Experimental study of viscoplastic surges down complex topographies and comparisons with numerical simulations.

In a context of climate change, sediment availability and occurrence of heavy rainfall episodes are tending to increase, resulting in a likely rise in frequency and magnitude of debris flows. For effective hazard management, predicting the velocity, depth, and run-out of these flows in realistic settings, while considering all relevant processes, is crucial. In particular, as debris flows may exhibit viscoplastic characteristics, a better understanding of the interplay between inertia, rheology and topographical features is necessary. We report on well-controlled laboratory experiments of viscoplastic surges flowing down a model topography. A volume of viscoplastic fluid (Carbopol) is released onto four 3D-printed topographies featuring different patterns of mounds and ridge. Using Moiré projection, we monitor flow depth with a temporal resolution of 250 Hz, exploring a wide range of configurations involving varying volumes and fluid rheological properties. This set-up enables us to investigate the front velocity, the deviation and accumulation of fluid induced by the obstacles and the shape of the deposit. Two distinct flow regimes are observed. Initially, a rapid regime develops with a high front velocity, and the fluid spreads in both transversal and longitudinal directions. This regime is driven mainly by inertial forces. Subsequently, the front velocity drops drastically and the fluid flows mainly along the slope. In this second regime, controlled by rheological effects, the flow is strongly influenced by the topography with various mechanisms depending on the case (deceleration, accumulation, channelization, etc.). These experimental results are then systematically compared with depth-averaged numerical simulations based on shallow-water hypothesis. We compare the outcomes of two models implementing different representations of the complex rheology. The experimental data serves as a benchmark to assess the predicting capabilities of the models and evaluate the resulting uncertainties. These findings offer new insights into the physical processes driving debris flows, and they will ultimately contribute to the enhancement of simulation tools used in hazard management.