Glacial CO₂ decrease and deep-water deoxygenation by iron fertilization from glaciogenic dust

Abstract

Abstract. Increased accumulation of respired carbon in the deep ocean associated with enhanced efficiency of the biological carbon pump is thought to be a key mechanism of glacial CO2 drawdown. Despite greater oxygen solubility due to seawater cooling, recent quantitative and qualitative proxy data show glacial deep-water deoxygenation, reflecting increased respired carbon accumulation. However, the mechanisms of deep-water deoxygenation and contribution from the biological pump to glacial CO2 drawdown have remained unclear. In this study, we report the significance of iron fertilization from glaciogenic dust in glacial CO2 decrease and deep-water deoxygenation using our numerical simulation, which successfully reproduces the magnitude and large-scale pattern of the observed oxygen changes from the present to the Last Glacial Maximum. Sensitivity experiments show that physical changes contribute to only one-half of all glacial deep deoxygenation, whereas the other one-half is driven by iron fertilization and an increase in the whole ocean nutrient inventory. We find that iron input from glaciogenic dust with higher iron solubility is the most significant factor in enhancing the biological pump and deep-water deoxygenation. Glacial deep-water deoxygenation expands the hypoxic waters in the deep Pacific and Indian oceans. The simulated global volume of hypoxic waters is nearly double the present value, suggesting that glacial deep water was a more severe environment for benthic animals than that of the modern oceans. Our model underestimates the deoxygenation in the deep Southern Ocean because of enhanced ventilation. The model–proxy comparison of oxygen change suggests that a stratified Southern Ocean is required for reproducing the oxygen decrease in the deep Southern Ocean. Iron fertilization and a global nutrient increase contribute to a decrease in glacial CO2 of more than 30 ppm, which is supported by the model–proxy agreement of oxygen change. Our findings confirm the significance of the biological pump in glacial CO2 drawdown and deoxygenation.