Qualitative analysis of rock avalanches propagation by means of physical modelling of non-constrained gravel flows

This paper presents an experimental study of rock avalanches run-out and propagation carried out with a small-scale physical model at the EPFL Rock Mechanics Laboratory. Fall height, volume, releasing geometry and materials used for testing have been varied and their influence on deposit characteristics (such as length, width, height and morphology) and run-out has been studied. The experiments have shown that deposit morphology is dependent on the type of material used: sand or gravel. Sand deposit shape, regular and compact, is in accordance with many tests described in the literature, while Aquarium gravel deposit is closer to real cases such as the Six des Eaux Froides event (Switzerland). There is also a considerable difference in deposit morphology and dimensions when the event is the consequence of one large volume released at once or when the same volume is released at different times. In this latter case the final deposit characteristics do not depend on the entire failed volume but on the individual smaller ones. This behaviour is in agreement with the Randa event (Swiss Alps) which occurred over several hours. For all-in-once releases, run-out distance depends primarily on the volume. Fall height and releasing geometry have a small influence on run-out, but on the other hand a higher fall height leads to a different lateral spreading and to the development of two different propagation mechanisms of the sliding mass.