A method of practical interest is presented in this study for predicting shallow landslide triggering due to expected rainfall scenarios. The proposed approach is based on some closed-form expressions derived using physically-based models and requires a limited number of material parameters obtained from conventional tests. Two typical triggering mechanisms are considered: in the first case, a landslide is caused by a reduction in suction due to rain infiltration in unsaturated soils; in the second one, a slope failure occurs due to the formation of a water table with consequent generation of positive pore water pressures within the slope. A simple criterion is also provided to establish which triggering mechanism occurs. For each considered mechanism, a rainfall threshold curve is derived, which is formally similar to several empirical intensity-duration relationships proposed in the literature for predicting the occurrence of shallow landslides. However, unlike these empirical relationships, the present threshold curves explicitly depend on basic parameters such as slope geometry, soil properties, and pore water pressure existing at the depth of the potential slip surface before the rainfall. The proposed approach is also applied to well-documented case studies to assess its predictive capacity.

This paper focuses on the numerical simulation of the deformation processes occurring in the slopes when soils with strain-softening behaviour are involved. In these circumstances, a progressive failure may occur with the consequent motion of the unstable soil mass in the post-failure stage. This problem can be only solved using advanced numerical techniques capable also of accounting for the occurrence of large deformations. However, the solution is generally mesh dependent and may be affected by lack of convergence. In this study, the material point method is employed to simulate the occurrence of large deformations, in conjunction with a strain-softening Mohr–Coulomb constitutive model, in which the shear strength parameters are reduced as a function of the accumulated deviatoric plastic strain, and a model parameter controlling the rate of strength decrease. To evaluate this parameter and to reduce the effects of the mesh dependency on the numerical solution, a novel procedure based on the results of direct shear tests is presented. This methodology is completely analytical and requires few parameters with a clear physical meaning and of simple experimental determination. Some simulations are performed to assess the reliability of the proposed procedure.