorphic ages of 244–221 Ma, consistent with the late prograde to early retrograde evolution of the garnet pyroxenite during subduction of the Yangtze craton. Thermodynamic modelling shows that gradients in μ H 2 O and μ K 2 O (where μ is chemical potential) generated during decompression from eclogite‐facies P–T conditions drove diffusive transfer of material from the gneiss to the garnet pyroxenite. This interpretation is consistent with the elevated contents of large ion lithophile and light rare earth elements in most of the leucosome pockets relative to host clinopyroxene. Thirteen leucosomes are nepheline normative and five are hypersthene normative, suggesting that partial melting of nominally anhydrous garnet pyroxenite produces alkaline to subalkaline basaltic melts. Our findings indicate that clinopyroxene‐breakdown melting is an intrinsic consequence of suprasolidus decompression from eclogite‐facies P–T conditions. Although the volume of melt generated in the garnet pyroxenite was minor, our study may have wider implications for the genesis of alkalic OIB magmas.

The transition to continental collision c. 650 Ma induced the bimodal hypsometry of the Arabian-Nubian Shield (ANS), and triggered the formation of voluminous post-amalgamation basins. The intermontane Kareim Basin is a voluminous post-amalgamation depocenter within the ANS. It comprises a thick siliciclastic fill (~ 7 km thick) that accumulated over tens of millions of years during late Neoproterozoic East African orogeny related to the amalgamation of West and East Gondwana. The basin fill consists of four main facies associations (FA1 to FA4) associated with 11 siliciclastic lithofacies and one volcaniclastic lithofacies, which are interpreted as alluvial fan to lacustrine deposits that accumulated under humid to semi-arid conditions. A conglomerate-dominated lithofacies characterizes proximal alluvial fan deposits (FA1), whereas mid to distal alluvial fan strata are represented by braided stream sandstone-dominated lithofacies and conglomerate (FA2-3). Distal fan deposits are composed mainly of sandstone and fine-grained lacustrine sediments (FA4). Tectonically-induced unconformities separate three depositional stages in the Kareim Basin. The lower stage comprises three sandstone-dominant cycles, locally separated by unconformities. The middle stage of the basin represents a stage of syn-depositional tectonic inversion, consistent with the presence of recycled basal boulders derived from the lower stage, and a divergence in the dominant paleo-current directions. Furthermore, thrust faults, tilted and overturned older strata, and the first occurrence of material derived locally from Pan-African volcanic rocks (the Dokhan Volcanic Suite) and basement gneiss domes are additional clues for the syn-depositional tectonic inversion. The upper stage comprises conglomerate-dominant cycles, and represents the transition to post-collisional extension and rapid subsidence. Detrital zircon U-Pb ages constrain the syn-depositional inversion of the basin to later than c. 635 Ma, likely coinciding with the onset of collision between West with East Gondwana.

Reliably reconstructing the pressure-temperature-time (P-T-t) history of Archean polymetamorphic terranes is key to gaining insight into the tectonic processes operating on the early Earth. In this regard, garnet is arguably the most important mineral. Not only does it capture information about both the timing and the conditions of metamorphism, but can also preserve this record through multiple tectonic cycles. Here, we used in situ Lu-Hf dating of garnet from two greenstone belts in the Western Dharwar Craton (WDC) in southern India to investigate their tectono-metamorphic evolution. Garnet ages reveal a record of two distinct medium-to high-pressure tectono-metamorphic events during the Archean eon. Whereas garnet from the Holenarsipur Greenstone Belt grew during late Neoarchean metamorphism at 2.53 Ga, garnet from the Kalyadi Greenstone Belt records a 2.96 Ga regional metamorphic episode that has not previously been recognized. Coupled with results from in situ monazite U-Pb dating and thermodynamic modelling, our data suggest that the two neighboring greenstone belts, which were previously thought to share a common tectonothermal history, underwent contrasting P-T-t evolutions throughout the Mesoarchean and Neoarchean, challenging existing models for the tectonic evolution of the WDC. Obtaining reliable age constraints in polymetamorphic terranes via in situ Lu-Hf dating of garnet that can be directly linked to P-T information represents a key advance in deciphering the cryptic record of crustal metamorphism throughout Earth history.

Petrochronological investigations of granulite-facies metapelitic rocks from the eastern contact of the Closepet granite in the Central Dharwar Craton (CDC), southern India, provide new pressure-temperature-age (P-T-t) constraints on two seemingly discrete Archean metamorphic events during the Neoarchean and late Paleoarchean eras. Phase equilibrium modelling and conventional thermobarometry coupled with in situ monazite and garnet geochronology constrain Neoarchean (ultrahigh-temperature; UHT) metamorphic peak conditions to similar to 930 degrees C and similar to 6.7 kbar at c. 2.63 Ga, then cooling and limited decompression to similar to 5.7 kbar at <= 810 degrees C. Monazite inclusions in garnet least affected by Neoarchean recrystallisation have distinct positive Eu anomalies and yield ages of c. 3.2 Ga, whereas garnet cores interpreted to have grown at the same time have slightly younger apparent ages of c. 3.1 Ga. We interpret this age mismatch to be the result of extensive resorption of garnet during Neoarchean UHT metamorphism, which led to widespread modification of the initial Lu-Hf systematics in garnet to produce younger apparent ages. The effect of retention and inward intracrystalline diffusion of Lu on the isotopic composition of garnet is most pronounced close to the resorbed grain margins and decreases towards the core, as reflected by progressively younger apparent single-spot garnet dates from core to rim. Despite extensive overprinting of the sample at c. 2.63 Ga, the trace element composition of Paleoarchean monazite indicates growth in equilibrium with garnet but in the absence of feldspar, which is predicted to occur over a broad stability range at P >= 8 kbar and T <= 700 degrees C. Such P-T conditions are uncommon in the metamorphic rock record prior to the Neoarchean, but are typical of Barrovian-type metamorphism, which is considered to be an expression of accretionary-to-collisional orogenesis. Rocks of similar age and metamorphic grade have been reported from the core of the Western Dharwar Craton (WDC) and may reflect regional subduction at the margins of the CDC and WDC since the late Paleoarchean. Neoarchean UHT metamorphism in the Dharwar Craton is coeval with a cluster of other UHT occurrences at 2.7-2.6 Ga, indicating the existence of globally elevated thermal gradients at that time.

In his classic contribution “A Heat Pipe Mechanism for Volcanism and Tectonics on Venus” [1989, JGR 94, B3, 2779-2785], Turcotte applied O’Reilly and Davies’ [1981, GRL 8, 313-316] model for Io’s volcanic heat transport to Venus, and further speculated that this mechanism might explain a thick lithosphere inferred for Archean Earth. The latter idea – to consider heat-pipe cooling for Earth – then lay fallow until roughly 15 years ago, when one of us (Moore) argued in talks and manuscripts (all rejected) that the cold, thick, and strong lithosphere generated by heat-pipe cooling might offer an alternative to subduction tectonics for generating the “too-cold” Hadean zircons reported by Hopkins et al. [2008, Nature 456, 493-496]. With the additional realization that the heat-pipe cooling mechanism might similarly account for the rocks preserved from the first half of Archean time, the concept of an early heat-pipe Earth finally received broad consideration just over a decade ago [Moore & Webb, 2013 Nature 501, 501-505]. Heat-pipe cooling is a hot stagnant-lid cooling mode based on our understanding of the active volcano-tectonics of Jupiter’s moon Io [O’Reilly & Davies, 1981]. Heat-pipes are not plumes: heat-pipes are conduits channeling melts upwards through lithosphere, whereas plumes commonly span the whole crust and mantle and accordingly have relatively complex histories. The heat-pipe Earth hypothesis posits that during the first third of Earth history, rapid volcanism dominated cooling from the end of the magma ocean period to the onset of (episodic?) plate tectonics. During the heat-pipe period, voluminous mafic volcanism resulted in protracted resurfacing, causing quasi-continuous burial of cold, hydrated surface materials that (1) cooled a single-plate lithosphere and (2) generated tonalite-trondhjemite-granodiorite melts deep in the lithosphere. It is noteworthy that the burial of surface materials to mantle depths – long seen as a distinguishing characteristic of plate tectonics – is a hallmark of heat-pipe cooling. The past decade has seen abundant explorations and tests of the heat-pipe Earth hypothesis, as well as renewed interest in the heat-pipe cooling mechanism for other terrestrial bodies. This presentation will review major results and consider key critiques. Highlights include demonstrations that heat-pipe cooling viably explains: (a) the early histories of the lithospheres preserved at Mercury, Venus, Mars, and the Moon, and thus can be hypothesized as a universal cooling mechanism for early / hot terrestrial bodies in our Solar System and others [Moore et al., 2017 EPSL 474, 13-19; Peterson et al., 2021 Sci.Adv. 7:eabh2482]; (b) the Eoarchean development of the Isua supracrustal belt of southern West Greenland [Webb et al., 2020 Lithosphere 12, 166-179 and a collection of subsequent works], which was previously understood exclusively via plate tectonic models; (c) the initiation of a global plate network, as thinning and corresponding warming of lithosphere during waning heat-pipe cooling caused thermal expansion which overcame the tensional strength of the lithosphere [Tang et al., 2020 Nat.Comm. 11:3621]; and (d) Earth’s detrital zircon depositional records older than ~3.3 Ga [Zuo et al., 2021 EPSL 575:117182].

Understanding how Earth's continental nuclei first formed in the Archean eon (4.0-2.5 Ga) underpins our notions of early Earth geodynamics. Yet, the nature of Earth's early protocrust and the primary mechanism for its transformation are poorly understood, as very ancient rocks preserving petrological evidence for these processes are incredibly rare. Here we report the discovery of a formerly melt-bearing amphibolite from the Sylvania Inlier of the Pilbara Craton in Western Australia. Radiometric dating of zircon and titanite in these rocks constrain the time of partial melting to 3565 Ma, providing evidence for a metamorphic event that predates most exposed rocks in the Pilbara Craton by ~30 million years. Low δ18O compositions and modelled melt compositions comparable to evolved Hadean rocks in the Acasta gneisses indicate Earth's oldest continental crust may have sourced rocks of a similar composition. Thermodynamic modelling suggests partial melting at temperatures of 680-720°C and pressures of 0.8-1.0 GPa, implying a maximum burial depth of ~30 km. These results support models of continental nuclei formation via shallow partial melting of hydrothermally altered mafic protocrust in high heat flow environments.

A newly identified tectonic sliver of low-grade Neoproterozoic continental crust comprising hybrid granitoids is exposed between Triassic ultrahigh-pressure (UHP) metamorphic rocks of the Sulu belt and the Cretaceous Laoshan granite. Outcrop evidence of magma mingling combined with the granitic mineral assemblages and primary igneous microstructures suggest >100 km difference in likely depth of subduction compared to the proximal UHP metamorphic rocks. The granitoids are cut by a brittle-to-ductile shear zone and extensional fractures that focussed fluid flow and allowed low-temperature fluid-rock alteration. Integrated accessory mineral geochronology from zircon, allanite and apatite records crystallization ages of c. 835-700 Ma and alteration ages of c.120-90 Ma. There is no evidence of the Triassic UHP metamorphic event recorded in the Neoproterozoic granitoids. The hybrid granitoids were likely generated during Neoproterozoic rifting of Rodinia, forming part of the northern margin of the Yangtze craton, but they did not experience deep subduction like other Neoproterozoic continental crust within the Sulu belt. The adjacent Cretaceous Laoshan granite was emplaced when all units were at shallow crustal depths, after >100 km of exhumation of the UHP eclogites. The Cretaceous ages retrieved from the hybrid granitoids date the brittle deformation and fluid alteration of these rocks in the Laoshan granite aureole. Cretaceous subduction retreat (slab roll back) of the paleo-Pacific plate caused extension and lithosphere thinning of the upper plate (eastern China), forming extensional shear zones and core complexes, associated with numerous granitoid intrusions, including the Laoshan granite. The extensional geological setting also facilitated preservation of this thin sliver of hybrid granitoids at the margin of the Laoshan granite adjacent to the UHP metamorphic terrane, and likely contributed to the final exhumation of the UHP metamorphic rocks of the Sulu belt. Our study shows that post-collisional extensional could be a common mechanism promoting final exhumation and exposure of deeply subducted terrains in orogens worldwide.

Continental crust is fundamental to planetary habitability, providing the geochemical reservoirs and physical interfaces that drive and regulate exchanges among the atmosphere, hydrosphere and biosphere. However, the evolution of Earth’s crust is uncertain owing to debate regarding the competing roles of internal versus external energetic drivers. In this Review, we examine the interplay between internal and external drivers of the production, modification and destruction of crust on the early Earth using geochemical, geological and geophysical data. Internal drivers are potentially linked to plate tectonics and processes such as subduction (dripping) or delamination. External drivers from large meteorite impacts likely influenced crust formation by inducing rapid decompression melting of the mantle to form basaltic protocratons, the early, mantle-derived crustal nuclei that preceded stable continental crust. On a planet covered by water, protocratons might have been transformed by intracrustal differentiation into evolved (continental) crust. Future research into the processes driving Earth’s early evolution and habitability should consider a wide range of temporal and spatial scales from seconds to millions of years and the subgrain to the galactic, to uncover the long-wavelength patterns, in mantle overturn rates and impact flux preserved in deep-time records. Continental crust is important for Earth’s habitability. This Review explores how the formation and stabilization of Earth’s early continental crust was modulated by internal and external factors such as subduction and bolide impacts, respectively.

Plate tectonics drives the compositional diversity of Earth’s convecting mantle through subduction of lithosphere. In this context, the role of evolving global geodynamics and plate (re)organisation on the spatial and temporal distribution of compositional heterogeneities in the convecting mantle is poorly understood. We test the hypothesis that an increase in the cumulative length of subduction zones associated with supercontinent assembly triggered geochemical enrichment of the convective mantle globally, in particular since the emergence of protracted, cold, deep subduction in the late Neoproterozoic. We compiled the trace element and Nd isotopic compositions of intracontinental basalts formed over the last billion years (1000 Myr). After careful filtering to eliminate samples with evidence for crustal contamination, the data show that intracontinental basalts formed before 300 Ma exhibit supra-chondritic initial 144Nd/143Nd values. Those with sub-chondritic initial 144Nd/143Nd values become common only after 300 Ma, broadly coeval with the global appearance of kimberlites with geochemically enriched isotopic signatures. We attribute these step-changes in the sources of intraplate magmatism to a rapid increase in the supply of deeply subducted lithosphere due to increased peri-continental subduction during the assembly of Pangea.

The operation of a hydrological cycle (i.e., exchange of water between the land, oceans, and atmosphere) has significant implications for the emergence of life. The oldest confirmed single-celled organisms at ~3.48 billion years ago (Ga) (Pilbara Craton, Western Australia) are thought to have formed in the presence of meteoric (fresh) water on emerged (subaerial) land in a hot spring environment. However, when widespread interaction between fresh water and emerged continental crust first began is poorly constrained. In this study, we use >1000 oxygen isotope analyses of Jack Hills detrital zircon to track fluid-rock interactions from the Hadean to the Paleoarchean (~4.4–3.1 Ga). We identify extreme isotopically light O (i.e., δ18O < 4.0 ‰) values older than 3.5 Ga. The data define two periods of magmatism with extreme isotopically-light O as low as 2.0 ‰ and –0.1 ‰ at around 4.0 and 3.4 Ga, respectively. Using Monte Carlo simulations, we demonstrate that such values can only be generated by the interaction of crustal magmatic systems with meteoric water. Our data constrains the earliest emergence of continental crust on Earth, the presence of fresh water, and the start of the hydrological cycle that likely provided the environmental niches required for a life less than 600 million years after Earth’s accretion.

The role of meteorite impacts in the origin, modification, and destruction of crust during the first two billion years of Earth history (4.5-2.5 billion years ago; Ga) is disputed. Whereas some argue for a relatively minor contribution overall, others have proposed that individual giant impactors (>10-50 km diameter) can initiate subduction zones and deep mantle plumes, arguably triggering a chain of events that formed cratons, the ancient nuclei of the continents. The uncertainty is compounded by the seeming absence of impact structures older than 2.23 Ga, such that the evidence for the terrestrial impact flux in the Hadean and Archaean eons is circumstantial. Here, we report the discovery of shatter cones in a complex, dominantly metasedimentary layer, the Antarctic Creek Member (ACM), in the centre of the East Pilbara Terrane, Western Australia, which provide unequivocal evidence for a hypervelocity meteorite impact. The shocked rocks of the crater floor are overlain by (unshocked) carbonate breccias and pillow lavas, stratigraphically constraining the age of the impact to 3.47 Ga and confirming discovery of the only Archaean crater known thus far.

Plate tectonics drives the compositional diversity of Earth’s convecting mantle through subduction of lithosphere. In this context, the role of evolving global geodynamics and plate (re)organization on the spatial and temporal distribution of compositional heterogeneities in the convecting mantle is poorly understood. Here, using the geochemical compositions of intracontinental basalts formed over the past billion years, we show that intracontinental basalts with subchondritic initial neodymium-144/neodymium-143 values become common only after 300 million years, broadly coeval with the global appearance of kimberlites with geochemically enriched isotopic signatures. These step changes in the sources of intraplate magmatism stem from a rapid increase in the supply of deeply subducted lithosphere during the protracted formation of Pangea following the widespread onset of “modern” (cold and deep) subduction in the late Neoproterozoic. We argue that the delay (~300 million years) in the appearance of enriched intraplate magmas reflects the time required for the sinking and (re)incorporation of slabs into the sources of mantle-derived magmas.