Résumé
Riverborne iron is a notable source for this biogeochemically key element to the oceans. Recent investigations have shown that its isotopic composition may vary significantly in oceanic waters. Hence, a proper understanding of the Fe cycle at the surface of the Earth requires a good characterization of the isotopic composition of its various reservoirs. However, as the database growths, it appears that the isotope composition of the riverborne Fe delivered to the oceans may be more varied than initially thought, in agreement with inferences from soil studies from different climatic contexts. It is therefore important to compare major rivers from different latitudes. We focused our attention on the Amazon River and its tributaries that represent ca. 20% of the freshwater delivered to the oceans by world rivers. Preliminary experiments suggest that water filtration may induce biases in stable Fe isotope composition. Therefore, we worked first on bulk waters, sampled during multidisciplinary field campaigns on the Amazon River and its tributaries, including the Solimoes, Negro, Madeira and Tapajos Rivers. Besides a complete sample physical-chemical characterization, Fe isotope determinations were conduced after water sample mineralization, iron purification and MC-ICP-MS analysis. Our first results reveal that most bulk water samples cluster close to the continental crust value (0.1ppm delta 57FeIRMM-14) with an overall range of 0.2ppm. This is consistent with the restricted range found in lateritic soils elsewhere that represent 80% of the Amazon basin surface. Only black water rivers flowing over the podzols of the northern portion of the Amazon basin tend to show lighter isotopic compositions, down to -0.18ppm. However, sediment analyses suggest that this light Fe isotopic is lost through sedimentation on the river bed, thereby leading the waters to have Fe isotope compositions remaining close to that of the continental crust. This constant isotopic signature holds whatever the relative proportion of dissolved Fe in the bulk waters budget, that ranges from 5 to 50% in these waters, whatever the sample depth and whenever the samples were taken in the river cycle. Hence, given that several studies have shown that Fe loss through flocculation in estuaries does not affect Fe isotope signatures, we conclude that the bulk waters from the Amazon River delivered to the ocean should have an isotopic composition close to that of the continental crust.