Higher future flows likely on the Colorado River, new study finds

After 25 years of declining flows on the Colorado River, a new study led by University of Colorado researchers offers hope for a partial turnaround.

River flows will most likely rise 5-7% in the quarter-century starting in 2026 compared to the far-below-normal flows of the last 25 years, it says.

Such an increase won’t come close to erasing the river’s nearly 20% flow decline since 2000. So efforts to improve water management and conservation should continue, the authors say. In Arizona, the river sends drinking water to millions of residents of the Tucson and Phoenix metro areas.

The study’s forecast is a dramatic change from current trends, and sharply contradicts four decades worth of studies that have predicted future, major river flow declines due to climate change.

This study finds that climate change will drive increasing flows, just as many scientists have blamed much of the recent declining flows on climate change.

But while other researchers have said human-triggered, warming temperatures have hurt the river, the new study says global warming will soon start boosting flows by triggering more rainfall.

It was published last month in the Journal of Climate.

“We find it is more likely than not that ... flows will be greater during 2026-2050 than since 2000 as a consequence of a more favorable precipitation cycle,” said scientist Martin Hoerling, the paper’s lead author. “This will compensate the negative effects of more warming in the near term.”

Hoerling is an affiliate with the University of Colorado’s Cooperative Institute for Research In Environmental Sciences. He had been a longtime National Oceanic and Atmospheric Administration researcher and meteorologist.

Criticized by other researchers

Many other researchers have concluded in recent years that warmer weather is a major factor lowering river flows since 2000 and will continue doing that for the foreseeable future.

But the river’s recent declines are due far more to a naturally-caused dry spell than to human-caused warming temperatures, a University of Colorado press release announcing the study said.

That and the study’s more optimistic future projections draw criticism from researchers who worked on the previous studies. They take issue with the study’s choices for computer models and with the very idea that computer models can accurately predict how climate change will affect future precipitation.

One critic, retired U.S. Geological Survey researcher Christopher Milly, charged the press release “publicly miscommunicated” a key finding and stated a conclusion the study didn’t spell out, let alone back up.

Hoerling and a co-author of the new study rebutted the criticisms in interviews with the Arizona Daily Star. But they acknowledged declining river flows over the next 25 years are still possible, just not nearly as likely as increases according to their research.

Based on the study’s worst-case scenario, river flows could fall another 20%, the researchers say. The best-case scenario is for a 40% increase. About 70% of the computer model runs they examined projected future precipitation increases. The rest projected decreases, they said.

“It comes with roses of optimism but lots of thorns around it,” co-author Balaji Rajagopalan said of the new study. “The takeaway lesson is that it is a nuanced, more complete picture of the basin. There is a chance of recovery, but that recovery also comes with uncertainty.”

The study comes as officials of the U.S. Bureau of Reclamation and the seven Colorado River Basin states including Arizona are locked in deadlocked negotiations over how to curb Colorado River water use to bring it in line with supplies. Their goal is to agree on revising the river’s federal operating guidelines that expire at the end of 2026.

Rajagopalan made the point in an interview that these conservation efforts need to continue if river flows rise. The decrease in flows has left the river’s biggest reservoirs, lakes Mead and Powell, only 35% and 34% full, respectively.

“You need to have management that makes sure the water supply is sustainable,” Rajagopalan said.

Climate-based computer models

Since 2000, 14 of 24 years had spring-summer runoff into Lake Powell below long-term averages. This year is on track to be number 15, with recent forecasts of runoff at 80% to 85% of average, says the federal Colorado Basin River Forecast Center. Last year’s runoff jumped to 166% of normal, but scientists and public officials have said that much runoff is unlikely to occur regularly.

Starting in 2017, three major studies predicted climate change will continue to significantly reduce future Colorado River flows. They followed several earlier, less detailed projections for river flow declines due to climate change dating back to 1983.

Two of the recent studies projected river flow declines of up to 20% to 30% and 14% to 31%, respectively, by about 2050.

A third, a 2022 study from Los Alamos National Laboratory, didn’t predict a specific flow decline but warned that continued climate change would transform the Northern Rockies’ wetter climate to resemble that of the arid Southwest.

The new study, relying heavily on climate-based computer models produced by the Intergovernmental Panel on Climate Change in 2021, looks at precipitation and temperature forecasts.

Many past studies overlook the “critical effect” of precipitation, the new study says. Decade to decade variations in precipitation have accounted for more than 80% of the historic range of changes in river flows at Lee Ferry, it finds. Lee Ferry, about 15 miles south of Glen Canyon Dam, marks the boundary between the river’s Upper and Lower basins.

Study co-author Rajagopalan noted that 15% of the river basin — the northern Rockies lying generally north of Denver and Salt Lake City — produces 85% of the river flows through snowpack.

It’s “quite understandable” that higher temperatures would put downward pressure on river flows, he said, partially through more evaporation and faster snowmelt.

But “all we are saying is that we need to look at both — what is the range of temperature and precipitation effects?” Rajagopalan said.

Looking at an average of all the climate model findings by the IPCC shows parts of the Upper Basin that include the Colorado River’s headwaters are going to get wetter, he said. The same models also project warmer temperatures over most of the basin — “that is where the conundrum is,” he said.

“The question is not will it be warmer. Of course it will be hot. The bigger question is, what is the snow going to do over the Rocky Mountains? If it continues to snow, (river flows) might improve even with a 4-degree warming,” Rajagopalan said.

Back to homepage

Subscriber Login

Keep reading with a digital access subscription.

Subscribe now

‘Willing to bet our water supply?’

Climate scientist Jonathan Overpeck reads the IPCC’s computer model projections on river basin precipitation differently.

While at the University of Arizona in 2017, he and Colorado State University researcher Brad Udall wrote one of the recent studies predicting major river flow declines.

Today, Overpeck is a professor and dean of the School for Environment and Sustainability at the University of Michigan.

In an interview Friday, he responded to the new study by pointing to maps from Chapter 4 of the latest IPCC report on climate impacts. He noted that much of the world, including the Colorado Basin, is covered in the maps with diagonal lines that indicate “no robust significant change.”

That means no statistically significant trends in future precipitation lie ahead in much of the world, he said.

“If you average large regions (the globe or a hemisphere or big latitude bands), you do get significant trends, but not at the scale of the Colorado River Basin,” Overpeck said.

The new study does a good job by showing a range of computer model simulations of possible future precipitation, he said. “You can sort of have your pick of precipitation estimates you want to go with. It’s a chance will get wetter. There’s also a chance it will get drier.”

But noting that Mother Nature brought the region a “really dry spell” this century, Overpeck said researcher Hoerling is saying, “’Don’t worry, the dry period will end. It will get wetter.’ He provides reasons why that might be true. Some of the models show that. Would you be willing to bet our your water supply on that?”

Udall goes further, expressing disdain for any precipitation-based climate computer models. Because he didn’t trust such models, his study with Overpeck didn’t even try to forecast future precipitation, he said.

“I do not believe they are reliable,” Udall said. “The precipitation amounts coming out of models on a regional level have all kinds of problems associated with them. Models are not to be trusted when it comes to future precipitation.”

Models that can show “physically-based” representations of future precipitation in specific areas are still several generations away, he said.

Also, while he hasn’t reviewed models used in the new study, he said that generally, “models for precipitation are biased, too much off in one direction all the time. In this case, the bias is to the positive side.”

‘Science by press release’

The 2020 study by Milly, the now-retired USGS researcher, did look at future precipitation as well as temperatures, and concluded that precipitation changes through 2050 would likely range from a 3% increase to a 40% decrease.

More prominently, that study said warming temperatures were responsible for nearly half of the river flow declines from 2000 through 2017. It said climate change slashed river flows by about a million acre-feet per year, or as much as the 336-mile long Central Arizona Project canal system has delivered most years to the state’s parched interior including Tucson. Udall back then called that finding “eye-popping.”

Regarding the new study, Milly attacked a University of Colorado press release declaring the study found precipitation was the biggest factor reducing river flows since 2000.

“I cannot find within the new paper a statement of the contrarian viewpoint expressed in the press release, nor any analysis supporting that viewpoint,” he said.

“Science by press release can be dangerous. With the livelihood of millions of people and the multi-trillion-dollar economy of the US Southwest depending on accurate information for management of its limited water resources, I think acceptance of this remarkable claim requires more than impromptu hand-waving in defense of a non-peer-reviewed press release that purports to represent the findings of a scientific paper,” Milly told the Star in an email.

In contrast to this study, Milly and fellow USGS scientist Kristen Dunne reached their conclusion about temperature impacts by observing not only precipitation, temperature and river flows, “but also of radiation and surface albedo, by paying special attention to the energy balance that drives evaporation,” and by using a more complex computer model analysis than the new study used to demonstrate climate impacts, he recalled. Surface albedo is an indicator of the fraction of solar radiation that is reflected at the Earth’s surface.

Authors respond

In response to critics, lead study author Hoerling noted the new findings drew partly on 38 individual computer models employed by the IPCC’s 2021 study. Those and other models used for the study produced, overall, “very similar statistics” about future precipitation trends, he said.

Replying to Udall’s criticism of precipitation-based computer models, Hoerling noted that six groups of models the new study’s researchers used all generally agreed on the level of variability in precipitation during the seven decades they covered, from 1950 through 2020. That’s the “most germane measure” of the models’ suitability, he said.

While not responding directly to Milly’s criticism of the study press release, Hoerling said the new study considered a range of potential impacts on future river runoff based on various temperature increases, including one that shows a bigger impact of warming on runoff than was shown in Milly’s 2020 study.

Summing up, he said the study’s core point is more about the wide range of possible future river flows than about computer model forecasts of precipitation changes.

“It is this range that provides the critical information on the ... risk going forward,” he said.

UA researchers react favorably

A longtime climate researcher at the University of Arizona’s Laboratory of Tree Ring Research, Connie Woodhouse, gave the study more favorable reviews than did other researchers who talked to the Star.

“I am definitely no expert in models, but I don’t see a reason to doubt what Hoerling et al. found in these models: increased precipitation, when all the models and their members are combined,” said Woodhouse, a professor of geography and development.

But the same time, a graphic illustrating the study’s findings contained a lot of markers indicating computer model runs showing decreasing precipitation from climate change, she said.. Another graphic that showed climate change’s likely impacts on temperature had no markers forecasting declines, she noted.

That suggests “the warming trend is more robust and certain than the wetting,” she told the Star in an email.

A second tree-ring lab researcher, David Meko, also found the new study’s forecast of higher possible river flows plausible.

“The findings rely on projections of increased basin precipitation, which is supported by climate models. The caveat is that projections vary greatly from one model to the next. Some models project wetting, a few project drying. We have to hope the ... mean across models is a reliable estimate for the future,” he said.