Laboratory-scale investigation of the 3-D geometry of landslide dams in a U-shaped valley

Landslide dams form when a landslide obstructs narrow and steep valleys in mountainous areas. Since these dams may fail over a short time and cause secondary geological disasters in the downstream area, the stability of such dams must be assessed in an accurate and timely manner. The three-dimensional (3-D) geometry of a landslide dam, which is one of the essential indices for stability analyses, is highly influenced by the landslide dam formation conditions, including the sliding surface dip-angle and the valley topography. However, few studies have focused on the correlation between the 3-D geometry and formation conditions of a landslide dam. A photogrammetric system was used to develop a laboratory apparatus to simulate the formation process of a landslide dam and construct a digital model for the 3-D geometry of a landslide dam in a U-shaped valley. Then, a series of experiments were designed to systematically investigate the formation process and the 3-D geometry of landslide dams under different combinations of valley bed inclinations (0, 4°, 8°, and 12°) and sliding surface dip-angles (30°, 45°, 60°, and 75°). The results indicate that the sliding surface dip-angle has a significant effect on the length and height of a landslide dam, whereas the valley bed inclination has a limited effect on the length and height compared with its effect on the relative angles of a landslide dam. In addition, the cross-sectional profiles of landslide dams can be classified into three categories, namely, symmetric, asymmetric and flat shapes, and these profiles are predominantly affected by the sliding surface dip-angle instead of the valley bed inclination. Subsequently, a mathematical model for predicting the 3-D geometry of a landslide dam was presented according to the experimental results, and it was verified by the experimental data and 42 historical cases of landslide dams. Based on the proposed model, the predicted 3-D geometry of the Tangjiashan landslide dam triggered by the Wenchuan earthquake of 2008 was consistent with its actual geometry, which confirmed the reliability of the prediction model in practical applications. The findings of our study contribute to the establishment of evidence-based disaster prevention and mitigation measures for landslide dams.