عنوان مقاله [English]
نویسندگان [English]چکیده [English]
A lot of watersheds in Iran do not have sufficient hydrometry stations. Also because of limited accessibility to hydrologic data in more stations, watershed modeling plays a vital role in development of water resources, and this is a subject to large uncertainties. In this research, Soil and Water Assessment Tool (SWAT) model was used for runoff simulation of Rood Zard basin. Model sensitivity analysis, calibration, uncertainty and validation analysis was carried out using Sequential Uncertainty Fitting, version2 (SUFI-2) program. The runoff data of hydrometric stations (Machin and Pol Manjanigh) were used for calibration and validation of the model.The model was run using daily precipitation and temperature data of seven rain gauge and synoptic stations for 15 years (1995-2009). The years (1997-2003) were used for calibration and of (2004-2009) were used for validation of the model. Uncertainty indexes (p-factor and d-factor), the coefficient of determination (R2) and Nash Sutcliffe (NS) were used to evaluate the ability of the SWAT model in simulation of runoff for both stations. These factors for calibration period at the Machin station were 0.92, 1.10, 0.85, and 0.85, respectively, and for the validation period, these statistics were 0.92, 0.85, 0.92, and 0.92 respectively.These factors at the Pol Manjanigh station for calibration period were 0.72, 0.99, 0.83, and 0.82, respectively, and for the validation period, these statistics were 0.70, 0.76, 0.79, and 0.78, respectively. The reasons for the weak simulation of runoff at some months of the year might be due to weak simulation of snowmelt in this mountainous watershed, disconformity of model’s suppositions to transfer flow in saturated and frozen soil layers, as well as uncertainty and lack of enough data. Overall results showed that the model can simulate the runoff properly in both stations. It is recommended to use this model instead of empirical models for estimating runoff and sediment basins country.