Nationally, more fires occur in shrublands and grasslands than in forests. During the 1984-2008 timeframe, 14.5 million hectares of shrublands and grasslands burned in the conterminous US, as opposed to 6.0 million hectares for forest lands. Unfortunately, available national wildfire fuel databases, such as LANDFIRE, do not have particularly accurate fuel data for shrub and grassland ecosystems. This is significant to the Department of the Interior (DOI), and is of special importance to the residential communities that are located within or near these ecosystems. We will demonstrate approaches that can be readily and operationally used to improve shrubland/grassland fuel products to support DOI fire risk assessment through the integration of remotely sensed data and biogeochemical modeling. This work extends previous investigations supported by the United States Geological Survey (USGS) in land cover classification, fire behavior fuel modeling, and carbon modeling. For the first year, we plan to: 1) use an improved classification method and vegetation change detection approach to update fuel maps on shrublands/grasslands; 2) create weekly fuel moisture data in shrublands/grasslands using a combination of Landsat and MODIS data; and 3) build a biogeochemical carbon modeling system to simulate vegetation carbon changes and update the fuel maps in shrublands/grasslands. From this work we will provide fuel load information in shrublands and grasslands ecosystems that is more current, accurate, and relevant than what is currently available. This information will be especially valuable to natural resource managers who manage and assess fire risk in the fire-prone shrublands/grasslands that dominate the western US.