Evolution of stony debris flows in laboratory experiments

Abstract

The run out and destructive potential of gravitational multi-phase flows is largely determined by the mixture composition, the material properties of the solid particles and the fluid. One instrument to expand the understanding of the governing processes of flow is laboratory experiments. In this study, we concentrate experimentally on landslide-induced stony debris flows as a particular type of flow-like mass movement. We aim to observe different natural flow types for varying initial and boundary conditions. In a laboratory flume, 12.0 m long, 1.3 m wide and 0.3 m deep, we initiate stony debris flows and measured flow variables such as flow depth, mass, bulk density, front velocity and front shape for varying particle size, solid volume fraction and basal roughness. Our experimental results reveal that flow type and evolution changes significantly for different solid volume fractions, as well as for different basal roughness. The particle size had a noticeable effect on flow velocity and front shape. The smooth surface facilitated rapid, shallow, and turbulent flows. In contrast, experiments with rough beds showed relatively lower velocities and dense flow behaviour. Although the flow parameters covered only a small spectrum of the naturally possible parameter space, flow phenomena such as phase-separation, longitudinal sorting, steep front surges, or front overtopping were observed. We group our observed flows regarding the flow properties and classify them into common flow types: (1) debris flood (hyperconcentrated flow), (2) debris flow, and (3) non-liquefied debris flow. To compare our results with natural events and other experimental results, we analyse the data with several dimensionless numbers. The flows were generally dominated by grain collision on the smooth surface. Naturally, frictional forces gain more influence on the rough surface but did not overrule collisional forces. Viscous forces played only a minor role in our experiments, due to the lack of highly viscous fluid. Overall, we infer, that our well-controlled experiments mimicked natural stony debris flow and give new profound insights into the causal relationship of how the initial and boundary conditions affect the flow evolution.