Carbon, geochemistry in a large hydroelectric reservoir of the Canadian boreal forest: isotopic approach

Average emission values of 1 to 2 g of CO2.m-2.d-1 have been reported for hydroelectric reservoirs located in northern Quebec [1]. Here, using isotopic approaches, we try to document the sources and pathways of this CO2 with special attention to seasonal patterns and to the cycling of organic carbon. The study site, the Robert-Bourassa reservoir (also known as LG-2), is located in the Boreal forest area, south-east of Hudson Bay and is part of a series of 8 reservoirs. LG-2 has a mean surface area of 2835 km2 and was flooded in 1979. The isotopic monitoring of the reservoir started in 1999, but we will refer here to data collected during the summers of 2001 and 2002. Three sampling strategies were retained: i) sampling in surface waters of ∼15 sites scattered across the reservoir, ii) sampling along the water column at three sites ranging from shallow to deep (∼ 10 to 150 m), and iii) sampling of inflow and outflow waters of the reservoir, once a month during 1 year. At each sampling site, in situ measurements included: water and air temperatures, pH, alkalinity and wind speed. Samples were collected at each site for the measurement of concentrations of dissolved organic carbon (DOC), C:N ratios of dissolved organic matter (DOM) and isotopic compositions of dissolved inorganic carbon (DIC), DOC and air CO2. Samples were also collected for the measurements of 14C activities in DOM. δ13C-values of DIC vary, throughout the reservoir, from -9%0 to -14%0 (vs. VPDB), i.e., from -13 to - 19%0, for the corresponding dissolved CO2. δ13C-values in the overlying air-CO2 vary from -9 to -11%o. Both show a shift towards more depleted values under windy conditions. δ13C-values in DOC vary little in the reservoir. They average -27.1 ± 0.2%o. DOM-C:N ratios vary between 12 and 38 with a mean of 30. The 14C activity of DOM, at the deepest sampling station, varies between 107.1 ± 0.76 % (vs. 'modern carbon'), at the water-air interface, and 110.4 ± 0.98 % at a depth of 70 meters. The isotopic signature of CO2 in the reservoir indicate mixing of an isotopically heavy CO2 (from the atmosphere) with an isotopically light CO2 ensuing from the in situ oxidation of dissolved organic matter (DOM), as suggested by the very strongly correlated negative relationship between δ13C-DIC values and DOC concentrations. However, isotopic data do not permit, at this stage of the study, to weight the relative roles of photolysis and of biodegradation-respiration in the production of this CO2. CO2-production areas in the reservoir have been defined and mapped based on δ13C-DIC-values. Low δ13C-DIC areas reflect an intense production of 13C-depleted CO2 (δ ∼ -27%o) from oxidation of DOM whereas high δ13C-DIC areas reflect a relatively more efficient exchange with atmospheric CO2. Radiocarbon data suggests that the DOM present in the reservoir is essentially derived from recently produced organic matter. The high C:N ratios of this young organic matter suggests that it is of terrestrial origin, likely from top soils in the drainage basin. Results of a monthly survey at the outlet of the reservoir, show a negative shift of δ13C-DIC values matching apCO2-increase at the end of the ice cover period. This suggests that part of the CO2. produced from the 13C-depleted DOM in winter, accumulates under the ice-cover. At ice-break up, δ13C-DIC returns to less negative values and the pCO2 drops down. Maximum CO2 emissions due to outgasing through the turbines of the Hydroelectric plan are recorded during this relatively short time interval. (authors)