关键词: 极端暴雨；城市内涝模拟；数据驱动模型；机理驱动模型

Abstract: Against the background of global climate change and urbanization, extreme rainstorm events are frequent, and urban waterlogging has become increasingly severe, threatening urban safety. In order to reduce the threat of waterlogging and improve the emergency management level of extreme rainstorm events, it has become an important research trend to analyze the urban waterlogging process under extreme rainstorm conditions with the help of simulation methods. Based on the analysis of the basic characteristics of extreme rainstorms, the main influencing factors of urban waterlogging accumulation were identified. The two main simulation methods of urban waterlogging under extreme rainstorms were systematically summarized, i.e., mechanism-driven models and data-driven models. The former has clear physical processes but long calculation time, while the latter meets the requirements of fast simulation prediction but lacks physical mechanism. On this basis, the future development trends of urban waterlogging simulation under extreme rainstorm conditions were prospected from five aspects: multi-index dynamic analysis of simulation results, improvement of simulation accuracy and efficiency, deep integration of urban-scale and basin-scale models, dynamic combination of mechanism models and numerical weather forecasts, and real-time coupling of mechanism-driven and data-driven simulations. The research results can provide a reference for the identification and management of urban waterlogging processes under extreme rainstorm conditions.

Key words: extreme rainstorm; urban waterlogging simulation; data-driven model; mechanism-driven model