Comparative Evaluation of Loss–Transform Methods in HEC-HMS for Flood Prediction in a Tropical Floodplain Catchment

Abstract

Flooding has become an increasingly severe hydrological hazard in Sibu, Sarawak, driven by the high discharge capacity of the Rajang River system and prolonged heavy rainfall during the monsoon season. The vulnerability of the area is further intensified by its low-lying floodplain setting, rapid urban expansion, and inadequate drainage infrastructure, resulting in recurring damage to lives and property. Despite the growing flood risk, localized rainfall–runoff modelling and flood-threshold assessment within the Rajang River catchment remain limited. This study, therefore, developed and validated rainfall–runoff models for the Rajang River using the Hydrologic Engineering Centre’s Hydrologic Modelling System (HEC-HMS) by evaluating six combinations of loss and transform methods. Model calibration demonstrated satisfactory hydrological performance across all six configurations, with each combination reasonably reproducing the observed hydrograph shape and runoff volume. Among the evaluated models, the combination of the Soil Conservation Service Curve Number (SCS CN) loss method and the Clark Unit Hydrograph (UH) transform method produced the best overall performance, achieving the highest Nash–Sutcliffe Efficiency (NSE) value of 0.69 and a coefficient of determination (R2) of 0.89. The results further revealed that the selection of the loss method exerted a substantially greater influence on peak discharge estimation than the transform method. The SCS CN approach generated significantly higher peak discharges, ranging from 8801.4 m3/s to 8984.7 m3/s, whereas the Initial and Constant loss method produced lower peak discharges, ranging from 5785.5 m3/s to 5806.8 m3/s due to greater infiltration losses. In addition, a flood threshold was established at an average total rainfall of 188.2 mm and a peak discharge of 5821 m3/s, effectively distinguishing flood events from non-flood events within the study area. These findings demonstrate the importance of appropriate loss-method selection in hydrological modelling and provide a scientifically grounded framework for improving flood forecasting, flood-risk assessment, and flood management planning in Sibu and similar tropical river catchments.