Detailed numerical simulation of landslide dam failure considering non‐uniform and multi‐layered sediments: A case study of the second Baige landslide dam in the Jinsha River, China

Abstract Prediction of the dam failure process and downstream flooding is crucial for hazard management and understanding the local morphological changes associated with landslide dam events. Indeed, numerous models exist to simulate landslide dam failures, but the characteristics of these dams, especially their very specific material composition, are often oversimplified or neglected. This paper presents a 2D numerical model that accounts for non‐uniform and multi‐layered sediments. It is applied to the failure of the second Baige landslide dam and the results are validated through comparisons with field observations. The effects of the upstream inflow rate and dam material composition on the failure process are investigated. Additionally, the numerical results are compared to those obtained from five other detailed, less detailed or empirical models. The results show that the proposed model accurately reproduces the dam failure process and captures the evolution of surface grain size distribution during the event. Increased upstream inflow rates significantly increase the peak discharge but have a limited effect on the final breach while lower inflow rates significantly extend the time to failure, allowing for emergency interventions and evacuation. Due to the high inflow discharge in the case of the Baige event and relatively fine dam material, all sediment sizes are mobilized, resulting in negligible differences in failure evolution between uniform and non‐uniform sediment scenarios. Furthermore, unlike empirical and simplified numerical models, the proposed model does not rely on limited case studies or predefined breach evolution assumptions, providing a more reliable simulation framework. It therefore offers a practical and flexible framework for hazard assessment and mitigation planning in mountainous regions worldwide.