Insights into the movements of landslides from combinations of field monitoring and novel direct shear testing

Abstract:

The mechanisms of pre- and post-failure movement of translational landslides remain surprisingly poorly investigated. Previous approaches have focussed on field monitoring, for example through high resolution automated surveying and/or GPS measurements, or from modelling using dedicated codes. There has been some experimental work too, most notably using ring shear devices, although there are limitations as to the type of analyses that can be completed in these devices. In recent years the author has been involved in a series of studies that have sought to understand pre- and post-failure behaviour in translational landslides using both high precision monitoring and experimental investigation using novel apparatus. The latter approach has involved the use of the back pressured shear box, a direct shear machine that allows near-infinite variation of the normal and shear stress state, and measurement and control of the pore water pressure. More recently, a more advanced version of this machine has been developed that allows dynamic loading of both direct and normal shear stresses. This paper presents key lessons learnt about the behaviour of translational landslides from these approaches. The data highlight a number of key elements:

1. The important differences in pre-failure behaviour for materials that show a brittle response compared with those that are ductile. In particular, some aspects of behaviour (e.g. the hyperbolic acceleration to failure) can only be replicated in materials that show brittle cracking processes;
2. In the post-failure domain, all materials show a high level of sensitivity to small changes in pore water pressure when the Factor of Safety is close to unity;
3. Rates of strain are not simply related to pore water pressure / stress state. In particular, some materials show a different deformation response during phases of increasing pore water pressure to that during periods of pore water pressure reduction. The reasons for this require further study;
4. Dynamic behaviour is complex, with variations in behaviour between different materials types being greater than expected.

These results show that the behaviour of materials in the post-failure domain is more complex than had been appreciated previously, suggesting that more work is needed to explain landslide behaviour in this regime.