In the late 13th century, Native Americans living in cliff dwellings in the what is now the Four Corners region of the U.S. Southwest abandoned their homes en masse. They moved into river valleys, closer to the water that had become a scarce resource in the midst of one of the worst “megadroughts” to hit the western part of the continent in the last millennium.

Scientists who study climate history have long been concerned that rising temperatures could help tip the already dry region into megadroughts (generally defined as droughts lasting 20 years or longer) again more easily in the future—a major worry in a rapidly growing part of the country that has already had trouble meeting water needs during nearly two decades of drought.

Park Williams, a bioclimatologist at Columbia University’s Lamont Doherty Earth Observatory, had always envisioned past megadroughts as having conditions unrecognizable to us today. But several years ago, he realized that the massive tree die-offs, bark beetle infestations and wildfire boom he envisioned for those past events had actually come to pass in recent years. The epiphany led him to look at how the region’s recent drought compares to the megadroughts of the last millennium—and he discovered it rivals them in severity. If it continues, it could be the first megadrought of the era of human-caused climate change. Scientific American spoke with Williams earlier this month about his research, at the annual meeting of the American Geophysical Union in Washington, D.C.

[An edited transcript of the interview follows.]

How do we know what past megadroughts were like?

The primary way that we know about the megadroughts is from tree ring records. Trees grow annual tree rings every year. During years of really good growth the ring is wide, and during years of bad growth the ring is narrow. And so in dry places where trees are struggling during the driest years, then droughts are preserved in the tree ring records as clusters of narrow rings.

There are other pieces of evidence about these megadroughts as well. In California’s Sierra Nevada [mountains], there were two periods between, say, 850 A.D. and 1300 A.D. , each about 100 to 150 years long, when it was so persistently dry that riverbeds that are now perpetually in water were dry enough persistently to be able to actually allow for trees to grow to maturity. And now we see in those riverbeds—and the bottoms of some lakes—the stumps of those trees. Scientists have gone and collected wood from these stumps and done carbon dating on them to place them during these megadroughts.

How do tree rings help compare past megadroughts to drought today?

In western North America, we’re really lucky that we have tree ring records from across the continent, and at a really high density—there’s more tree ring records in this small area than any other area in the world. Many of them have preserved records of drought within them with extremely high fidelity. The reason for that is western North America is just a dry place, so trees in general are going to store drought information really well in their tree ring records. So we took advantage of this really high density of drought-sensitive tree ring records to reconstruct how soil moisture has varied over the last 1,200 years. And we came up with one single record of drought variability from 800 A.D. to present for the western U.S., using those tree ring records. Then we were able to compare this current 19-year drought to all other 19-year periods since 800, in order to assess how bad is this drought—not only in the century-long context but in the millennial-long context.

When did you start thinking that this current drought might be on par with these past megadroughts?

In 2012 I did a study using tree records across the Southwest. I saw that the 13-year drought from 2000 through 2012 was standing out like a sore thumb—like far worse than anything in the last 400 years. It was looking a lot like the first 13 years of the megadrought in the late 1500s. So that’s how I got thinking about this. And actually, if we now retrospectively still look at 2000 through 2012, that was a really, really intense 13 years. So we found that this 19-year period comes in fourth [in terms of severity] in the last 1,200 years—but that 13-year period might actually come in first in [all] 13-year periods when we look across the entire western U.S.-northern Mexico region.

Did climate change contribute to this recent drought, and could it cause it to continue?

Climate change can affect droughts in a couple of ways. One is if precipitation were to change due to climate change—then droughts would undoubtedly change. The other is if the evaporative demand of the atmosphere were to change, then droughts can change even if there’s no long-term change in precipitation. And that’s what we’ve seen in western North America, where the last century … has not seen really big trends in precipitation. Climate models are kind of all over the place, in terms of whether or not they simulate changes in precipitation to arise as a function of increasing greenhouse gas in the atmosphere. But what has happened is that it has warmed rapidly in the last century—the western U.S., northern Mexico has warmed by about just over a degree and a half Celsius over the last 120 years. And climate models say that all of that warming is due to human-caused climate change. Warming enhances the evaporative demand to the atmosphere, causing essentially every raindrop or snowflake to be a little less valuable because it’s going to be more easily evaporated back to the atmosphere. That alone can drive soil moisture down, and that’s what we’ve seen so far.

We can retroactively calculate how soil moisture would have been impacted if the long-term warming trends that were caused by humans had never occurred. And in doing that, we see that this 19-year drought still would have occurred. It would still have been bad; it would compete with the worst droughts of the last century. But it wouldn’t be able to compete with the droughts of the last millennium [without climate change]. We calculate that about 38 percent of the severity of the current 19-year drought was driven by the long-term warming trend, and then the other 62 percent of the drought would have still occurred anyway, just because of a [random] reduction in precipitation.

Could this drought eventually rival past megadroughts in duration as well?

In western North America, generally La Niña-type conditions [when the tropical Pacific Ocean is cooler than average] will promote drought—they aren’t a guaranteed cause of drought, but generally our big droughts coincide with La Niña events. And oftentimes droughts are ended by big El Niño events [when the tropical Pacific is unusually warm] that result in high precipitation. So it’s entirely possible that in the next few years there will be a big El Niño event—or even a 10- or 20-year period of El Niño-type sea surface temperatures—and that could end the drought. But precipitation is increasingly competing with temperature in terms of its ability to dictate the beginning and ends of drought, and as temperatures are going up, then it’s going to be harder and harder for an El Niño event to end this current drought. So it’s entirely possible this drought will end in the next few years or decade—but it is increasingly probable the drought will persist.