Influence of spatial variability of soil strength parameters on probabilistic seismic slope displacement hazard analysis

Permanent sliding displacement analysis is commonly used to evaluate the seismic performance of slopes. Specifically, the fully probabilistic seismic slope displacement hazard analysis (PSSDHA) has attracted increasing attention recently, since it appropriately accounts for various sources of uncertainties associated with the estimation of sliding displacement. Within this analysis, the uncertainty of soil strength parameters (e.g., effective cohesion c' and effective internal friction angle phi') is commonly quantified by a logic-tree method, in which the spatial variability of c' and phi' is simply ignored. This study aims at proposing an extended PSSDHA framework by rigorously accounting for the spatial variability of c' and phi', and then investigating its influence on the resultant slope displacement hazard curves based on the Newmark rigid-block model. Two different approaches, namely an analytical-based and a simulation-based method, are proposed separately. The results of the analytical-based method are found to be consistent with those obtained by the simulation-based one. It is also shown that the displacement hazard would be underestimated if the spatial variability of c' and phi' is ignored; such underestimation is more significant for slopes with soil parameters exhibiting weak c'-phi' correlation and strong spatial variability. Besides, it suggests that the proper configuration of logic-tree framework plays an important role in developing displacement hazard curves, and a relatively large range of discrete strength values is recommended to better account for the effect of the spatial variability of soil strength parameters.