Flood Risk Assessment and Mitigation for Urban Rail Transit Stations under Heterogeneous Information and Multisource Uncertainty

Abstract Urban rail transit stations (URTS) are increasingly exposed to flood risks due to more frequent extreme rainfall events and rapid urbanization. However, reliable flood risk assessment remains challenging given the complexity of risk information and the diverse sources of uncertainty. To address this issue, this study proposes a flood risk assessment framework capable of jointly handling heterogeneous information and multisource uncertainty by integrating stochastic multicriteria acceptability analysis (SMAA) with evidential reasoning (ER). By independently sampling objective factor values and subjective parameters within their feasible domains, the framework captures multiple sources of uncertainty and their combined effects on flood risk assessment. The ER rule is used to fuse heterogeneous risk information, and Monte Carlo simulations are conducted to obtain stable risk assessment results. To support targeted mitigation, Shapley additive explanations (SHAP) is introduced to interpret the assessment results and identify key drivers and unfavorable feature patterns associated with high-risk stations. The proposed framework is demonstrated using 256 URTS in Chongqing, China. The results showed that high-risk stations are mainly concentrated in the urban core and exhibit clear spatial clustering. These stations are generally characterized by low elevation, high imperviousness, intense rainfall exposure, low drainage pipe density, and underground station structures. A mitigation scenario analysis further showed that enhancing drainage capacity and reducing imperviousness around stations can effectively reduce URTS flood risk. This study provides a methodological reference for URTS flood risk assessment and mitigation under heterogeneous information and multisource uncertainty.