John Schults / Reuters

Identifying a killer can be difficult when it seems like every murder weapon imaginable has been used in the crime, and when the victim is the entire planet. About 252 million years ago, a rich and wonderful world was annihilated in the worst mass extinction ever: the end-Permian, a catastrophe with no close competitor in Earth’s history. Volcanoes of a truly preposterous scale erupted in Siberia over many thousands of years, loosing all manner of chaos on the world. Rounding up, everything died.

Diagnosing the particular flavor of chaos responsible for this mass death has proven elusive. The Siberian Traps, now long retired as a vast swath of basalt plateaus in the far northern reaches of Russia, might have poisoned the world with mercury. Or maybe they destroyed the ozone layer by incinerating huge underground layers of ancient evaporites. Or perhaps they acidified the planet with sulfur dioxide and carbon dioxide, stripping vegetation, killing corals worldwide, and altering the chemistry of the planet’s soil so that dirt would have tasted like vanilla. Or maybe the Siberian Traps wracked the planet with brief volcanic winters, or they poisoned the planet with carbon dioxide itself, or the oceans became stagnant and poisoned with toxic hydrogen sulfide. Maybe that’s what killed everything.

Or maybe the oceans became stratified and nutrient-starved, and phytoplankton suffered. Or perhaps it was a lack of oxygen in the ocean that suffocated everything, or maybe it got just too damn hot. Some paleontologists have thrown up their hands and yielded to this overdetermination of kill mechanisms at the end-Permian, proposing an inelegant Murder on the Orient Express hypothesis of mass extinction that implicates the entire suite of killers. A new study, though, claims to pinpoint the primary killer from this murderers’ row: Among the slew of Very Bad Things implicated in the worst calamity the Earth has ever known, it was the global-warming-driven ocean anoxia that stands out as the primary agent of Armageddon. And in this reaper of the Paleozoic, the study’s authors see a future menace.

“I think this study shows the end of the road that we’re heading down,” says the lead author, Justin Penn of the University of Washington, about our modern, warming, and increasingly suffocating oceans.

As we’ve known since the 1860s, carbon dioxide is a very important greenhouse gas, and if you increase the amount of it in the atmosphere, it makes the planet warmer. At the end of the Permian, the volcanoes of Siberia, by burning through a giant basin filled with coal and other carbon-rich rocks, emitted enough carbon dioxide to warm the planet by a sweltering 10 degrees Celsius or so. Coincidentally, this is about the same magnitude of warming predicted for humanity in a modern burn-it-all scenario.

“If we are truly the stupidest intelligent species ever, we probably could do the same thing,” says Curtis Deutsch, a co-author on the study. “As it is, we’re headed toward 3 to 4 degrees Celsius of warming by the end of the century, which is nothing to sniff at. But 10 degrees isn’t that off the charts.”

As the world warmed 10 degrees more than 250 million years ago, up to 96 percent of species in the ocean went extinct (for comparison, roughly zero percent of modern species in the ocean have gone extinct so far in what’s been called “the sixth extinction”). For all the planet knew, it was time for complex life to close up shop after a nice quarter-billion-year run. After several millennia of eruptions from the Siberian Traps, the fossil record remained startlingly impoverished for almost 10 million years, before limply convalescing in the ensuing Triassic period.

To tease out what was primarily responsible for all this death in the oceans—in this, “the Great Dying”—Penn and his colleagues developed a climate model, not dissimilar to those used to project future warming on our own modern world. Only in Penn’s model, the continents were reunited. This was Pangaea, the mythic supercontinent that reached its apotheosis in the Permian period, joining Morocco with New Jersey and India with Antarctica. Then the team lit this ancient world on fire.

By jacking up the CO2 in their model high enough, Penn and his colleagues were able to re-create the scorching temperatures of the end-Permian mass extinction, searing their ancient climate model maps in worrying shades of red. The flip side of a hot ocean is one with less oxygen, and it has long been known that the oceans of the end-Permian were gasping for the stuff. This paleoceanographic fact has been uncovered by geologists who have found the sickly presence of laminated, pyrite-rich ocean rocks in end-Permian rock outcrops around the word—from the Salt Range in Pakistan to the old whaling redoubt of Spitsbergen in the Arctic Ocean. Even subtler, uranium isotopes in rocks from the catastrophe whisper dark rumors about the asphyxiation of the entire ancient ocean.

By driving up the temperature in his model, Penn re-created this end-Permian oxygen loss as well. This is because oxygen is less soluble in hot water, so heat alone can cause oxygen to plummet in the ocean. But a hotter ocean is also a more sluggish and stratified one, so the depths become starved of oxygen’s delivery as well. Even worse, the hotter it gets, the more oxygen animals need to power their metabolism, so hot water quickly creates a crisis of supply and demand. By populating his model ancient ocean with modern creatures, such as sharks and crabs and corals—creatures with a variety of tolerances to things like deoxygenation, temperature, and pH—and then letting all climate hell break loose, Penn found that it was the heat-driven loss of oxygen in the ocean, more than any other factor, that could explain the end of the world.

While essentially nothing was spared in this mass extinction, it appears from the fossil record that life at the poles was especially marked for destruction. And this makes sense in a world losing its oxygen. There’s more oxygen in colder waters at the poles, so the creatures there have adapted to it—and are especially sensitive to its diminishment. Meanwhile, tropical creatures are adapted to warmer, more oxygen-poor waters. So if anyone is to survive on a world that’s about to become outrageously hot and lose three-quarters of its oxygen, it will be those hardy inhabitants of the low latitudes. And that’s what’s seen in the fossil record. The tropics still faced an unthinkable cataclysm, but they fared mildly better. The predicament for polar creatures was not unlike those luckless creatures that today inhabit the tops of mountains, and that will have no higher altitude to which to escape in the coming decades. “They’re shit out of luck,” Deutsch says.

When the ocean loses its oxygen, it’s something of a great leveler compared with other hypothesized mass-extinction mechanisms. Ocean acidification, for instance (what happens when too much CO2 reacts with seawater), has previously been proposed as the great killer of the end-Permian. But while acidification can have a surprisingly variable effect on the survival of different kinds of sea life, there is hardly any selectivity at all when oxygen disappears from an ecosystem. Everyone dies, matching the near-universal signal of slaughter in the ancient ocean.

“This study suggests we should be worrying much more about hypoxia than about ocean acidification,” Deutsch says. “There’s vastly more resources being put into [studying] organisms’ responses to pH in seawater than there is into understanding temperature-dependent hypoxia. I think that the field has basically allocated those resources in exactly the wrong way.”

The modern oceans have already lost 2 percent of their oxygen since 1960, a remarkable loss driven mostly by coastal nutrient pollution and global warming. It’s an environmental problem that promises to worsen in the warmer world of the coming centuries, just like it did in the end-Permian. And if Earth’s past is any indication of its future, this asphyxiation could be truly world changing. The prospect has led dozens of paleoclimatologists, geochemists, and oceanographers to sign the Kiel Declaration on Ocean Deoxygenation, developed this September to raise global awareness of a problem with increasingly worrying geological precedent.

“This study shows that we’re on that same road toward extinction, and the question is how far down it we go,” Penn says.

To head off warming and drastic deoxygenation of the oceans, humanity can either stop burning fossil fuels or artificially dim the amount of sunlight reaching Earth’s surface—a kludgy solution that goes by the name of geoengineering. One of the most hotly debated methods of geoengineering would involve shooting sulfate aerosols out of planes and into the atmosphere to reflect sunlight back into space. Unfortunately, says the City College of New York geoscientist Benjamin Black, the end-Permian mass extinction represents something of a proof of concept for the idea, and it’s not pretty.

In a paper published last week in Nature Geoscience, Black and his co-authors describe not only the extreme warming of the climate around 252 million years ago, but also the extreme climate whiplashes that throttled the planet on much shorter timescales. These environmental seesaws, the researchers show, were induced by volcanic sulfur dioxide from the Siberian Traps, which would have briefly cooled the planet and masked the warming before being rained out of the atmosphere and allowing the planet to jump back into the high-CO2 inferno. Similarly, if humans embark on a geoengineering project of shooting sulfur dioxide into the air to head off all of the problems of warming, without drastically reducing the concentration of CO2 in the atmosphere—and if they ever give up on this project after it starts, for whatever reason, at any time in the next few thousand years—Black can imagine a similar disaster unfolding.

“So if you think about the Siberian Traps and the end-Permian mass extinction as a feasibility study for global warming and sulfur geoengineering, the results of that study were … not the best,” he said, referring to, quite literally, the worst thing that’s ever happened. “It’s worth asking ourselves whether we really want to make our current predicament even more like the end-Permian mass extinction.”

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.