Monitoring Water Availability in Africa with Land Surface Models and Remote Sensing

Severe droughts in 2015 and 2016 in Ethiopia and Southern Africa negatively impacted agriculture and municipal water supplies resulting is food and water insecurity. Information from remotely sensed data and field reports indicated that the Famine Early Warning Systems Network (FEWS NET) Land Data Assimilation (FLDAS) accurately tracked both the anomalously low soil moisture and runoff conditions. This work presents efforts to more precisely track water storage, as estimated by the FLDAS outputs, to improve early warning of water availability anomalies. Preliminary results indicate that routed FLDAS runoff is well correlated with observed streamflow where irrigation and other channel modifications are not present. FLDAS evapotranspiration (ET) anomalies are also well correlated with ET anomalies from the Operational Simplified Surface Energy Balance model (SSEBop) in Eastern and Southern Africa. In locations where precipitation, runoff and ET estimates are accurate and where we have information on per capita water use, we evaluate how and why water storage is changing over time. We then use these results to better estimate changes in water storage in regions where information on the water balance is lacking.