As they say, ‘everyone knows’ that the dinosaurs were snuffed out, except, of course, for those that had evolved to become birds and somehow survived. When it happened is known quite precisely – at the end of the Cretaceous (66.043 ± 0.011 Ma) – and there were two possible causal mechanisms: emissions from the Deccan Trap flood basalts and/or the Chicxulub impact crater. But what was the Cretaceous-Palaeogene (K-Pg) boundary event actually like? Many have speculated, but now there is evidence.
In 2016 a deep-sea drilling rig extracted rock core to a depth of 1.35 km beneath the sea floor off Mexico’s Yucatan Peninsula, slightly off the centre of the circular Chicxulub structure (see K-T (K-Pg) boundary impact probed, November 2016). This venture was organised and administered jointly by the International Ocean Discovery Program IODP) and the International Continental Scientific Drilling Program (ICDP) as Mission Specific Platform Expedition no. 364. Results from the analysis of the cored rock sequence have been generating pulses of excitement among palaeontologists, petrologists and planetary scientist on a regular basis. The science has been relatively slow to emerge in peer-reviewed print. Appetites have been whetted and the first substantial paper is about the bottom 130 metres of the core (Gulick, S.P.S. and 29 others 2019. The first day of the Cenozoic. Proceedings of the National Academy of Sciences. 9 September 2019; DOI: 10.1073/pnas.1909479116). It might seem as though the publication schedule has been stage managed to begin with, literally, the ‘bang’ itself.
The deepest 20 m thick layer is mainly silicate glass. It was formed in the seconds after the 12 km-wide impactor arrived to smash through the water and sea-floor sediments of the early Caribbean Sea, at speed of around 20 Km s-1. It vaporised water and rock as well as shoving aside the surrounding sea and blasting debris skyward and outward. In an instant a new hole in the crust was filled with molten rock. The overlying rock is a veritable apple-crumble of shattered debris mixed with and held together by glass, and probably formed as water flowed into the crater to result in explosive reaction with the molten crystalline crust beneath. The fragments lessen in size up the core, probably reflecting ejected material mixed in the displaced seawater. Impact specialists have estimated that this impactite layer formed in little more than ten minutes after collision. The glass-laden breccia is abruptly capped by bedded sediments, considered to have been delivered by the backwash of a huge, initial tsunami. In them are soils and masses of charcoal, from the surrounding land areas, scorched and burnt by the projectile’s entry flash, inundated by the tsunami and then dragged out to sea as it receded. These are the products of the hours following the impact as successive tsunamis swashed to and fro across the proto-Caribbean Basin; hence ‘The first day of the Cenozoic’, of Gulick et al.’s title.
Other cores drilled beyond the scope of the Chicxulub crater during offshore oil exploration show a sequence of limestones with thick beds of gypsum (CaSO4.2H2O). Yet the crater debris itself contains no trace of this mineral. Around 325 Gt of sulfur, almost certainly in the form of SO2, entered the atmosphere on that first day, adding to the dust. Ending up in the stratosphere as aerosols it would have diffused solar radiation away from the surface, resulting in an estimated 25°C global cooling that lasted 25 years. The sulfur oxides in the lower atmosphere ended up in acid rain that eventually acidified the upper ocean to devastate shallow-marine life.
Predictably, the dialogue between the supporters of the Deccan Trap flood basalts and the Chicxulub impact as triggers that were responsible for the mass extinction at the end of the Mesozoic Era (the K-Pg event) continues. A recent issue of Science contains two new approaches focussing on the timing of flood basalt eruptions in western India relative to the age of the Chicxulub impact. One is based on dating the lavas using zircon U-Pb geochronology (Schoene, B. et al. 2019. U-Pb constraints on pulsed eruption of the Deccan Traps across the end-Cretaceous mass extinction. Science, v. 363, p. 862-866; DOI: 10.1126/science.aau2422), the other using 40Ar/39Ar dating of plagioclase feldspars (Sprain, C.G. et al. 2019. The eruptive tempo of Deccan volcanism in relation to the Cretaceous-Paleogene boundary. Science, v. 363, p. 866-870; DOI: 10.1126/science.aav1446). Both studies were initiated for the same reason: previous dating of the sequence of flows in the Deccan Traps was limited by inadequate sampling of the flow sequence and/or high analytical uncertainties. All that could be said with confidence was that the outpouring of more than a million cubic kilometres of plume-related basaltic magma lasted around a million years (65.5 to 66.5 Ma) that encompassed the sudden extinction event and the possibly implicated Chicxulub impact. The age of the impact, as recorded by its iridium-rich ejecta found in sediments of the Denver Basin in Colorado, has been estimated from zircon U-Pb data at 66.016 ± 0.050 Ma; i.e. with a precision of around 50 thousand years.
Because basalts rarely contain sufficient zircons to estimate a U-Pb age of their eruption, Blair Schoene and colleagues collected them from palaeosols or boles that commonly occur between flows and sometimes incorporate volcanic ash. Their data cover 23 boles and a single zircon-bearing basalt. Sprain et al. obtained 40Ar/39Ar ages from 19 flows, which they used to supplement 5 ages obtained by their team in previous studies that used the same analytical methods and 4 palaeosol ages from an earlier paper by Schoene’s group.
The zircon U-Pb data from palaeosols, combined with estimates of magma volumes that contributed to the lava sequence between each dated stratigraphic level, provide a record of the varying rates at which lavas accumulated. The results suggest four distinct periods of high-volume eruption separated by long. periods of relative quiescence. The second such pulse precedes the K-Pg event by up to 100 ka, the extinction and impact occurring in a period of quiescence. A few tens of thousand years after the event Deccan magmatism rose to its maximum intensity. Schoene’s group consider that this supports the notion that both magmatism and bolide impact drove environmental deterioration that culminated in mass extinction.
The Ar-Ar data derived from the basalt flows themselves, seem to tell a significantly different story. A plot of basalt accumulation, similarly derived from dating and stratigraphy, shows little if any sign of major magmatic pulses and periods of quiescence. Instead, Courtney Sprain’s team distinguish an average eruption rate of around 0.4 km3 per year before the K-Pg event and 0.6 km3 per year following it. Yet they observe from climate proxy data that there seems to have been only minor climatic change (about 2 to 3 °C warming) during the period around and after the K-Pg event when some 75% of the lavas flooded out. Yet during the pre-extinction period of slower effusion global temperature rose by 4°C then fell back to pre-eruption levels immediately before the K-Pg event. This odd mismatch between magma production and climate, based on their data, prompts Sprain et al. to speculate on possible shifts in the emission of climate-changing gases during the period Deccan volcanism: warming by carbon dioxide – either from the magma or older carbon-rich sediments heated by it; cooling induced by stratospheric sulfate aerosols formed by volcanogenic SO2 emissions. That would imply a complex scenario of changes in the composition of gas emissions of either type. They suggest that one conceivable trigger for the post-extinction climate shift may have been exhaustion of the magma source’s sulfur-rich volatile content before the Chicxulub impact added enough energy to the Earth system to generate the massive extrusions that followed it. But their view peters out in a demand for ‘better understanding of [the Deccan Traps’] volatile release’.
A curious case of empiricism seeming to resolve the K-Pg conundrum, on the one hand, yet pushing the resolution further off, on the other …