Evaluating snow weak-layer failure parameters through inverse finite element modelling of shaking-platform experiments

Snowpack weak layers may fail due to excess stresses of various natures, caused by snowfall, skiers, explosions or strong ground motion due to earthquakes, and lead to snow avalanches. This research presents a numerical model describing the failure of "sandwich" snow samples subjected to shaking. The finite element model treats weak layers as interfaces with variable mechanical parameters. This approach is validated by reproducing cyclic loading snow fracture experiments. The model evaluation revealed that the Mohr–Coulomb failure criterion, governed by cohesion and friction angle, was adequate to describe the experiments. The model showed the complex, non-homogeneous stress evolution within the snow samples and especially the importance of tension on fracture initiation at the edges of the weak layer, caused by dynamic stresses due to shaking. Accordingly, a simplified analytical solution, ignoring the inhomogeneity of tangential and normal stresses along the failure plane, may incorrectly estimate the shear strength of the weak layers. The values for "best fit" cohesion and friction angle were ≈1.6 kPa and 22.5–60°. These may constitute valuable first approximations in mechanical models used for avalanche forecasting.