Application of a Probabilistic Assessment of the Permanent Seismic Displacement of a Slope

Abstract

Predicting the seismic performance of slopes involves an assessment of the expected permanent sliding displacement induced by ground shaking. Often, this analysis uses a deterministic approach that predicts sliding displacements using one level of ground shaking and the best-estimate slope properties (e.g., soil shear strengths). However, this approach does not consider the variability inherent in the prediction of sliding displacement, nor the uncertainties regarding slope properties. A probabilistic framework computes a displacement hazard curve using: (1) the entire ground motion hazard curve from a probabilistic seismic hazard analysis, (2) a model for predicting the dynamic response of the sliding mass, and (3) a model for predicting the sliding displacement of the sliding mass. A logic tree can also be included, which incorporates the uncertainties in the input parameters for the displacement calculation. This probabilistic framework is demonstrated through application to a slope at a site in California. The results of this analysis show that the probabilistic approach predicts displacements larger than the deterministic approach, which is a direct result of accounting for the uncertain inputs into the analysis. Incorporating these uncertainties and variability into an analysis via the probabilistic approach better captures the seismic risk associated with seismic slope movements and provides a mechanism to account for a reduction in seismic risk when uncertainties are reduced through additional data collection.