Hydrological real-time modeling using remote sensing data

Abstract

Reliable real-time forecasts of the discharge canprovide valuable information for the management of a riverbasin system. For the management of ecological releaseseven discharge forecasts with moderate accuracy can be be-neficial. Sequential data assimilation using the Ensemble Kalman Filter provides a tool that is both efficient and robustfor a real-time modelling framework. One key parameter in ahydrological system is the soil moisture, which recently canbe characterized by satellite based measurements. A fore-casting framework for the prediction of discharges is devel-oped and applied to three different sub-basins of the Zam-bezi River Basin. The model is solely based on remote sens-ing data providing soil moisture and rainfall estimates. Thesoil moisture product used is based on the back-scatteringintensity of a radar signal measured by a radar scatterome-ter. These soil moisture data correlate well with the mea-sured discharge of the corresponding watershed if the dataare shifted by a time lag which is dependent on the size andthe dominant runoff process in the catchment. This time lagis the basis for the applicability of the soil moisture data forhydrological forecasts. The conceptual model developed isbased on two storage compartments. The processes modeledinclude evaporation losses, infiltration and percolation. Theapplication of this model in a real-time modelling frameworkyields good results in watersheds where soil storage is an im-portant factor. The lead time of the forecast is dependent onthe size and the retention capacity of the watershed. For thelargest watershed a forecast over 40 days can be provided.However, the quality of the forecast increases significantlywith decreasing prediction time. In a watershed with littlesoil storage and a quick response to rainfall events, the per-formance is relatively poor and the lead time is as short as 10 days only.