Abstract :
[en] During this study, we focused on the biogeochemical cycles of carbon and nitrogen in two tropical lakes, Lake Kivu and one of its bays, Kabuno Bay, located on the border between the Democratic Republic of the Congo and Rwanda, and a small temperate lake located in a limestone quarry in Belgium (Lake Dendre). Seasonal and inter-annual variations in methane (CH4) and nitrous oxide (N2O) fluxes to the atmosphere in Lake Kivu were calculated based on the concentrations of these two elements measured monthly for almost two consecutive years. These data show that Lake Kivu is a low CH4 emitter throughout the year (mean of 86 µmol m-2 d-1), in proportion to the high levels of CH4 present in its deep waters, and alternates between a source and a sink of N2O. The oxidation of CH4 has been proposed to explain these low emissions to the atmosphere. This study concentrated particularly on the detection of the anaerobic oxidation of CH4 (AOM) within the water column and on the identification of the potential electron acceptors: nitrate (NO3-), nitrite (NO2-), sulfate (SO42-), iron (Fe) and manganese (Mn). Significant levels of AOM, up to 16 and 75 μmol m-2 d-1, were found in the water column of Lake Kivu and Kabuno Bay, respectively, but the identification of the potential electron acceptors is not so obvious. At Kabuno Bay, which is considered as a ferruginous basin, Fe seems to be the most likely electron acceptor, since NO3-, NO2- and Mn concentrations are very low, and the sulfur cycle seems to be not really developed. Despite the high concentrations of SO42- measured in oxic waters (up to 600 μmol L-1), concentrations of sulfide (HS-) remained very low in anoxic waters (<1 μmol L- 1), suggesting a poor occurrence of SO42- reduction. In Lake Kivu, the main electron acceptor is most likely SO42- in view of the high concentrations recorded (up to 225 μmol L-1) compared to the concentrations of the other elements (NO3-: <10 μmol L-1, NO2- <1.5 μmol L-1, Mn and Fe total <15 μmol L-1), and high HS- concentrations (up to 120 µmol L-1 at 80 m depth) suggesting the occurrence of significant SO42- reduction. However, some vertical profiles observed in the rainy season, which showed that AOM levels were higher in the NO3- accumulation zone, suggest that NO3- could be an electron acceptor for AOM, but at a low extent. NO3- concentrations were mostly too low to explain the AOM rates observed, and competition with other processes is most likely too high. Indeed, during this study, we also highlighted the occurrence of denitrification, the dissimilatory reduction of NO3- to ammonium (NH4+) (DNRA) and the anaerobic oxidation of NH4+ (Anammox), which compete for substrates. Finally, since Lake Dendre shares some characteristics with Lake Kivu (ie mainly they are both meromictic and have high CH4 concentrations in their anoxic waters), we also measured CH4 oxidation within the water column and put in evidence AOM rates up to 14 μmol L-1 d-1. Despite these high oxidation rates, Lake Dendre was a large emitter of CH4 for the atmosphere. SO42- was likely the primary electron acceptor, but high concentrations of NO3- (up to 80 μmol L-1) suggest that they could also be used for AOM, since AOM coupled to denitrification is thermodynamically much more favorable than the AOM coupled to the reduction of SO42-.
