BILLINGS, Mont. — As large swaths of Yellowstone National Park burned during the destructive fires of 1988, a small group of fire analysts descended on the park for a firsthand look at how massive fires burn.

From those observations, U.S. Forest Service researcher Richard Rothermel crafted a technique to track some of the largest and most dangerous types of wildland blazes known as crown fires.

Twenty years later, Rothermel's fire behavior formulas are still in use today and stand as a pillar of wildland fire science, said Lloyd Queen, a University of Montana professor who directs the National Center for Landscape Fire Analysis in Missoula.

"The paradigm shifted in 1988," said Queen, a graduate student at the time of the Yellowstone fires. "Rothermel's equations are the basis for virtually all of the fire spread and fire behavior models we have today."

Unlike surface fires, which move through a forest at ground level, occasionally torching entire trees as they pass, crown fires roll through the dense upper reaches of the forest canopy. Leaping from treetop to treetop, their flames can top 100 feet in length.

Fully exposed to the wind, crown fires can race through a forest — making them both difficult to track and extremely dangerous for crews fighting fires on the ground.

Previous studies of crown fires had been driven by laboratory modeling, augmented by after-the-fact observations of forests that had already burned.

But from their satellite research station set up in West Yellowstone that summer, Rothermel and his fellow analysts had the ideal living laboratory — a front-row perspective on fires that ultimately burned more than a million acres in and around the park.

"I remember talking to some of the guys I was with and saying we really need a way of understanding these crown fire behaviors so we can get some kind of handle on what they're going to do," said Rothermel, now retired and living in Missoula.

The model he developed was an attempt to forecast a fire's mood swings at the landscape level — offering clues, for example, about when a fire will explode up a mountain valley or how long it will take to reach a residential subdivision.

To predict those activities, fire managers insert into a graphical equation what's known about conditions on the ground. From those bits of data — how dense the forest canopy is, what direction the winds are coming from, how steep the surrounding hillsides are — emerges a prediction of where and when a crown fire will spread.

Such knowledge can give fire bosses enough warning of dangerous conditions so they can gauge whether to put their own firefighters at risk.

Rothermel learned this firsthand, when he was out in the field studying a crown fire and a sudden shift in winds caused the blaze to lurch toward him and his fellow researchers.

"These are very dangerous fires," he said. "There were so many fires in Yellowstone, it was very difficult for the fire managers to organize any kind of suppression effort."

U.S. and Canadian forestry officials later collaborated on an effort to manufacture crown fires in a controlled setting. Beginning in 1997 and lasting several years, researchers from the two countries closely recorded the movements of a series of deliberately set fires in Canada's remote Northwest Territories, said Marty Alexander, senior fire behavior research officer for the Canadian Forest Service.

From those experiments, the analysts refined their understanding of crown fires. That is reflected today in the increasingly complex computer algorithms that govern fire forecasting, said Chuck Bushey, president of the International Association of Wildland Fire.

Yet the insights gleaned from Yellowstone's summer of fire proved to be essential.

"We've done a lot of cutting edge work to try to fill in the blanks," said Colin Hardy, program manager at the Missoula Fire Sciences Laboratory. "The work that Rothermel did there still prevails."




© Copyright Associated Press. All Rights Reserved. This material may not be published, broadcast, rewritten or redistributed.