The massive fires that burned a third of Yellowstone National Park in 1988 captured national headlines, introducing many across the country to the destructive potential of large wildfires in the West.
But as the growing blazes captured the nation’s attention during that long, hot summer, they also demonstrated to fire scientists the shortcomings of their ability to predict the spread and behavior of large wildland fires.
In 1988, as now, the mathematical formulas for predicting wildfire behavior relied mostly on an underlying model developed in 1972 by Richard C. Rothermel, a former aeronautical engineer tapped by the U.S. Forest Service to help direct research at the Northern Forest Fire Laboratory in Missoula.
But while the Rothermel model could reliably simulate the future behavior of a wildfire burning in grass, brush or downed timber, it had only limited use once the flames made the jump to the forest canopy and became a crown fire.
“It’s almost like two different kinds of fires,” said Matt Jolly, a research ecologist with the fire lab, now called the Missoula Fire Sciences Laboratory. “A surface fire that’s basically attached to the ground isn’t able to pull air in from underneath the way an elevated crown can … The flame lengths (on a crown fire) are 20 times or greater than those you would see with a surface fire.”
The fires in Yellowstone that year weren’t the first to prominently feature crown fires raging across a landscape. But the destructive force of the runaway blazes inspired Rothermel and others who witnessed them to develop better tools for firefighters to understand what they were up against once the flames began reaching into the treetops.
“I remember Yellowstone provided us with one of the first opportunities to actually look at how we could model crown fire behavior,” said Greg Poncin, a deputy incident commander for one of the nation’s 16 Type 1 incident command teams.
Poncin, the area manager of the Montana Department of Natural Resources and Conservation's land office in Kalispell, spent the better part of the past decade at the helm of his Type 1 team, leading its response to recent high-profile fire complexes in Glacier National Park in 2015 and the Lolo Peak and Rice Ridge fires near Missoula last year. But as a young firefighter in 1988, he was sent to Yellowstone as a strike team leader trainee.
“We got to see it kind of in this laboratory that was Yellowstone Park,” he said. “The inversions would break around 11 or noon and shortly after that we’d be able to observe the fire growing into these lodgepole crowns and making these big runs.”
Since that summer, additional models developed by Rothermel and other fire scientists have helped fire behavior specialists more accurately predict crown-fire movement.
And the means for running those models have also changed dramatically. Jolly’s current desktop computer, armed with 140 gigabytes of RAM, has roughly 2 million times the computing capacity of the Commodore 64 machine he was using 30 years ago.
John Krebs, who retired in 1995 as a fire management officer with the Clearwater National Forest in Idaho, was a fire behavior analyst in 1988 serving on one of the Type 1 teams fighting the Yellowstone fires.
“When I started out, they were using nomograms to help determine how fires would spread,” Krebs said, referring to a diagram used to sketch out complex formulas on paper. “Then they turned to the computer stuff, and that was on a TI-59, a calculator that you held in your hand. They figured out how to put a chip in there and do the same thing.”
With advances in the field of meteorology, modern firefighting teams can also rely on detailed forecasts to predict more than two weeks’ worth of weather in and around the burn area, Jolly said. And not least of all, fire behavior specialists can now access models that combine satellite mapping with ecological data collected on the ground to create high-resolution maps of vegetation and fuels present on the landscape.
“If we don’t have landscape-scale descriptions of the fuels, then it’s just a guess,” he said.
Yet while the science of firefighting has developed substantially since the ’88 fires, much of the tactics remain the same. The military-style “incident command” structure that firefighting teams use today has been in place since a few years before the Yellowstone fires broke out.
“The actual tactics of how they’re going to engage the fire are still determined from what they see, the observations and likelihood that something might change,” Jolly said.
Even Rothermel, who had a habit of showing up to large wildfires to study fire behavior in its natural environment, understood the limited value his models could provide without solid data gathered from the front lines.
Working on a separate fire burning near Glacier National Park, Krebs said he unexpectedly bumped into the computer-toting engineer walking around with one of his colleagues and punching numbers into the portable device.
“I was in charge of prescribing when they would do a burnout on this fire, and here come these two guys from the fire science lab,” Krebs said. “I said, ‘You’re scaring the heck out of me, man, with that little black box in your hand.’ … He said, ‘John, it’s not what comes out of this black box, it’s what goes into it that makes the difference. You’re not going to get an answer unless you know what goes into getting that answer.’”
But while the science that firefighters use to plan their approach has progressed dramatically since 1988, the type of fire behavior seen in Yellowstone that year has simply become more common in the West.
“Those of us that were out there during that summer of ’88 thought that this would be one of the summers to end all summers. It was just incredible how much fire was out there on the landscape,” Poncin said. “Looking back, it was more of a foreshadowing of what we’re going to see more and more of. That was probably one of my big takeaways.”