Collapse, fragmentation, high-speed boulders, and dust cloud: analysis of the 2017 Pousset (Cogne, Val D’Aosta) rockslide in Northern Italy

The dynamics of rock fragmentation during the collapse of a rock avalanche, a rockfall, or an extremely energetic rockfall, is insufficiently known (De Blasio et al., 2018). Fragmentation especially at the base of a rock avalanche may affect on the one hand the dynamics of the rock avalanche and the geometry of the final deposit. On the other hand, fragmentation in the upper layers produces a dust of rock particles which: i) impacts energetically with the surrounding areas, and in a later stage, ii) propagates as a dust cloud. Although such dynamics are commonly observed, they are still inadequately addressed.

Recently, a rock avalanche in the Italian Alps occurred in November 2017, giving us the possibility to investigate these phenomena in better detail. In particular, we analysed a  8,000 m³ collapse of serpentinites and metabasics (Grivola-Urtier metaophiolitic Unit) from the Pousset peak (Aosta Valley Region in Western Italian Alps). The peak collapsed from an average height of 2800 m a.s.l. to the foot of the slope 800 m below, where it completely disintegrated. The impact on the ground produced a rock dust cloud which subsequently flowed downstream over the successive few minutes.  The site was visited immediately after the event, and it was possible to investigate the fresh deposit of rock dust before alteration by climate or weathering. This collapse thus represents an interesting case study for trying to determine the energy threshold required for fragmentation and dust cloud formation, the redistribution of the kinetic energy after impact and the amount related to cloud generation within the energy balance.

After identifying in situ the main characteristics of the collapse, we then concentrated our efforts on a more quantitative understanding of the event via numerical calculations. We reproduced the blocks trajectories and computed the impact points where a strong energy dissipation occurred by using the 3D rockfall simulator code HY-STONE (Crosta & Agliardi 2004; Frattini et al. 2012). In these points, the block fragmentation has been taken place and the formation of dust occurred. Through laboratory analysis of dust samples collected from the few centimetres thick deposits on trees and paths, we determined the particle size frequency curves for each location. The fragmentation energy was then estimated by integrating the spectrum of the grains assuming that the fragmentation energy is proportional to the area just created.

Once obtained the fragmentation energy, we estimated the maximum speed and runout of the dust cloud and the settling time using a simple model for suspension flows. From the analysis of the results obtained in the three described procedures, the fragmentation energy was found to be a relatively small fraction of the initial energy of the landslide, and the calculated flow rate of the suspended powder was found to be compatible with the one observed, even though flowage parameters for the cloud still need to be understood from first principles. In conclusion this case study, even if volumetrically small (or perhaps because of it), may add interesting information on the ongoing debate about rock fragmentation in catastrophic events.