Abstract

Since 1860, the atmospheric CO2 has increased from 275 ppm to over 400 ppm, strongly enhancing the associated greenhouse effect. The global ocean has mitigated the impacts, sequestering around a third of the human emitted carbon dioxide, with the highest uptake of anthropogenic CO2 (Cant) occurring in the Southern Ocean and North Atlantic. However, oceanic Cant cannot be measured directly, being generally inferred by using indirect methods with a nominal range of uncertainties included between ±10% and ±20%. To better constraint these estimates, I focus on the North Atlantic Mode waters as they store the largest amount of Cant per unit area. In these water masses, the transit time distribution (TTD) gives the lowest Cant while the ∆C gives the highest estimate, so I concentrate on these two techniques. For both, a total uncertainty higher than previously suggested is quantified, reducing the level of Cant confidence to ±29 – 31%. However, the Cant range of uncertainties depends on regions, datasets, and timeframes: between 1992 and 2010, observations allow to reliably decrease it to ±13 – 15% in the subtropical North Atlantic (20 – 30◦N). Here, the Mode waters Cant content increases at +0.5 (TTD) and +0.8 (∆C) ± 0.2μmol kg−1 yr−1, thus the anthropogenic carbon estimates diverge over time. This divergence is ascribable to unsteadiness in the air-to-sea CO2 disequilibrium (+0.7 ± 0.2μmol kg−1 yr−1), and biogeochemical changes, as suggested by the increasing (+0.3μmol kg−1 yr−1) dissolved inorganic carbon from remineralised soft-tissue: both alterations are unequally captured by the TTD and the ∆C techniques. Changes in the ocean biogeochemistry are further explored using the output of a CM2 Mc control simulation over 2kyr. I attribute the accumulation of remineralised carbon to increasing surface productivity (18%), decreasing wind stress curl (13%), weakening meridional mass stream function (22%), hence increasing water mass mean residence time (86%). Ocean acidification additionally alters (68%) and it is influenced by the remineralisation increase, potentially shifting the oceanic carbon cycle towards new equilibria, and building a stronger North Atlantic carbon sink than previously thought.

Turning attention to the Southern Ocean, three questions remain open: (1) how accurate is the estimation of anthropogenic carbon in this basin? (2) Is our understanding of the Southern Ocean anthropogenic carbon pool impacting on other components of the carbon cycle (e.g. pH)? (3) Are our predictions of the future Southern Ocean carbon uptake and associated climate change reliable?