Comparison of 2D and 3D assessment of slope stability using the Local Factor of Safety concept

Landslides occur as a result of failure in mechanical balance within slopes and any factor that influences the stressdistribution can change the slope stability. It is well-known that slope stability is influenced by subsurface hydrologicalprocesses, since changes in water content of the soil as a resultof infiltration impact the matric suction and weight of the slope material and therefore the effective stress distribution.Consequently, slope morphology, soil depth, and soil hydraulic and mechanical properties are among the keyfactors determining slope stability. Most of the available models that couple hydromechanical processes use 1Dor 2D representations of the subsurface, which may lead to an overly simplified representation of hydromechanicalprocesses in the case of more complex subsurface layering. Therefore, the aim of this study was to develop athree-dimensional coupled hydromechanical model that is able to predict the spatial distribution of water contentand stress at each point within the 3D hillslope and to compare slope stability predictions obtained from 2D and3D models. For coupling the hydrological and mechanical processes, we relied on COMSOL Multiphysics. TheRichards equation was used to describe transient unsaturated water flow, and the Local-Factor-of-Safety (LFS)concept was used to assess slope stability assuming linear elastic behavior until the point of slope failure (i.e.plastic deformation was not considered). In a first step, the 2D and 3D model implementations were successfullyverified with an available 2D benchmark model that provides the LFS for a simple slope geometry. In a next step,slope stability predictions for 2D and 3D slopes with two layers of variable thickness and hydraulic parameters arecompared in terms of water content and stress distribution and the resulting LFS distribution. The results show thatsimplified 2D simulations of hydromechanical processes lead to under- or overestimation of slope stability