Exploring glacial-interglacial nutrient conditions in the Antarctic Zone: Insights from a one-dimensional water column model

In the Antarctic Zone (AZ), deep nutrient-rich waters ascend to the surface, feeding the Southern Ocean's overturning circulation cells. However, the rate of upwelling exceeds the capacity of phytoplankton to fully consume the gross nutrient supply to the AZ surface, leading to the release of previously sequestered CO₂ into the atmosphere. During ice ages, enhanced nutrient utilization has been proposed as a mechanism that could contribute to lower atmospheric CO₂ concentration. Fossil-bound δ¹⁵N records in the AZ point to a more complete nitrate consumption in surface waters during ice ages. This increase in nitrate utilization coincides with reduced export production, suggesting a lower gross nitrate supply to the surface and, therefore, a reduction in the exchange of water between the surface and the deep ocean. Preliminary reconstructions indicate more than a 5-fold reduction in the rate of gross nitrate supply to match paleo proxy data and near complete nitrate consumption at the surface. Model simulations are ambiguous, but none show more than a ≥ 2-fold reduction in water exchange in the AZ during ice ages.

One hypothesis for this discrepancy is the progressive depletion (“mining-out”) of nutrients from the AZ upper ocean. Reduced glacial upwelling, combined with repeated summer nitrate consumption and the export of assimilated nitrate as sinking organic matter, followed by deep winter mixing, could gradually deplete the upper water column’s nutrient reservoir. This process would lower the shallow subsurface nutrient concentrations and elevate nitrate δ¹⁵N relative to the deep ocean. As a result, the nutrient supply per volume of upwelled water would decline, aligning better with model simulations. 

To test this hypothesis, we developed a 1D advection-diffusion-reaction model of the water column, accounting for surface nitrate consumption and isotope fractionation. The model was calibrated using Argo floats data and high-resolution hydrographic nitrate isotopes transect in the AZ (GO-SHIP SO4P 2018), successfully matching depth and seasonal profiles. We also applied the model to the western subarctic Pacific, which exhibits a similar observation pattern for fossil-bound δ¹⁵N and export production during ice ages but contrasts in the ratio between advective and diffusive nutrient supply.

Our results highlight the critical role of nutrient mining in driving isotopic changes during ice ages. With reduced upwelling, nutrients are progressively depleted in the upper AZ. However, even under this mechanism, a substantial reduction in upwelling (more than a twofold decrease) is still required to achieve observed glacial δ¹⁵N values – though less extreme than previous estimates. Nevertheless, in reduced upwelling scenarios, the glacial surface nitrate concentration is significantly higher than previous estimates. This supports the potential of nutrient mining in matching paleo-data with less drastic changes to the Southern Ocean.