Spatiotemporal Evolution of Urban Flood Risk in a Mid-Sized City Cluster: A Three-Stage Assessment of Six Northern Henan Cities Around the July 2021 Extreme Event

Post-disaster research on the July 2021 (“7·20”) Henan flood has concentrated on the megacity Zhengzhou, leaving six severely hit mid-sized cities unassessed. This study assesses urban flood risk for Anyang, Hebi, Xinxiang, Jiaozuo, Puyang, and Jiyuan over 2018–2024 (42 city-year samples) on a 12-indicator Hazard–Exposure–Sensitivity–Adaptive capacity panel, objectively weighted (CRITIC × entropy, game-theory combination) and graded by set pair analysis. The design prioritizes verification over novelty: indicators survive first-difference collinearity screening (16 to 12); official disaster data are reserved as external validation criteria; results are stress-tested by seven robustness scenarios (including a threshold-free TOPSIS benchmark) and a cloud-model operator check reproducing the city ranking (ρ = 0.975). Exposure, not hazard, structures the ranking—Xinxiang highest (0.565) and Jiyuan lowest (0.446). The 2021 shock raised risk in all six cities (+0.027 to +0.087), and risk proves sticky: Xinxiang and Jiyuan had not returned to pre-disaster levels by 2024. Obstacle profiles are city-specific and exposure-led, with drainage shortfall acute in the two sticky cities. On a three-city validation sample, the panel reproduces official hazard and affected-population rankings exactly (Spearman ρ = 1.0; illustrative given n = 3). Indicator-panel screening offers mid-sized city clusters a data-light route to flood-risk triage beyond megacity attention.