Frequent urban inundations from changing rainfall patterns highlight the need for reliable, timely flood forecasting. This study evaluates inundation prediction under varying storm sewer network simplification and surface grid resolutions using InfoWorks ICM, which supports 1D–2D coupled simulations. Five storm sewer configurations, determined by cumulative watershed area, and five grid resolutions yielded 25 scenarios. Results show that the number of two-dimensional mesh elements influences simulation time more than the number of storm sewer pipes. Coarser grids significantly reduce computation but may oversimplify topographic variability, reducing inundation depths and overestimating inundation extents. Conversely, increasing network detail raises inundation area and depth as well as computation time. For small- and medium-scale basins (e.g., Sinlim4, Daelim), a simplification threshold of 2.25 ha was optimal, while for large basins, 1 ha was best. These thresholds strike a balance between modeling accuracy and simulation speed. A sensitivity analysis was performed to examine changes in inundation depth and area for each simplification level, revealing high sensitivity in topographically complex areas. Additionally, the need for historical flood event validation—including the August 2022 flood in Dorim-river—is outlined for future work. The findings offer practical guidelines for real-time urban flood forecasting system design.
Kim et al. (Thu,) studied this question.