Hillslope morphology drives variability of detrital 10Be erosion rates in steep landscapes

Abstract:

The measurement of in situ produced ¹⁰Be in quartz-bearing river sediment emerged as the primary tool for characterizing patterns in erosion rates across landscapes. It captures temporal and spatial scales relevant for studying a wide range of Earth surface processes. Detrital ¹⁰Be measurements are fundamental to applications such as studying the connection between tectonics and topography, interpreting climate controls on surface processes, and calibrating hillslope erosion models. Yet, the degree to which ¹⁰Be faithfully represents erosion rates is debated, particularly in steep landscapes subject to landsliding relevant to tectonic geomorphology and hazards. Here we compile new (n = 44) and published (n = 73) detrital ¹⁰Be measurements from watersheds in the San Gabriel Mountains, CA, ranging from 0.02–200 km². We analyzed the relationship between ¹⁰Be erosion rates and lidar-derived topographic metrics within and across regions with variable degrees of deep-seated landslide influence. Hillslope morphology emerges as a robust indicator of ¹⁰Be erosion rate across catchment mean slopes ranging from 15–47°, despite increasing bare bedrock exposure and deep-seated landslide impact above mean slopes of 30°. Although more nonlinear than the relationship between channel steepness and erosion rate, there is no evidence for a threshold in catchment-mean hillslope angle. Whereas the relationship between channel steepness and erosion rate breaks down in catchments smaller than 3 km² due to the increasing impact of low-concavity headwater channels, erosion rate is sensitive to mean hillslope angle for catchments as small as 0.3 km². Analysis of small (<10 km²) headwater catchments reveals that although ¹⁰Be erosion rates can vary up to 5-10x at the km scale, this often reflects topographic variability. Comparing watersheds with similar topography, the presence of more deep-seated landslide deposits weakens the correlation between hillslope morphology and ¹⁰Be erosion rate but does not result in systematically higher or lower erosion rates. In contrast to models that recommend targeting large watersheds to dampen the effect of landslides on detrital ¹⁰Be concentrations, our results indicate that sampling small headwater catchments may be more appropriate for understanding hillslope erosion processes.