Parameter uncertainty of Land Use Change Impact Prediction on the Hydrology of the Upper Blue Nile Basin, Ethiopia

AbstractLand use land cover (LULC) has changed significantly due to population growth and causing a huge environmental problem at global and regional level, especially in countries where their economy depends on agriculture. Previous studies have used hydrological models to assess the impact of LULC change on the hydrology of Upper Blue Nile (UBN) basin, but may not have addressed a wide range of uncertainties which arises due to parameter, error in input data, model structure uncertainty and level of understanding of the user. The aim of this research was to estimate the effect of parameter uncertainty on land use change impact prediction on UBN basin, Ethiopia. SWAT model was chosen and set up in the UBN basin and Sequential Uncertainty Fitting (SUFI2) technique was used for calibration, sensitivity and uncertainty analysis of the model. Parameter uncertainty range, prediction change uncertainty and the impact of land use change on the UBN basin hydrology was assessed using two hypothetical land use scenarios (Scenario 1: convert half of forested area in to agricultural land andScenario 2: completely change the forest cover into agriculture). The results of the simulation indicated that SWAT model simulate the flow for the UBN basin with p-factor of 0.94 and 0.85 for calibration and validation respectively, while the r-factor were 0.86 and 0.71 for calibration and validation, respectively. The goodness of fit results between the observed and simulated flow showed that NSE for calibration and validation period were 0.87 and 0.89, respectively, and PBIAS for calibration and validation period were − 2.7% and 1.4%, respectively. Average daily stream flow was increased by 5.5% and 11% for scenario 1 and scenario 2, respectively compared to the baseline. For both hypothetical scenarios, the parameter uncertainty shows an effect for the predicted flow and based on coefficient of variation CN, CANMX and ALPHA_BF have relatively shows higher uncertainty. Additionally, predicted change uncertainty for the minimum flow is higher than the average and maximum flow. The predicted effect of land use change on the basin is significant and there is also a chance of land use change to occur in the UBN basin, due to that soil and water management practices should be done in the basin to prevent environmental and socio-economic problems which arise due to land use changes. Uncertainties on model prediction should not be ignored in LULC change studies to get a better understanding for decision making and to avoid over and underestimation of possible pressures.