Structural analysis and analogue modeling of the kinematics and dynamics of rockslide avalanches

We present a structural analysis of subaerial natural and analogue rockslide avalanches. Such deposits often have well-developed faults, folds, and hummocks. These structures can be used to determine the kinematics and dynamics of emplacement. Large-scale terrestrial rockslide avalanches show large runout distances compared to their fall height. Most attempts to explain this phenomenon invoke fluidizing mechanisms or lubricating agents to reduce forces opposed to momentum, especially at the base. However, the properties and mechanics of low friction are still poorly understood. Any model for motion and emplacement must integrate geometric, morphologic, and structural features, all crucial in constraining kinematics and dynamics. Here we first examine the morphological and structural features displayed by 13 natural rockslide avalanche deposits; we then use simple and well-constrained analogue models involving the slide of stratified granular material down smooth, curved ramps. These differ from previous analogue models in that we concentrate on observing the structures produced by brittle deformation and use a low-friction sliding surface. Models show that variations in the sliding surface curvature, lateral profile, roughness, and modifications in material cohesion can successfully reproduce the majority of rockslide-avalanche deposit features. After discussing the geometrical and dynamic similarity between experiments and natural examples, we propose a model for structure formation and a fourfold classification based on model and natural deposit morphology and dynamics: hummocky, nonhummocky, dominantly extensional, and dominantly compressional rockslide avalanches. The models require a brittle core and surface that spreads and contracts by adjustment on large numbers of faults that bottom into a low-friction decollement layer. Spreading is accommodated by normal and strike-slip faults, while on deceleration, thrust faulting generates thickening. To be realistic, any physical predictive model must take into account these fundamental kinematic and structural aspects.