Precise radiometric age establishes Yarrabubba, Western Australia, as Earth’s oldest recognised meteorite impact structure

The ~70 km-diameter Yarrabubba impact structure in Western Australia is regarded as among Earth’s oldest, but has hitherto lacked precise age constraints. Here we present U–Pb ages for impact-driven shock-recrystallised accessory minerals. Shock-recrystallised monazite yields a precise impact age of 2229 ± 5 Ma, coeval with shock-reset zircon. This result establishes Yarrabubba as the oldest recognised meteorite impact structure on Earth, extending the terrestrial cratering record back >200 million years. The age of Yarrabubba coincides, within uncertainty, with temporal constraint for the youngest Palaeoproterozoic glacial deposits, the Rietfontein diamictite in South Africa. Numerical impact simulations indicate that a 70 km-diameter crater into a continental glacier could release between 8.7 × 1013 to 5.0 × 1015 kg of H2O vapour instantaneously into the atmosphere. These results provide new estimates of impact-produced H2O vapour abundances for models investigating termination of the Paleoproterozoic glaciations, and highlight the possible role of impact cratering in modifying Earth’s climate.

present high-resolution orientation mapping and correlated in situ U-Pb analysis to investigate the microstructure and age of shock features in zircon and monazite in target rock and impact melt from the Yarrabubba structure in Western Australia. These results establish Yarrabubba as the oldest preserved impact structure on Earth.
In Barlangi granophyre, zircon textures range from unshocked grains preserving primary igneous growth zoning, to grains containing clear evidence of impact metamorphism, such as planar microstructures, polycrystalline aggregates and grains with ZrO 2 inclusions ( Fig. 2c; Supplementary Fig. 3) 34 . Neoblasts Composite aeromagnetic anomaly map of the Yarrabubba impact structure within the Yilgarn Craton, Western Australia, showing the locations of key outcrops and samples used in this study. The image combines the total magnetic intensity (TMI, cool to warm colours) with the second vertical derivative of the total magnetic intensity (2VD, greyscale) data 81 . The demagnetised anomaly centred on the outcrops of the Barlangi granophyre is considered to be the eroded remnant of the central uplift domain, which forms the basis of the crater diameter of 70 km 18 . Prominent, narrow linear anomalies that cross-cut the demagnetised zone with broadly east-west orientations are mafic dykes that post-date the impact structure.
within polycrystalline zircon aggregates contain systematic misorientation relationships with one another of either 65°about <110> or 90°about <110>, which can only be caused by recrystallisation after formation of {112} twins and the highpressure polymorph reidite, respectively, and are unambiguous indicators of shock metamorphism [34][35][36] . While the ZrO 2 inclusions index as baddeleyite (monoclinic-ZrO 2 ), crystallographic orientation relationships among transformation twins demonstrate they originally formed from tetragonal-ZrO 2 parent grains 37 . Thermal dissociation of zircon to tetragonal-ZrO 2 only occurs in silica-saturated melts above 1673°C 34 , unequivocally indicating the Barlangi granophyre was a super-heated impact melt. Monazite grains in Barlangi granophyre preserve a similar range of impact-related features to those from Yarrabubba monzogranite, including crystal-plastic strain, deformation twins diagnostic of shock conditions, and strain-free neoblastic domains ( Fig. 2d; Supplementary Fig. 4).  Table 2). These ages may represent either formation during a post-crystallisation metamorphic event or partial radiogenic Pb-loss during the impact event or a subsequent thermal event. In contrast, spots from low-strain, randomly oriented neoblasts cluster around concordia, and have a weighted mean 207 Pb/ 206 Pb age of 2227 ± 5 Ma (n = 12, MSWD = 0.89) (Fig. 3b). Barlangi granophyre zircon 207 Pb/ 206 Pb ages also show a bimodal age distribution. We interpret the oscillatory-zoned cores with apparent ages of 2781 ± 14 to 2319 ± 28 Ma to represent inherited (pre-impact) zircon grains that were incorporated into the Barlangi granophyre as xenocrysts, consistent with zircon ages determined previously 24,25 . These results indicate the presence of a significant source component in the Barlangi granophyre that predates the 2.65 Ga Yarrabubba monzogranite. Individual analyses from polycrystalline zircon domains are variably discordant and yield 207 Pb/ 206 Pb ages from 2259 ± 30 to 2156 ± 52 Ma (Supplementary Table 1). We interpret the data array ( Fig. 3a) to be a function of near-recent Pb-loss resulting from exposure to surface fluids. The data array from recrystallised zircon domains yields an upper intercept age of 2246 ± 17 Ma (n = 13, MSWD = 1.2), and two rim analyses with baddeleyite intergrowths are collinear with this regression. We interpret the upper intercept to reflect both new zircon growth and near complete resetting of U-Pb systematics in pre-existing domains during shock metamorphism. This date is within uncertainty of a single previously reported U-Pb zircon analysis from the Barlangi granophyre of 2234 ± 28 Ma 24 , which was inferred to indicate a Palaeoproterozoic impact age 18  Step size = 0.5 μm Step size = 0.5 μm Step size = 0.4 μm Step  Table 2).
When combined, all neoblastic monazite domains from both Barlangi granophyre and Yarrabubba monzogranite define a cluster centred on concordia (Fig. 3a) and yield a weighted mean 207 Pb/ 206 Pb age of 2229 ± 5 Ma (n = 26, MSWD = 1.4, Fig. 3c), which we interpret to record monazite recrystallisation during shock metamorphism and the best estimate of the Yarrabubba impact event (Fig. 3). The weighted mean 207 Pb/ 206 Pb age for neoblastic zircon of 2246 ± 17 Ma (n = 13, MSWD = 1.2) overlaps with the monazite age, but is less precise. The new Yarrabubba impact age of 2229 ± 5 Ma determined here extends the terrestrial record of impact craters by 200 million years 8 , and demonstrates the potential for discovery of ancient impact structures on Archaean cratons.
Implications for the Palaeoproterozoic Earth. The age constraints presented here establish Yarrabubba as the first recognised meteorite impact to have occurred during the Rhyacian period, a dynamic time in the evolution of Earth following the transition from the Archaean to the Proterozoic eon. At least four glacial diamictite deposits, three of which are found on multiple cratons, are recognised between 2.4 and 2.2 Ga 38 . Of these deposits, the >2.42 Ga Makganyene diamictite from the Kaapvaal craton of southern Africa has been interpreted to represent lowlatitude glaciers that may signify global ice conditions 39 . The youngest Palaeoproterozoic glacial deposit, the Rietfontein diamictite within the Transvaal basin of South Africa, has a minimum depositional age of 2225 ± 3 Ma based on the overlying Hekpoort basalt (Fig. 4) 38 , the global δ 13 C trend of carbonates is modified from ref. 51 , and the lull in magmatic activity is modified from ref. 44  years ( Fig. 4) 41,42 . What caused the extended absence of glacial conditions after ca. 2225 Ma is debated 43 . The end of the Palaeoproterozoic glaciations at 2225 Ma occurred within an apparent~50 Myr lull in global magmatism from 2266 to 2214 Ma 44 , making it difficult to appeal to volcanic outgassing as having played a significant role in forcing the glacial termination. Therefore, other mechanisms such as impact cratering need consideration. Radiometric age data presented here demonstrate synchronicity, within uncertainty, between the 2229 ± 5 Ma Yarrabubba impact event and the termination of glacial conditions (i.e. Rietfontein diamictite) at 2225 ± 3 Ma. The geographic extent of the Rietfontein diamictite is poorly constrained, and it is not yet known if global glacial conditions existed at this time. Nonetheless, we apply numerical simulations below to explore the potential effects that a Yarrabubba-sized impact may have had on climactic conditions. Several factors caused by the Yarrabubba impact event could have triggered a change in regional or global climate. Depending on the ambient climate state and palaeogeographic nature of the northern Yilgarn craton at the time of impact (e.g., ice cover, shallow ocean or carbonate platform overlying silicate basement), which is unknown, significant amounts of CO 2 , water vapour or other greenhouse gases could have been released into the relatively oxygen-poor Palaeoproterozoic atmosphere 45 by the impact event. Given that the age of the Yarrabubba impact overlaps with the youngest Paleoproterozoic glacial deposits, we explore scenarios where the Yarrabubba impact site could have been covered by a continental ice sheet at the time of impact. Numerical models using the iSALE shock physics code [46][47][48] (see Methods) demonstrate that the formation of a 70-km-diameter impact crater into a granitic target with an overlying ice sheet ranging from 2 to 5 km in thickness results in the almost instantaneous vaporisation of 95-240 km 3 of ice and up to 5400 km 3 total melting (Fig. 5). The vapourised ice corresponds to between 9 × 10 13 and 2 × 10 14 kg of water vapour being jetted into the upper atmosphere within moments of the impact (Fig. 5d). Impact-generated water vapour in the lower atmosphere would have condensed and rapidly precipitated as rain and snow with no significant long-term climate effects, or could have even triggered widespread glacial conditions via cloud albedo effects during interglacial periods 49 . However, ejection of high-altitude water vapour has potential for greenhouse radiative forcing, depending critically on atmospheric residence time 50 . Nonetheless, uncertainties in the structure and composition of Earth's Palaeoproterozoic upper atmosphere mean that the precise nature of atmospheric interactions of the collapsing vapour plume is inherently difficult to model 50 . Nevertheless, considering that Earth's atmosphere at the time of impact contained only a fraction of the current level of oxygen 51 , a possibility remains that the climatic forcing effects of H 2 O vapour released instantaneously into the atmosphere through a Yarrabubba-sized impact may have been globally significant. Understanding the residence times of impact-produced water vapour in a cold Palaeoproterozoic atmosphere, e.g. ref. 50 , and the complex interplay of radiative versus insulative effects of clouds, e.g. ref. 52 , during glacial conditions requires further investigation. The effects of impact cratering have long been recognised as drivers of climate change 2,53 . Many studies have described the atmospheric effects of the end-Cretaceous Chicxulub impact structure in Mexico 49,54,55 , which resulted in global cooling of oceans and production of widespread acidic rains 56,57 . While the Yarrabubba structure dated at 2229 ± 5 Ma represents the Earth's oldest dated impact crater, its coincidence with termination of Palaeoproterozoic glacial conditions prompts further consideration of the ability of meteorite impacts to trigger climate change.

Methods
Samples and preparation. Two samples from the Yarrabubba impact structure were selected for analysis (Fig. 1): a sample of the shocked Yarrabubba monzogranite (14YB07) and a sample from the Barlangi granophyre (14YB03). The Yarrabubba monzogranite was sampled from a small outcrop approximately 1.2 km WSW of Barlangi Rock. The Barlangi granophyre was sampled 2.7 km NNW of Barlangi Rock from an apophyse intruding the Yarrabubba monzogranite along what appears to be a shallowly dipping fault plane or fracture.
Thin sections were prepared from each hand sample to allow petrographic characterisation of the lithological fabric and identification of shock features. Zircon and monazite grains were separated from each sample. To separate zircon and monazite grains~1 kg splits of each sample were processed with a Selfrag electric pulse disaggregator in the John de Laeter Centre (JdLC), Curtin University, Western Australia. The heavy mineral fraction was then separated using the heavy liquid methylene iodide. Further concentration of zircon and monazite was achieved with a Frantz isodynamic magnetic separator. Grains were then handpicked and mounted in a 25.4 mm epoxy round. The epoxy rounds were given a mechanical polish to 1 µm with diamond paste before a final chemical-mechanical polish with a colloidal dispersion of 5 nm silica in NaOH. After polishing, monazite and zircon grains were imaged using backscatter electron (BSE) atomic contrast imaging and cathodoluminescence (CL) imaging; images can be found in Supplementary Figs. 1-4. All scanning electron microscope (SEM) analyses were undertaken on the Tescan Mira3 field emission-gun (FEG) SEM at the Electron Microscopy Facility, within the JdLC. BSE photomicrographs were collected using an accelerating voltage of 15 kV, and CL images were collected with an accelerating voltage of 10 kV.
Electron backscatter diffraction microstructural analyses. Shock-deformed monazite and zircon grains were mapped by electron backscatter diffraction (EBSD). Electron backscatter patterns (EBSPs) were collected from the monazite and zircon in orthogonal grids using a Nordlys Nano high-resolution detector and Oxford Instruments Aztec 2.4 acquisition software package on the Mira3 FEG-SEM. EBSD analyses were collected with a 20 kV accelerating voltage, 70°sample tilt,~20 mm working distance and 18 nA beam current. EBSPs were collected with the following parameters; an acquisition speed of~40 Hz, 64 frames were collected for a background noise subtraction, 4 × 4 binning, high gain, a Hough resolution of 60 and band detection min/max of 6/8. Maps were collected with a step size between 1.0 and 0.12 µm. Mean angular deviation values of the electron backscatter patterns for the maps ranged between 0.81 and 0.29. Individual zircon grains were mapped using the match unit Zircon 5260 based on the unit cell parameters of Hazen et al. 58 after the methods of Reddy et al. 59 . Monazite grains were mapped with the match unit described in ref. 60 , which originates from crystallographic data of ref. 61 . From the Yarrabubba monzogranite (14YB07) seven zircon and four monazite grains were analysed, while eight zircon and seven monazite grains were analysed from the Barlangi granophyre (14YB03).
Post-processing the EBSD data was undertaken with Oxford Instruments Channel 5.11 software suite. All EBSD data were given a wild-spike noise reduction and a six nearest neighbour zero-solution correction. EBSD maps were produced using the Tango suite of Channel 5, while pole figures were processed in the Mambo suite of Channel5. EBSD maps and pole figures (as equal area, lower hemisphere projections) of the shocked monazite and zircons can be found in Supplementary Figs. 1-4. Using Tango the following maps were produced for the shocked zircon and monazite grains: (1) Inverse pole figure (IPF) maps of crystallographic orientations of zircon (Fig. 2a, c). (2) All Euler crystallographic orientation map of shocked monazite (Fig. 2b, d).
(3) Grain misorientation map, using the grain rotation orientation direction (GROD)-angle function of Channel5, which helps visualise the substructure of the grains by plotting the deviation angle of each pixel from the mean grain orientation, grain boundaries are defined as >10°. Blue domains are low strains, while warm colours represent higher degrees of misorientation ( Supplementary Figs. 1-4).
Secondary ion mass spectrometry U-Pb age analyses. Following EBSD mapping of monazite and zircon, in situ U-Th -Pb isotopic measurements targeting specific shock microstructural domains were carried out using the SHRIMP-II secondary ion mass spectrometer (SIMS) at the JdLC. Operating procedures for uranium, thorium and lead isotopic measurements on zircon are based on those described by Compston et al. 62 and Claoué-Long et al. 63 73 . A~10 μm diameter primary beam was employed with an intensity of~0.5 nA. Ion microprobe analyses of monazite are affected by an uneven background spectrum of scattered ions 74 , which can be reduced effectively by use of the SHRIMP retardation lens system, which is set at 10 kV. This discriminates against low-energy ions entering the collector. Each analysis consists of six cycles through the isotopic masses in the following sequence: 202 (species [ 139 La 31 P 16 75,76 . This interference has been observed to correlate with thorium content 74 . Excess 204 Pb+ counts are corrected against the India monazite standard assuming 206 Pb/ 238 U-207 Pb/ 235 U age-concordance of the standard at a known thorium concentration. Fractionation of the 207 Pb/ 206 Pb ratio is typically observed when the retardation lens system is at operating voltage during monazite analysis. Fractionation of the 207 Pb/ 206 Pb ratio is monitored and corrections were applied, if necessary, by reference to the GM3 monazite standard 77 which was run as an unknown. Uncertainties associated with this correction are added in quadrature to the uncertainties of 207 Pb*/ 206 Pb* ratios and dates. The common-Pb correction was based on measured 204 Pb and Stacey and Kramers 71 crustal Pb composition appropriate for the age of the sample. Data were reduced using SQUID 2 software and plotted using Isoplot 3.66 (ref. 66 ).
Nine SHRIMP analyses from 7 zircon grains and 16 SHRIMP analyses from four monazite grains were collected from the sample of the Yarrabubba monzogranite 14YB07. Nineteen SHIRMP analyses from 8 zircon grains and 19 SHRIMP analyses from 7 monazite grains were collected from the sample of the Barlangi granophyre 14YB03. Selection criteria for the analytical spots were based on the EBSD data, and both the strained domains and the strain-free neoblastic domains of the shocked monazite and zircon were targeted.
U-Pb isotopic data are provided in Supplementary Tables 1 and 2 Hydrocode impact simulations. To better constrain the environmental effects of a Yarrabubba-sized impact structure into a continental ice sheet, the formation of various impact craters were computed using the iSALE shock physics code [46][47][48] , which is based on the original SALE algorithm 78 . The initial conditions were a granitic target covered by an ice sheet of varying thickness (2-5 km). For all ice thicknesses considered here, an impactor diameter of 7 km produced a final crater of~70 km diameter. The impactor was prescribed an impact velocity of 17 km/s, the current best estimate for the average speed of bolides striking Earth. The ANEOS equation of state for granite 79 was used to model both the impactor and the granitic target and the 5-phase SESAME equation of state for ice 80 was used to model the overlying ice sheet. A computational cell size of 125 m was used, resulting in a resolution of 28 cells per projectile radius. The end product was an impact crater with a~70 km diameter impact structure, with a zone 40-50 km in diameter in the granite that experienced high shock pressures (>50 GPa) and is heavily fractured, which in some simulations produces an uplifted region. The mass of ice that vapourised during the impact was determined by comparing the entropy of Lagrangian tracer particles in the ice sheet to the critical entropy for incipient (entropy required to go to liquid and vapour state) and complete vaporisation (entropy required to go to a completely vapour state) for ice. In the model with a 2 km thick ice sheet, 1.0 × 10 15 kg of H 2 O was above the critical entropy for incipient vaporisation, and 8.7 × 10 13 kg of H 2 O was completely vapourised. The mass of ice vapourised in each scenario with different ice thicknesses can be seen in Fig. 5d. Snapshots of the iSALE model at time 0, 36 and 780 s can be found in Fig. 5.

Data availability
All data used in this manuscript are included in the Supplementary Information file. For access to geologic specimens contact author T.M.E.

Code availability
The Isoplot and Squid programs used for U-Pb data processing are available through the Berkley Geochronology Center http://www.bgc.org. iSALE is a widely used shock physics code distributed to registered academic non-commercial users via a GitHub repository. Scientists interested in using or developing iSALE should see http://www.isale-code.de for a description of application requirements.