Stability analysis of submarine slopes with asymmetric shear band propagation

Submarine landslides often occur on gentle cohesive slopes with inclinations of less than 5°, which cannot be adequately reproduced using traditional limit equilibrium methods. The shear band propagation method, considering the strain-softening behavior of cohesive soils, provides a rational explanation for such failures. Previous shear band propagation methods mostly focused on planar or symmetrical S-shaped slopes, where the symmetric propagation of shear bands is simplified as a unidirectional problem. However, most natural slopes are characterized by asymmetric profiles. A theoretical method for asymmetric shear band propagation is proposed, where the evolution of shear stress and related variables is solved based on deformation compatibility along the entire slip surface to assess slope stability. Using the asymmetric Gaussian slope as an example, the progressive failure process triggered by local disturbances is investigated. The results show that when the shear band length exceeds a critical value, catastrophic failure occurs. As the curvature variation of the slope decreases, the critical gravity shear stress at the steepest point of the potential slip surface approaches the peak undrained shear strength of soils. Based on these findings, a simplified failure criterion is proposed and its general applicability and accuracy are verified through finite element analysis. A case study of the St. Niklausen landslide in Switzerland is conducted to evaluate the prediction performance/predictive capability of different methods. Limit equilibrium methods and conventional symmetric shear band propagation methods that approximate the slope as a symmetric shape overestimate slope stability. The proposed asymmetric analysis method can reasonably reproduce the occurrence of the St. Niklausen slide, with predicted safety factors and shear band locations closely matching finite element results. The developed simplified failure criterion provides quick stability evaluations with a conservative bias.