Large-deformation analyses of seismic landslide runout considering spatially random soils and stochastic ground motions

Abstract

Landslides, among the most common natural catastrophes, pose significant risks to life and property. Uncertainties in soil strength and ground motions are widely reported to notably affect the landslide runout process. Existing research has predominantly focused on the influence of either non-uniform soil strength or stochastic ground motions, often limited to two-dimensional (2D) analyses. This research thus introduced a three-dimensional (3D) finite-element computational framework to explore the combined effect of these two factors on landslide runout distance, via coupling the coupled Eulerian–Lagrangian (CEL) large-deformation technique, random field approach and stochastic vibration theory. The results demonstrate that stochastic ground motions contribute to significant randomness in the runout distance, and soil heterogeneity further amplifies both mean value and variation of runout distance. This underscores the importance of considering the combined effect of random soil strength and stochastic ground motions on landslide runout. Furthermore, our comparison between 2D and 3D random analyses suggests that 2D random analysis tends to yield conservative results for a non-uniform soil slope, emphasizing the advantage of the established 3D large-deformation modelling of landslides. This research provides some valuable insights into the risk assessment of landslides, considering both non-uniform soil strength and stochastic ground motions.