UCAR > Communications > Staff Notes > August 2000 Search

August 2000

Burning clues: Wildfire research team lights up Washington

A moment in the evolution of a wildfire, as simulated by the NCAR coupled atmosphere-fire model.

Janice Coen. (Photo by Carlye Calvin.)

Experts from NCAR and the U.S. Forest Service (USFS) headed to Washington on 22 June to offer "Burning Clues: The State of Wildfire Research," a timely luncheon briefing for congressional staff organized by UCAR's Office of Government Affairs and held in the Rayburn Building. The session was sponsored by Representative Scott McInnis (R-Colo.). Washington staffers heard from Don Latham of the USFS Fire Sciences Laboratory's Rocky Mountain Research Station, Janice Coen and Terry Clark (MMM), and Larry Radke (ATD). The next day, the team made a similar presentation at NSF. Back in Boulder, Janice gave Staff Notes Monthly an overview of the Washington trip.

Latham led off with a quick course in wildfire elements, including fuel components and heat transfer processes. After outlining his lab's research program, he zeroed in on the problem with the algorithms (mathematical problem-solving procedures) firefighters now use to predict fire behavior: the atmospheric component is not interactive. The winds in these algorithms change the speed and direction of fire spread, but they don't incorporate the effects the fire is having on the atmosphere, which in real life feed back into the fire's behavior. Without realistic interactions among fire, wind, and terrain, Latham emphasized, the models cannot predict the rapid, severe changes in fire behavior that put lives and land in jeopardy.

Terry followed with a look at observational techniques the NCAR research group is using to inform, initialize, and validate the models. To pierce the thick smoke that obscures many features in the visible spectrum, the team has aimed an infrared (IR) camera from ground stations and the NSF/NCAR C-130 aircraft at wildfires and planned burns during several experiments. From videotaped data gathered by the camera, they've been able to derive the spatial and temporal structure of fire winds and heat flux profiles, as well as vertical vorticity (circulation). "By tracking features from frame to frame we can discover the velocities within the fire itself, which hadn't been done until these techniques were developed by Terry here at NCAR," says Janice.

The team's search for fire fingers—a deadly feature reported anecdotally but never before observed scientifically—was rewarded in IR data from a C-130 flight over a 1998 wildfire in Glacier National Park, Montana. In a one- to two-second interval, "a finger of flame shot out from the fire about 100 meters [330 feet] and pulled back, leaving a trail of burning embers, which then propagated farther," Janice says. This capture of a fire finger in action confirmed a method of propagation not reported in the literature, Janice adds. "Fire spreads itself in much more dynamic ways [than previously described], with these forward bursts and big balls of fire shooting forward, especially in crown fires" (fires that spread through the treetops).

On the modeling front, Janice described the work she and Terry have done to couple a wildfire model to an atmospheric model that includes the passage of large atmospheric features, such as cold fronts. The wildfire model incorporates the algorithms developed by the USFS that factor in the local winds and the type and amount of fuel available. The heat and water vapor released by the fire are fed back into the atmosphere, producing the intense horizontal winds and updrafts that drive the fire line itself. "So they're tied very closely together and you get an interactive effect," Janice says. The model's regional resolution can be telescoped down to focus on fires in one valley. Collaboration among Terry, Janice, Larry, and Don Middleton (SCD) has yielded an image-flow analysis software package and dramatic visualizations of the modeling effort (see Staff Notes Monthly ).

The dramatic features revealed in the observational data have inspired several modeling experiments that reproduce the behavior of real fires when crossing a ridge: stalling, spilling over, backing down, or veering left or right. Small parts of the fire line, called "escapes," can be seen taking off in new directions. A new animation shows heat, rotation, and fuel (burned and unburned) as fire propagates over a hill. Another animation simulates a blowup, in which changes in atmospheric conditions on an uphill slope cause a sudden increase in both burning intensity and spread rate—conditions similar to what researchers believe happened in the fatal fire on Storm King Mountain in Colorado in 1994.

Larry focused on economic and environmental impacts associated with wildfires and the team's hopes for real-world applications of their research. The costs to the United States for wildfire suppression and resource loss exceed $10 billion a year. On the environmental side, atmospheric pollution sequestered in vegetation is re-released into the atmosphere during a wildfire. Mercury and other toxins are transported long distances in pyrocumulus clouds that build high into the troposphere, and carbon dioxide released by the fire contributes to the total budget of greenhouse gases. Larry also mentioned the environmental benefits of wildfires, including transport of released nutrients and opening of seed pods in fire-dependent plant species.

Putting the research to work

The ability to model and predict the behavior of real wildfires in real time, including picking out the one in 10,000 fires that poses a severe threat, could have a significant impact on the nation's well-being. The team's goal is to create a system in which fire managers can try out different suppression strategies before deploying personnel and equipment. Prescribed burns could also be modeled in advance to assure favorable conditions and predict the track of smoke plumes.

Another application the group foresees is assistance with long- term land management. Many fires spread because there's an abundance of undergrowth and dead material acting as fuel; planned burns, as well as thinning and clearing programs, are designed to reduce this risk. But there's an ecological benefit to allowing vegetation to decay on the forest floor. "The modeling could tell us whether, if you thinned out the fuel in one limited area, it could prevent a wildfire from spreading over tens of thousands of acres," Janice says. That kind of simulation might help resolve conflicts between competing land-management philosophies.

At the briefing, staffers asked what the team needs to reach its goals. "We didn't try to portray our work as something immediately helpful to their constituents, because there are many steps in between," Janice says. An operational model is several years down the road. Creative use of information technology to handle the computationally heavy work of data collection and modeling is one key. Increased collaboration between NCAR and the USFS on research and training is another important ingredient that could help make better prediction and management of wildfires a reality.

• Zhenya Gallon

In this issue... Other issues of Staff Notes Monthly

Edited by Bob Henson, bhenson@ucar.edu
Prepared for the Web by Jacque Marshall
Last revised: Fri Aug 11 15:01:08 MDT 2000