Background: Accurate forecasting and communication of water and water-related hazards in developing regions could save untold lives and property. To this end, the proposing team has developed and implemented the CREST (Coupled Routing and Excess Storage) hydrologic model over East Africa. Since 2009, CREST has found uses in dozens of countries as a user-friendly, flexible, and highly extensible platform for monitoring water resources, floods, droughts, and landslides. One of the major extensions to the CREST family of modeling tools is EF5, the Ensemble Framework for Flash Flood Forecasting. The EF5 framework enables ensemble forecasts of hydrologic variables via the use of multiple hydrologic models including CREST. The EF5/CREST modeling infrastructure has been extensively tested through a series of training workshops and capacity building activities in USA, Africa, Mesoamerica, and South Asia over the past decade and is thus particularly well-suited for hydrologic capacity building in emerging countries. Finally, EF5/CREST can be coupled to other software programs for additional functionality11Research: Building on past activities and incorporating recent advances, this proposal seeks to continue our fruitful partnership with the Eastern and Southern African hub of SERVIR RCMRD (the Regional Centre for Mapping of Resources for Development) as well as across SERVIR’s global network. The central goal of this project is to develop an ensemble hydrologic prediction system, forced by weather and climate forecasts in a single continuum, to communicate forecasts on scales ranging from sub-daily to seasonal and in formats designed for better decision making about water and water-related disasters.Outputs: This proposal will deliver an updated hydrologic ensemble modeling framework (EF5) to RCMRD with new model physics and better forecasts of streamflow, soil moisture, and other hydrologic states. The specific outputs include: 1) Ensemble forecasts of flood impacts at the local and basin scale using updated hydrologic models; 2) Flood inundation forecasts visualized in RCMRD’s Flood Mapper tool; 3) Ensemble forecasts of water resources at seasonal scales; 4) An automated service for the communication of outputs from the proposed ensemble forecasting system to national users; 5) Improved collection of reports from end users for assimilation into the forecasting system; and 6) Training and capacity building activities at SERVIR hubs and other regional partners and stakeholders upon request.Relevance: This proposal directly responds to two of this SERVIR solicitation’s major foci: water and water-related disasters as well as weather and climate. The proposal relies heavily upon NASA remote sensing data, including NASA TRMM/GPM, SMAP, and SRTM. The proposal also builds upon past and current SERVIR projects at RCMRD, including CREST, Flood Mapper, Frost Monitor, and the GeoSocial API, and has been developed and refined with RCMRD’s long-term strategic goals in mind.Intrinsic Merits and Broader Social Impacts: EF5/CREST is a proven performer at getting researchers and officials in emerging regions excited about and confident in their ability to independently monitor, forecast, and understand water and water-related disasters. To date, few regional or continental scale systems enable ensemble forecasts of floods in one system from near real-time to seasonal time scales. The inclusion of dissemination and observation collection services will connect output from this project to a wide audience. The results of the project, though initially designed for deployment in eastern and southern Africa, will be applicable to other hubs of the SERVIR global network. Broader societal impacts include improved understanding and monitoring of flooding conditions and the ability of national authorities and non-governmental organizations to deploy resourcesand respond to disasters threatening human lives, livelihoods, and health.