Carbon-13 Observations Necessitate Reduced Ocean Ventilation at the Last Glacial Maximum, but Biogeochemical Processes Explain Most of the Atmospheric Carbon Dioxide Decrease

While the driver of the 80ppm decrease in atmospheric CO2 (pCO2atm) at the last glacial maximum (LGM) is likely oceanic, its attribution remains challenging. Viable hypotheses must explain a ~50ppm decline; since subsequent deep ocean carbonate compensation reduces pCO2atm further. Additionally, numerous observations have highlighted a 50% increase in the surface-deep gradient in the isotopic composition of oceanic dissolved inorganic carbon (δ13CDIC) at the LGM. As such, the widely measured variability in δ13CDIC is an important constraint on hypotheses that seek to explain LGM pCO2atm. Here we show that by using a state-of-the-art global ocean general circulation and biogeochemical model with different LGM circulation schemes (generated by a fully coupled ocean-atmosphere model under LGM forcing) and LGM dust input, only reduced ocean ventilation can be reconciled with 133 LGM δ13CDIC observations. However, while ocean circulation provides an explanation for a variety of paleoceanographic proxies, it explains less than 10% of LGM pCO2atm. Nevertheless, the impact of increased dust deposition, as well as changes in phytoplankton stoichiometry, are amplified by a favourable circulation to reduce pCO2atm by up to 25ppm, or one-half of the required 50ppm glacial-interglacial variability (prior to carbonate compensation). An analysis of our results that is constrained by two observational datasets (pCO2atm and δ13CDIC) suggests that although two-thirds of the change in LGM δ13CDIC gradient is controlled by ocean circulation, over 90% of the pCO2atm decline is biogeochemically driven.