Numerical Simulation of Landslide Dam Overtopping Failure Considering the Headcut Erosion Process

Rapid and accurate prediction of landslide dam failures is extremely important for making accurate disaster prevention decisions. However, current models do not consider the phenomenon of headcut erosion and the effect of sediment siltation on the downstream slopes of the dam when simulating the longitudinal evolutionary processes of landslide dams. Based on the theories of hydraulics, hydrology, and sediment dynamics, this study aims to establish a two-stage numerical model that considers the headcut erosion process and the bed slope changing process of diversion channels to predict the landslide dam failure process and related hydrological parameters more accurately. To investigate the changing process of the diversion channel slope, we propose an empirical model to predict the location of the rotation point at the downstream and final slopes of the diversion channel based on the collected experimental data and information. The phenomenon of headcut erosion is considered in the model. To verify the effectiveness of the proposed dam-break model, we considered the 11.03 Baige landslide dam located at the border of the Tibet Autonomous Region and Sichuan Province in China as an example. The results showed that the relative errors of the simulated peak discharge, time to peak, and total flood volume were 4.67%, 4.98%, and 0.63%, respectively, satisfying the accuracy requirements of hydrological forecasting. This study demonstrates that neglecting headcut erosion overlooks the sudden change in the breach and leads to an overestimation of the flood peak discharge; the longitudinal evolution process calculated by the numerical model considering the downstream dam slope rotation point and the final slope of the diversion channel is more in line with the actual process.