Erosion des paysages montagneux par les coulées de débris: des études morphométriques au modèles d’évolution du paysage

Soutenance de thèse d’Aude Lurin

Le jury est composé de:
Simon MUDD, Rapporteur, University of Edinburgh
Wolfgang SCHWANGHART, Rapporteur, University of Potsdam
Hélène ROUX, Examinatrice, Institut de Mécanique des Fluides de Toulouse
Laure GUERIT, Examinatrice, Géosciences Rennes
Odin MARC, Directeur, GET
Sébastien CARRETIER, Co-directeur, GET
Patrick MEUNIER, Invité, Laboratoire de Géologie, ENS-Paris

Abstract :

This thesis investigates the critical role of debris-flow incision in landscape evolution, addressing a significant gap in current geomorphological models. Debris flows are stochastic flows of sediment and water, which have been shown to incise bedrock on their path. Studies have shown that in active mountainous area, a distinctive signature of debris-flow erosion could be observed in up to 80% of the upstream channel network. Debris-flow incision therefore may be pivotal in shaping mountain ranges, and controlling their reaction to tectonic and climatic changes, including their maximum altitude and lifespan. Despite their importance, debris-flow erosion has been largely overlooked in landscape evolution models, and debris-flow channels have been considered as part of hillslope in most morphometric studies. This is mostly due to the lack of high resolution topographic data and to the difficulty of integrating stochastic processes in landscape evolution models.

To provide empirical constraints on prospective erosion laws, I employ high-resolution topographic data from kilometer-scale mountain catchments with documented debris flow activity, alongside erosion rates derived from cosmogenic radionuclides. I developed a  novel automatic channel head extraction algorithm designed to work regardless of the process shaping channels and on the erosion rate, which influence the roughness of the terrain. I used this method to precisely identify first-order debris flow channels in 38 catchments in the US and France, and extract key topographic metrics, such as their initial slope and concavity. I find a significant positive correlation between slope and erosion rate across the catchments. On the contrary, the concavity of debris flow channels remains consistent across different geological and climatic settings.
These finding offer new insights into the mechanics of erosion. I compared them to a simple one-dimensional, single-channel model of debris flow erosion, which allowed me to constrain the parameters of the debris-flow incision law. Subsequently, I developed one of the first three-dimensional landscape evolution models incorporating debris flows. I built a debris-flow erosion component in the python landscape evolution modelling library Landlab, by adapting an existing model of landslide triggering and deposition to account for longer debris-flow runouts and compute incision along the debris-flow path. Although this model still need major improvement, initial results confirm that debris flows have the capacity to shape drainage networks by carving channels upstream of fluvial systems, therefore influencing overall drainage density and landscape connectivity. Debris flows also seem to limit the formation of sediment dams by redistributing landslide sediments further downstream in fluvial networks.

Overall, this thesis underscores the necessity of integrating debris-flow dynamics into landscape evolution models to enhance our comprehension of geomorphological processes and their broader climatic and tectonic implications. The new landscape evolution model could be used to revisit studies such as the controls on the altitude and lifespan of mountain ranges, the onset of asymetric mountain ranges, and the influence of stochastic extreme events on landscapes. As for the morphological and mechanical analysis, future research directions include exploring the slope changes at tributary junctions to further constrain the erosion law and the impact of debris-flow frequency on erosion, and computing the total extent of the debris-flow dominated domain in order to understand the competition between debris-flow erosion and other processes.