BELCANTO (BELgian research on Carbon uptake in the ANTarctic Ocean) is a longterm project aiming at applying and developing process-level studies, geochemical proxy tools and numerical tools for assessing and understanding the present-day functioning of the CO2 biological pump in the iron-limited Southern Ocean (S.O.) and for predicting its evolution in response to scenarios of increasing atmospheric CO2. Over the last four years BELCANTO implemented and developed several multi-proxy approaches for assessing nutrient consumption and carbon fluxes, including 234Thdeficit and natural silicon isotopic composition. High quality results were obtained for whole water column δ29Si-silicate and δ29Si-opal, including first ever results for the low silicate Subantarctic region. Results show that the δ29Si signature of diatoms appears to be homogeneous in the mixed layer and between diatom species. The preferential uptake of light isotope by diatoms is well reflected in the vertical distribution of silicate δ29Si, as well as in the upper ocean silicate δ29Si values which increase northwards in parallel with the decrease of silicate concentration. However, the relationship between δ29Si and silicate concentration appears complex and depends on variations in vertical and horizontal supply of silicate. Results based on other proxy tools (Baxs 234Th-deficit, f-ratio & new production) indicate relatively high particulate carbon export and absence of strong mesopelagic mineralization in the Subantarctic Zone, but relatively low export and enhanced mesopelagic mineralization further south in the Polar Front Zone and the southern ACC. Furthermore, remineralization was clearly enhanced during summer as compared to spring. Our observations of 234Th-based carbon export and new production along cross frontal transects also appear to challenge the widely accepted idea of enhanced primary and export production in the Subantarctic Zone and the Polar Frontal Zone, compared to more southern areas. Laboratory-controlled experiments on two widespread phytoplankton species (diatoms-Thalassiosira gravida and Phaeocystis Antarctica) showed an effect of Fe addition on the morphological form, enhancing the presence of the colonial form compared to the free-living cells for Phaeocystis and increasing the appearance of long chains of diatoms vs. free living cells. Under Fe enriched conditions, the maximum photosynthesis and growth rates increased. In addition the quality of the organic matter was modified, enhancing bacterial remineralization. Consequently, Fe addition may impact on the fate of the phytoplankton in the planktonic food web and the resulting efficiency of the biological pump. Accordingly, during the EIFEX large scale iron enrichment experiment, iron addition fosters a diatom bloom which appeared to broke up rather fast and to sink rapidly accordingly to 234Th measurements. This in turn triggered mesopelagic mineralization as evidenced in Baxs measurements. N-uptake experiments with Fe-limited and Fe-replete natural algal communities indicate that the effects of ammonium and iron on f-ratio and new production are not simply cumulative and that the enhancement of the f-ratio due to Fe addition depends on the ambient ammonium concentration. However the relationship between this enhancement and ammonium concentration is at present not fully understood. Our results imply that there is no simple relationship between export production and iron availability. Ammonium appears to counter the effects of iron addition on export production, particularly for HNLC areas such as the Southern Ocean. The meso-scale iron enrichment experiment performed during the EIFEX cruise, clearly showed that iron addition induced a diatom bloom but that the latter did not persist for long and collapsed soon after, thereby inducing a massive export of carbon as witnessed from significant deficits in 234Th activity. This export in turn triggered mesopelagic organic carbon mineralization as evidenced by excess Baxs contents. Measurements of pCO2, scaled using remote sensing data (sea surface temperature, chlorophyll and wind stress) and related CO2 fluxes, suggest that previous budgets of atmospheric CO2 uptake in the S.O. may be overestimated. However, our observations indicate that sea ice cover can act as an additional CO2 sink not taken into account in previous CO2 budgets. This sea-ice CO2 sink function results from physical and biogeochemical processes prevailing within the sea ice itself. Results obtained with SWAMCO-4, an idealized 1-D model of the marine planktonic system calculating C, N, P, Si, Fe cycling within the upper ocean, the export production and air-sea CO2 fluxes, suggest that the sink for atmospheric CO2 will increase in response to the raising atmospheric CO2 concentration. SWAMCO-4 simulations show that the amplitude of the predicted CO2 sink displays large regional and inter-annual variations which are related to local hydrodynamics and the dominant phytoplankton species. This is particularly acute in the Phaeocystis-dominated marginal ice zone of the Ross Sea. There the predicted annual CO2 sink appears positively related to the length of the sea ice cover period. This results from the accumulation of iron within the ice and its sudden release in the water column at the time of ice melting, favoring algal growth. The past and future evolution of the sea ice cover was conducted using the ORCA2- LIM and ECBILT-CLIO 3D ice-ocean models. The influence of the Southern Annular Mode on zonal integrated sea surface temperature and ice concentration seems to be small due to counteracting effects. Over the last 250 years, the annual mean ice coverage decreased in response to both natural and anthropogenic forcing, with the impact of the latter forcing clearly being enhanced over the last 150 years. Nowadays, the decrease of ice cover is more acute in the northern hemisphere as compared to the southern hemisphere, and this difference is due to thermal inertia of the S.O. However, model outputs predict a more abrupt decrease of S.O. sea ice extent in the future, resulting in similar decreases of annual mean sea ice extents in both hemispheres by the end of the century. In parallel, the seasonal amplitude of sea ice extent will increase. Preliminary runs with the coupled ice-ocean-biogeochemical model ORCA-LIMSWAMCO- 4, predict that the efficiency of the S.O. biological pump will be very sensitive to changes in the seasonal amplitude and the mean extent of the ice cover.