Failure process and hydrologic response of a two layer physical model: Implications for rainfall-induced landslides

Abstract

Field observations and theoretical analysis have been used in the literature to assess slope instability caused by permeability variations. This investigation aims to study the influence of permeability variations on slope behaviour by experimental means. It focuses particularly on the pore water pressure generation in the vicinity of soils with different permeabilities, and the corresponding failure mode. A series of generated failures in a model with 2 soil layers was performed by means of a flume device. The soil layers were made of a medium-sized sand and a fine sand, placed in horizontal layers. A combination of photography and pore water pressure measurements was used to examine the relationship between the pore water pressure generation and failure modes. Experiments were conducted for different arrangements of soil layers (by changing the soil layer position), and infiltration direction (downward infiltration by sprinkling water on the soil, and upward infiltration from the bottom of the lower soil layer). The results revealed no clear link between the failure mode and recorded pore water pressure. Instead, the failure mode was seen to depend mostly on the relative layer position, and the recorded pore water pressures on the imposed infiltration direction. Failure was not confined to a single failure mode, but ranged instead from retrogressive slides and lateral spreads to seepage erosion. Regarding the hydrologic response, interesting results were recorded for experiments performed by downward infiltration, where perched water tables were formed. The granular nature of both soils and the absence of an impermeable barrier at the downslope end of the model seem to have favoured water seepage as the controlling failure mechanism, enhancing seepage erosion and sliding, and restricting other failure modes that require a high soil saturation such as flow failure. As the drainage conditions were not controlled, these experiments provided a unique opportunity to study the effect of water seepage as a trigger. In some experiments, failure evolution was seen to be dominated by sand washout, which started after the pore water pressure reached its maximum, suggesting that the dragging effect of seepage forces have a minor impact on triggering failure. An extra set of experiments conducted in a triaxial apparatus supported the efficacy of pore water pressure as a trigger rather than the dragging effect of seepage forces. The results obtained here provide an insight into the pre-failure mechanisms and processes of heterogeneous natural slopes. © 2005 Elsevier B.V. All rights reserved.