Hydrological Modeling Using Observed And Satellite Percipitation Data: The Case Study Of Wabe Watershed, Omo-Gibe River Basin, Ethiopia.

Abstract

Precipitation is the most significant atmospheric input to land surface hydrological models. But the rainfall stations are scarce and distributed unevenly, making it difficult to collect precipitation data, which leads to insufficient the simulation of hydrological processes. Alternatives such as satellite-based rainfall estimates can be useful. There are several satellite based datasets accessible for use in modeling; however, selecting the most accurate and reliable satellite product that should be well matched to the intended area should be considered. Therefore, there is a need to evaluate the accuracy of satellite rainfall product. The aim of this research is hydrological modeling using observed and satellite (GPM_IMERG and CHIRPS) precipitation data for Wabe watershed. Both continuous statistical and hydrologic modeling approaches were used for the performance evaluation from the most representative rain gauge for five stations at daily and monthly time steps. Intercomparison between satellite rainfall product and observed data were done using point to grid method by selecting five representative meteorological stations. For continuous statistical evaluation, CHIRPS is performed better than GPM_ IMERG for daily and monthly timescales in detecting and estimating rainfall for the basin. The performance of HEC-HMS model was evaluated using, Net-Sutcliffe Efficiency (NSE), Root Mean Square Error (RMSE), a coefficient of determination (R2 ) and Percent Bias. The model Calibration and Validation results of observed station showed (NSE = 0.55, R 2 = 0.571, RMSE = 0.7, PBIAS = -0.64) and (NSE = 0.58, R 2 = 0.6, RMSE = 0.7, PBIAS = -6.71), throughout the periods respectively. For CHIRPS, the model Calibration and Validation results showed (NSE = 0.532, R 2 =0.51, RMSE = 0.5, PBIAS = 2.48) and (NSE = 0.51, R 2 = 0.547, RMSE = 0.6, PBIAS = 12.04), throughout the periods respectively; whereas for that of GMP_IMERG the model Calibration and Validation results showed (NSE = 0.5, R 2 = 0.516, RMSE = 0.6, PBIAS = 8.33) and (NSE = 0.44, R 2 = 0.458, RMSE = 0.7, PBIAS = 11.08) throughout the periods, respectively. The study shows that the HEC-HMS model is satisfactory in hydrological modeling. However, the finding indicated that observed rainfall is more suitable than Satellite Rainfall for hydrological modeling of wabe watershed. The research outcome is more important to policy planner for efficient water resource management.