A Hybrid-fidelity Modeling Approach for Evaluating Flooding Risk in Large Urban Infrastructures

As climate change and urbanization intensify, urban infrastructures face escalating flood risks, necessitating robust assessment methods for effective prevention and control measures. The complexity of large-scale urban infrastructures, with their numerous entrances and exits, demands comprehensive flood risk modeling that can balance computational efficiency with simulation accuracy. However, traditional high-fidelity flood simulations at the urban basin scale are often constrained by data limitations and computational demands. This study proposed a hybrid-fidelity modeling approach to address these challenges. At the basin scale, the method integrates rainfall-runoff hydrological modeling with one-dimensional river hydrodynamics. At the neighborhood scale, where infrastructure entrances are concentrated, it couples one-dimensional sewer hydrodynamics with two-dimensional flood propagation modeling. This multi-scale approach enables the lower-fidelity basin models to provide boundary conditions for higher-fidelity neighborhood simulations, optimizing computational efficiency while maintaining assessment accuracy. The hybrid-fidelity modeling approach was applied to flood risk assessment of urban infrastructures in a planned region. Model simulations determined maximum inundation depths at infrastructure entrances under various rainfall scenarios, establishing a scientific foundation for developing targeted flood prevention and control strategies. This hybrid-fidelity approach offers a practical solution for comprehensive urban flood risk assessment in large infrastructures, balancing model complexity with computational feasibility.