Modelling rock avalanche propagation onto glaciers

Abstract

Ice-rock avalanches which occur in glacial environments are controlled by the presence of snow and ice in the moving material and by possible propagation onto icy basal surfaces. All these factors contribute to enhancing the flow mobility. Mixing with ice and snow hampers block collisions and favours dense flow behaviour. Ice melting reduces granular friction by saturation of the basal material and fluidization effects. Propagating onto glaciers offers a smooth surface with low shear resistance.This work is a review of the best documented ice-rock avalanches and focuses on evaluating their mobility for hazard analysis purposes by providing a set of calibrated cases. The rock avalanches have volumes ranging from 5*106 m3 to 25*106 m3. We replicate these events by using SPH and FEM numerical methods, assuming frictional and Voellmy basal rheologies. The Voellmy rheology best performs at replicating the landslide propagation. Among the back analyzed cases, the frictional coefficient ranges in the interval 0.03–0.1, the turbulent coefficient within 1000 m s−2–2000 m s−2. The bulk basal friction angle ranges within 2.75° and 14° with values inversely related to event volumes. Forward selection of the basal friction angle based on event volume, allows the replication of the Mount Cook ice-rock avalanche predicting a maximum runout which is less than 4% larger than observed. In the perspective of forward modelling, large uncertainty is related to the reconstruction of the post-event topographies, particularly for the sliding surface.Mixing with ice and snow reduces basal friction proportionally to ice and snow content. Pure ice has a basal friction which is reduced by about 75% than basal friction of pure rock. Melting of ice during rock avalanche propagation has been evaluated for the Sherman event. The frictional heat generated at the glacier surface results in the melting of 86.2 ± 5.9 kg m−2, which could have contributed to a minimum 20–35% (±10%) reduction of the material friction angle through the sole pore pressure generation within a 40 and 20 cm thick shear layer, respectively. The largest uncertainty is related to the area of contact between rock and ice.