Purpose Suspended particulate matter (SPM) transport through rivers is a major vector of nutrients and pollutants to continental shelf areas. To develop efficient sediment management strategies, there is a need to obtain quantitative information on SPM sources. For many years, the geochemical properties of SPM have been commonly used as tracers to identify sediment sources. In large watersheds, with numerous sources, the expected alteration of tracers during their transport requires that their reactivity be taken into account.Materials and methods To overcome this issue, we tested the use of major and trace element signatures in the residual fraction of SPM, using two different extraction methods. This original fingerprinting approach was applied to the Upper Rhone River basin (similar to 20,000 km(2)) in order to assess the respective SPM contributions of its main five tributaries (Arve, Ain, Fier, Guiers, and Bourbre Rivers) for contrasted hydrological conditions (base flow periods, flood events and dam flushing). By incorporating element concentrations previously corrected from particle size distribution in a mixing model coupled to Monte Carlo simulations, we estimated the associated uncertainties of the SPM contributions from each tributary. The relative SPM contributions obtained using this fingerprinting approach were compared with those calculated with a 1-D hydro-sedimentary model.Results and discussion The use of element concentrations, such as Zn, P, Cu, Pb, Mn, or Sr, in the total fraction of SPM as conservative fingerprinting properties was not suitable, since they are mainly bound (> 50%) to reactive carrier phases. By using Ba, Ni, Fe, Mg, Cu, Sr, and V concentrations in the SPM residual fraction, the apportionment modeling of SPM sources was successfully assessed. The fingerprinting approach showed that, in base flow conditions, SPM originated mainly from the Arve River. During dam flushing event, the fingerprinting approach consistently estimated that re-suspended sediments came from the Arve River, while the 1-D hydro-sedimentary model estimated a proportion of re-suspended sediment originating within the Rhone River.Conclusions This original fingerprinting approach highlighted the relevance of using geochemical properties in the non-reactive fraction of SPM in order to obtain reliable information on spatial sources of SPM in a large river basin. This methodology opens up promising perspectives to better track SPM sources in highly reactive environments such as estuaries/delta or in historical sediment archives.