Operational Application of Satellite Remote Sensing for Hydrological Prediction in Ungauged African Basins

Floods are among the most catastrophic natural disasters around the globe impacting human lives and infrastructure. Implementation of a flood prediction system can potentially reduce these losses. Typically, the set up and calibration of a hydrologic model requires in situ observations (e.g. rain gauges and stream gauges). Satellite remote sensing data have emerged as viable alternatives or supplements to in situ observations due to their coverage over ungauged regions. The focus of this study is to utilize the best available satellite products and integrate them in a state-of-the-art hydrologic model to characterize the spatial extent of flooding and associated hazards in sparsely-gauged or ungauged basins in Africa.

We will present a methodology based entirely on satellite remote sensing data to calibrate a hydrologic model, simulate the spatial extent of flooding, and evaluate the probability of detecting inundated areas. MODIS- and ASTER-based flood inundation maps are retrieved over the region and used to benchmark the NASA-OU CREST distributed hydrologic model simulations of streamflow and inundation areas in Lake Victoria Basin. Furthermore, Aqua AMSR-E Microwave based stream flow estimates are also used to validate the CREST model parameter transferability in the Kavango basin. The analysis has showed the applicability of integrating satellite data products as input for a distributed hydrological model as well as direct estimation of flood extent maps. The quantification of flooding spatial extent through orbital sensors can help to evaluate hydrologic models and hence potentially improve hydrologic prediction and flood management strategies in these ungauged basins. The ultimate goal is to apply the above research results to the regional operational decision-making support system SERVIR-Africa (www.servir.net).