A176-0012
Exploring wildfire management applications for advanced earth observation techniques

Tuesday, 15 December 2020
Poster
Mary Ellen Miller1, Nancy H F French2, Charlotte Weinstein3, Peter R Robichaud4, Sam Batzli5, Matthew B Dickinson6, William J Elliot7, Joseph Paki2 and Michael Billmire8, (1)Michigan Technological University, Michigan Technology Research Institute, Houghton, MI, United States, (2)Michigan Technological University, Houghton, MI, United States, (3)Michigan Technological University, Ann Arbor, United States, (4)USDA Forest Service Rocky Mountain Research Station, Moscow, ID, United States, (5)University of Wisconsin, Madison, WI, United States, (6)US Forest Service, Delaware, OH, United States, (7)Rocky Mountain Research Station Moscow, Moscow, United States, (8)Michigan Technological University, Research Institute, Houghton, MI, United States
Abstract:
Earth observations are a vital component of wildfire management. Before the flames, earth observations are needed for mapping fuels and monitoring fire danger. Thermal imagery during a wildfire is critical for understanding fire location, intensity, and spread rates. Post-fire land management depends heavily on earth observations to quantify the effects of fire on the landscape in order to help predict flooding, erosion, slope stability and tree mortality. Forecasting emissions requires detailed spatial information on fuels as well as fire location and severity. As a part of the FASMEE Western Wildfire Campaign, and using FIREX-AQ airborne data, we are exploring the potential for applying new sensors and techniques for fire management. The focus will be the 2019 Williams Flat Fire that burned in Washington state. Evaporative Stress Index products from NASA’s ECOSTRESS (ECOsystem Spaceborne Thermal Radiometer) are being compared with fuel moistures derived from Remote Automated Weather Station (RAWS) measurements. Individual stations show some correlation, indicating potential for improving maps of pre-fire danger and enhancing predictions of post-fire emissions that rely on fuel moisture. We are also exploring the potential benefits of incorporating ECOSTRESS products into a machine learning model designed to map fire danger and forecast potential burn severity. In order to provide rapid post-fire assessment of burned areas we are exploring the use of VIIRS (Visible Infrared Imaging Radiometer Suite) derived burn severity maps. VIIRS has a lower spatial resolution than Landsat data, but higher temporal resolution offering the potential for more rapid indication of potentially hazardous post-fire conditions. We compared model predictions of post-fire erosion and runoff derived from traditional Landsat derived burn severity with the new VIIRS product and found a high degree of correlation. New tools and datasets also need to meet the needs and formats used by fire and land managers.