Investigating the failure mechanisms of cascade landslide dams under overtopping conditions: an experimental approach

Under strong earthquakes and extreme rainfall conditions, landslide clusters near rivers may block incoming water flows, forming cascade landslide dams. When subjected to extreme hydrological conditions, these cascade landslide dams can experience overtopping, which erodes and entrains material from the dam surface, compromising dam stability, which can potentially lead to outburst flooding. Although existing research provides valuable insights into landslide dam failures, it falls short in addressing the complexities of cascading failures in dam groups. The influence of various factors on this intricate process is still only partially understood, highlighting the need for a comprehensive understanding of dam surface evolution and flood flow dynamics specifically within the context of cascading failures in landslide dams. In this study, we conduct physical model experiments to categorize the stages of cascade landslide dam failure, identify typical failure modes, and investigate the mechanisms affecting breach morphology evolution and flood flow development during cascading failures. Our findings reveal that cascade landslide dams can amplify floods during cascading breaches, with peak flow rates increasing progressively from upstream to downstream. In cascade systems, downstream dams experience accelerated breach processes, where layer erosion becomes predominant. Significant fluctuations in river water levels between cascades can lead to slope failures in both upstream and downstream dams. Specifically, downstream dams exhibit more rapid breach widening, fewer instances of collapse at the breach slope compared to upstream dams, and earlier peak flow timing, making cascade landslide dam breaches more challenging for emergency response. These observations highlight the notable differences between cascade landslide dam failures and single dam failures. Based on these findings, we provide insights into developing mathematical models for cascade landslide dam failures to improve physical-based flood predictions, which will assist the development of early warning strategies for cascade landslide dam breach floods.