Fighting fires with geospatial data

Starting with a 6 a.m. daily briefing, John Truett’s team members — he calls them his troops — use maps for just about everything they do. Maps help them locate their drop points and visualize an area’s topography when dense smoke blocks all visibility. “If it wasn’t for mapping, we’d be lost,” said Truett, an operations section chief for the California Interagency Incident Management Team 2 (CIIMT2). The team helped battle a 68,000-acre blaze in Elko, Nev., earlier this summer. “Maps are what drive the operations section. If I don’t have a map, I can’t point out to these guys what our strategy and tactics are.”The wildland fires raging throughout parts of the United States at a near-record pace this year are also burning through federal land management agencies’ budgets. Because of the situation, fire managers plan to seek more money from lawmakers and rely on geographic information system maps to help them battle the blazes. GIS maps let managers layer data about weather patterns, fuel sources and an area’s fire history on a topographical map of the fire zone. Increasingly, fire managers use GIS maps to manage fires rather than suppress them. Maps help them make better predictions about how a fire will spread and better decisions about how much money to spend fighting it. The Bureau of Land Management and the Forest Service have traditionally focused on fire suppression rather than fire management, a practice congressional auditors criticized in a June report. The Government Accountability Office report encouraged agencies to follow 2004 guidelines that require fire managers to consider all available options when selecting a firefighting strategy, including less aggressive approaches. Choosing fire management when possible, instead of active suppression, can lower costs, GAO said. Many Forest Service fire managers agree with GAO that technology can help firefighters manage some fires that they might otherwise try to suppress. “We’ve proven in spades that we can’t stop fires or prevent them,” said Mark Finney, a research forester at the Forest Service’s Rocky Mountain Research Station. “They are part of nature.” Finney has designed several predictive fire programs that use GIS data. Nevertheless, active suppression will always be necessary in cases of fires ignited by a person or blazes that threaten to harm communities, fire managers say. And for fighting large blazes — which requires communication and coordination among different federal agencies and state and local officials — GIS data and maps provide a common view of the fire zone.Truett’s operations section is a part of CIIMT2, which includes logistics, finance and planning sections. Mapping professionals and fire behavior analysts are included in the planning section. All team members report to an incident command staff. Authorities can call in CIIMT2 to manage disasters when state or local officials are unable to respond adequately. CIIMT2 includes about five trained mapping or GIS specialists, depending on the size of the incident. Corey Ferguson, CIIMT2’s permanent GIS specialist for the planning team, said he constantly makes new maps for public briefings and firefighting operations.Team members gather new GPS information daily and combine it with infrared and other data overnight. Planes fly over the fire during the night to gather infrared data. Team members add that data to information field observers gathered during the day. By 6 a.m., team members present fire managers with a new master briefing map and projections for the day.“We used to have field observers actually walking the fire perimeter,” Truett said. “Now, with GIS, you have all of these layers so you can pull up, say, infrastructure. You can pull up v getation types, you can pull up fire history. All those layers add up to make our job more effective in terms of building a strategy.” Teams such as the 45-member CIIMT2 usually spend a few weeks at a fire scene. Incoming GIS specialists typically spend a day in briefings with their outgoing counterparts. Local jurisdictions also often give the map-makers several layers of administrative data about the fire zone.The GIS specialists also create their own geospatial data. Ferguson’s team maps the fire line and various types of point data, such as drop off points and staging areas. Once on the scene, team members use GIS software to analyze the data and make predictions about the path of the fire.Even as analytic tools continue to improve, experts say, the effectiveness of GIS maps is still largely dependent on the accuracy of spatial data files. And until recently, the accuracy of that data was problematic.“For years and years, decades in fact, people had worked on fire modeling, and their models were basically point models,” Finney said. “They were one-dimensional models.” The older models didn’t allow people to see how fires would expand in different environments and under varying conditions. But as models improved, fire managers still faced the problem of inconsistent data. Information from local authorities often would conflict with national or state data, for example. Variations in data had a ripple effect when mixed with other spatial data necessary for making predictive and operational maps. The problem was particularly frustrating for fire managers who tracked data on factors such as vegetation and terrain that affect the movement of fires.“My state counterparts…and my Bureau of Land Management partners have different data at different resolutions, and we’ve struggled tremendously,” said Berni Bahro, the Forest Service’s regional fuels planning specialist for the Pacific Southwest Region. “The last decade has been just about building seamless data that’s consistent for the things that affect fire.”Officials say they’ve made great progress in recent years because of cross-jurisdictional initiatives to standardize data. As part of the Landfire project, for example, the U.S. Geological Survey, the Forest Service and the Interior Department have teamed to produce geospatial data products that describe factors such as the composition and structure of existing vegetation.Another challenge has been to get data that is as current as possible, Bahro said. “If your data is four years old, and you are trying to make decisions in 2007, then you run the risk of making some assumptions about that data that aren’t legitimate.”The Landfire project is developing standardized data for fire managers. The project has completed datasets for 11 western states, and fire managers are using those this fire season. New predictive programs that use probability statistics need that data.“None of those programs could run without geospatial data, and Landfire provides the basic data layers for those programs to run,” said Dan Crittenden, the Forest Service’s National Landfire business leader. Experts say it is crucial that agencies devise a strategy for updating Landfire data regularly. Agencies are reviewing a schedule, and a plan for the revisions should be established in the next few months, Crittenden said.If Landfire data does not reflect current fuel conditions, land managers and incident management teams are unlikely to trust fire-spread probability maps, the GAO audit report states. And without accurate maps, fire managers are likely to select more aggressive and more costly firefighting strategies than necessary, the auditors said. Fire managers increasingly use GIS maps to monitor rather than fight blazes. Dan Mindar, a fire behavior analyst, worked with Shelly Crook, a long-term analyst, at the Fool Creek wildfire in Montana this summer. Unlike the Elko inferno, the 11,000-acre Fool Creek blaze a the time, was deemed a useful fire. Mindar and Crook’s job was to perform GIS mapping activities t at helped ensure that the fire did not burn out of control.Within the fire management community, the focus has shifted recently from primarily suppressing and fighting fires toward accepting fire as a necessary part of the ecosystem, Crook said. That shift coincides with the proliferation of increasingly sophisticated predictive computer programs that can forecast a fire’s path. Commanders can use that information to reduce risks by restricting access to particular forests if conditions appear too dangerous. For example, the Farsite program, which Finney helped create, uses GIS data and weather forecasts to calculate how quickly and in what direction a fire will spread in 24 to 72 hours. Farsite has been around since the 1990s. But Finney recently developed another program, FSPro, which allows fire managers to make predictions as long as 30 days in advance by running thousands of Farsite simulations using Landfire data. “What we are doing now is adapting the same fire growth models for problems looking out 14 days or 21 days and accommodating the uncertainty in weather,” Finney said. “What you are getting is a probability map…that summarizes thousands of possible outcomes.” Farsite and FSPro are in use today.Experts are quick to point out that even though predictive technology is improving, the public should not expect fire officials to eliminate forest fires or even wish that forest fires would never occur. Finney said that when fires are not allowed to burn, the result is a build-up in vegetation fuels, which cause fires to burn uninterrupted for longer distances.