Click the link to read the article on the Yale Climate Connections website (Bob Henson):

August 13, 2025

Climate change appears to have driven an ongoing 25-year shortfall in winter rains and mountain snows across the U.S. Southwest, worsening a regional water crisis that’s also related to hotter temperatures and growing demand. Multiple studies now suggest that human-caused climate change is boosting an atmospheric pattern in the North Pacific that favors unusually low winter precipitation across the Southwest. 

This weather pattern – known to scientists as a negative mode of the Pacific Decadal Oscillation, or PDO – is one phase of a slow-moving swing between warm and cool temperatures in the northeast and tropical Pacific Ocean. The PDO’s monthly value for July was the lowest in 171 years of data (see Fig. 1 below).

Climate change was already implicated in warming temperatures that pull more moisture from the landscape and shorten periods of mountain snow cover, thus exacerbating the impacts of dry spells. But scientists had previously assumed that the PDO’s variations over decades, which affect the rainfall and snowfall itself, were largely natural.

A study published in Nature on Wednesday, August 13, finds that emissions of climate-warming greenhouse gases and tiny sun-blocking particles called aerosols have driven long-term PDO changes over the last few decades, depriving the Southwest of much-needed winter rain and snow.

Using new techniques to extract signal from noise in model output, the researchers found that “observed PDO impacts – including the ongoing multidecadal drought in the western United States – can be largely attributed to human activity.” [ed. emphasis mine}

For the past quarter-century, precipitation across the Southwest has been on par with the driest periods in modern history. As the landscape dries, sunshine is able to warm it more effectively, helping boost temperatures even more and worsening the drying effects on the rivers, reservoirs, and landscapes crucial for the Southwest’s growing population.

Until now, those temperature effects were believed to be the main human-caused climate factor in the mix.

But scientists looked more closely at the PDO in part because of its relationship to a better-known pattern, the El Niño and La Niña oscillations in the tropical Pacific that influence weather across the world. Shorter-term La Niña events, lasting 1 to 3 years, are more common and can be stronger when the longer-term PDO phase is negative, and both of these patterns strongly favor drier-than-usual winters across the Southwest. 

During the last 25 years, La Niña has been in place for 12 winters versus just eight winters for El Niño, a tilt that has helped to reduce winter precipitation in the Southwest. The latest outlook from NOAA predicts a near-even chance of La Niña conditions yet again in 2025-26.

The Southwest’s largest two reservoirs, Lake Powell and Lake Mead, were both running at less than a third of capacity as of August 3, and total inflow for the water year ending this summer was expected to be only about 50% of average.

In recent years, the Southwest’s normally scorching heat has intensified to levels that are smashing record after record. On August 7, Phoenix reached 118 degrees Fahrenheit, the highest reading ever observed there so late in any summer in data going back 130 years.

Experts in the Phoenix area have documented a major spike in heat mortality over the past decade, as population and vulnerabilities increase along with the heat itself. More than a thousand heat-related deaths were recorded in 2023 and 2024 alone. 

Even more disconcerting is what the new work suggests for the Southwest going forward. The Nature study warns that as long as human-produced greenhouse gases and aerosols continue to produce these effects, “the PDO will remain persistent in its negative state, driving continued precipitation deficits in the western U.S.”

Confounding expectations

The puzzling behavior of the Pacific over the last several decades has drawn increasing scrutiny, especially since it’s long been expected that 21st-century warming would lead to an El Niño-like pattern. Instead, the Pacific has behaved in the opposite fashion. It’s been unclear why model projections of the PDO have been off track for so long.

“I don’t think we’ve untangled all this yet, but I think this opens up new possibilities for what models are missing,” said Jeremy Klavans of the University of Colorado Boulder, lead author of the Nature paper.

Read: A mystery in the Pacific is complicating climate projections

Figure 3. A schematic showing where sea surface temperatures are generally above and below average during the positive (warm) phase of the Pacific Decadal Oscillation. The “warm” refers to the horseshoe-like arc of warmer-than-average readings from the Gulf of Alaska along the west coast of North America and into the eastern tropical Pacific, where it often coincides with an El Niño event. Since 2000, most years have featured the opposite pattern, with the horseshoe in blue instead of red and the eastern Pacific often in a La Niña mode. (Image credit: Adapted by NOAA Climate.gov from original by Matt Newman based on NOAA ERSSTv4 data.)

Plucking the signal of climate change out of decades of noise

The large year-to-year and decade-to-decade variability in the PDO makes it hard to detect subtle but important longer-term trends. Moreover, climate models tend to exaggerate the peaks and valleys in the PDO’s natural variability. 

Scientists increasingly study questions like the PDO’s recent behavior using model ensembles – dozens of simulations from the same model for the same period, with tiny variations in the starting-point data that account for inherent uncertainty in models and observations. Klavans and colleagues found that at least 70 simulations were needed in order for a model ensemble to extract the longer-term climate-change influence from the natural variations. Their project ended up drawing on 572 ensemble members from 13 separate models. 

Like a sound mixer at a recording studio boosting an instrument that would otherwise be drowned out, the researchers amped up the strength of the PDO’s longer-term climate change signal while retaining its shorter-term variability. After this adjustment, the models ended up doing a much better job of replicating the recent multi-decade drop in winter precipitation across the Southwest. This finding suggests that the fainter, longer-term signal, obscured until now, is actually a crucial part of what’s happening.

Based on prior work in the Atlantic Ocean, it appears that the climate-change impact on the PDO stems from greenhouse gas increases as well as the global evolution of sun-blocking air pollution over the last few decades.

“We’ve now demonstrated the signal-to-noise problem in both the North Atlantic and North Pacific,” Klavans said. In both cases, the signals of longer-term climate change in atmospheric patterns were getting drowned out by the noise of natural variability. The techniques employed to get around this problem are helping to reveal strengths in model performance that can now be accessed, according to Klavans: “We think this example is just scratching the surface of what models can tell us more broadly about regional climate impacts.” 

The biggest El Niño events can sometimes push the PDO into a positive mode that can persist for years or decades, but the strong El Niño of 2023-24 didn’t accomplish that feat. Next time around, Klavans will be watching intently: “If the eastern equatorial Pacific starts warming, if we get an El Niño-like response, does it flip the PDO?”

More sleuthing bolsters the case

Another recent paper, published last month in Nature Geoscience, reinforces the idea that climate change itself has pushed the Southwest into a lower-precipitation mode since the 1980s. Using a variety of model simulations, the authors show that sun-blocking aerosol emissions appear to have teamed up with the influence of human-produced warming in the tropics to favor persistently high pressure in the North Pacific. In turn, this negative-PDO-like pattern has helped steer wintertime precipitation away from the Southwest. 

Climate scientists refer to these chains of events as “forcings”, meaning that something other than natural variability has driven, or forced, changes to weather and climate. Forcings can be anything from a one-time massive volcanic eruption to decades of sun-blocking pollution or centuries of greenhouse-gas emissions.

“The main takeaway is that there’s this forced signal in historical droughts for the Southwest since 1980, not only in temperature but also in the precipitation changes,” said lead author Yan-Ning Kuo of Cornell University.

There’s been some research suggesting that the long-expected climate-change trend toward El Niño-like patterns in the Pacific could finally emerge later this century as the world continues to warm. But even if that occurs, “it is unlikely to substantially alleviate the currently projected future drought risk,” Kuo and colleagues warn in their new paper.

“For the longest time, we chalked these precipitation changes up to natural variability,” said Cornell’s Flavio Lehner, a co-author on the paper. “I think we’re revisiting that, and it heightens the stakes. If indeed the forcings continue to act in this way, then precipitation decline in the Southwest may continue. It makes a much stronger case for human influenceRead: Wet winter won’t fix Colorado River woes

Clues from 6,000 years ago

Yet another just-published study – this one looking back thousands of years – suggests that a warming planet itself, even without human-added greenhouse gases, can help push the PDO into its drier-in-the-Southwest mode for many years. This paper, also published in Nature Geoscience last month, focuses on the mid-Holocene period, about 6,000 years ago. 

At that point, Earth’s 23,000-year precession cycle (basically a wobble around Earth’s rotation axis) had lined up Northern Hemisphere summer with perihelion, the planet’s closest approach to the Sun. As a result, winters were generally colder and summers warmer than today. Also, the current Sahara Desert had been layered with vegetation for millennia; it would be hundreds of years more before it would start morphing into the arid landscape that “Sahara” brings to mind.

Although the causes were different from today, the climate was relatively warm across the world, making this study period useful for shedding light on what’s happening now, said the study’s lead author, Victoria Todd of the University of Texas at Austin.

When a set of 23 paleoclimate simulations from 17 models replicated this period, they produced a long-lived negative-PDO-like pattern. This matches up with winter precipitation records for the Southwest, inferred from new leaf-wax isotope records from sites in New Mexico and Colorado that extend back 12,000 to 14,000 years. 

“We found that Northern Hemisphere warming in the past, and what we see in the future projections, really does keep the North Pacific in this persistent sea surface temperature pattern that resembles the negative phase of the PDO, and that this drives long-term drought in the Southwest U.S.” Todd said.

Todd and co-authors end their paper with a stark warning that captures the mood of all three recent studies:

“models may be underestimating the severity of future winter precipitation changes and the future risk of drought in the Southwest United States.”

Clara Deser, a senior scientist at the NSF National Center for Atmospheric Research and a longtime researcher on variability and change in the Earth system, is among the coauthors on the papers led by Klavans and Kuo. “I still think there is a role for both natural variability and anthropogenic [human-related] influences on PDO trends over the past 30 years or so,” Deser said. “But the new research (which comes from independent lines of evidence) is pointing to a relatively larger role for the latter compared to the former.”

Dive deeper: What exactly does “drought” mean?

The term “drought” is often used in multiple and overlapping ways that can get confusing. When precipitation is below average for an extended period, that’s meteorological drought. When such a dry period affects soils and crops, it’s agricultural drought, and when it hits water supplies, it’s hydrological drought. More recently, the term ecosystem drought has come into use, referring to more general landscape drying.

The U.S. Southwest has dealt with all of these unwelcome guests over most of the last quarter-century. A number of high-profile studies have classified the period since 2000 as a megadrought, which refers to an intense, multi-decade drought – in this case, an especially stark one in its impacts on the environment and society.

An analysis led by Park Williams (University of California, Los Angeles) deemed the period from 2000 to 2021 as the worst megadrought in at least 1,200 years for a broad region from southern Idaho and Oregon to northwest Mexico.

What about the drought subtypes? Precipitation has fallen persistently short of average in this megadrought period, with 17 out of 25 water years from 1999-2000 to 2024-25 running drier than the 20th-century average. Looking purely at meteorological drought, this has been a prolonged, high-impact event, yet it’s not completely unprecedented. Across the Southwest climate region (Arizona, Colorado, New Mexico, and Utah), total water-year precipitation from 1999-2000 through 2024-25 averaged 13.53 inches, according to NOAA. These values were actually a touch lower during several periods in the mid-20th century, including 13.42 inches from 1942-43 through 1966-67.

It’s all too clear what has pushed this dry period into truly historic territory: a warming climate. Distinctly hotter temperatures across the Southwest – rising about 2.6 degrees Fahrenheit over the past 130 years, close to the rate of global-scale warming – have drawn more and more moisture out of the landscape.

In their 2022 study noted above, Williams and colleagues based their worst-megadrought designation on soil moisture, reconstructed over the past 1,200 years using proxy data from tree rings, whose width corresponds closely to annual moisture.

We can’t know for sure how much rain or snow fell across these 1,200 years. But Williams and colleagues estimated that without human-caused climate change, “the turn-of-the-twenty-first-century drought would not be on a megadrought trajectory in terms of severity or duration.” Based on model output, they attributed 42% of the 22-year drought (as defined by soil-moisture loss) to climate change. One could imagine that percentage going higher if the most recent PDO-related research above were taken into account.

Jeff Masters contributed to this post.