THE ISOTOPIC SIGNATURE OF LANDSLIDES: MODELING THE EFFECT OF VARYING LANDSLIDING RATES ON THE IN SITU 14C AND 10BE COSMOGENIC NUCLIDE CONCENTRATIONS

In steep mountains, landslide activity regulates sediment delivery from hillslopes into the fluvial network, influencing the degree to which landscapes record tectonic and climatic forcings, and ultimately controlling the extent and persistence of soils. However, the dynamics of steep hillslopes characterized by stochastic transport events is difficult to fully capture at geologically-relevant timescales, hampering our predictions of hazard potential and regolith dynamics in tectonically active regions. We formulate a numerical model that evolves a synthetic landscape that combines steady (i.e., soil creep), and stochastic (i.e., landslides) erosional processes that occur at user-defined rates to understand how these parameters modulate soil thickness on hillslopes and predict the resulting in-situ ¹⁴C-¹⁰Be disequilibrium in detrital sands. Our modeling results indicate that: (i) ¹⁴C provides a more sensitive tool than ¹⁰Be to address long-term erosion rates in landslide-dominated catchments, provided that floodplain sediment storage is negligible, (ii) there is no systematic relationship between drainage area and ¹⁰Be-derived denudation rates, consistent with empirical studies performed over the last decades, (iii) varying landsliding rates define distinct domains in the ¹⁰Be vs ¹⁴C/¹⁰Be space, allowing us to infer millennial-scale landsliding rates from ¹⁴C and ¹⁰Be measurements in detrital sands collected from landslide-prone regions, and (iv) under a broad combination of stochastic and steady erosion rates, soils may never reach steady-state thicknesses and thus soil production rates measured at the regolith-bedrock interface are likely to produce incorrect estimates. Taken together, these results support our modeling scheme, illustrate the advantages of combining ¹⁴C-¹⁰Be to understand tectonically-active landscapes, and provide fresh insights for studies aiming to query landslide-dominated regions. More importantly, the results raise the following questions: is there a functional relationship between rates of landsliding and (steady-state) erosion? If so, what form does it take?