Unraveling the factors behind self-reported trapped incidents in the extraordinary urban flood disaster: a case study of Zhengzhou City, China

With the rapid development of urbanization, urban flood disasters caused by extreme rainfall have become increasingly frequent, causing great losses to residents' lives and property. Understanding the spatial disparities of residents trapped in floods and its influencing factors at the fine-scale is crucial for realizing urban risk mitigation. This study analyzed 994 self-reported trapped incidents from social media platforms during the 7 & sdot;20 extreme flood disaster in 2021 in Zhengzhou City, China. Using the Kernel density estimation (KDE) method and Geographical random forest (GRF) model, we explored the spatial disparities and influencing factors of these incidents. Results showed that self-reported trapped residents mainly concentrated in the central urban areas characterized by high population density and frequent socio-economic activities, exhibiting spatial heterogeneity characteristics. In addition, precipitation, road density, density of vulnerable groups, per unit GDP, water depth and river density had the significant impact on spatial disparities of residents trapped in floods. The spatial heterogeneities of local coefficients (%IncMSE) in the GRF model showed that these influencing factors played different importance within the city. Based on the perspective of detailed trapped incidents in floods at a fine scale, this study can provide vital implications for building resilient cities in other areas worldwide.