Lake Mead, the massive reservoir created in the late 1930s by Hoover Dam on the Arizona–Nevada border, has dropped to its lowest level ever, it was reported earlier this month. The lake has been steadily growing shallower since drought began reducing the flow of its source, the Colorado River, starting in 2000 due to below-average snowfall in the Rockies.

It is still too early to know whether the situation at Lake Mead and recent droughts throughout the U.S. Southwest are due to anthropogenic global warming, says Aiguo Dai, an atmospheric scientist in the Climate and Global Dynamics Division of the National Center for Atmospheric Research in Boulder, Colo. There is not enough data to rule out natural variability as the fundamental cause. But the decadelong dry spell is consistent with the predictions of models used in the Intergovernmental Panel on Climate Change (IPCC) Fourth Assessment Report (AR4) in 2007, which projected that warming of the planet would lead to long-term drying over the subtropics—the climatic regions adjacent to the tropics, ranging between about 20 and 40 degrees north and south latitude, which includes the Southwest.

It's also consistent with a new analysis, authored by Dai, which forecasts that increasing dryness over the next several decades will eventually become devastatingly severe, with long-lasting drought predicted for most of Africa, southern Europe, the Middle East, most of the Americas, Australia, and Southeast Asia. Dai computed global values through the end of the century for a commonly used index of drought severity, using data from 22 models used by IPCC-AR4, under a middle-of the-road emissions scenario that assumes human-generated greenhouse gas emissions start reducing about 2050, and that atmospheric CO2 increases to 720 parts per million (we're currently at around 380) by the end of the century.

Dai's projections are helpful because they begin to bring into focus some of the water-related global warming consequences that may be upon us relatively soon, says Richard Seager, a senior research scientist at Columbia University's Lamont–Doherty Earth Observatory. Seager, who was not part of the study, was a co-author of a 2007 study in Science that analyzed the findings of the IPCC-AR4 models. "When the IPCC report came out in 2007, there was relatively little that looked at how these climate changes developed within the coming decades," he says. But in Dai's new figures, "you can see that even in the coming decades or so we're already getting into some trouble in this regard."

Desiccation nations
The new report, published online October 19 in Wiley Interdisciplinary Reviews: Climate Change, draws a more quantitative picture of how bad global warming–related droughts may actually be—and when they might occur. To calculate severity, Dai employed a widely used drought index, known as the Palmer Drought Severity Index (PDSI). On the PDSI scale, 0 is normal, whereas negative numbers indicate the degree of dryness and positive numbers wetness. The new projections suggest that by 2030, the subtropical regions will be overtaken by droughts of even greater severity than the one in the Sahel region of western Africa in the 1970s, the most severe dry spell in recent history, which scored around –3 on the PDSI scale. By the end of the 21st century, the outlook is even worse, with PDSI values near –10 for much of the U.S., southern Europe, Southeast Asia, Brazil, Chile, Australia and most of Africa.

These results go hand-in-hand with previous drought predictions, which have generally projected more frequent, longer-lasting droughts due to global warming. But the new forecast is more detailed, Dai says. Previous studies have been limited, he says, because they used models to analyze changes in soil moisture only. Drying is a more complicated outcome, dependent on both the supply of soil moisture and the atmosphere's demand for water vapor, which will increase if the atmosphere heats up. Because it accounts for increasing temperature, "the PDSI actually gives you a sense of how severe this drought might be in 30 or 50 years," Dai says.

The PDSI is not without limitations. It does not consider the essential role that plants play in evaporation, and for colder regions, it does not account for the fact that snow and ice accumulate in the winter and melt in the spring—rather, it treats all precipitation as moisture that is immediately available for evaporation. But despite these caveats, Dai notes that in many regions of the world PDSI values correlate strongly with other drought gauges such as stream-flow data and measured soil moisture content during warmer months.

Near-term consequences
So why can't it be said definitively that the current conditions in the Southwest are due to anthropogenic global warming? "It gets pretty dodgy when you start looking to see whether these things are occurring," Seager says, because the changes scientists expect over the coming decades are much bigger than any effects global warming may have had already. "So when you are looking at the last 50 years, you are often looking at natural variability, because this forced anthropogenic signal is still small compared to the amplitude of the year-to-year, decade-to-decade variability."

This also is the case for dropping water levels in the Colorado River and Lake Mead. There is not enough data yet to be able to discern whether or not a long-term trend exists that would implicate global warming, Dai says, but eventually the relationship will become clearer. "Obviously, a drier western U.S. will lead to lower lake levels, and I feel that this might already be happening over the southwest U.S.," Dai adds. "We will need longer records to be sure."