Study on the morphological characteristics of landslide dam in valley deposition

The morphological characteristics are a fundamental prerequisite for the development, evolution, mitigation, and prevention of landslide dam hazards, necessitating urgent theoretical quantification. A comprehensive experimental setup was designed to simulate the entire process of landslide-induced river blockage and dam formation. Structure from Motion of three-dimensional reverse reconstruction technology (Structure from Motion, SfM) was employed to quantitatively analyze the morphological characteristics. Based on typical cross-sectional and longitudinal profiles, a computational method for landslide dam volume was derived using 10 characteristic points and 12 parameters. Experimental results and theoretical analysis reveal that fine particles in the landslide dam are predominantly concentrated near the sliding-side accumulation zone, while coarse particles gradually increase toward the distal sliding side and the valley’s free surfaces along the sliding direction. At smaller sliding angles, the source material tends to accumulate on the near-slide side, with the lowest point of the dam situated on the distal side. When the sliding distance is short, the highest point of the landslide dam is located on the near-slide side of the valley; as the sliding distance increases, the peak elevation shifts toward the distal side of the river channel. The cross-section of a landslide dam exhibits a trapezoidal shape, while the longitudinal profile can be categorized into three types: left-high, central-peak, and right-high. The deposit morphology was generalized using 10 characteristic points. A volume calculation method was derived using the cut-and-fill technique, and a quantitative model for morphological characteristics was established by incorporating the static angle of repose, sliding distance, and sliding angle. Experimental data demonstrate that the computational model yields a relative error range of -3.22% to 14.96% for forward volume calculations, while the inverse predictions of height, base width, and length all exhibit relative errors below 15%. These results confirm that the proposed quantitative model effectively captures the influence of varying static angles of repose, sliding angles, and sliding distances on the morphology of landslide dams. This advancement provides a robust basis for assessing and predicting the disaster-scale potential of landslide dam.