The need for a second nuclide: former glaciations impact the validity of 10Be-derived denudation rates in the Vosges Mountains (NE France)

<p><span>Assessing the rates at which mountain ranges erode is fundamental to disentangle interactions between the various factors controlling their denudation. Measurements of </span><span><em>in situ</em></span><span>-produced cosmogenic nuclides in modern riverborne sediments allow inferring catchment-wide denudation rate averaged over several thousand years. This well-established method allows determining the main factor(s) controlling the denudation in those mountain ranges displaying strong gradient(s) in e.g., lithology, morphometry or climate. However, the validity of this method depends on several assumptions. One of these is the </span><span><em>cosmogenic</em></span> <span><em>steady-state </em></span><span>for which it</span><span> is assumed that the nuclide production equals the nuclide removal by erosion. Paired-nuclide analysis consisting in measuring the concentrations of two different nuclides from the same sampling material allows testing the validity of this assumption. Although this was commonly undertaken in the early days of the method, the overwhelming majority of cosmogenic-derived denudation rates now settle on a single nuclide, which is </span>^{<span>10</span>}<span>Be.</span></p><p>&#160;</p><p><span>Quaternary erosion of </span><span>the Hercynian low-mountain ranges strewn across </span><span>the </span><span>European</span><span> alpine foreland</span><span> (where much of the recent work was undertaken) has received little attention so far. Among them, the </span><span>Vosges Mountains (VM) are particularly suitable for assessing denudation via cosmogenic nuclides since </span><span>they display a strong N-S gradient for potential factors controlling denudation. First, the geological basement allows a bipartite subdivision: the heterogeneous crystalline core in the southern part contrasts with the much more homogeneous sandstone cover in the northern part. Second, a clear topographic and morphometric gradient is reflected by steep and gently-sloping hillslopes in the southern and northern part, respectively. Third, a sharp N-S precipitation gradient (>1000mm/yr) is recorded, well matching the imprint left by former glaciations: whereas the southern part hosted abundant valley glaciers, the northern part was free of ice. </span></p><p>&#160;</p><p><span>This study thus aims to test the validity of the cosmogenic-based approach by sampling 22 catchments draining the whole VM for a paired-nuclide analysis (</span>^{<span>26</span>}<span>Al-</span>^{<span>10</span>}<span>Be). Lithological, morphometric and climatic characteristics were also quantified for each catchment. Our results show that almost half (10/22) of the samples violate the </span><span>steady-state assumption</span><span>. Interestingly, a vast majority of these </span><span><em>unsteady</em></span><span> catchments are located in the south of the massif which was massively and repeatedly glaciated during cold stages. The impact of former glaciations on the cosmogenic steady-state was confronted with the surface of glacial and fluvio-glacial deposits in each catchment. The negative relation suggests a complex exposure history in the formerly glaciated catchments (i.e., bedrock inheritance and/or sediment reworking). Whilst this </span><span><em>unsteadiness</em></span><span> most probably prevents the observation of a N-S gradient in denudation rates, it importantly emphasises the importance and need for using a second nuclide to infer reliable denudation rates at the massif-scale when glacial erosion is involved.</span></p><p>*Georges Auma&#238;tre, Didier L. Bourl&#232;s, Karim Keddadouche</p>