Analysis of mechanism and optimal value of urban built environment resilience in response to stormwater flooding

The concept of urban resilience focuses on understanding the process and mechanisms of disaster occurrence, providing innovative approaches to address stormwater flooding. However, existing studies primarily concentrate on enhancing overall system resilience, with limited research examining the temporal progression from stormwater disturbance to flood generation. To fill this gap, this study categorizes the development process of stormwater flooding into three periods: disturbance resistance (DR), adjustment and adaptation (AA), and rapid recovery (RR). Using the SWMM (Storm Water Management Model) software, 27 representative parcels in the Beijing-Tianjin-Hebei region of China were simulated. By sequentially considering single-indicator control variables, resilience indicators that significantly impact the three periods were identified through the construction of a stormwater flooding resilience indicator library. Subsequently, resilience models for each disaster phase were constructed using the BP (Back Propagation) neural network, and genetic algorithms were employed to optimize the models and determine the optimal values of resilience indicators for each period. Finally, the research findings were summarized into a resilience design method for the built environment to address stormwater flooding, accompanied by a guide for improving stormwater flooding resilience.The study reveals the following key findings: (1) the influence of physical and spatial elements in the built environment on stormwater flooding formation varies across different stages of the disaster process; (2) distinct resilience indicators operate at different times and in different ways throughout the entire stormwater flooding resilience process; (3) enhancing stormwater flooding resilience in the built environment does not necessarily require setting specific threshold values for each influencing indicator; instead, an optimal single value emerges when multiple indicators interact. Moreover, when multiple indicators interact, an optimal combination module with the best value for a single indicator exists. This study investigates the complete cycle from storm disturbance to flood disaster formation, offering both solutions for cities to mitigate storm flood disasters and advancing theoretical research on urban storm flood resilience while fostering interdisciplinary integration.