EVALUATION OF SATELLITE RAINFALL PRODUCTS FOR RIVER FLOW PREDICTION (THE CASE OF GENALE WATERSHED, ET

Abstract

In the present day, prediction of a river flow during early condition becomes an important input to the efficient operation of any water resource system. And Satellite rainfall products are recognized as an essential source of rainfall data especially for a region (including Genale watershed) having sparse and unevenly distributed rain gage observations. Therefore, the objective of this study is to evaluate the applicability of satellite rainfall products (TAMSATv-3 and CHIRPS) for Genale river flow prediction using HEC-HMS model. This is done using performance indices and pairwise comparison statistics in simulating the observed flow data. The evaluation was based on daily time scale for the period of nine years (2000-2008). Significant improvements in model stream flow simulations are obtained when the biases of these satellite products were corrected by bias factor. The rainfall estimates were initially validated against gauge data. Accordingly, TAMSAT showed a greater correlation coefficient than CHIRPS. The estimates were also tested by using them as input to a HEC-HMS model. Comparing the overall performance of both satellite, CHIRPS exhibited higher skills in detecting the peak river flow but not good in detecting the timing of peak. On the other hand, TAMSAT has shown an excellent score of percent of bias and best score of efficiency than CHIRPS. This result indicates the rainfall estimates of TAMSAT are best choice for Genale river flow prediction. The predicted river flow using TAMSAT were also tested using relative mean absolute error (RMAE) and root mean squared to standard deviation ration (RSR) performance measures. The result indicates that the use of for river flow prediction in Genale watershed exhibits un-satisfactory output.