The Temporal Variability of Anthropogenic Carbon Storage in the Labrador Sea

Abstract

The Labrador Sea is a key region for the transport of anthropogenic carbon (Cant) into the ocean interior. The formation of Labrador Sea Water (LSW) provides a direct path for atmospheric gases to be exchanged with the deep ocean. However, model projections suggest there is potential for significant future reduction of convection depth and LSW formation in response to increased freshwater input from Greenland. In order to predict how physical changes will affect the fate of Cant in the Labrador Sea, it is crucial to understand how the storage of Cant has been evolving in the region. In this PhD thesis I present the first multi-decadal (1993-2016) estimate of Cant in the Central Labrador Sea based on two indirect methods (Transit Time Distribution; TTD further extended to 1986, and extended Multiple Linear Regression; eMLR), and I compare these estimates to Dissolved Inorganic Carbon (DIC) observations. I first focus on the quality of the carbonate chemistry data by assessing the internal consistency of the carbonate system. I recommend best practices to perform calculations of the carbonate system parameters and I highlight the importance of conversion to in situ conditions of temperature and pressure for comparisons with measurements performed with autonomous sensors. For the application of the TTD method I calculate the saturations of CFC-12 and SF6 from observations and show their strong temporal variability in this region. Therefore I provide reconstructed saturations of these tracers to accurately estimate Cant with the TTD method here. By using the eMLR method with different time intervals and starting years, I show that the choice of starting year can greatly affect the estimates of Cant storage rate and therefore mislead the interpretation of the role of the Labrador Sea in sequestering Cant. Both methods to estimate Cant highlighted the importance of assumptions implied in a proxy method, they revealed that the storage of Cant is non-steady over time and that the temporal variability of the Cant storage in the region appears to be associated to the strength of convection. Overall this thesis highlights the importance of a multi-disciplinary long-term monitoring program and represents a solid base for validation of biogeochemical models.