We currently witness in the Arctic: 1) a decrease in summer ice cover that exposes sea surface to solar radiation and physical forcings, 2) permafrost thawing and increased river runoff, both leading to an increase in the export to ocean of organic carbon previously sequestered in the Tundra, and 3) an increase in ultraviolet radiation. These three phenomena potentially favour a growing mineralization of organic carbon through photo-oxidation taking place in the surface layer of the ice-free Arctic Ocean waters, resulting in an acceleration of the increase in atmospheric CO2. At the same time, the exposure of a larger fraction of ocean surface to sun light and the possible increase in nutrients brought by rivers lead to larger autotrophic production and sequestration of organic carbon. The general objective is to determine the absolute significance of organic matter photo-oxidation and of primary production, and their relative balance in the Arctic Ocean. Diagnostic (Part 1) and prognostic (Part 2) approaches will be developed to address the problem quantitatively. Part 1: Based on existing and new data collected in situ in different regions of the Arctic Ocean, we will developed 1) a model for light propagation through atmosphere and within ocean, 2) diagnostic models for the two light-related processes of interest, and 3) algorithms for the processing of ocean colour remote sensing data specifically in the Arctic Ocean. Organic matter photo-oxidation and primary production will then be estimated using different sources of remote sensing information (UV, clouds, ice, optical properties of surface waters) at various space and time scales. Part 2: The diagnostic models will be integrated into a coupled physical-biological ecosystem model including bacterial activity, and applied using Global Climate Model outputs to assess the fate of the associated carbon fluxes in the Arctic Ocean during the next decades under different climate change scenarios.