ABSTRACT Flood risk assessment serves as a critical tool, providing theoretical foundations for minimizing flood damage and effective flood management. In June 2023, the Kakhovka Dam failure highlighted the need for accurate flood risk assessment under different breach scenarios. This study assessed the spatiotemporal flood risk to population and built‐up areas under partial (S1) and complete (S2) dam‐breach scenarios using the HEV (hazard–exposure–vulnerability) framework. Hotspot analysis was used to identify high‐risk zones, and classification differences between HEV and DV methods were compared. The results are as follows: Flood propagation was simulated using the TELEMAC‐2D model, yielding an NSE of 0.98 based on observed water depths. Spatial validation produced a precision ( P ) of 70.7%, a false positive rate (FPR) of 22.9%, a false negative rate (FNR) of 3.2%, and an accuracy of 84.6%. For population flood risk, the total number of people at risk was approximately 101,000 under S1, increasing sharply within the first 41 h. Under S2, the total rose to approximately 296,000, with a rapid increase observed in the first 20 h. For built‐up area, the total flood‐affected extent was 51.6 km 2 under S1, showing a sharp increase in the first 41 h. Under S2, the total affected area expanded to 157.4 km 2 , with a rapid rise during the first 20 h. High‐risk clusters of both population and buildings were mainly located near the dam site and in the midstream right‐bank urban area. New high‐risk zones also emerged in the southern part of the left riverbank under S2. Although the DV‐based method produced a wider spatial extent of flood risk, it underestimated risk levels in both scenarios. The results provide a practical basis for flood risk assessment and emergency management.
Zhang et al. (Wed,) studied this question.