A hydrologic calibration of the SWAT model in Kranji Catchment, Singapore

Abstract

The Public Utilities Board (PUB) of Singapore wishes to expand recreational activities in Singapore's surface waters through the Active, Beautiful, and Clean Waters Program (ABC Waters). One area of concern with ABC Waters is microbial pollution. Pathogens pose an immediate and substantial risk to human health when humans come into contact with contaminated water. In order to open surface waters for public recreation, PUB must insure that these waterways are free of pathogens. The Massachusetts Institute of Technology in the United States and Nanyang Technological University in Singapore are working together on several water quality research projects related to ABC Waters. A surface runoff model for Kranji Catchment, located in northwest Singapore, was created in the Soil and Water Assessment Tool (SWAT) by Granger (2010) and improved by Bossis (2011). The model incorporates hydrology, land use, weather, soil, and slope to calculate surface flows and water quality constituents. The model was originally created to be a prediction tool for bacteria concentrations in outflows throughout the Kranji Catchment. Earlier versions of the model, however, grossly underpredict flows. In an effort to improve the accuracy of the SWAT model, the surface hydrology was calibrated to flows measured at five stream gauge stations in Kranji Catchment. Precipitation within the SWAT model was changed to equal a rainfall record at one of the stream gauge stations. A sensitivity analysis was also conducted for surface flows. Surface flows were most sensitive to changes in six input parameters, and two of these parameters were changed in the final, calibrated model. The soil evaporation compensation factor (ESCO) was increased, and the curve number for moisture condition II (CN2) was decreased. These three changes in the model greatly improved the way the SWAT model predicts surface outflows, and with these changes, surface outflows predicted by the model match measured values very closely at all five stream gauge stations. While hydrologic calibration greatly improved surface flow calculations in the SWAT model, predictions of bacteria concentrations did not improve. The SWAT model still requires further calibration work to bring bacteria concentrations closer to measured values.