Threshold bedrock channels in tectonically active mountains with frequent mass wasting

Models of how mountain belts grow and erode through time largely rely on the paradigm of fluvial bedrock incision as the main motor of response to differences in rock uplift, thus setting base levels of erosion in tectonically active landscapes. Dynamic feedbacks between rock uplift, bedrock river geometry, and mass wasting have been encapsulated within the concept of threshold hillslopes that attain a mechanically critical inclination capable of adjusting to fluvial incision rates via decreased stability and commensurately more frequent landsliding. Here we provide data that challenge the widely held view that channel steepness records tectonic forcing more faithfully than hillslope inclination despite much robust empirical evidence of such links between bedrock-river geometry and hillslope mass wasting. We show that the volume mobilized by mass wasting depends more on local topographic relief and the sinuosity of bedrock rivers than their mean normalized channel steepness. We derive this counterintuitive observation from an unprecedented inventory of ~300,000 landslides covering the tectonically active Japanese archipelago with substantial differences in seismicity, lithology, vertical surface deformation, topography, and precipitation variability. Both total landslide number and volumes increase nonlinearly with mean local relief even in areas where the fraction of steepest channel segments attains a constant threshold well below the maximum topographic relief. Our data document for the first time that mass wasting increases systematically with preferential steepening of flatter channel segments. Yet concomitant changes in mean channel steepness are negligible such that it remains a largely insensitive predictor of landslide denudation. Further, minute increases in bedrock-river sinuosity lead to substantial reduction in landslide abundance and volumes. Our results underline that sinuosity (together with mean local relief) is a key morphometric variable for determining the fraction of landslide-affected terrain throughout Japan. We infer that neither average nor extreme bedrock channel steepness is a sufficiently sensitive measure for gauging landslide activity in terms of volumes mobilized as opposed to bedrock river sinuosity. This new and largely underexplored channel planform metric may reflect spacing of discontinuities in rock mass that constrains the susceptibility to large rock-slope failures, and may contain valuable compound information on both fluvial and hillslope erosion dynamics in tectonically active landscapes.