Multi‐Scale Hydrodynamic Simulation Framework for Flood Risk Assessment of Urban Metro Stations

ABSTRACT Urban metro stations are highly vulnerable to pluvial flooding, yet accurately evaluating internal risks remains challenging due to the disconnected modeling approaches between surface inundation and internal intrusion processes. This study proposes a practical multi‐scale 2D hydrodynamic modeling framework that bridges overland flooding with multi‐level internal metro station inundation. Using a unified, performance‐portable shallow water equation solver (SERGHEI‐SWE) and a cascading boundary condition strategy, the framework provides an efficient screening‐level alternative to 3D computational fluid dynamics models. The methodology was applied to a metro station in Shanghai under a 100‐year design rainstorm to demonstrate its capabilities. Results demonstrate that local topography controls the relative timing and estimated magnitude of water ingress, producing strong heterogeneity among station entrances. Static or spatially uniform entrance boundary conditions can substantially distort internal hazard assessment. In particular, using spatially averaged entrance flood depth reduces peak hazardous area in the station by 92.8% while reducing peak water volume by only 7.4%, highlighting the nonlinear relationship between flood intrusion and evacuation hazard. The results also identify narrow corridors and stair‐connected passageways as critical hydraulic bottlenecks. By capturing continuous fluid dynamics from street level to underground platforms, this framework offers a computationally efficient, scalable tool to support the preliminary assessment of underground vulnerabilities and inform flood mitigation strategies.