How debris flows shape mountain catchments? Insights from high-resolution topography.

Abstract

Landscapes are shaped by the interaction of diverse erosion processes, such as hillslope processes, fluvial erosion, debris-flow erosion. The efficiency of each of these processes depends differently on slope, on over- and underground water flow, on bedrock material properties and sediment grain size. Therefore, the competition of erosion processes structures landscapes into different domains where one process dominates over the others. These domains are characterized by a specific topographic signature, such as the slope-area power-law which characterizes fluvial domains or the convex profile of diffusive hilltops. In mountain landscapes high-resolution topographic data has only recently allowed researchers to study the topography of headwater catchments, where debris flows occur. Therefore, the topographic signature of debris-flow channels as well as their relationships with the hillslope and fluvial domain are still very partially understood. Applying the CO²CHAIN method of Lurin et al (2023) on high-resolution topography, we studied the debris-flow domain in various mountain catchments in France and the United States, where debris-flow evidence has been found. First, combining the CO²CHAIN and DrEICH methods to detect channel heads and the fluvial upstream limit, respectively, we studied the extent of the debris-flow domain and its dependence to basin characteristics. Our results suggests that the debris-flow domain extends further both upstream and downstream when erosion rate increases, which is consistent with an analytical prediction of the channelization area building upon recent modeling work (McGuire et al., 2023). This also allowed us to constrain the morphology of convergent hillslopes upstream of debris-flow channels.Then, a closer analysis of slope throughout the debris-flow channel network allowed us to study what constrains slope gradient within these channels. Overall, we found that the average gradient of the debris-flow domain increases with catchment-wide erosion rate, consistent with previous studies. Focusing on downstream gradient evolution, we found that gradient decreases sharply at channel confluences, while its dependence with drainage area in between confluence is much weaker. This suggests that the number of debris-flow sources upstream, and thus the frequency of debris flows is a key control of channel incision in the debris-flow domain, rather than sediment supply or flood discharge.These results give us insight into the processes shaping the bedrock channels and could allow us to test new models for debris-flow erosion.