Event-based rainfall-runoff simulation using different precipitation loss methods: case study in tropical monsoon catchment

In the context of monsoon-dominated catchments, it is important to investigate runoff simulation responses to highly intense rainfall events which produce high-magnitude floods. In the process of calibrating parameters to a particular flood-prone area, a comparative analysis of loss methods which estimates the volume of runoff must be conducted first to determine the optimal set of parameters that assist accurate rainfall-runoff simulation. As such, this study aims to evaluate the runoff simulation performance of HEC-HMS event-based hydrological model setup using three precipitation loss methods: initial and constant, SCS curve number and Green and Ampt by calibrating their parameters against the observed discharge data in monsoon-dominated Gin catchment (972 km(2)) in Sri Lanka which frequently subjected to flooding. Four peak rainfall events covering the two monsoon seasons in 2016 and 2017 were used for the model calibration and validation. Nash-Sutcliffe efficiency (NSE), coefficient of determination (R-2) and percentage bias (PBIAS) were used as the objective functions. Overall, the SCS curve number loss method showed the best results for both Thawalama and Baddegama discharge gauging stations for calibration (NSE > 0.5, R-2 > 0.80, and PBIAS <= +/- 5.5%) and validation (NSE > 0.8, R-2 > 0.80, and PBIAS <= +/- 8.5%) in combination with Clark's unit hydrograph transform method, recession baseflow method and Muskingum routing method. By simulating the magnitude of peak runoff and the timing of peak runoff generation to an acceptable level, the study revealed the successful applicability of the HEC HMS event-based modelling with the SCS curve number precipitation loss method to simulate monsoon peak flows. Calibrated model with its optimized parameters can be used for hydrological investigations in monsoon-dominated catchments with similar characteristics, especially for flood simulations.