During Marine Isotope Stage 3 (MIS-3; 57–29 ka) Antarctic ice cores reveal a glacial climate state punctuated by millennial-scale warming events and pulses of CO2. Changes in iron-fertilised export production and ocean circulation/upwelling, interpreted from South Atlantic sediment cores, suggest that the Southern Ocean is a conduit for the storage and release of CO2 from the deep ocean. However, it is unclear whether this occurs throughout the Southern Ocean as these processes have not previously been investigated in the southwest Pacific . Here we describe localised iron limitation linked to glaciation changes in New Zealand, which reduced export production during early MIS-3 (60–48 ka) and caused decreases/increases in export production during late MIS-3 (48–29 ka) millennial-scale warming/cooling. Consistent decreases in foraminifera-bound δ15N during all MIS-3 warming events may reflect changes in the supply of nitrate to the subantarctic Pacific, possibly from increased wind-driven upwelling in the Antarctic and northward eddy-driven transport and/or shifting SO fronts. Concomitant decreases in bottom water oxygen and increases in the 14C age of deep waters suggest that old, nutrient-rich waters influenced upper circumpolar deep water in the southwest Pacific during warming events. This signature may reflect an expansion of Pacific Deep Water into the Southern Ocean as Southern Ocean overturning strengthens. Iron-limitation of export production, the expansion of Pacific Deep water, and increased wind-driven upwelling would all work to contribute to increasing atmospheric CO2 through reduced drawdown, and increased outgassing from the Pacific carbon reservoir during the millennial-scale warming events of MIS-3.

Model intercomparison studies of coupled carbon-climate simulations have the potential to improve our understanding of the processes explaining the pCO2 drawdown at the Last Glacial Maximum (LGM) and to identify related model biases. Models participating in the Paleoclimate Modelling Intercomparison Project (PMIP) now frequently include the carbon cycle. The ongoing PMIP-carbon project provides the first opportunity to conduct multimodel comparisons of simulated carbon content for the LGM time window. However, such a study remains challenging due to differing implementation of ocean boundary conditions (e.g. bathymetry and coastlines reflecting the low sea level) and to various associated adjustments of biogeochemical variables (i.e. alkalinity, nutrients, dissolved inorganic carbon). After assessing the ocean volume of PMIP models at the pre-industrial and LGM, we investigate the impact of these modelling choices on the simulated carbon at the global scale, using both PMIP-carbon model outputs and sensitivity tests with the iLOVECLIM model. We show that the carbon distribution in reservoirs is significantly affected by the choice of ocean boundary conditions in iLOVECLIM. In particular, our simulations demonstrate a ~250 GtC effect of an alkalinity adjustment on carbon sequestration in the ocean. Finally, we observe that PMIP-carbon models with a freely evolving CO2 and no additional glacial mechanisms do not simulate the pCO2 drawdown at the LGM (with concentrations as high as 313, 331 and 315 ppm), especially if they use a low ocean volume. Our findings suggest that great care should be taken on accounting for large bathymetry changes in models including the carbon cycle.

The Southern Ocean is the largest region in which iron limits the growth of phytoplankton. However, a phytoplankton bloom thousands of square kilometres in area forms each spring-summer in the Indian sector of the Southern Ocean, both above and to the east of the Kerguelen Plateau. The central region of the Kerguelen Plateau hosts the volcanically active islands, Heard and McDonald (HIMI), the former of which is largely covered by glaciers. The sources and processes governing supply of iron from HIMI to the region are relatively unknown. In the austral summer of 2016, the first voyage to focus on biogeochemical cycling in the HIMI region was undertaken (GEOTRACES process study GIpr05). Using iron redox measurements, we show here that each of the adjacent islands are strong sources of dissolved iron(II) (DFe(II)), though controlled by different supply mechanisms. At Heard Island, the greatest DFe(II) concentrations (max 0.57 nmol L) were detected north of the island. An inverse correlation of DFe(II) concentrations with salinity suggests the origin is from a sea-terminating glacier on the island. At McDonald Islands, the greatest DFe(II) concentrations (max 1.01 nmol L) were detected east of the islands which, based on DFe(II) profiles from five targeted stations, appears likely to originate from shallow diffuse hydrothermalism. Elevated DFe(II) around HIMI may increase Fe availability for biota and indicate slower oxidation kinetics in the region, which has implications for transport of Fe away from the islands to the broader northern Kerguelen Plateau where the annual plankton bloom is strongest.

Quantitative knowledge about the burial of sedimentary components at the seafloor has wide-ranging implications in ocean science, from global climate to continental weathering. The use of 230 Th-normalized fluxes reduces uncertainties that many prior studies faced by accounting for the effects of sediment redistribution by bottom currents and minimizing the impact of age model uncertainty. Here we employ a recently compiled global dataset of 230 Th-normalized fluxes with an updated database of seafloor surface sediment composition to derive global maps of the burial flux of calcium carbonate, biogenic opal, total organic carbon (TOC), non-biogenic material, iron, mercury, and excess barium (Baxs). The spatial patterns of burial of the major components are mainly consistent with prior work, but the new quantitative estimates allow evaluations of global deep-sea burial. Our integrated deep-sea burial fluxes are 136 Tg C/yr CaCO3, 153 Tg Si/yr opal, 20Tg C/yr TOC, 220 Mg Hg/yr, and 2.6 Tg Baxs/yr. Sedimentary Fe fluxes reflect a mixture of sources including lithogenic material, hydrothermal inputs and authigenic phases. The fluxes of some commonly used paleo-productivity proxies (TOC, biogenic opal, and Baxs) are not well-correlated geographically with satellite-based productivity estimates. Our new compilation of sedimentary fluxes provides more detailed information on burial fluxes, which should lead to improvements in the understanding of how preservation affects these paleoproxies.