NSF Award
1249627
#1249627<br/>Francis Fendell<br/><br/>Indiscriminate fighting of wildland fire can be costly, unsafe, ineffective, counterproductive (may result in soil-destruction), and ecologically more harmful in its execution and consequences than the fire itself. At a large-area wildland-fire incident, under the Incident Command System, the highest-priority information sought by the key decision-maker regarding response, the Incident Commander (IC), concerns firespread. Given where the actively flaming firefront is currently located, and given locally pertinent information on topography and vegetation and meteorology, where will the firefront be located six-to-seven hours in the future, both in the absence of intervention with firefighting countermeasures and with such intervention? Better guidance is continually sought from the in-place Wildland Fire Decision Support System, created under the lead of the United States Department of Agriculture Forest Service, with assistance from the United States Department of Interior land-management agencies. Can more adequate objective prognostic aids be generated for real-time support of decision-making by responders to wildland fire? Here, a systematic approach with three fairly autonomous components is outlined: (1) collection of data on topography, vegetation, and meteorology, with emphasis on meteorology, typically the crucial, short-time-scale variable; (2) systematic laboratory-scale testing under well-characterized, well-controlled conditions in a novel facility dedicated to measurement of the quasisteady rate of firespread as a function of the just-discussed data; and (3) utilization of level-set methods to evolve, in real time, the firefront evolution under the just-discussed semi-empirical rate of firespread, so that an ensemble of realizations permits the uncertainty (robustness) of the forecast to be assessed. A treatment with a one-way interaction between the weather and the firefront, idealized as without structure, adequately encompasses the preponderance of wildland fires encountered by wildland-fire responders. The third component, utilizing limited but already-in-hand spread rates, is to be undertaken first. The goal is to generate and validate an openly accessible, fast-running, robust, readily portable, well documented, easily modified, fire-growth-prediction computer code, useful to both a novice and a specialist in wildland-fire behavior.<br/><br/>By developing a computer code that applies well-suited level-set methods to firefront evolution, many basic issues impeding a practically important implementation of what remains a mathematical prescription will be addressed. An accurate, efficient quantitative aid for wildland-fire growth has multiple pre-incident applications (e.g., siting of roads, reservoirs, and firebreaks), during-incident applications (e.g., comparison of alternate firefighting strategies and justification of evacuation orders), and post-incident applications (e.g., certification of personnel and assistance with wildfire-related litigation). The project is transformational re: use of laboratory testing in firespread research; adoption of the ?fire-use? (restrained-response) option by decision-makers; and expediting operational firespread forecasts.
