Evaluation of satellite precipitation products using HEC-HMS model

AbstractAccurate measurement of precipitation is vital to investigate the spatial and temporal patterns of precipitation at various scales for rainfall-runoff modeling. However, accurate and consistent precipitation measurement is relatively sparse in many developing countries like Ethiopia. Nevertheless, satellite precipitation products may serve as important inputs for modeling in an area with scarce field data for a wide range of hydrological applications. In this study we evaluate the high-resolution satellite rainfall products for hydrological simulation, the Climate Hazards Group Infrared Precipitation with Stations (CHIRPS) and Tropical Rainfall Measuring Mission Multisatellite Precipitation Analysis (TMPA_3B42v7) satellite rainfall products for stream flow simulation at daily temporal and 0.25° × 0.25° spatial resolution. The study area is located in Dabus watershed, Abbay basin, Ethiopia. We applied a nonlinear power law to remove the systematic error of satellite precipitation estimates for input into HEC-HMS hydrological model for runoff generation. The performance of the satellite rainfall and hydrological model was evaluated using Nash–Sutcliffe efficiency (ENS), coefficient of determination (R2), relative volume error (RVE), and percentage error of peak flow objective functions. The result of HEC-HMS model performance revealed R2 of 0.78, ENS of 0.69 for CHIRPS_2 and R2 of 0.79, ENS of 0.76 for TMPA_3B42v7 satellite rainfall products during calibration periods. Our result indicated that the HEC-HMS model well predicated catchment runoff for both satellite precipitation products. The study shows that the model performance was significantly improved when bias-corrected satellite rainfall input replaced than the original uncorrected satellite products. Overall, our study showed that gauge-based simulation outperformed than satellite in terms of all objective functions over the study area.