Development of a SWMM-RAS Mutual Feedback-Based Technique for Inland-River Flooding Prediction Using Forecast Rainfall

Recent increases in localized heavy rainfall have intensified compound flooding in urban river basins. However, previous studies have mainly relied on one-way boundary conditions from rivers to urban areas, and therefore, their ability to account for interactions between urban drainage discharge and river water levels is limited. This study developed a 10-minute mutual feedback framework linking EPA-SWMM and HEC-RAS for the Oncheoncheon watershed in Busan, Korea (NSE = 0.97), using 60 minutes of observed rainfall and 30 minutes of forecast rainfall for real-time flood prediction. The framework enabled prediction of pluvial flooding and river water levels over the following 30 minutes and identification of increasing fluvial flood risk. A depth-weighted flood index was also applied to assess pluvial flood risk. Under an extreme rainfall event, water levels exceeded the design flood level in some downstream sections, demonstrating that the coupled framework can capture the interaction among urban drainage discharge, river water levels, and pluvial flooding. Expanded application of flood-mitigation facilities reduced pluvial flooding but increased river water levels under certain conditions, highlighting the need for integrated flood prediction and facility operation considering both pluvial and fluvial flood risks.