Early deglacial Atlantic overturning decline and its role in atmospheric CO<sub>2</sub> rise inferred from carbon isotopes (δ<sup>13</sup>C)

Abstract

Listen

Translate article

Abstract. The reason for the initial rise in atmospheric CO2 during the last deglaciation remains unknown. Most recent hypotheses invoke Southern Hemisphere processes such as shifts in midlatitude westerly winds. Coeval changes in the Atlantic meridional overturning circulation (AMOC) are poorly quantified, and their relation to the CO2 increase is not understood. Here we compare simulations from a global, coupled climate–biogeochemistry model that includes a detailed representation of stable carbon isotopes (δ13C) with a synthesis of high-resolution δ13C reconstructions from deep-sea sediments and ice core data. In response to a prolonged AMOC shutdown initialized from a preindustrial state, modeled δ13C of dissolved inorganic carbon (δ13CDIC) decreases in most of the surface ocean and the subsurface Atlantic, with largest amplitudes (more than 1.5‰) in the intermediate-depth North Atlantic. It increases in the intermediate and abyssal South Atlantic, as well as in the subsurface Southern, Indian, and Pacific oceans. The modeled pattern is similar and highly correlated with the available foraminiferal δ13C reconstructions spanning from the late Last Glacial Maximum (LGM, ~19.5–18.5 ka BP) to the late Heinrich stadial event 1 (HS1, ~16.5–15.5 ka BP), but the model overestimates δ13CDIC reductions in the North Atlantic. Possible reasons for the model–sediment-data differences are discussed. Changes in remineralized δ13CDIC dominate the total δ13CDIC variations in the model but preformed contributions are not negligible. Simulated changes in atmospheric CO2 and its isotopic composition (δ13CCO2) agree well with ice core data. Modeled effects of AMOC-induced wind changes on the carbon and isotope cycles are small, suggesting that Southern Hemisphere westerly wind effects may have been less important for the global carbon cycle response during HS1 than previously thought. Our results indicate that during the early deglaciation the AMOC decreased for several thousand years. We propose that the observed early deglacial rise in atmospheric CO2 and the decrease in δ13CCO2 may have been dominated by an AMOC-induced decline of the ocean's biologically sequestered carbon storage.