Since the industrial revolution, humans have caused profound climate changes, primarily by releasing geological carbon into the atmosphere and increasing atmospheric CO₂, with current levels reaching >400ppm, a concentration unprecedented in the last 800ka. This has led to far-reaching socioeconomic consequences for human society and risks for all levels of ecosystem. A better understanding of rapid climatic changes is desperately needed in order to inform mitigation and adaptation strategies for future climate change.

The last glacial cycle experienced orbital and millennial scale climatic variability, as indicated by high latitude ice core records and many other high-resolution marine and terrestrial records. These climatic changes included, but were not limited to, changes in atmospheric CO₂, temperature, the hydroclimate, sea surface temperature (SST), ocean circulation and ocean biogeochemistry. The ocean is thought to have played a key role in controlling and modulating these changes through its impacts on both heat transport and the carbon cycle. High resolution marine sediment cores can be used to reconstruct these changes and may help to elucidate the mechanisms behind them. To date, most studies have focused on the deglaciation, and only sparse, low-resolution records exist for Marine Isotope Stage (MIS) 4, a key paleoclimatic interval for the last glacial inception. MIS 4 is characterised by a rapid CO₂ drop of ~40ppm, which is comparable in duration and magnitude to the first rapid increase seen during the last deglaciation. It also involved a large drop in temperature, as indicated by Greenland and Antarctic ice cores, a decrease in sea level, and a possible slowdown of Atlantic Meridional Overturning Circulation (AMOC) as reconstructed from various proxy records. Several millennial events occurred during MIS 4, including Heinrich Stadial 6 and Dansgaard-Oeschger (DO) events 16-19. MIS 4 is thus an ideal interval to study and disentangle, glacial-interglacial and millennial variability. It also provides a window into the mechanisms of rapid CO₂ change and their contribution to longer-term (orbital) climate change. Furthermore, the termination of MIS 4 allows for a comparison with the last deglaciation. In this thesis, I collect paleoceanographic data to improve coverage of this important interval from a suite of sediment cores retrieved from the Iberian Margin in the Northeast Atlantic, and a single core from the deep Sub-Antarctic Atlantic core site. This thesis ultimately aims to enhance the current understanding of the ocean’s role in and response to abrupt and orbital-scale climate changes during MIS 4 and to draw lessons on its wider implications for climate variability. Ultimately, this may contribute to our understanding of natural carbon cycle-climate feedbacks that will play a role in anthropogenic climate changes in the future.

High resolution planktonic foraminifera Mg/Ca-based SST reconstructions from the Iberian Margin during MIS 4 show that certain aspects of the surface ocean response may not always track Greenland temperature and that Greenland ice core records do not serve as a universal template for climatic variability across the whole of the North Atlantic, likely due to the seasonal habitat biases associated with SST reconstructions. A strong hydroclimate signal is shown in planktic foraminifera δ¹⁸O from the Iberian Margin, whereby glacial (MIS 4) hydroclimate variability is coupled to a combination of the high-latitude North Atlantic changes and low-latitude tropical hydroclimate. Furthermore, for the first time, a high-resolution Mg/Ca-based SST record from the Iberian Margin, covering the last 85ka, demonstrates clear similarities between MIS 4 and MIS 2. This includes a similar decoupling of sub-tropical summer SST from Greenland temperatures recorded in ice core records during pre-HS 6 MIS 4 and the Last Glacial Maximum (LGM). The record also emphasises that the most severe (coldest and driest) conditions occurred in the midlatitude North Atlantic during Heinrich Stadials, rather than the ‘peak’ glacial conditions of MIS 4 or the LGM.

The deep ocean likely played a key role in modulating CO₂ on millennial and astronomical timescales, for example through changes in its respired carbon inventory. Conservative parameters that are indicative of deep-water hydrography, and by extension circulation, are deep water temperature (T_dw) and associated δ¹⁸O_dw. Reconstructed T_dw changes from the Iberian Margin show a larger influence of southern sourced waters during MIS 4 and particularly during HS 6. Atlantic sector Southern Ocean (SO) T_dw closely follows Antarctic temperature, atmospheric CO₂ and the mean ocean temperature, implying that the deep SO contributed significantly to the global ocean energy budget on multi-millennial time scales across MIS 4, likely mediated by buoyancy forcing in the SO. This in turn was likely linked to sea-ice expansion at the MIS 5a/4 transition, aided by a parallel shoaling of North Atlantic Deep Water (NADW) as suggested by the North Atlantic Tdw record. Together with (arguably smaller) contributions from reduced air-sea gas exchange efficiency in the SO, these changes would have lowered atmospheric CO₂ during MIS 4, through their impact on the solubility- and soft tissue “pumps” (i.e. the ocean’s disequilibrium and respired carbon budgets).

Because the amount of respired carbon in deep-water broadly scales with the dissolved oxygen concentration, bottom water O₂ reconstructions, [O₂]_bw, were investigated for a depth transect from the Iberian Margin and for the Atlantic sector of the Southern Ocean. Qualitative benthic foraminiferal assemblage records from a depth transect on the Iberian Margin show that shifts in oxygenated environments are primarily controlled by the quality and/or quantity of C_org reaching the sea floor, rather than [O₂]_bw. There are distinct shifts in assemblages associated with more periodic and/or degraded C_org flux during MIS 4 and an indication of low [O₂]_bw during HS 6 at the mid-depths, however no significant changes in the flux of C_org (i.e. ‘export production’) were found. Multi-proxy foraminiferal geochemical [O₂]_bw reconstructions from the Iberian Margin show a gradual decrease in [O₂]_bw at the mid-depth North Atlantic during MIS 4 with a minimum during HS 6, likely controlled by ventilation changes (i.e. changes in ocean circulation, including water mass sourcing combined with active but diminished transport, or altered preformed ‘end-member’ values). In the meantime, the [O₂]_bw record from the South Atlantic closely follows atmospheric CO₂, likely indicative of ocean ‘ventilation’ impacts on ocean-atmosphere carbon exchange. Indeed, the Southern Ocean appears to have represented a significant reservoir for sequestering CO₂ away from the atmosphere during MIS 4, as indicated by the respired- and equilibrium carbon inventory changes that are implied by the [O₂]_bw and T_dw reconstructions.