Forecasting Fires
New Database Aids Firefighters and Forest Ecologists
By Dan Ferber
As wildfires blazed through northern Montana during the summer of 2003, one fire threatened to jump a reservoir and make a beeline for the town of Hungry Horse. At a nearby laboratory, two retired Forest Service scientists quickly downloaded data on conditions of the fire, the forest and the weather and ran computer models to forecast what the fire would do.
The Beta Fire burns near Hungry Horse, Montana
AP/Daily Inter Lake, Robin Loznak
The resulting maps, says Joe Ferguson, supervisor of local firefighting efforts, for the first time “gave us very precise information on where the problem was going to be.” And Ferguson’s fire crews were there waiting for it. They contained the fire—which could have burned thousands of acres—to just 20 acres, protecting key power lines as well as Hungry Horse, which did not have to be evacuated.
The information that helped protect the town came from the prototype of a new database. Called LANDFIRE, the database is the result of a five-year, $40 million collaboration between The Nature Conservancy, the Forest Service and the Department of the Interior.
The database combines the agencies’ mapping and survey data with the Conservancy’s ecological modeling expertise to provide information that can help fight uncharacteristic, devastating fires and bring ecosystems into ecological balance, says landscape ecologist Ayn Shlisky, who leads efforts on the project for the Conservancy’s Global Fire Initiative.
Many ecosystems, including the ponderosa pine forests of the Rocky Mountains, have evolved to rely on frequent fires, which clear out underbrush, enable seedlings to sprout and rejuvenate the forest. But decades of excluding even harmless forest fires have caused flammable underbrush and dead wood to accumulate. When this tinder ignites, ordinary forest fires can become infernos, threatening homes and the ecosystems themselves.
To predict such fires, the two agencies and the Conservancy are compiling almost 100 types of data and models, including satellite maps, vegetation types, historical fire frequency and historical weather patterns.
“We can play what-if scenario games,” says Kevin Ryan of the Forest Service’s Missoula Fire Lab and acting program manager for LANDFIRE. In northern Montana, for example, fire commanders used LANDFIRE-generated models to predict the behavior of 50 wildfires in the summer of 2003.
Land managers can also use LANDFIRE to decide where and how to remove excess tinder. In southern Utah’s Dixie National Forest, a team led by assistant fire-management officer Taiga Rohrer has used prototypical LANDFIRE data in current fire-behavior models to plan how to remove fuels from a fire-suppressed ponderosa pine forest while preserving the canopy structure needed by northern goshawks.
LANDFIRE can give more precise predictions than previous fire-behavior models did, allowing land managers to let beneficial fires burn, Shlisky says.
The database has been tested in the northern Rockies and in central Utah; now it will expand nationwide. As more land managers plan for both fire prevention and forest health, Shlisky says, LANDFIRE will help them “justify what they’re doing on an ecological basis.”
For More Information
- United States Geological Survey (USGS): LANDFIRE
LANDFIRE is a wildland fire, ecosystem, and fuel assessment-mapping project designed to generate consistent, comprehensive, landscape-scale maps of vegetation, fire, and fuel characteristics for the United States.
