by Cary Shimek

When the Northern Rockies blazed last summer during the worst fire season in 90 years, University of Montana researchers helped fire managers make better decisions through the use of new imaging and reconnaissance techniques.

LLoyd Queen, a UM forestry associate professor and remote-sensing scientist, used technology developed for NASA satellites to help planners peer under the smoke to map fires, monitor fire behavior and predict what the blazes would do next.

“I heard comments like, ‘No one at any fire anywhere has ever had data like this,’” Queen says.
Queen is part of UM’s Numerical Terradynamic Simulation Group, which crafted software for the Terra environmental satellite that NASA launched in December 1999. The instrument NTSG programmed scans the entire world every one to two days, with the ability to hone in on details a kilometer wide. The satellite monitors a multitude of earthly conditions — everything from vegetation and snow cover to fire activity.

Queen says he had hoped the 2000 fire season would offer five fires of at least 100 acres so UM researchers could test the fire models and algorithms they had developed. What actually took place went way beyond his expectations. He says there were so many fires over such a large area for such a long period of time that traditional resource-assessment tools needed a helping hand.

Queen works to develop practical applications for the remote-sensing tools pioneered by NASA and UM. He says there is a national technology infrastructure in place that is powerful but wasn’t designed to meet an average fire manager’s needs in western Montana during the summer of 2000.

“The picture we were getting with the satellites was too big,” he says. “We were looking at the entire forest. We couldn’t make a specific fire intervention based on this data. At some point we needed to see the trees.”

When the fire season began to seriously heat up in mid-July, Queen and his fellow researchers teamed up with area Forest Service employees to form an ad hoc group called FIST — the Fire Intelligence Support Team. FIST members gathered around the table to discuss the information and assessment needs of fire managers and what sort of real-world applications UM could offer firefighters immediately.

“It was sort of like, ‘Hey, someone does care about what we are doing,’” Queen says. “I knew then that it was possible that what I do would help an assessment specialist or a planner have better data to make better decisions about fire priorities.”

He says the UM team volunteered to help out in any way it could, offering the equivalent of two full-time workers around the clock until the fires were put out. About six UM employees found themselves working 18- to 20-hour days during the peak six weeks of the fire season.

Queen says the first thing they did was put in place some computing resources for Forest Service Region 1 to access. UM system administrator Saxon Holbrook brought together server space, a Web site and computers that provided a clearinghouse for NTSG data.

“What Saxon put together was kind of an extension of the things we need for our research activity,” Queen says. “We opened up our servers so people in the fire community could have access to our resources. And it’s all in one spot.”

He says this computing infrastructure allowed them to make available many aspects of Terra science, such as fire detection and monitoring, surface-moisture index modeling and some vegetation/productivity assessments. The UM group also had its computers busily ingesting satellite data and mapping fire locations at least once — and usually twice — a day.

Queen says the team members also adjusted the global algorithms used for Terra to the regional setting of the Northern Rockies. In general they tweaked the models to be conservative on estimating fuel danger and conditions but more liberal when it came to fire activity and spotting.

“Part of adjusting the software is understanding what kind of error you can live with,” he says. “A model won’t be perfect — it’s an abstraction of reality. But you don’t want to have a model that doubles the number of fires that are actually burning. That doesn’t help fire managers with priorities. It leads to chaos.”

Queen says FIST also got involved with tactical-level fire reconnaissance using an airplane wired with a digital infrared scanning system. Often used by the military, the system allowed team members to take fire snapshots with a 1-meter resolution. Queen and Colin Hardy of the Forest Service Intermountain Fire Sciences Laboratory contracted with Minnesota-based Airborne Data Systems to provide the high-tech imagery.

“The first day ADS was here, Colin and I got into this two-engine plane and flew 600 miles of reconnaissance,” Queen says. “We identified hot-spot activity and priority areas for data collection.”

The ADS imagery allowed them to see through smoke and create incredibly detailed maps of the current fire situation. Queen says the data collected from the plane was processed using the algorithms designed for the Terra satellite. Fresh information was in the hands of fire managers two hours after the plane returned from its flyover. Queen says such close-range fire reconnaissance tools have not been available before.

“We put that aircraft in some very challenging situations,” he says. “We sent the plane to areas of very complex terrain because — as you can imagine — getting a good map in very steep terrain is difficult to do. We had them fly in areas of thick, intense smoke, over heavy fuels, light fuels, closed canopy, open canopy, areas that have been harvested, not harvested. ... We were just trying to cover a range of settings so we could test the performance of our system.”

He says it was challenging to collect data over very hot fires since intense burns could saturate the detector and blind it — like looking directly into the sun. “We had to figure out things like: If we’ve got a fire backing downhill at 3:30 a.m. at 18 percent relative humidity on a southwest-facing slope in lodgepole pine with a 33-percent canopy cover and 100 tons per acre of 1,000-hour fuels, how do you not burn out? We had to adjust a lot of dials.”

Queen says his team developed a good working relationship with Wayne Cook, the incident commander on the Wilderness Complex fires southwest of Conner. Cook would tell them of areas where he needed fire reconnaissance.

“So we would take the coordinates he gave us, plot them, do out a flight plan and the ship leaves,” Queen says. “The ship comes back. Two hours after it hits the ground, the (geographic information system) analysts on Wayne Cook’s team down at the West Fork Ranger Station are transferring files from our server to their computers. We could tell them exactly where their hot spots were in terms of (latitude/longitude) coordinates.”

Cook, who has managed fires for 25 years, says the data UM and ADS provided him was extremely useful.

“I hope this is a snapshot of the future as far as technology to help fire managers do our job better,” he says. “That’s why we came together — to test this technology and determine if it has potential. I hope this will be a stepping stone to the future.”

The images obtained by the ADS plane and processed by UM computers are often stunning. A picture of the Alder Fire southeast of Missoula, for example, shows a wooded area shrouded in a haze of smoke. Look at the same image in the infrared, however, and a large burning mass is revealed.

Pointing at a computer screen, Queen says, “The fire we see here can’t be seen with the naked eye.”
He says fire managers use a program called Farsite that tries to predict where a fire is going and at what intensity. The improved data provided by ADS and the FIST team allowed managers to hone their models and the performance of the Farsite program.

“One of the neatest things that came out of all this was that we were able to turn the science we are doing here at UM to practical uses,” Queen says. “That’s something that university researchers don’t always get a chance to do.”

For more information, contact LLoyd Queen at (406) 243-2709 or e-mail lpqueen@ntsg.umt.edu.