Liquefaction-induced lateral spreading deformations can significantly affect the seismic performance of bridge pile foundations, shallow foundation systems, and critical underground infrastructure. Current simplified approaches for predicting lateral spreading displacements largely neglect complex ground motion, material, hydraulic, and topographic factors that influence them. Newmark sliding block analyses based on back-calculated liquefied shear strengths have also been proposed for prediction of lateral spreading displacements. However, certain assumptions of the sliding block method (e.g., deformations along a discrete failure surface, rigid perfectly-plastic soil behavior, and constant shearing resistance) are inconsistent with the actual mechanics of lateral spreading. In this study, the applicability of sliding block analyses to lateral spreading displacement prediction was assessed in terms of biases in displacement predictions and uncertainty in both predicted displacements and the back-calculated shear strengths upon which they are based. A probabilistic analysis using the sliding block-based framework indicated that significant uncertainties, primarily related to characterization of the liquefied soil and record-to-record ground motion variability, resulted in extremely low precision in both predicted displacements and back-calculated shear strengths. Furthermore, a comparative analysis between sliding block, empirical, strain potential-based, and numerical methods showed that the sliding block model generally produced significantly lower displacement predictions than the other approaches. These sources of uncertainty and biases in the sliding block framework have a strong impact on the evaluation of lateral spreading in both traditional and performance-based frameworks.

Introduction

Liquefaction-induced lateral spreading has caused significant damage to bridges, embankments, wharves, pipelines and other important elements of infrastructure in many past earthquakes. Lateral spreading occurs when liquefaction is triggered in soils beneath sloping ground surfaces or under flat ground adjacent to slopes such as riverbanks, shorelines, and embankments. The ground deformations associated with lateral spreading are often irregular in amplitude and location, and can impose significant deformation demands on structures supported on, or on foundations extending through, liquefiable soil deposits. Geotechnical engineers are often called on to estimate permanent deformations caused by lateral spreading, either for the resilient design of structures it may affect, or for the design of soil improvement measures that may be used to mitigate the lateral spreading hazard. Because the mechanics of lateral spreading are so complex, lateral spreading deformations have historically been estimated by empirical methods based on correlation to case history observations. More recently, methods based on both simple and more complex dynamic analyses have been used for estimation of lateral spreading displacements. The simple methods take the form of sliding block analyses, which have been proposed, most recently [18], for use with sliding resistances tied to the residual strength of the liquefied soil.