HYDROLOGICAL SIMULATION IN DABKA WATERSHED USING SWAT MODEL

Abstract

The present study is taken up to test the suitability of SWAT model for estimation of runoff and to understand sensitiveness of model parameters in a predominantly forested watershed in Kumaun region of Himalaya. The study area Dabka micro watershed (69.41 km2) lies in North West of Nainital in Uttarakhand state, receives abundant rainfall but the complex environmental factors like high seasonal variations in precipitation, temperature and other meteorological parameters and watershed features such as geology, steep and inaccessible slopes, degradation of forests and natural hazards like landslides, earthquakes etc. creates tremendous constraint in collecting reliable long term hydro meteorological data. This calls for its rainfall- runoff modelling by a properly calibrated hydrologic model in order to provide systematic and consistent information on water availability. Among the available rainfall- runoff models, the SWAT (Soil and Water Assessment Tool), developed by Agricultural Research Services of USDA, has gained popularity in the recent past, and is being used in the present study to quantify basin runoff. SWAT model is a comprehensive conceptual model and uses several parameters varying widely in space and time while transforming input into output. For model application, the catchment area is divided into 4 sub catchments. The whole catchment area is classified into five land use categories. Available hydrologic data (i.e. from January 2005 to May 2007), split into two groups is used for calibrating and validating parameters of the model. The model is calibrated at an upstream intermediate gauging site Bagjhala draining approximately an area of 65.78 km2 on monthly basis due to non availability of observed data at main outlet. The study concluded that model performed well with RMSE value 0.242 for calibration and 0.81 for validation. NSE for calibration and validation period was obtained as 0.77 and 0.73 respectively whereas R2 for calibration and validation period was found as 0.86 and 0.90 indicates a good model performance both for calibration and validation period