Several countries in SERVIR hub regions lack real-time information about hydrologic conditions such as streamflow—the amount of water flowing in their streams and rivers. For example, water managers in eastern Africa are interested in the availability of water in the region in multiple time ranges: historical, current, near-term (next 72 hours), and long-term. They need this information to predict where droughts or floods might occur, and decision-makers and disaster-response organizations need it to take effective advance action to save lives, crops, and property.

The CREST (Coupled Routing and Excess Storage) Hydrologic Modeling Tool provides water managers in the SERVIR hub regions with information about recent precipitation amounts and the resulting flood likelihood for watersheds. This information enables better decision-making related to water resources, floods, and agriculture. Currently, SERVIR is running the tool in several renditions, including historical model runs, near real-time, short-term forecast, and long-term forecast to quantify the impacts of climate scenarios on regional water resources.

The CREST modeling tool integrates NASA TRMM (Tropical Rainfall Measuring Mission) satellite data and local rainfall information, taking into account evaporation, transpiration, soil porosity, soil moisture, and land shape/elevation to produce streamflow estimates. CREST and TRMM melded data are available from 2002 to the present. Eastern Africa: SERVIR runs the CREST model regionally, using near real-time satellite-derived rainfall and KMD (Kenya Meteorological Department)-generated 72-hour rainfall forecasts to estimate streamflow. By contrasting these streamflow estimates against historical modeled estimates, SERVIR-Eastern and Southern Africa forecasts the likelihood of floods for every square-kilometer region. SERVIR-Eastern and Southern Africa also runs the model for specific watersheds of interest to the Kenya Department of Water Resources and sends those streamflow data to appropriate personnel daily.