Between 1545 and 1548, an epidemic swept through the indigenous people of Mexico that is unlike anything else described in the medical literature. People bled from their face while suffering high fevers, black tongue, vertigo, and severe abdominal pain. Large nodules sometimes appeared behind their ears, which then spread to cover the rest of their face. After several days of hemorrhage, most who had been infected died.

The disease was named cocoliztli, after the Nahautl word for “pest.” By contemporary population estimates, cocoliztli killed 15 million people in the 1540s alone—about 80 percent of the local population. On a demographic basis, it was worse than either the Black Death or the Plague of Justinian. For several centuries, its origin remained a mystery.

Then, about two decades ago, researchers began to compare the known cocoliztli outbreaks with clues etched in the tree rings of modern-day Mexico. They found that cocoliztli struck during an apparent “mega-drought,” a decades-long period with little rain. Central Mexico suffered two mega-droughts in the 16th century, but, paradoxically, 1545 was a comparatively wet year in the drought. Cocoliztli itself also presented a problem: Unlike smallpox, which devastated the indigenous Mexican population starting in 1520, cocoliztli’s symptoms don’t resemble a known Old World disease.

So researchers advanced a hypothesis: Cocoliztli was some kind of animal-spread hantavirus or arenavirus normally contained in Mexico’s highlands. When a brief wet period allowed the population of rodents (or otherwise) to boom, cocoliztli was able to take hold. The disease may still lurk in the highlands, waiting for an opportunity to arise.

That, at least, is the hypothesis. Researchers are hampered in part because the 16th century is the last time that the deserts of southern North America experienced a mega-drought. Alas, they may get another opportunity soon.

A new study, published last week in Science Advances, says that climate change will make a similar mega-drought far more likely in the American Southwest. In fact, this kind of phenomenon could become a near certainty: If carbon emissions continue unabated, the risk of a mega-drought could exceed 99 percent.

“This will be worse than anything seen during the last 2,000 years and would pose unprecedented challenges to water resources in the region,” says Toby Ault, a professor of earth science at Cornell University and one of the authors of the study, in a statement. “As we add greenhouse gases into the atmosphere—and we haven’t put the brakes on stopping this—we are weighting the dice for mega-drought conditions.”

Here’s what “weighting the dice” looks like from a scientific perspective: The study found that if carbon emissions continued on their current trajectory, and if global warming did not create a generally rainier climate in the American West, then the chance of a mega-drought would exceed 90 percent. But if climate change decreases precipitation—which is what most models predict—then the rate would sit at 99 percent, making mega-drought a virtual certainty.

Somewhat counterintuitively, even if climate change makes the West rainier, a mega-drought is still more likely than it would be otherwise. A warmer world will put more demand on trees and other plants, requiring them to pull more water out of the ground; water will also evaporate faster from reservoirs and the soil. So even if global warming increases the amount of rain, then the chance of a mega-drought would still exceed 70 percent.

Jonathan Overpeck, a professor at the University of Arizona, invented the word mega-drought in the 1990s with his colleague Connie Woodhouse. He was not involved with this paper. He told me that this new paper actually researches an especially intense form of the phenomenon, only examining a mega-drought that would spell infernal heat and terrific dryness for at least 35 years.

“This isn’t like the drought of our grandfathers,” he says. “It’s a drought that everyone would agree would be devastating as all heck.”

Overpeck said that the paper laid out two scenarios for the American Southwest. In the first, the world continues to pump carbon dioxide into the atmosphere, and a post-2050 “searing mega-drought” becomes a near certainty. In the second, carbon emissions soon begin to decrease, and the region sees about a 66 percent chance of a “warm mega-droughtin the second half of this century.

The “searing drought” possibility scares him most. Most research predicts that a climate-addled “searing mega-drought” would be much worse than anything described in the 2000-year-long store of evidence left behind in lake beds and tree rings. Toxic dust storms could rage across the region, making driving extremely dangerous. The vast majority of trees in the region would die. Agriculture would become all but impossible.

“The Southwest could be a really difficult place to live and make a living, at least as we know it today,” he told me. “It’s scary. I just cannot imagine a drought that long and that hot.”

A warm mega-drought, on the other hand, could probably be endured. (The water journalist John Fleck talked to Vox last month about how that could be done.) And even in a world with no climate change, mega-droughts happen every couple centuries, and the “natural” risk of one occurring in any given year is about 10 percent.

But making even the “warm mega-drought” a possibility will require a historic effort. Last week, the Paris Agreement, the first global treaty to limit carbon emissions, entered into force. But even if every country upholds its promises to restrict greenhouse-gas emissions, the world could still overshoot its 2-degree “emissions budget” by 2030.

So we’re left thinking about the cocoliztli hypothesis. That idea speaks to the devastating consequences that can follow from even a modest disruption of the normal climate. It suggests that a climatic tragedy can trigger a human one, and a lack of rain can bring about consequences far more dire than a forest of dead trees. A drought, in other words, is never just a drought.