Simulation of Hydrologic Processes through Calibration of SWAT Model with MODIS Evapotranspiration Data for an Ungauged Basin in Western Himalaya, India

224Hydrological modeling is widely practiced to understand the hydrological processes and its response to environmental changes. To obtain realistic results, sensitivity analysis and calibration are performed to identify the key parameters and their optimal values. Availability of discharge data makes the calibration process easier for gauged basins whereas for ungauged basins, the process is a challenging but essential task. Although regionalization approaches are popularly applied for ungauged basins, these approaches have some limitations and data availability issues. Satellite-based evapotranspiration can be effective for calibration of physically based models. The present study describes a simple and effective approach to calibrate hydrological model for areas with sparse data. The study used the Soil and Water Assessment Tool (SWAT) model on an ungauged river basin (Sirsa River) in northwest Himalaya, India. This model was parameterized through sensitivity analysis and manual calibration. The model’s simulated actual evapotranspiration (ET_a) was compared with a Moderate Resolution Imaging Spectroradiometer (MODIS) ET_a data product (MOD16A2) at daily (8-day composite) and monthly time-step for calibrations (2004–2006) and validations (2007–2008). The sensitivity of key parameters on selected five components (stream flow, surface runoff, base flow, deep aquifer recharge, and ET_a) was tested by manually changing parameters one at a time. The simulated ET_a on iteration was verified with four statistical parameters to choose the optimal parameter value. After optimization, the model was run to simulate hydrological components for the period 2003–2008. The calibration and validation results showed “good” and “very good” performance of the model for daily and monthly comparisons, respectively. Overall, the model overestimated ET_a in premonsoon and monsoon periods and underestimated it in postmonsoon and winter periods. This study showed that satellite-based evapotranspiration data can be effectively used for calibration of distributed hydrological model in data-limited regions worldwide.