RESEARCH MASS EXTINCTION To estimate the eruption date and associated uncertainty for each sample, we developed an approach using Bayesian statistics to account U-Pb constraints on pulsed eruption for the probability distribution of zircon dates and their analytical uncertainties (20) (fig. S6). Although we considered alternative data inter- of the Deccan Traps across pretations (13), they do not affect the conclusions of this study. Twenty-one of 24 dated horizons the end-Cretaceous mass extinction are from five stratigraphic sections along prom- inent roads in the Western Ghats, providing 1 1 2 3 completecoverageoftheupperfourDeccanfor- Blair Schoene *, Michael P. Eddy , Kyle M. Samperton , C. Brenhin Keller , Gerta Keller1, Thierry Adatte4, Syed F. R. Khadri5 mations (Fig. 1 and figs. S1 and S2). The remaining three samples span the lower five Deccan for- mations, where redboles are rare and less likely Temporal correlation between some continental flood basalt eruptions and mass to contain zircon. extinctions has been proposed to indicate causality, with eruptive volatile release When compiled into a composite stratigraphic driving environmental degradation and extinction. We tested this model for the section (Fig. 1), almost all samples follow anti- Deccan Traps flood basalt province, which, along with the Chicxulub bolide impact, cipated “younging-up” temporal order based on is implicated in the Cretaceous-Paleogene (K-Pg) extinction approximately the independently defined regional stratigraphy 66 million years ago. We estimated Deccan eruption rates with uranium-lead (14–17) (figs. S2 and S7). The exception is the (U-Pb) zircon geochronology and resolved four high-volume eruptive periods. Katraj Ghat south of Pune city, where two sam- According to this model, maximum eruption rates occurred before and after the Downloaded from ples from what was mapped as upper Poladpur K-Pg extinction, with one such pulse initiating tens of thousands of years prior to both Formation are ~100 ka younger than samples the bolide impact and extinction. These findings support extinction models that near the Poladpur-Ambenali contact in other incorporate both catastrophic events as drivers of environmental deterioration sections. To resolve this discrepancy, we placed associated with the K-Pg extinction and its aftermath. the Poladpur-Ambenali contact in the Katraj Ghat sectionas~100mlowerthanpreviouslymapped. ontinental flood basalt provinces are char- within the Deccan Traps volcanic province, India This simple adjustment does not violate geo- http://science.sciencemag.org/ acterized by eruption of >1 million km3 of (Fig. 1). The province is temporally correlated chemical or geological observations in the strat- basalt over a period of <1 million years to the K-Pg mass extinction, in which roughly igraphy, as the Poladpur-Ambenali contact is (1, 2), representing the largest volcanic three-fourths of life on Earth was eradicated, in- geochemically transitional in published datasets C 6 14 events on Earth. Four of the five most cluding non-avian dinosaurs ( ). Previous at- ( ). Furthermore, our placement of the contact severe Phanerozoic mass extinctions [~541 mil- tempts to constrain eruption rates were limited is consistent with geochemical studies of the lion years (Ma) ago to the present] coincided by poor stratigraphic coverage and/or high nearby Sinhagad Fort section suggesting that with emplacement of one of these provinces analytical uncertainties (7–12). We used U-Pb the Poladpur Formation is relatively thin just (3). Although the temporal link between flood geochronology by isotope dilution–thermal ion- south of Pune (14). basalts and extinctions is well established, the ization mass spectrometry (ID-TIMS) (13), which To further refine the composite stratigraphic mechanisms by which eruptions drive extinc- provides analytical uncertainties (±2s)aslowas age model, we used a Bayesian Markov chain tion are poorly understood (4). Two models of 40,000 years (40 ka) for individual dated zircons. Monte Carlo (MCMC) model in which strati- on March 18, 2019 environmental change from volcanic activity Our sampling covers the nine major Deccan for- graphic superposition is imposed on U-Pb zircon relate to eruptive volatile emissions (1, 4). The mations in the Western Ghats, where the most dates (13, 21) (Fig. 1). The result is a deposition first is volcanogenic CO2 release, with associated voluminous (>90% total volume) and complete age estimate for each dated horizon, incorporat- global warming, ocean acidification, and carbon Deccan exposures are preserved (14–17) (Fig. 1). ing dates from all beds above and below each cycle disruption. The second is SO2 injection into We sampled both coarse-grained basalts and sample to produce an internally consistent age the stratosphere and its conversion to sulfate sedimentary beds between basalt flows that model (Fig. 1). The accuracy of refined age esti- aerosols, causing global cooling, acid rain, and infrequently contain zircon-bearing volcanic mates depends solely on sample placement in ecosystem poisoning (5). The predicted time ash (11) (fig. S1). These beds, locally termed proper stratigraphic order and is independent of scales of these perturbations contrast sharply. “redboles,” range from oxidized volcaniclas- samples’ exact stratigraphic heights. The emission of SO2 from a single eruption would tic material with visible lithic fragments and To calculate volumetric eruption rates through produce years of cooling, whereas accumulated phenocrysts to paleosol-type horizons produced the Deccan Traps, we adopted the volume model greenhouse warming from CO2 can be sustained by in situ weathering of flow tops (18, 19). Of 141 of Richards et al.(22), in which units of the for many thousands to tens of thousands of sampled redboles and coarse-grained basalts Wai subgroup (i.e., the Poladpur, Ambenali, and years. Testing the effects of this interplay on (Fig. 1 and figs. S1 and S2), 23 redboles and Mahabaleshwar Formations) were interpreted ecosystems thus requires precisely calibrated one basalt sample yielded sufficient zircon (≥5 as more voluminous than is apparent from volcanic eruption rates that can be correlated crystals) to estimate an eruption age, includ- their proportionate thickness in the Western to high-resolution climate proxy and biostrat- ing four distinct bole horizons and one basalt Ghats. This assertion carries nontrivial uncer- igraphic data. previously presented by Schoene et al.(11). tainties, but we believe it is justified given the We applied U-Pb zircon geochronology to Pristine volcanic crystal morphology indicates correlation of these formations to basalt flows construct a precise temporal record of eruption minimal transportation or reworking of zircon on the province’s periphery, including massive in a sedimentary environment. Consequent- flows that traveled ~1000 km to India’s eastern ly, we inferred that this volcaniclastic, zircon- shore (23, 24). Although different volume models 1Department of Geosciences, Princeton University, Princeton, NJ, USA. 2Nuclear and Chemical Sciences Division, Lawrence bearing material was incorporated into redboles produce changes in the magnitude of calculat- Livermore National Laboratory, Livermore, CA, USA. as air fall tuff, consistent with some redboles ed eruption rates, the timing of peak eruption 3Berkeley Geochronology Center, Berkeley, CA, USA. 4ISTE, containing a high-SiO2 (nonbasaltic) com- rates is unaffected by either the volume model Institut des Sciences de la Terre, Université de Lausanne, 5 ponent (19), and that these zircons provide or the interpretation approach of the zircon GEOPOLIS, Lausanne, Switzerland. Department of Geology, 13 Amravati University, Amravati, India. a robust means for dating Deccan eruptive data ( ) (figs. S8 and S9). Additional uncer- *Corresponding author. Email: [email protected] stratigraphy. tainty relates to the unconstrained mass and Schoene et al., Science 363, 862–866 (2019) 22 February 2019 1of4 RESEARCH | REPORT age of Deccan basalt that is currently submerged from ~66.3 to 66.15 Ma ago; the second to the estimate a date of 66.016 ± 0.050 Ma ago for and inaccessible off India’s western shore. We Poladpur Formation from ~66.1 to 66.0 Ma ago; the KPB [95% credible interval, internal un- consider this uncertainty to be intractable be- the third to the Ambenali Formation from ~65.9 certainties only (13)]. Comparison of our data cause current volume models cannot account to 65.8 Ma ago; and the fourth and final to the with recently published 40Ar/39Ar geochronology for this mass component of the province. Con- uppermost Mahabaleshwar Formation, from from the Deccan Traps and the Chicxulub im- sequently, all eruption rates are likely mini- ~65.6 to 65.5 Ma ago. pact (12, 26) is currently not possible at the mum estimates, although we also cannot assess OurDeccaneruptionmodel(Fig.2)constrains necessary level of precision given systematic whether the offshore component erupted during the volcanic tempo with high resolution, provid- bias between the two dating methods, pri- the same time intervals as that of the Western ing a means to correlate eruption records with marily related to uncertainty in ages of 40Ar/ Ghats. biostratigraphic and climate proxy data across 39Ar fluence monitors and the values of the We converted our age model into a proba- the K-Pg extinction. Our model places the second 40K decay constant and physical constants (13). bilistic estimate of volumetric flux of basaltic