The 1991 Randa rockslide in the Swiss Alps involved several complex mechanisms relating to geological, mechanical and hydrological processes for which no clear triggering mechanism can be ascertained. This paper investigates the concept of progressive failure and the numerical modelling of rock mass strength degradation in natural rock slopes using the Randa rockslide as a working example. Results from continuum (i.e. finite-element) modelling are presented to illustrate an interpretation suggesting that initiation of a progressive rock mass degradation process, ultimately leading to failure, was initiated following deglaciation of the valley. Discontinuum (i.e. distinct-element) modelling is then applied to investigate the underlying mechanisms contributing to the episodic nature of the rockslide. Finally, the use of hybrid methods that combine both continuum and discontinuum techniques to model fracture propagation are discussed in the context of modelling progressive slip surface development linking initiation and degradation to eventual catastrophic failure.
In many rock slope stability analyses, the failure surface is assumed to be structurally-controlled and is predefined as a continuous plane or a series of interconnected planes. The reasons for this are partly due to post-failure observations where fully persistent discontinuities are fitted to the failure surface to explain its origin in a geological context, and partly due to the constraints of the analysis technique employed, many of which require the input of fully persistent discontinuities (e.g. limit equilibrium wedge or planar analysis, distinct-element method, etc.). Such assumptions are often valid, but only in cases where the volume of the failed block is relatively small (e.g. 1000''s of m3) or where major faults and/or bedding planes are present. In massive natural rock slopes and deep engineered slopes (e.g. open pit mines), it is unlikely that such a network of fully persistent natural discontinuities forming a complete 3-D outline of the unstable mass exists. Terzaghi (1962), Jennings (1970), Einstein et al. (1983) and others suggest that the persistence of key discontinuity sets is limited and that a complex interaction between existing natural discontinuities and brittle fracture propagation through intact rock bridges is required to bring the slope to failure. Eberhardt et al. (2001) argue that such processes must be considered to explain the temporal nature of massive natural rock slope failures. For example, in small engineered slopes, the rock mass may be continuously disturbed by blasting and fully persistent discontinuities may be exposed/day lighted during excavation enabling kinematic feasibility. However, natural rock slopes do not experience such rapid changes to their kinematic and have stood relatively stable over periods of several thousand years. This is not to say that in a natural rock slope, a system of natural discontinuities may not be interconnected forming a significant portion of what will eventual1y be the failure surface, but that a component of strength degradation with time must also occur within the rock mass. As such, massive rock slope instability requires the progressive degradation of cohesive elements, for example intact rock bridges and interlocked joint asperities, to bring the slope to catastrophic failure.