Fire managers will soon have ready access to weather and fire models to predict the behavior of fires whose locations are known. The models result from many years of research and development, but they require substantial computing support, which has become feasible for fire applications in recent years. Currently, spatial decision support systems (DSS) used by the USDA Forest Service incorporate fire prediction models such as FARSITE and Wildland Fire DSS (WFDSS) and various remote sensing data to help forest managers with the tools to effectively monitor and manage fuel buildup. A major source of error in prediction models that impacts the DSS function and in turn the fire management, is the spatial distribution of fuel loads. Recent innovations in active remote sensing technology and development of new techniques to accurately quantify the distribution of forest fuel loads may radically improve the models? performance and directly impact the efficiency of the WFDSS. In this proposal, we study the feasibility of integrating the fire fuel loads estimated from high resolution L-band synthetic aperture radar (SAR) data with the FARSITE model. The study will be performed with data acquired over the Yellowstone National Park (YNP) as part of an earth science application research in 2004 and the FARSITE and WFDSS models already developed for the region. The proposed work will include: 1. Development of canopy fuel parameters from AIRSAR and ALOS data using algorithms developed by Saatchi et al. (2007). Backscatter data from L-band polarimetric SAR images will be processed and used to create a wall-to-wall distribution of canopy fuel loads at 25 m spatial resolution. 2. Assessment of improvements of FARSITE and WFDSS models using the estimated fuel parameters as input data layers. The performance of the model using the radar derived parameters will be compared with the fuel load lookup tables used for vegetation types of the park. 3. Development of measurement requirements for future NASA missions such as DESDynl by quantifying spatial resolution and precision of fuel parameters needed in fire models. The model simulations will be evaluated in a risk assessment framework to predict the probability of fire spread, given a fire event at a known location in the park. We will examine the performance in terms of the estimation errors associated with canopy fuel loads and the spatial resolution of the products. The results of the study will benefit the fire managers and the decision making process for hazard assessments by providing a more realistic state of the fuel loads. The research will allow us to develop specific requirements for DESDynl mission to be used effectively in forest fire prediction and management.