Models Predict “Megadrought” Risk for American Southwest This Century

Near the beginning of the American Association of the Advancement of Science (AAAS) conference in San Jose, CA, on a winter day that happened to be warm, dry, and sunny, research scientists held a press conference to announce the conclusions of their work on predicting the risks of future “megadroughts.” They published their paper with the ominous title, “Unprecedented 21st-Century Drought Risk in the American Southwest and Central Plains,” in the first issue of Science Advances, a new digital, open-access journal. (The publisher, Marcia McNutt, gave brief opening remarks about how the journal will “showcase new and exciting research.”)

Benjamin Cook (NASA Goddard Institute), Toby Ault (Cornell University), and Jason Smerdon (Columbia University) obtained surprising results from computer model simulations. According to Cook and Smerdon, who videoconferenced with a shared microphone, previous models—such as those used for the IPCC’s 4th Assessment Report—underpredicted drought risks. Cook and his colleagues used drought records documented in more than 1800 tree-ring chronologies over the the past millennium, where ring width decreases in dry years, to develop 17 model projections of 21st century climate in the American Southwest and Central Plains. Their disturbing findings include predictions of megadroughts, lasting 35 years or longer, in both regions worse than any seen in the last 1000 years. In short, they expect climate change to increase drought length and severity in the coming decades.

Mean summer soil moisture and Palmer Drought Severity Index out to 2099. Courtesy: Cook et al. (2015)

Mean summer soil moisture and Palmer Drought Severity Index out to 2099. Courtesy: Cook et al. (2015)

The drought risk is twofold, due to reduced precipitation and to warmer temperatures drying out soils of rivers and lakes, in which models predict increasing evaporation. Long droughts due to climate variations have occurred in the past, such as those occurring during the 12th and 13th centuries (the Medieval Climatic Anomaly) that serve as important benchmarks. But with their tree-ring based reconstruction of the climate history, in a “business-as-usual” emissions scenario, they predict a “persistent shift in the future toward longer droughts” that could exceed even those of these extremely dry centuries.

Ault described how risk assessments are made, in terms of the magnitude of impact and the likelihood. “The levels [of risk] that we see are striking,” with an 80% or higher risk of a drought 35 years or longer in duration by the end of this century if climate change is not mitigated. He described the situation like a golf course, which an initial 10% of it consisting of sandpits. If climate change continues unmitigated, the golf course will gradually become almost entirely sandpits!

Risk (%) of decadal and multidecadal drought calculated from three sets of models. Courtesy: Cook et al. (2015)

Risk (%) of decadal and multidecadal drought calculated from three sets of models. Courtesy: Cook et al. (2015)

In the paper, Cook and his co-authors comment on the difficulties people in the Southwest and Central Plains will face when attempting to adapt to these climate conditions. In particular, the current depletion of nonrenewable groundwater reservoirs “will likely exacerbate the impacts of future droughts.” They discussed implications for both the water supply and food supply in the press conference, considering the dependence on agriculture in California and the breadbasket in the Central Plains. “Water security is food security,” as Ault put it, and people need to “take a no-regrets attitude toward preparedness.” Droughts will inevitably occur, and some of them could be as destructive as large earthquakes and hurricanes.

Their ongoing research will focus on examining the future severity, persistence and geographical scope of droughts, and they will attempt to improve the spatial resolution of their simulations, which currently employ coarse-grained averaging. They also plan to consider hydrology and snowpack, in addition to soil moisture. In any case, the soil moisture metrics and PDRI all point to one conclusion: unless people find a way to substantially mitigate climate change and prevent rising temperatures, the American Southwest and Central Plains can expect to face megadroughts like they’ve never seen before.

Finally, if you’re interested, below you can see my photo from the press conference (and I’m the one in the lower left with a cellphone in front of his face.) In other coverage of these results, see this nicely written Science article by Emily Underwood, and I saw a Washington Post reporter writing an article about it too (but I forgot his name), so watch for that.


Water Policy Issues, with a Focus on the US Southwest

Water policy issues are very important, but we haven’t discussed them much on this blog yet. Much of my information here comes from Ellen Hanak and other analysts of the Public Policy Institute of California (PPIC), analysts from the Union of Concerned Scientists (UCS), a recent article by Christopher Ketchum in Harper’s, a book by Robert Glennon (Unquenchable), and other sources. I’m not an expert on water policy, and any errors are my own. As usual, please let me know if you notice any errors, and I’m happy to hear any comments. I’ll focus on the southwestern US (mainly because I grew up in Colorado and now live in California), but many of these issues apply elsewhere as well. And while the Southwest is dealing with drought and water scarcity, other places, such as the UK and the Midwest US, are dealing with flooding.


According to the Worldwatch Institute, already some 1.2 billion people live in areas of physical water scarcity, while another 1.6 billion face “economic water shortage”. By 2025, almost half of the world will be living in conditions of water stress. Some analysts predict that water wars (see Vandana Shiva’s book) and conflicts will increase in the future. Considering that we need water to live, it’s not surprising that the United Nations General Assembly voted in a resolution declaring that access to clean water and sanitation is a fundamental human right.

At least conditions on Earth are not as bad as Mars, which has experienced 600 million years of drought and which probably hasn’t supported life, at least on its surface. But water scarcity is an extremely important problem that we’re probably not taking seriously enough; as Stephen Colbert put it, “if the human body is 60 percent water, why am I only two percent interested?”

The Southwest and California in particular are experiencing their worst recorded drought (for example, see the NASA satellite images below). In response, the California state legislature and Gov. Brown passed a drought relief package last month, while Sen. Feinstein and others are seeking to pass a bill in Congress to aid drought-stricken states.


Now here’s some historical and legal context. The Colorado River Compact of 1922 was negotiated by members of the upper-basin states (Colorado, New Mexico, Utah, Wyoming) and the lower-basin states (Arizona, California, Nevada), and it was an agreement for hydraulic management of the Southwest. According to the US system of water rights, however, the person who first made “beneficial use” of a stream or river had first right to it. Under this doctrine, the earliest users of the Colorado River (California) could legally establish a monopoly over regional water supply, even though most of that water came from another state (Colorado). A major problem was that because 1922 happened to occur during an unusually wet period, people assumed that the Colorado held more water than it really did: its annual water flow as estimated to be 17-18 million acre feet, though it was later more accurately estimated at 14 million acre-feet (17 billion cubic-meters) on average. It was therefore already overallocated from the start. The lower basin (including southern CA) is now overusing its share of the Colorado River, and it’s not a sustainable situation. A court case (Arizona v. California) that was decided by the Supreme Court in 1963 affirmed that Arizona was owed 2.8 million acre feet of water annually, but under the doctrine of prior appropriation, Arizona’s rights would remain secondary to California’s.

For water use, it’s useful to distinguish between water withdrawal (from surface or ground sources) and the consumption of water already withdrawn. Consequently, as argued by Ellen Hanak at a recent PPIC event in Sacramento, we need to consider not just water supplies but also water management and (in)efficient water consumption. Although one usually thinks of water for drinking, washing, cleaning, and other residential uses, much more water is used for irrigation (agriculture), industry, and power plants; according to the UCS, power plants account for 41% of freshwater withdrawals in the US. It’s also useful to distinguish between direct and indirect water use, and I’ll get into that more below.

Water shortages, already a critical issue in the Southwest, are likely to become far worse with climate change (although the extent to which it’s due to climate change is still debated). Rivers such as the Colorado, which is primarily supplied by snowmelt and is already overallocated, are particularly vulnerable. For the past fourteen years, the Colorado River has been at its lowest level since the ninth century. According to Tim Barnett from UC San Diego’s Scripps Institution of Oceanography (SIO), with climate change, currently scheduled water deliveries from the Colorado River are unlikely to be met by mid-century. Rising air temperatures due to global warming will result in reduced snowfall: by the end of this century, California’s ski season could disappear with a 80% loss of Sierra snowpack, and Washington and Oregon would experience reduced snowfall as well. In addition, although per capita water use has been gradually decreasing, population growth in the Southwest is likely to increase urban water demand in some regions. In a high carbon emissions scenario, annual losses to agriculture, forestry, and fisheries could reach $4.3B in California alone, and the prices of fresh fruit, vegetables, dairy, and fish, will rise. There will be more competition between human water use and water needed to support fish and other wildlife, and potential solutions will involve difficult trade-offs. (The following figure from the EPA summarizes climate impacts on the hydrologic cycle.)


In the studies mentioned above by SIO scientists, the Colorado River’s average annual flow could decline by as much as 30% by 2050. As a result, without massively reducing water usage, Lake Mead has a 50% chance of declining to “dead pool” by 2036. At that level, water deliveries to millions of people in California and Arizona and to millions of acres of farmland will cease, and hydroelectric production at the dam will already have stopped. It is incredible to consider that this could happen in our lifetime, as the Colorado is the same river that carved the Grand Canyon over tens of millions of years, and it is one of the rivers on which the Ancient Puebloan depended until around 1300, when drought, decreased rainfall, and a drop in water table levels appeared to drive the people away from their civilization. (See also this article in National Geographic about ancient “megadroughts” in human history.)

The largest fraction of water consumption is due to agriculture, power plants, and industry. Considering the fact that we indirectly need water because of our need for energy, this points to the issue of the “water-energy nexus.” The average U.S. family of four directly uses 400 gallons of freshwater per day, while indirectly using 600-1800 gallons through power plant water withdrawals. We need energy for water production and distribution (and the desalination plant being constructed near San Diego will require quite a bit), and we also need water for energy-related infrastructure. Coal and nuclear power plants use large amounts of freshwater to cool the plants: for example, a typical 600-MW coal-fired plant consumes more than 2 billion gallons of water per year from nearby lakes, rivers, aquifers, or oceans. In addition, as we discussed in my previous blog post, fracking techniques for extracting shale gas require millions of gallons of water to be injected into a well, and they can contaminate groundwater as well. Fortunately, wind turbines and solar photovoltaic modules require essentially no water at all, but other renewable energies, like hydroelectric, bioenergy, and geothermal, can be water intensive. As argued by Laura Wisland, since we expect climate change to increase the frequency and severity of droughts in California, it will be important to hedge our electricity supplies with predictable, renewable resources, especially wind and solar.

What can be done? As a “silver lining” of the current situation, the ongoing drought in the Southwest provides a window for reform, and here are a few ideas. We should shift toward less water-intensive sources of energy such as wind and solar. Water should cost more: we should modernize water measurement and pricing with better estimates of water use and prices that reflect water’s economic value. We could learn from cities in dry places elsewhere (such as Australia) about how to make urban areas more water efficient, and we could have tiered water rates with higher prices for greater use. In agriculture, crops that cannot be grown without subsidies should not be grown. We need improvements to local groundwater management. Since surveys show that most Californians believe that there are environmental inequities between more and less affluent communities in the state, it’s also important to consider environmental justice issues while developing new water policy programs (see this article, for example). We need to develop more reliable funding (through state bonds or local ratepayers), especially for environmental management, flood protection, and statewide data collection and analysis. Finally, as argued in this PPIC report, water management agencies at all levels should aim to develop more coordinated, integrated approaches to management and regulatory oversight, drawing on scientific and technical analysis to support sound and balanced decisions.