Refined community-scale urban flood inundation model with real-time dynamic backflow simulation

The combined effects of climate change and urbanization have led to an increase in urban flooding due to extreme precipitation events, which show significant spatial variation within cities. Therefore, the development of a fine community-scale flooding model is crucial for urban flood prevention and control. This study presents the Urban Flood Inundation Model (UFIM), a refined model that integrates high-resolution topographic data, dynamic real-time 1D–2D coupling, and backflow effects to simulate flood inundation at community scale. The key findings are as follows: (1) Due to its built-in one-to-two-dimensional (1D–2D) connection modeling approach, the model can consider backflow effects, rendering water accumulation simulation more realistic and enhancing simulation accuracy. (2) UFIM achieved high accuracy with a precision of 0.81, recall of 0.77, Probability of Detection (POD) of 94%, and Nash-Sutcliffe Efficiency (NSE) between 0.785 and 0.800 under various actual precipitation events. (3) The model effectively simulated urban flood inundation for different return periods (1-, 5-, 10-, 20-, and 50-years), and different community types (including residential, functional and comprehensive communities). These results highlight the UFIM's potential to support refined flood risk assessment across diverse community types, inform adaptive urban drainage design, and enhance decision-making for resilient planning and emergency response.