Interactive statistical analysis of predisposing factors for earthquake-triggered landslides: a case study of the 2013 Lushan, China Ms7.0 earthquake

Remote sensing data and field investigations indicate that at least 3,878 landslides were induced by the 2013 Lushan, China Ms7.0 earthquake. This study selects a square area of 100 km × 100 km around the epicenter to make an interactive statistical analysis to these landslides. Percentages of landslide number and landslide concentration are calculated based on a GIS platform, so as to find out the relationship between the distribution of landslides and six predisposing factors including elevation, slope angle, slope aspect, slope curvature, geologic ages of underlying bedrocks, and peak ground acceleration (PGA). In general, the correlations from interactive statistical analysis show a good consistency with the results from the bivariate statistics. However, several abnormal values are observed. For example, the relationship between slope aspect and co-seismic landslides is more easily influenced in the areas of slope angles larger than 40º or with the PGA value range of 0.36–0.44 g. The correlation between slope angle and co-seismic landslides is stronger in areas with underlying bedrocks of the Triassic age. Interactive statistical analysis of co-seismic landslides controlled by the factors in this study can effectively compensate the results from bivariate statistical analysis, which allows us to better understand spatial distribution patterns of earthquake-triggered landslides.