Think back to any dinosaur illustration you saw as a kid. The background was almost certainly one of two things: an asteroid streaking across the sky or a volcano blowing its top. (If the illustrator was feeling extra dramatic, maybe both.)
A 6-mile-wide asteroid, which hit the coast of the Yucatán Peninsula 66 million years ago, obliterated any nearby dinos and filled the sky with material that plunged the planet into a species-dooming winter. But don’t sell those volcanoes short. A growing body of geological evidence is suggesting that the dinosaurs were already enduring climatic chaos before the asteroid, thanks to huge, relentless volcanism in India’s Deccan Traps.
For 300,000 years before impact, and for another 500,000 after it, these volcanoes emitted vast clouds of carbon dioxide and sulfur dioxide. Even when they weren’t actively erupting, they were doing “pre-eruptive” degassing. The CO2 heated the planet—as humanity’s emissions are doing today—and the SO2 cooled it by reflecting some of the sun’s energy back into space. The back-and-forth created a climatic whiplash that drove a mass extinction. So the asteroid wasn’t a singular Grim Reaper for the dinosaurs, but the coup de grâce that sealed their fate. At least, that’s how the theory goes.
Scientists are still debating just how much of a role this volcanism might have played in the Cretaceous–Paleogene extinction, compared to the obvious ouchie of an asteroid slamming into the planet. Now, a machine-learning computer model has weighed in, finding that the CO2 and SO2 gas required to cause the extinction of the dinosaurs is consistent with the output of the Deccan Traps.
“Our findings specifically lend credence to the idea that volcanism was disturbing the atmosphere and the climate way before the asteroid,” says Dartmouth College computational geologist Alexander Cox, lead author of a new paper in Science. “You can actually recreate the environmental conditions that could cause a dinosaur extinction solely by volcanism, as if the asteroid weren’t there. But of course, we can't discount the fact that the asteroid definitely didn't cheer up the dinosaurs.”
“This study is really interesting. No one has really done anything like this before,” says University of Florida geochronologist Courtney J. Sprain, who studies the extinction but wasn’t involved in the research. “In the last couple of years, there's really been a shift into reexamining the role of volcanism in the mass extinction.”
The Deccan Traps take their name from the Swedish trappa, meaning “stair,” due to the step-like outcroppings there. Over the course of almost a million years, their eruptions produced a million cubic kilometers of lava, burping up 10.4 trillion tons of CO2 and 9.3 trillion tons of SO2. For perspective, between 2000 and 2023, humans emitted 16 billion tons of CO2 per year, which is around 100 times the rate from the Deccan Traps. So this volcanism was a slower release of planet-warming gas, but it happened for hundreds of thousands of years. This rise in greenhouse gases, of course, warmed the climate, although the SO2 had a contradictory cooling effect.
Geologists already have historical climate data thanks to a proxy: tiny oceanic organisms known as foraminifera, which built shells of calcium carbonate, died, and sank to the seafloor to become rock. By looking at the different isotopes of carbon and oxygen in these ancient shells, scientists can determine both atmospheric carbon concentrations and ocean temperatures millions of years ago.
Instead of working from this geological record, the new model worked toward it. It ran 300,000 scenarios of how different levels of CO2 and SO2 might together produce a climate that matches the objective data from the shell fossil record.
Using an array of 128 computer processors crunching in parallel, the scientists could play with random atmospheric concentrations of volcanic CO2 and SO2, and see what values of carbon and oxygen isotopes that led to. The model could then compare those values to the actual data in the fossil record and give itself a score. “It might say, ‘Let’s try adding a bit of CO2 here and a bit less SO2 here,’” says Cox. “We let the model run wild—it keeps going again and again and again.”
The 128 processors could also compare their scores to one another. “They each choose a solution from a processor that they think is doing particularly well,” says Cox. “And it just keeps going, like how you might copy off your classmates at school if you think they're smarter than you. That way, all the processors can kind of drag everyone else up to get to a solution a lot quicker.” In the end, running those 300,000 simulations in parallel took a matter of tens of hours, even with so much data to churn through.
“The beauty of what these folks have done is that they can feed it really objective data that nobody would dispute, and come up with some surprisingly detailed inferences,” says geologist Paul Renne, director of the Berkeley Geochronology Center, who studies volcanoes and mass extinctions but wasn’t involved in the paper. One of those inferences, he says, is “that the carbon dioxide and the sulfur outputs are decoupled, which is something that I've been arguing for a long time.”
In other words, the volcanoes didn’t belch up CO2 and SO2 in equal proportions all the time. An uneven mix resulted in a gnarly climate seesaw that ultimately wiped out organisms. A species might have been able to adapt to rising temperatures, only to be doomed by falling temperatures 50,000 years later, or vice versa. (CO2 also lasts way, way longer in the atmosphere than SO2 does, so it could accumulate over thousands of years.)
The volcanism would have also spawned acid rain and ocean acidification. Those factors alone helped kill off plants, then the herbivores, and then the carnivores that fed on them. But it also skewed the planet’s carbon cycle as a whole. For example, as the oceans acidified, creatures like the foraminifera, which constructed their shells of carbon, would have struggled to survive. And if fewer of their shells were available to store carbon and sequester it on the ocean floor, says Sprain, “it's perturbing the system even more.”
Then came the asteroid. The fireball and shock wave instantly wiped out nearby organisms. To make things extra bad, it slammed into a sulfur-rich part of Earth. Unlike the more gradual volcanism of the previous 300,000 years, the impact immediately fired all that sulfur into the atmosphere, along with tiny bits of rock and glass. It created a haze that enveloped the Earth, blotting out the sun.
The climate chaos was now full-tilt climate anarchy. “The Deccan Traps are still erupting after that event,” says Yale University geochronologist Jennifer Kasbohm, who studies volcanoes’ influence on climate change but wasn’t involved in the new paper. “It was a really bad day, but things continue to be kind of rough on planet Earth for a few hundred thousand years more. And then maybe you're getting back to normal, although missing some of your old friends, in terms of species that used to be around.”
The asteroid’s impact released so much energy, in fact, that it may have triggered more volcanic activity at the Deccan Traps. The shock could have essentially jostled the plumbing in the volcanoes, driving magma to the surface. “There would be a huge amount of energy being propagated around the world,” says Renne. “When you perturb that system violently, you can actually do several things, one of which is you induce gases to be released from the liquid. I hate to use such a simplistic analog, but it's like shaking a can of soda. That triggers eruptions.”
However, not everyone in the scientific community is Team Volcano. “There are people who have argued—and continue to argue—very, very fiercely about this,” says Peter Roopnarine, curator of geology at the California Academy of Sciences, who wasn’t involved in the research. “I would say that the asteroid definitely played a major, if not the major, role. Really, the question is, to what extent did the volcanism also play a role?” There are still many uncertainties, Roopnarine says, such as the timing of the volcanic gas release over those hundreds of thousands of years.
An asteroid also causes a distinctly different kind of climate trauma than a slow gas leak. “In the opinion of many of us, the major kill mechanism—if you will—from the impact would have been darkness, not cooling,” says Roopnarine. “You would have enough material injected into the atmosphere that it would have blotted out the sun up to, let's say, 10 years or so.”
All of this has added a great deal of subtlety to the asteroid vs. volcanoes debate. This new modeling tries to set human biases aside and let machines do the number crunching. So far, the answer seems to be that both natural disasters played a role, in a kind of one-two punch. “To many people who aren't really following the field that closely, it seems like, ‘What are you guys doing? You’re flailing around. Get to the answer, goddammit,’” says Renne. “Twenty years ago, people would stop at ‘Well, was it an asteroid or was it the volcanism?’ And that was as far as the argument went. Now, people are much more concerned about the nuances.”
Update, 10/3/23, 2:15 pm ET: This story was updated to correct the use of the term "machine learning.”
