Southern Ocean Ecosystem Response to Last Glacial Maximum Boundary Conditions

Abstract

Sarmiento & Gruber, 2006) by exporting carbon from the surface into the deep ocean via the biological pump. The net efficiency of the biological carbon export depends partly on the relative competitive fitness of a variety of phytoplankton functional types (Kvale et al., 2015b;Kvale et al., 2019). Major groups of these plankton include diazotrophs, coccolithophores, and diatoms. These groups differ in size, shape, and cell wall composition, which affect their sinking velocities (Collins et al., 2014;Klaas & Archer, 2002;Miklasz & Denny, 2010) and thus the amount of carbon exported into the deep ocean (DeVries et al., 2012). Calcifiers (e.g., coccolithophores) and silicifiers (e.g., diatoms) are two functional types thought to exert a dominant influence on global carbon cycling (Matsumoto et al., 2002) due to both their relatively efficient carbon export properties as well as their dominance in the Southern Ocean. Both diatoms and coccolithophores photosynthesize, which leads to oceanic CO 2 uptake. However, coccolithophores also produce calcite platelets, which leads to a decrease in surface alkalinity, that induces a net CO 2 outgassing (a mechanism also called the calcium carbonate counter-pump). Diatoms tend to dominate export production (EP) in the High Nutrient Low Chlorophyll (HNLC) regions. As EP is limited by the availability of Fe in HNLC regions, it has been hypothesized that iron fertilization of HNLC regions might be an efficient mechanism for enhancing ocean sequestration of carbon both in the modern climate as well as in the past (Martin, 1990).It is therefore important to better constrain the response of ecosystems to climatic changes, and in particular the response of diatoms and coccolithophores in the Southern Ocean. One example of significantly

Impact of iron fertilisation on atmospheric CO2 during the last glaciation

“…In this section, we describe the simulated physical and biological conditions at 70 ka BP, at the onset of MIS4 (Figure 3). In our 70ka-control simulation, the globally averaged ocean temperature is ∼2.8 • C, about 0.6 • C lower than in PI-control, but 0.9 • C higher than in a LGM simulation integrated with the same model (Saini et al, 2021). The globally averaged annual mean SST is 17.1 • C, 0.8 degrees lower than in the PI-control and 1.1 degrees higher than in the LGM simulation.…”

“…The colder conditions, more extensive sea-ice cover, and changes in the strength of the deep water masses impact marine productivity (Saini et al, 2021). Globally, NPP decreases by ∼9.7%, while EP decreases by ∼3% in our 70ka-control simulation compared to PI-control (Table 2).…”

“…The changes in NPP and EP are spatially heterogenous and sensitive to shifts in plankton distributions, changes in the length of the growing season, sea ice cover, and nutrient availability. In the Southern Ocean, the ∼7.5% reduction in NPP and 2% reduction in EP might be due to a shorter growing season for diatoms and coccolithophores (Saini et al, 2021),…”

“…To understand the resulting ocean carbon changes, we start by describing changes in the simulated ecosystems for one of our sensitivity experiments, lambfe3%, compared to 70ka-control. We choose lambfe3% because earlier research shows that the lambfe Southern Ocean dust mask provides a better fit with available glacial dust flux proxy records than the glacfe mask (Saini et al, 2021). Furthermore, a solubility factor of 3% corresponds to a likely estimate of iron solubility at 70 ka (see Methods) and has also been used in previously studies (Tagliabue et al, 2014;Lambert et al, 2015;Muglia et al, 2017;Yamamoto et al, 2019).…”

“…The aeolian iron fluxes based on glacfe and lambfe dust patterns compared to the PI iron flux (pife), and assuming a 1% solubility, are shown in Figure 2, while the full dust deposition maps for PI and the two LGM reconstructions are available in the supplementary material of Saini et al (2021). The integrated aeolian iron input for all experiments is listed in Table 1.…”