isotope fractionation in the Amazon River basin controlled by the weathering regimes, Geochimica et Cosmochimica Acta (2015), doi: http://dx. AbstractWe report Li isotope composition (δ 7 Li) of river-borne dissolved and solid material in the largest River system on Earth, the Amazon River basin, to characterize Li isotope fractionation at a continental scale. The δ 7 Li in the dissolved load (+1.2 to +32 ) is fractionated toward heavy values compared to the inferred bedrock (-1 to 5 ) and the suspended sediments (-6.8 to -0.5 ) as a result of the preferential incorporation of 6 Li into secondary minerals during weathering. Despite having very contrasted weathering and erosion regimes, both Andean headwaters and lowland rivers share similar ranges of dissolved δ 7 Li (+1.2 to +18 ). Correlations between dissolved δ 7 Li and Li/Na and Li/Mg ratios suggest that the proportion of Li incorporated in secondary minerals during weathering act as the main control on the δ 7 Li diss across the entire Amazon basin. A "batch" steady-state fractionation model for Andean and lowland rivers satisfactorily reproduces these variations, with a fractionation factor between weathering products and dissolved load (α sec−dis ) of 0.983. Two types of supply-limited weathering regimes can be identified for the lowlands : "clearwaters" with dominant incorporation of Li in secondary minerals, and "black waters" (e.g. Rio Negro) where dissolution of secondary minerals enhanced by organic matter produces low δ 7 Li. Apart from the black waters, the δ 7 Li of Andean and lowland rivers is negatively correlated to the denudation rates with the lowest δ 7 Li corresponding to the rivers having the highest denudation rates. In contrast, the main tributaries draining both the Andes and the lowlands have higher δ 7 Li compared to other rivers. We propose that part of the dissolved Li derived from weathering in the Andes is re-incorporated in sediments during transfer of water and sediments in floodplains and that this results in an increase of the dissolved δ 7 Li along the course of these rivers. Unlike other rivers, the dissolved δ 7 Li in the main tributaries is best described by a Rayleigh fractionation model with a fractionation factor α sec−dis of 0.991. Altogether, the control imposed by residence time in the weathering zone and floodplain processes results in (i) a non-linear correlation between dissolved δ 7 Li and the weathering intensity (defined as W/D) and (ii) a positive relationship between the dissolved Li flux and the denudation rate. These results have important implications for the understanding of past ocean δ 7 Li and its use as a paleo weathering proxy.

International audienceThe chemical weathering of rocks with sulfuric acid is usually not considered in reconstructions of the past evolution of the carbon cycle, although this reaction delivers cations and alkalinity to the ocean without involvement of atmospheric CO2. The contribution of sulfuric acid as a weathering agent is still poorly quantified; the identification of riverine sulfate sources is difficult. The use of δ34S and δ18O of dissolved sulfate allows us to demonstrate that most of the sulfate in surface waters of the Mackenzie River system, Canada, derives from pyrite oxidation (85% ± 5%) and not from sedimentary sulfate. The calculated flux of pyrite-derived sulfate is 0.13 × 1012 mol/yr, corresponding to 20%–27% of the estimated global budget. This result suggests that the modern global ocean delivery of sulfide-derived sulfate, and thus chemical weathering with sulfuric acid, may be significantly underestimated. A strong correlation between sulfide oxidation rates and mechanical erosion rates suggests that the exposure of fresh mineral surfaces is the rate-limiting factor of sulfide oxidation in the subbasins investigated. The chemical weathering budget of the Mackenzie River shows that more than half of the dissolved inorganic carbon discharged to the ocean is ancient sedimentary carbon from carbonate (62%) and not atmospheric carbon (38%). The subsequent carbonate precipitation in the ocean will thus release more CO2 in the atmosphere-ocean system than that consumed by continental weathering, typically on glacial-interglacial time scales