Evaluating the effect of urban flooding on spatial accessibility to emergency shelters based on social sensing data

Abstract

Urban flooding is a growing source of natural hazards, significantly threatening the safety of sustainable development in cities. The distribution of flood risks is heterogeneous, so it is crucial to allocate emergency resources reasonably. This article develops an analytical framework to evaluate the effect of urban flooding on emergency responses based on social sensing data. Initially, we designed a Weibo search pattern and used natural language processing technologies to get high-risk flood points. Then, assuming that such high-risk flood points can disrupt traffic, we assessed urban emergency shelter accessibility after flooding events. Finally, we carried out an evaluation of spatial fairness between population and emergency shelter accessibility through spatial correlation analysis. We analyzed the urban area of Nanjing as a case study, extracting 37 high-risk flood points from the past 5 years. The results highlight that existing emergency shelters fall short in accommodating the needs of urban residents under disaster conditions. This disparity is notably amplified in high-risk flood points. By measuring the impact of flooding quantitatively, we expect to promote a more comprehensive management on urban flood risks.