Relevance of using the non-reactive geochemical signature in sediment core to estimate historical tributary contributions

International audienceFluvial suspended particulate matter (SPM) fluxes transport large amounts of contaminants that can affect water quality and river ecosystems. To better manage these inputs in river systems, it is essential to identify SPM and sediment sources. Many studies have applied a fingerprinting method based on using metals integrated into a numerical mixing model to estimate source contributions in a watershed. Most fingerprinting studies use contemporary SPM to trace historical inputs, whereas their metal concentrations were modified over time due to anthropogenic inputs. Moreover, total concentrations of these properties are subject to change due to diagenetic processes occurring in stored sediments. The aim of this study was to assess the relevance of using the non-reactive fraction of metals (i.e. metals and metalloids) in fingerprinting studies to estimate the historical contributions of SPM tributary inputs in a sediment core. To assess metal concentrations in the ‘conservative’ (i.e. non-reactive) fraction, SPM (samples of sources) and sediment core layers (targeted sediments) were subjected to total mineralization and soft extraction, and the non-reactive fraction was obtained by calculating the difference between the two extractions. This approach was applied on a sediment core from the Upper Rhône River (France), using geochemical signature in contemporary SPM of three major tributaries. We showed that the non-reactive fraction retains a higher number of metals in the range test for the deepest layers, which are characterized by significant anthropogenic inputs. Through apportionment modelling using Monte Carlo simulation, we demonstrated that the tributary contributions computed using the non-reactive fraction are more consistent with historical flood and water flow data and have lower uncertainties than with the total fraction. Working with the non-reactive fraction made it possible to decipher historical inputs of SPM using contemporary SPM samples. This approach enables robust identification of sub-catchment areas liable to provide large quantities of SPM. The non-reactive fraction can be used in a variety of environmental conditions and at various spatial and temporal scales to provide a robust quantification of sediment sources