Smokey the Bear is trading in his shovel for a laptop and a satellite dish.
At fire camps around the country, portable satellite terminals and computers are joining bulldozers and chain saws in the battle against forest fires. Meanwhile, in labs from Missoula, Mont., to Helsinki, Finland, scientists are developing ways to use space-based remote-sensing data to detect and track wildfires, as well as forecast their behavior.
Researchers hope such tools not only will help move fire fighters into a trouble spot more quickly but will also allow fire officials to adjust their strategy for battling a blaze almost as fast as the wind shifts. "We want to base the field of fire management on technical information rather than solely on personal judgment," says Mark Finney, a scientist in the fire-sciences lab at the US Forest Service's Intermountain Research Station in Missoula.
Fires have kept state and federal forestry officials exceptionally busy this year. Since January, wildfires in the US have charred some 2.8 million acres, more than three times the average for the same period over the past five years. Last week alone, 16 fires burned through more than 135,000 acres in California, Utah, Nevada, New Mexico, and Arizona. These included a 48,000-acre blaze in the Kibab National Forest, near the north rim of the Grand Canyon. By the weekend, cooler, wetter weather had helped firefighters contain many of the fires.
Many of the components for detecting and tracking fires have been evolving for 20 years or more, researchers say. But only recently have computers become small enough, powerful enough, and inexpensive enough to blend data from these components - and from weather instruments - in new ways that are practical for field use.
Traditionally, foresters have relied on spotters in mountain-top fire towers to record the bearings of lightning strikes or smoke plumes in their areas. Today, much of the country is blanketed with electronic lightning detectors "that give a real-time picture of lightning activity in each fire suppression district," says Phil Sielaff at the National Interagency Fire Center (NIFC) in Boise, Idaho. The center, run by the US Bureau of Land Management, coordinates firefighting efforts for six federal agencies.
The lightning data, along with other weather information and descriptions of vegetation in the area, can be fed into a relatively simple computer program that yields the potential for ignition.
But Mr. Finney says the program, known as BEHAVE, can't "grow" a fire. Over the past five years, he has been developing an approach that takes vegetation, terrain, weather conditions, and an ignition point to map how a fire may grow. First distributed in 1994, the program is being used by the USFS and the California Department of Forestry. His colleague Don Latham says he currently is folding data from the lightning network into the program, and hopes to bring the software to the point where "it will connect ignition to the surviving fire" and show more complex fire behaviors.
Yet lightning data do little to tip fire officials off to blazes started by people, particularly in remote areas. "If a system keys off of lightning and man starts the fire, we wouldn't know until a plume of smoke gets high enough to see or until it takes off in a barn-burner," the NIFC's Mr. Sielaff says.
Since 1993, the Finns have been refining a system that uses infrared information from US weather satellites in orbit around the earth's poles. During last year's experiments, the system detected 85 fires in forests in Russia, the Baltic states, as well as in Finland. The 16 Finnish fires included 11 prescribed burnings, one forest fire, one unknown hot spot, and a steel mill.
In the United States, the Defense Department and the Bureau of Land Management have been experimenting with infrared data from US reconnaissance satellites in low-Earth orbit as a means of detecting and tracking fires. With the Pentagon's help, "you'd see a hot spot the size of the back end of a pickup truck very, very quickly," Sielaff says.
Such techniques would augment the Forest Service's fleet of three corporate-type aircraft, outfitted with infrared sensors. Flown at night when the ground has cooled and the fire lines have stabilized somewhat, the detectors can see a 9-inch hot spot from 13,000 feet.
Meanwhile, researchers at the National Center for Atmospheric Research in Boulder, Colo., and at the Los Alamos National Laboratory in New Mexico are reporting progress in developing detailed computer models of how fires behave. The ultimate goal is an artificial-intelligence adviser that can take any fire's starting conditions and accurately forecast its behavior several hours later to within a few yards of any location, given a range of fire-suppression approaches.
Currently, the information used as the basis for fire-behavior models is based on field observations, says Frank Harlow, a physicist at Los Alamos. "We're taking a computer model and trying to get into the physics and chemistry of a fire" and let that determine its evolution.