In partnership with the Department of Interior's Bureau of Land Management (BLM), we propose to build and evaluate the prototype RECOVER decision support system. RECOVER will be an automatically deployable, site-specific multi-criteria decision aid that brings together in a single application the information necessary for Burned Area Emergency Response (BAER) teams to plan reseeding strategies and monitor ecosystem recovery in the aftermath of savanna wildfires. RECOVER will use state-of-the-art cloud-based data management technologies to improve performance, reduce cost, and provide site-specific flexibility for each fire. Customized RECOVER instances will be automatically deployed in the Amazon EC2 Cloud when a fire is detected. RECOVER's decision products will be dynamically assembled from the existing network of data resources. RECOVER will automatically generate and refresh derived fire severity, fire intensity, and other products throughout the burn so that when the fire is contained, BAER teams will have at hand a complete and ready-to-use RECOVER system customized for the target wildfire. Since BAER remediation plans must be completed within 14 days of a wildfire's containment, RECOVER has the potential to significantly improve the BAER decision-making process. Most work in this area focuses on forest wildfires. RECOVER adds an important new dimension to post-fire decision-making by focusing on ecosystem rehabilitation in semiarid savannas. A novel aspect of RECOVER's approach involves the use of soil moisture estimates, which are an important but difficult-to-obtain element of post-fire rehabilitation planning. We will use downscaled soil moisture data from the three primary observational sources currently available to begin evaluating the use of soil moisture products in BAER decision-making and to build the technology needed for RECOVER to manage future SMAP products. As a result, RECOVER, BLM, and the BAER applications community will be ready customers for data flowing out of new NASA missions, such as NPP, LDCM, and SMAP.