Scientists Improve Wildfire Forecasts for Western United States

National Geographic News

July 18, 2001

Borrowing statistical methods long used by climate forecasters,
scientists in California have developed a modeling tool to help
officials better predict the wildfires that devastate large parts of the
western United States every year.

In an initial study of the
application, they produced the first comprehensive forecast for
wildfires in the western United States. The results showed that this
year's fire season is likely to be much milder than in 2000. The
wildfire season generally extends from May to October.

Last year the region suffered the worst season of destructive land and forest fires in the past half-century, according to official reports. As many as 90,000 outbreaks of fire burned 7 million acres, causing U.S. $1.6 billion in damage. Thirty thousand firefighters were called on to combat the blazes.

The scientists say that after further refinements, the new analytic technique should be highly useful in fire-fighting preparation and response.

"This promises to be a valuable tool for scheduling fire management," said Anthony Westerling of the University of California's Scripps Institution of Oceanography in San Diego, who led the team that developed the new forecasting model.

"The western United States has been pooling its resources together for fire suppression," he said. "A lot of important decisions are made early in the year about pre-positioning equipment and personnel. Anything we can do to provide a forecast of where fires are likely to occur is going to save a lot of money, time, and effort."

The Fuel Factor

Traditional methods of wildfire prediction have been based on forecasts of the summer weather during an impending fire season.

The new approach takes into account a much more extensive set of factors and conditions, viewed over much longer periods. The data come from records compiled over the past 20 years by the U.S. Forest Service, the Bureau of Land Management, and the National Park Service.

Westerling said the forecasting model "is based on our understanding of fuels that are available for fires to burn." That entails calculating the amount of vegetation that's likely to exist in an impending wildfire season, based on weather conditions in the preceding years.

The scientists look at correlations between levels of drought—an indicator of the moisture content of soil and vegetation—and acres burned, or fire frequency. "The previous climate tells us how much fuel has been produced, and how moist that fuel is going into the fire season, so that's what's driving this forecast," Westerling said.

In California, for example, fires are more likely to occur after a two-year cycle of a wet winter followed by a dry winter. This happens because a high-moisture winter increases vegetation, and as that vegetation dries out in the subsequent winter it makes conditions ripe for the spread of wildfires.

The situation is different in the Great Basin area, where the climate is drier than on the West coast. The vegetation will usually always be dry enough to burn, Westerling explained, so "it's just a matter of enough fuel becoming available to burn." In those drier Western states, therefore, the most important factor in predicting the severity of wildfires is the levels of moisture—or "fuel" production—in the winter and spring of the year preceding a fire season.

The scientists' comprehensive forecast for the 2001 fire season in the West, from the West Coast to the Rocky Mountains, indicates that fire severity has tapered off after two years of dry conditions. In other words, the fuel is dry, but there's less fuel to burn.

Tailored Applications

The wildfire forecasting research is sponsored by the U.S. National Oceanic and Atmospheric Administration and the Department of Energy. It's administered by a Scripps-based initiative called the California Applications Program, which works to increase local use of forecasting methods to improve the management of natural resources and planning for natural disasters.

Dan Cayan, the director of the applications program, said the next step in the project is to work more closely with California's Department of Forestry and Fire Protection to tailor the forecasting methodology to the agency's needs.

The still-experimental tool doesn't provide highly specific information such as how particular fires will behave and how much they will burn. But future versions will become more and more sophisticated.

Eventually, managers may be able to predict wildfires within shorter time frames, such as in early or late phases of a fire season, and with greater lead time. Incorporating more extensive data, especially detailed ecological and climate information, should also make it possible to better pinpoint fire-prone areas across the West so officials can act to prevent wildfires or minimize their damage.

The research was done by scientists from the Scripps Institution, the U.S. Geological Survey, and the Desert Research Institute, which is based in Reno, Nevada.