Lakes are significant sources of CO2 to the atmosphere ranging between 0.14 à 0.17 PgC/yr globally. This emission of CO2 is comparable to the one from rivers of 0.34 PgC/yr and from estuaries of 0.32 PgC/yr. Africans lakes are characterized by partial pressures of CO2 (pCO2) twice higher than the global average (2300 ppm versus 1060 ppm). Also, African lakes represent about 10% of the total lake surface area (225,000 km2 versus 2426,000 km2). The emission of CO2 is attributed to the net heterotrophy of these systems sustained by the organic carbon inputs from the watershed. However, several unknowns remain on the CO2 dynamics in lakes, in particular African ones :
1 - few simultaneous and integrated studies of CO2 dynamics and metabolic performance are available,
2 - African lakes are under-sampled in relation to temperate and boreal lakes,
3 - most pCO2 estimates in lakes are based on pH and alkalinity measurements with unkown quality,
4 - seasonal and diurnal pCO2 variations in lakes are significant but not well constrained,
5 - spatial variability of pCO2 in lakes is strong but not well documented.

Situated in the western Rift Valley between Rwanda and the Democratic Republic of Congo, Lake Kivu is one of the large East African rift lakes with a maximum depth of 485 m, a volume of 550 km3 and a surface area of 2370 km2. It is situated upstream of Lake Tanganyika and has a relatively small catchment area of 7200 km2. Lake Kivu is geologically young when compared with other large lakes, especially to the older Lake Tanganyika. It is about 15,000 years old and was formed during the Pleistocene volcanic events of the Virunga Mountains which blocked the connection between the south and the basin of the River Nile to the north. Two active volcanoes are situated near the northern shore, Nyiragongo which erupted in January 1977 and January 2002 and Nyamulagira which erupts every few years.

Lake Kivu is a meromictic lake with a relatively shallow euphotic layer (~18m) usually smaller than its oxic mixolimnion (20-60 m), and with a weak thermal gradient in the mixolimnion. Lake Kivu has unique limnological characteristics, with temperature and salinity increasing in the deep water layers, due to the input of geothermal sources at the bottom of the lake. As in all Eastern Africa, the Lake Kivu region is characterized by a windy dry season (from June to September) and a calmer rainy season (October-May). In the Kivu region, however, there is a short dry season around January when the winds can be stronger than in typical rainy season conditions. During dry season, the daily range of air temperature increases, minimum temperatures occur (at night), but the mean air temperature is higher. The rest of the year, temperatures are more stable, around 19°C. Another feature of the dry season in relation to the hydrodynamic regime of Lake Kivu is the south-eastern dominant wind that reaches maximum velocity around July.

Because of its relatively high altitude (1463 m), the surface waters of Lake Kivu are 2 to 4°C colder than those of other large lakes of the region: 23.0 to 24.5 °C compared with 26.0 to 29.0°C in lakes Tanganyika, Mobutu and Edward. The hydrodynamics and vertical structure of the water column of the lake depend on meteorological and seasonal climate. During the dry season when the wind speed increases while air temperature decreases, the density gradient in the mixolimnion is reduced and this allows a deep mixing of the water column to occur. However, vertical mixing does not take place deeper than 60 m depth, where the chemocline is located. Plankton dynamics in Lake Kivu are driven essentially by these external physical factors.

The deep monimolimnion is rich in dissolved gases, particularly carbon dioxide (CO2) and methane (CH4). The release of a fraction of these gases, which could be triggered by a magma eruption within the lake, would have catastrophic consequences for the two million people living on its shore. Bacteria produce CH4 in the lake sediment both by decomposing settled organic material and using magmatic CO2 and hydrogen. In the surface layers, CH4 is oxidized using oxygen and sulfate as electron acceptors. Two transport mechanisms bring the CH4 to the surface: (1) a deep water input leads to a slow upwelling over the whole lake area which transports CH4 upward, and (2) turbulent mixing. It has previously been assumed that the gases are mainly transported to the surface by turbulent diffusion with a residence time of 400 years in the deep water.

With an annual average chlorophyll a in the mixed layer of 2.2 mg/m3 and low nutrient levels in the euphotic zone, the lake is clearly oligotrophic. Kivu has a higher algal biomass than in the larger lakes Malawi and Tanganyika, and a slightly higher primary production. Primary production varies between 0.37 and 1.50 g C/m2/d, and seasonal averages are significantly different, with higher values during the dry season (0.80 ± 0.32 gC/m2/d), than during the rainy season (0.67 ± 0.21 gC/m2/d). Annual primary production ranges between 223 and 286 in gC/m2/yr.

A significant change in zooplankton community occurred with the introduction of the sardine Limnothrissa miodon. The most obvious consequence was the disappearance of the large grazer, Daphnia curvirostris, and a decrease of the average total zooplankton biomass. The removal of this large cladoceran likely resulted in a reduction of the grazing pressure on that fraction of the phytoplankton that is edible by filter-feeding zooplankton, mostly the nanoplankton size class. This single fact may explain why present-day Chl a concentration in Lake Kivu is higher than in Lake Tanganyika despite similar surface light, daylight duration and nutrient conditions. It is worth noting that, despite a 2 or 3 fold difference in algal biomass, rates of phytoplankton production and average daily primary production lie in the same range in both Lake Kivu and Lake Tanganyika. So, the higher algal biomass in Lake Kivu is related to the fate of the algal biomass - particularly less grazing due to the disappearance of the sole large grazer, and possibly, to lower edibility of the phytoplankton - rather than to higher primary production.

The aim of this research is to determine with an integrated approach the cycles of nutrients and carbon in the surface waters of Lake Kivu. Dissolved inorganic carbon dynamics (spatial and temporal variability of pCO2 and air-water CO2 fluxes) will be investigated by the Chemical Oceanography Unit (COU-ULg). Nutrient dynamics, phytoplankton identification, primary production and vertical organic carbon export will be studied by Laboratory of Freshwater Ecology (URBO-FUNDP). Bacterial biomass and production will be studied by Laboratory of Ecologie des Systèmes Aquatiques (ESA-ULB).