Resumen
Precipitation is a significant input variable required in hydrological models such as the Soil & Water Assessment Tool (SWAT). The utilization of inaccurate precipitation data can result in the poor representation of the true hydrologic conditions of a catchment. SWAT utilizes the conventional nearest neighbor method in assigning weather parameters for each subbasin; a method inaccurate in representing spatial variations in precipitation over a large area, with sparse network of gauging stations. Therefore, this study aims to improve the spatial variation in precipitation data to improve daily streamflow simulation with SWAT, even pre-model calibration. The daily streamflow based on four interpolation methods, nearest neighbor (default), inverse-distance-weight, radial-basis function, and ordinary kriging, were evaluated to determine which interpolation method is best represents the precipitation at Yongdam watershed. Based on the results of this study, the application of spatial interpolation methods generally improved the performance of SWAT to simulate daily streamflow even pre-model calibration. In addition, no universal method can accurately represent the long-term spatial variation of precipitation at the Yongdam watershed. Instead, it was observed that the optimal selection of interpolation method at the Yongdam watershed is dependent on the long-term climatological conditions of the watershed. It was also observed that each interpolation method was optimal based on certain meteorological conditions at Yongdam watershed: nearest neighbor for cases when the occurrence probability of extreme precipitation is high during wet to moderately wet conditions; radial-basis function for cases when the number of dry days were high, during wet, severely dry, and extremely dry conditions; and ordinary kriging or inverse-weight-distance method for dry to moderately dry conditions. The methodology applied in this study improved the daily streamflow simulations at Yongdam watershed, even pre-model calibration of SWAT.