The Monitoring Trends in Burn Severity (MTBS) project was established, in part, to help assess the effectiveness of national fire policies; provide essential spatial fire disturbance data for update of regional/national land cover characterization databases; and provide a comprehensive and consistent dataset for regional planning and resource monitoring. We propose to automate portions of the burn severity assessment process using a combination of 1) satellite-based fire detections to improve the fire occurrence database and 2) automated query and processing of the Landsat image archive to speed identification of appropriate Landsat imagery. Once identified, further automated image processing will be developed to generate fire perimeters and apply default burn severity thresholds derived from the analysis of over 13,000 fires in the MTBS database. The proposed project aims to create an automated process that will aid in mapping previously unreported fires; and partner with federal and state agencies that use MTBS products to better determine wildfire risks and model landscape responses to fire. MTBS is mandated to map all fires greater than 405 hectares in the Western US and greater than 202 hectares in the Eastern US (from 1984 to present). At its inception (2005), the MTBS program was guided by federal agency fire records. Additional sources of fire occurrence records, particularly at the state level, have added to the work load. The program has mapped over 5000 undocumented fires that were discovered while mapping reported fires. We hope to determine how many other unreported fires exist. Further, the sheer number of fires to assess is beginning to overwhelm MTBS¿ current production capability. Increasing numbers of federal and state users are utilizing MTBS data and it is important that MTBS continues to increase its capacity to identify and map all significant fires. MTBS will use current and historical remotely sensed fire detections from MODIS Aqua, MODIS Terra, AVHRR, and GOES satellites compiled by the Hazard Mapping System and MODIS Active Fire Detects to drive an automated process that selects potential pre- and post-fire Landsat images and generates subsets (centered on the fire detection) for analyst review. Once an MTBS analyst has determined whether the fire detection meets MTBS criteria and has selected the most appropriate Landsat scene(s) for mapping the fire, a second automated process will create a thematic burn severity image product (low, moderate, and high burn severity) and a fire perimeter for each fire. Analysts subsequently review the MTBS products for accuracy and correct them as needed. The proposed procedures will provide a more complete fire occurrence database and drastically reduce the amount of analyst time spent to identify fires from state and federal records, manually search and select appropriate Landsat scenes, and map the fires. MTBS will therefore be better able to ensure that all large fires on public and private land within the US are assessed and mapped quickly and consistently. MTBS will increase production of products and collaborate with identified state and federal users to provide more accurate estimates of fire acreage and severity on public and private lands. A more complete characterization and understanding of current and historical fires will help MTBS and collaborators. Overall, an improved MTBS dataset will lead to better prediction of wildfire hazards, ecological modeling, and reduced threat to people and property. All members of the MTBS group have extensive experience in creating MTBS products. In addition, group members have developed similar automated scripting processes for MTBS and other projects that utilize programming languages (Perl, Python, etc.). Team members are working to identify appropriate state and federal collaborators for this proposal.