Estimation of Ground Water Recharge Using SWAT Model

In this study the SWAT model was applied to estimate a shallow groundwater recharge at the watershed scale. The SWAT model is a semi-distributed comprehensive surface and subsurface flow model with the capability of simulating sediment and agricultural chemicals. The study site is Bocheongcheon watershed which is one of IHP experimental watersheds in Korea and the rainfall and stream flow have been monitored since 1984. The dominant land use types of the watershed are the mixed forest and agricultural land. The input data for SWAT model were prepared using the digital land use and soil maps with daily rainfall measured at ten rain gauge stations and the meteorological variables such as daily wind speed, relative humidity, solar radiation, and temperature collected within the watershed. The SWAT model was calibrated based on four years of daily stream flow data using the shuffled complex evolution global optimization method. Since there was no information for soil hydraulic properties, the various published pedotransfer functions were used for the specification of soil hydraulic conductivity. The annual recharge calculated from SWAT model ranged from 125 mm to 191 mm depending on the selected pedotransfer functions. Although the estimates of the regional groundwater recharge vary with the selection of pedotransfer functions, the performance measures between the simulated and measured daily stream flow are appeared to be similar. The result indicates that it is very difficult to identify a unique parameter set and the proper identification of spatially consistent soil hydraulic conductivity has an important implication for modeling groundwater recharge at the watershed scale using the spatially distributed watershed model such as SWAT.