Hydrodynamics of Cascading Dam Failures: Flood Wave Propagation and Spillway Gate Operations in the Edenville–Sanford Sequential Breach in Michigan

Aging high-hazard dams with limited spillway capacity face increasing risk from extreme hydrologic events. The May 2020 Edenville–Sanford cascading dam failure in Michigan exemplifies these vulnerabilities. In this study, we conducted a forensic and counterfactual hydraulic analysis using a calibrated two-dimensional hydrodynamic model to reconstruct the breach sequence and evaluate how alternative spillway gate operations influenced hydraulic conditions related to breach initiation and downstream flood hazard. Our study examines four spillway gate operation scenarios: (1) all gates closed; (2) gates open to 1.5 m; (3) gates open to 2.1 m; and (4) gates fully opened in advance of flood peak. The model was calibrated against the satellite-derived flood extent and validated with high-water mark evidence. The model achieved an F1 score of 0.66, demonstrating moderate spatial agreement in inundation extents. Although some overprediction occurred in marginal floodplains, close agreement of inundation extent and water surface elevation confirmed the model’s reliability for scenario analysis. Modeling results show that the Edenville Dam breach could generate total peak flow of 4,323 m3/s, attenuated by 35%, with peak arrival delayed approximately 3.5 h before reaching Sanford Dam, located about 18 km downstream. Results inform that all gates closed and partially open scenarios produce water surface elevation exceeding the breach-initiating thresholds at both dams, whereas full gate opening could maintain safety margins and prevented breach entirely. The fully open scenario can reduced downstream flood area by 31% and maximum depths from more than 10 m to approximately 7 m compared to partial opening scenarios. Manning’s roughness variations showed less than 5% change in breach peak flows, with breach timing and flood extents remaining consistent across all scenarios. These findings demonstrate that partial gate openings were insufficient to prevent failure, while full opening was critical for cascade prevention. Our results provide actionable guidance for spillway gate operation protocols, emergency action planning, and dam safety reviews at aging embankment dams with limited spillway capacity in sequential systems.