Mapping fire fuels through detection of canopy biomass loading in Juniper, sagebrush, and Gambel oak communities

Fire fuel inventory processes are customarily labor intensive endeavors. There is a growing need for an increase in accuracy of these inventories at a landscape level, due in large part to the ever increasing development of Wildland Urban Interface (WUI). More accurate inventory and mapping of wildland fuels will facilitate a more accurate simulation of wildfire behavior and analysis of fire behavior given a myriad of fuels treatments. This paper examines one approach to inventorying fire fuels at a landscape level and developing fuel model maps to be utilized in landscape level fire behavior simulations for use by land managers in making fire and fuels related decisions. Three dominant vegetation classes are examined: Juniper, Gambel oak, and Big Sagebrush. Data was gathered and analyzed for Army Garrison Camp W.G. Williams, Utah. IKONOS multispectral data was used to develop several spectral derivatives such as texture and Normalized Difference Vegetation Index (NDVI). These coupled with gradient data were used to develop a regressive prediction model, to predict above-ground biomass for use in fuel model assignment. It was shown that this approach was ineffective in assessing fuel load and developing fuel maps. Several other approaches are discussed as alternatives.