Developing Alternative Fuel Models with WorldView-3 to Estimate Fire Behavior in a Semi-Desert Grassland

Holly Eagleston

Holly Eagleston US Fish and Wildlife Service

22E

Developing fuel models to help predict fire behavior and hazard for desert grasslands is challenging due to the spatial and temporal variability of fine-fuels. We developed custom fuel models specifically tailored to semi-desert grasslands in southern Arizona and compared them to LANDFIRE (LF) fuel models. We used our own fine-fuel biomass and cover type layers developed from field plots and 2015 Worldview-3 and Landsat 8 satellite imagery using machine learning algorithms to assign the standard fire behavior fuel models. LF (2001 and 2014) and custom fuel models (2015) were used in fire behavior software to compare 20 randomly assigned fire starts run under 90th percentile fire weather conditions. Custom models differed significantly from LF model outcomes showing up to 26 times greater fire size, particularly in areas with higher non-native plant cover. Modeled fire-line intensity and actual burn severity data from four wildfires and two prescribed burns indicated that the custom model more closely follows burn severity patterns within the study area. Semi-desert grasslands dominated by non-native perennial grasses with more contiguous fuel-bed structure are likely to experience larger fire size. Fuel models developed from data representing these conditions produced more realistic fire behavior that can improve targeting for fuel hazard mitigation and long-term efforts to restore native grassland composition. Accurate fuel models are critical for wildland fire management, they allow fire managers to determine how hazardous fuel conditions are, and the likelihood of a large fire that could endanger wildlife habitat and human infrastructure.

10:45 Developing Alternative Fuel Models with WorldView-3 to Estimate Fire Behavior in a Semi-Desert Grassland, Holly Eagleston

January 30 @ 10:45
10:45 — 11:00 (15′)

Mineral B

Holly Eagleston

Add to Google Calendar