A Rapid Flood Inundation Model for Urban Flood Analyses

An efficient inundation model is required for flood early warning systems in urban areas. A 2D flood model uses a governing shallow water equation, and this model is computationally expensive despite benefiting from parallel computing techniques. As an alternative to conventional flood models, cellular automata (CA) and DEM-based models (DBMs) have been studied. CA flood models simulate floods efficiently. However, a small time step is required to ensure model stability when the grid size decreases due to its diffusive characteristics. Conversely, DBM models produce results quickly, but they only show the maximum flood extent. Additionally, pre- and postprocessing are required, which take considerable time. This study proposes a hybrid inundation model that combines the two alternative approaches, and it efficiently produces a highresolution flood map without complex pre- and postprocessing. The hybrid model is also integrated with a 1D drainage module, and thus, the model reliably simulates urban area floods.