Role of Southern Ocean stratification in glacial atmospheric CO2 reduction

Paleoclimate proxy data at the glacial period shows high salinity of more than 37.0 psu in the deep South Atlantic. At the same time, data also indicate that the residence time of the water mass was more than 3000 years. These data implies that the stratification by salinity was stronger in the deep Southern Ocean (SO) in the Last Glacial Maximum (LGM). Previous studies using Ocean General Circulation Model (OGCM) fail to explain the low glacial atmospheric carbon dioxide (CO₂) concentration at LGM. The reproducibility of salinity and water mass age is considered insufficient in these OGCMs, which may in turn affect the reproducibility of the atmospheric CO₂concentration. In coarse-resolution OGCMs, The deep water is formed by unrealistic open-ocean deep convection in the SO. Considering these facts, we guessed previous studies using OGCM underestimated the salinity and water mass age at LGM.

This study investigate the role of the enhanced stratification in the glacial SO on the variation of atmospheric CO₂ concentration by using OGCM. In order to reproduce the recorded salinity of the deep water, relaxation of salinity toward value of recorded data is introduced in our OGCM simulations. It was found that deep water formation in East Antarctica is required for explaining the high salinity in the South Atlantic. In contrast, it is difficult to explain the glacial water mass age, even if we assume the situation vertical mixing is very weak in the SO. Contrary to previous estimate, the high salinity of the deep SO resulted in increase of Antarctic Bottom water (AABW) flow and decrease the residence time of carbon in the deep ocean, which increased atmospheric CO₂ concentration. On the other hand, the weakening of the vertical mixing in the SO contributed to increase the vertical gradient of dissolved inorganic carbon (DIC), which decreased atmospheric CO₂ concentration.

Adding the contribution of the enhanced stratification in the glacial SO, we obtained larger reduction in atmospheric CO₂ concentration than previous studies. However, we still fail to explain the full amplitude of recorded glacial reduction of atmospheric CO₂ concentration. The carbonate compensation process, which is not incorporated in our simulations, might be required for further reduction in atmospheric CO₂ concentration.