Abstract:
Soil erosion, land degradation and loss of agricultural soils are a major problem in Upper Blue Nile Basin of Megech and Koga watershed, Ethiopia. The parameter efficient semi-distributed watershed model (PED) and soil and water assessment tool (SWAT) are tested for the prediction capability of sediment yield in Megech and Koga watersheds. The objective of this study is to evaluate sediment prediction capability of two hydrological models (PED-WM and SWAT) in a range of watershed sizes in the Upper Blue Nile Basin. The selected watersheds were Koga (231Km2) and Megech (507 Km2). Model calibration and validation were carried out using suspended sediment yield for Koga watersheds from 1998-2008 and 2009-2013, respectively and for Megech watershed from 2000-2009 and 2010-2014, respectively. For evaluation of the model, the method was used model performance evaluation criteria such as NSE, R2 and PBIAS. The model efficiency on monthly time step for sediment yield simulation during calibration of PED-WM was (NSE = 0.81, R2 = 0.92, and PBIAS = 7.5) and for SWAT (NSE = 0.77, R2 = 0.81, and PBIAS=9.3) and during validation of PED-WM (NSE =0.79, R2 = 0.93, and PBIAS = -5.6) and for SWAT (NSE = 0.68, R2 = 0.73, and PBIAS= 13.2) were obtained for Koga watershed. In Koga watershed, the average daily predicted suspended sediment yields were 0.5g/l using SWAT and 0.42 g/l using PED-WM during calibration period. The positive PBIAS in both the SWAT and PED-WM indicate that the model under estimated the Koga and Megech watershed during the calibration period. Similarly the model efficiency on monthly time step for sediment yield simulation during calibration of PED-WM was (NSE = 0.7, R2 = 0.88, and PBIAS = 9.18) and for SWAT (NSE= 0.63, R2 = 0.77, and PBIAS = 11.2) and for validation of PED-WM (NSE, R2, PBIAS=0.7, 0.87, 13.3) and SWAT (NSE =0.67, R2 = 0.78, and PBIAS = 15.3) for Megech watershed. In Megech watershed the average daily predicted suspended sediment yields using SWAT and PED-WM were 0.53g/l and 0.34 g/l, respectively, during calibration period. The model performance measure such as NSE, R2 and PBIAS result for the two watersheds were the range of satisfactory to a very good agreement between the observed and simulated suspended sediment yield in both the SWAT and PED-W model. The overall model performance indicated that PED-WM model was the most appropriate model to predict sediment yield than SWAT model.
