Dimensionality and directionality effects in Newmark sliding block analyses

Newmark sliding block analyses represent a useful and practical tool for evaluation of seismic slope stability hazards. The analysis requires assumptions about material and failure surface behavior that have been well documented in the literature. This paper investigates the effects of assumptions about the manner in which loading is applied on the results of the analysis. The common practice of applying a one-dimensional input motion parallel to the plane of the sliding block model was compared with the application of two-and three-dimensional input motions for sliding block models with frictional and cohesive failure surfaces; the differences were found to be small for some cases and large for others-in both conservative and unconservative directions. The effect of slope azimuth on Newmark sliding block displacements was also investigated. Computed displacements were found to be very sensitive to the assumed azimuthal direction of the input motion, particularly for cases of high yield acceleration. Procedures for estimating azimuthal variability are presented.