Advances and challenges in modelling and prediction of urban flooding: a comprehensive review on recent progresses

Urban flooding has been an increasingly significant challenge for many cities due to rapid urbanization and climate change, leading to more frequent and intense of extreme events. Recent advances in modelling and prediction of urban flooding have focused on application of machine learning and other AI-oriented methods in modelling and prediction of urban flooding. This paper is a review of the recent development trends and results in modelling and prediction of urban flooding, categorizing them into four sub-topics: 1). Coupled Hydrological-Hydraulic Models, 2). Integration of High-Resolution Data and Remote Sensing in the model simulations, 3). AI-based Flood simulation and Prediction and 4). Climate Adaptation and resilient approaches. These new or improved approaches have been developed and demonstrated in the real case studies internationally. It is concluded that the recent development in urban flooding modelling reflects a shift towards more integrated, data-driven, and climate-resilient approaches. The combination of advanced machine learning and other AI technologies with coupled 1D& 2D hydrological-hydraulic models, high-resolution data can improve the flood simulation accuracy and computation efficiency, and further, the community involvement is enhancing the ability of cities to predict, mitigate, and adapt to urban flooding challenges.