Block-level spatial integration of population density, social vulnerability, and heavy precipitation reveals intensified urban flooding risk

Abstract

Under the context of global warming and rapid urbanization, cities worldwide confront the pressing problem of urban waterlogging, hindering progress towards Sustainable Development Goals. Effective planning and mitigation of urban flooding require a comprehensive understanding of the spatial and temporal patterns of rainfall and risk heterogeneity. However, evaluating urban water-logging risk is challenged by the need for city-scale hydrological simulation and generally lacks comprehensive metrics integrating fine-scale datasets. To address these gaps, we developed a simulation method for urban flood hazards by integrating hydrological models and Random Forest algorithms. We then took Shenzhen, a megacity in China, as a case study, and investigated the spatial patterns of urban flooding risk and its determinants at the block level based on the risk assessment framework represented by Hazards-Exposure-Vulnerability (H-E-V) dimensions. We found that socio-economic indicators exhibited spatial clustering, while hazard-related indicators displayed more dispersed patterns. High-risk areas exhibited a highly heterogeneous spatial pattern, predominantly influenced by vulnerability and exposure factors, as well as the spatial mismatch among the three dimensions. Our results emphasize the importance of integrating spatial heterogeneity of exposure and vulnerability into climate adaptation resource allocation, addressing both current and future demands for effective climate mitigation.