3D stability analysis of submarine slopes: a probabilistic approach incorporating strain-softening behaviour

Submarine landslides exhibiting extreme geometrical and run-out characteristics have been identified and mapped along most continental margins; raising concerns about potential risks to populations should similar events occur. Hazards associated with such events have frequently been assessed using approximations, resulting in data unsuitable for mitigation strategies. Three approximations appear consequential: (i) addressing the problem in two dimensions, thereby neglecting the effect of complex morphology; (ii) employing a deterministic approach that disregards uncertainty related to the heterogeneity of sediment properties; and (iii) treating the sediment as a perfectly elastic–plastic material, simplifying the mechanical behaviour and overlooking the degradation of sediment mechanical properties (strain softening) during different phases of slope movement. Here, we introduced the strain-softening behaviour into a 3D slope stability model. Identification of the critical failure surface was conducted in terms of the probability of failure, considering the influence of sediment parameter variability and uncertainty on the likelihood of failure. The developed model was then used to assess the slope stability of a well-studied example from the literature, the Nice slope (SE France). Our findings indicate that neglecting lateral morphological changes leads to an overestimation of the probability of failure. Additionally, we demonstrated that strain-softening behaviour could significantly affect the factor of safety and the probability of failure for the studied slopes. We argue that a risk assessment and definition of a mitigation strategy require well-advanced characterisation of the mechanical behaviour of sedimentary layers and an analysis incorporating the complex morphology of submarine slopes.