Assessing transient changes in the ocean carbon cycle during the last deglaciation through carbon isotope modeling

Atmospheric carbon dioxide concentration (pCO2) has increased by approximately 80 ppm from the Last Glacial Maximum (LGM) to the early Holocene. The change in this atmospheric greenhouse gas is recognized as a climate system response to gradual change in insolation. Previous modeling studies suggested that the deglacial increase in atmospheric pCO2 is primarily attributed to the release of CO2 from the ocean. Additionally, it has been suggested that abrupt change in the Atlantic meridional overturning circulation (AMOC) and associated interhemispheric climate changes are involved in the release of CO2. However, understanding remains limited regarding oceanic circulation changes and the factors responsible for changes in chemical tracers in the ocean during the last deglaciation and their impact on atmospheric pCO2. In this study, we investigate the evolution of the ocean carbon cycle during the last deglaciation (21 to 11 ka BP) using three-dimensional ocean fields from the transient simulation of the MIROC 4m climate model, which exhibits abrupt AMOC changes similar to those observed in reconstructions. We investigate the reliability of simulated changes in the ocean carbon cycle by comparing the simulated carbon isotope ratios with sediment core data, and we examine potential biases and overlooked or underestimated processes in the model. Qualitatively, the modeled changes in atmospheric pCO2 are consistent with ice core records. For example, during Heinrich Stadial 1 (HS1), atmospheric pCO2 increases by 10.2 ppm, followed by a reduction of 7.0 ppm during the Bølling–Allerød (BA) period and then by an increase of 6.8 ppm during the Younger Dryas (YD) period. However, the model underestimates the changes in atmospheric pCO2 during these events compared to values derived from ice core data. Radiocarbon and stable isotope signatures (Δ14C and δ13C) indicate that the model underestimates both the activated deep-ocean ventilation and reduced efficiency of biological carbon export in the Southern Ocean and the active ventilation in the North Pacific Intermediate Water (NPIW) during HS1. The relatively small changes in simulated atmospheric pCO2 during HS1 might be attributable to these underestimations of ocean circulation variation. The changes in Δ14C associated with strengthening and weakening of the AMOC during the BA and YD periods are generally consistent with values derived from sediment core records. However, although the data indicate continuous increase in δ13C in the deep ocean throughout the YD period, the model shows the opposite trend. It suggests that the model either simulates excessive weakening of the AMOC during the YD period or has limited representation of geochemical processes, including marine ecosystem response and terrestrial carbon storage. Decomposing the factors behind the changes in ocean pCO2 reveals that variations in temperature and alkalinity have the greatest impact on change in atmospheric pCO2. Compensation for the effects of temperature and alkalinity suggests that the AMOC changes and the associated bipolar climate changes contribute to the decrease in atmospheric pCO2 during the BA and the increase in atmospheric pCO2 during the YD period.