Abstract:
This research was aimed at conducting catchment modeling for better understanding of hydrologic functioning and runoff generation mechanisms of Gilgel Abay River basin and to evaluate the performance of ANN model and to compare with previous studies and finally to develop rainfall-runoff model that can be used in the design, planning, and management of water resources in Gilgel Abay River basin. The rainfall runoff models chosen for this study was ANN model. The input data for this model i.e. precipitation, temperature, potential evapotranspiration and flow data was treated from their gaps and the quality of these data (homogeneity, consistency and outlier by rainbow software and double mass curve analysis was done) checked. The model was applied to test the catchment, using split record evaluation, involving the calibration and validation periods (about 70% for calibration and 30% for validation). The performance of model was done based on hydrographs, scatter plots, performance evaluation parameters (NSE, RMSE, R2 and PBIAS) and residual analysis in the calibration and validation periods for Gilgel Abay River basin and ANN model was better performed during the calibration and validation periods. This study tries also to differentiate the performance of ANN on daily or month basis and it performs better in daily basis than monthly basis. Further study is recommended by producing more accurate satellite rainfall data (due to their obvious advantage of high spatial & temporal resolution and large areal coverage) and spatial data integrating with models through field measurement which is important for the catchment in assessing available water resource & for its irrigation potential.