The Amazon. Bio-geochemistry applied to river basin management

A hydrochemical model, using hydrograph separation, developed for the Niger basin, has been proposed as a strategic tool for studying the watershed dynamics at any time and space scales. The model is applied to the Amazon basin, including the main channel and its major tributaries. The database corresponds to a sampling and analytical program developed over 8 cruises at 9 stations (about 70 samples), collected in the framework of the CAMREX Project (1982-1984). The model, based on a hydrograph separation of 3 reservoirs, is successful in extrapolating and predicting the geochemical and environmental behaviour of such large basins, naturally submitted to large secular or annual, regular or even catastrophic climatic oscillations. Several topics have been considered. (1) Coherence among the physico-chemical analyses: dissolved species (pH, NH4+, Na+, K+, Ca2+, Mg2+, NO3-, HCO3-, Cl−, DOC−, SO42-, HPO42-, SiO2, O2 and CO2), and inorganic or organic suspended load (fine and coarse fractions FSS, CSS, POCF, POCC). (2) Hydrograph separation in 3 reservoir contributions: RS, the superficial or rapid runoff, RI, the hypodermic or intermediate runoff, including the flood plain contributions, and RB the ground water or base flow. (3) Estimation of the isotopic and physico-chemical features of each of the 3 flow components: RS, RI, and RB. (4) Determination of the 3 hydrological parameters (size of the reservoir, drying up coefficient, and residence time of water), characterizing each of the 3 flow components (RS, RI, and RB), in each of the 9 basins considered. (5) Hydrological and geochemical balances for all the parameters analysed either (a) cruise by cruise for all tributaries and the Amazon River at Obidos, or (b) among each of the 3 river flow components. (6) Isotopic data set of δ18O in waters, tests of coherence of the hydrograph separation model. (7) Relationships between isotopic signatures and morphological or hydroclimatical parameters characterizing the river-soil-vegetation systems. The developed procedure presents a new tool in environmental predictions, emphasizing the potentiality of geochemical interpretation of complex hydrochemical data sets.