Displacement hazard curves derived from slope-specific predictive models of earthquake-induced displacement

Probabilistic hazard curves for earthquake-induced permanent slope displacements incorporate the important variabilities associated with predictions of slope movements, such that a rational assessment of slope performance can be obtained. To date, displacements estimated from sliding block analyses predominantly have been used to compute displacement hazard curves, despite the fact that nonlinear finite element analysis is becoming the preferred method to evaluate the performance of slopes. This paper extends the probabilistic approach for use with displacements from finite element analyses. We develop slope-specific displacement prediction models using displacements computed by nonlinear finite element analyses. Various models are developed that utilize different ground motion intensity measures and each model is used to compute a displacement hazard curve for a site in Northern California. The computed hazard curves reveal that the different models generate a large range of epistemic uncertainty, and the model with the smallest standard deviation does not necessarily predict the smallest displacement hazard due to the influence of the shape of the ground motion hazard curve and of the displacement model. As a result, multiple displacement models that incorporate different intensity measures should be considered in any assessment of the seismic performance of slopes, in the same way that multiple ground motion models are incorporated into ground motion hazard analyses.