Quantification and uncertainty assessment of long-term erosion in the Campine area, NE Belgium, based on 10Be data

Beerten, Koen Vanacker, Veerle [UCL] Laloy, Eric Rogiers, Bart

The Campine Basin, NE Belgium, is situated between the uplifting Ardennes Massif and rapidly subsiding Roer Valley Graben. It contains a thick series of marine, estuarine and continental Neogene and Quaternary sediments, locally more than 300 m. As a result of relief inversion during the Quaternary, the Campine Plateau is nowadays a distinct morphological feature in this basin. Its surface elevation dips from 100 m in the south to 30 m in the north over a distance of about 60 km, which is the result of differential uplift. The Campine Plateau is covered by Early and Middle Pleistocene erosion-resistant fluvial sediments from the Rhine and Meuse and can thus be regarded as a fluvial terrace. The timing of terrace abandonment and amount of posterior erosion have not yet been resolved by direct numerical dating. In this study, we apply the cosmogenic radionuclide (CRN) profiling technique that, in ideal circumstances, allows one to constrain the exposure age and post-depositional erosion rate of the landform. Samples were taken from a 3.5 m deep cross-section in coarse river sands that were deposited by the river Rhine, and now situated at an altitude of about 50 m (a.s.l.). Nine of them were prepared for CRN measurements according to state-of-the-art techniques. The in-situ 10Be concentration of the samples was determined using accelerator mass spectrometry (ETH, Zurich). The 10Be concentrations are 1.5x105 atoms/g for the uppermost sample (at 0.3 m depth) and 0.9x105 at/g for the lowermost sample (at 3.1 m depth), yielding an estimated 0.6x105 at/g of radionuclide accumulation following sediment deposition. Next, we solved for the exposure duration and erosion rate that best fit the measured in-situ 10Be depth profile data, nuclide inheritance and their associated analytical uncertainties. When taking previous geological age constraints of the sediments (between 0.6 and 1 Ma) into account, the model optimum occurs with a terrace erosion rate of 10 to 25 m/Ma. The uncertainty on the model fit calls for a dense sampling scheme to capture the full spectrum of internal variability in CRN concentrations, likely related to the highly polycyclic nature of the fluvial sedimentary environment. Probabilistic inverse modelling of the 10Be concentration was subsequently performed in order to derive the posterior model parameter (exposure age, erosion rate and actual burial depth) distribution. This multivariate distribution encodes the parameter uncertainty given the available measurement data. Sampling from this distribution then allowed us to quantify the uncertainty associated with the simulated amount of erosion for the northern part of the Campine Plateau.