A first step towards a complete Southern Ocean proxy compilation for the Last Glacial Cycle: glacial-interglacial changes in Sea Surface Temperature

Previous research has suggested that different processes in the Southern Ocean contributed to the drawdown and release of atmospheric carbon dioxide (CO2) during the last glacial cycle (0130 ka), yet their dynamics and interplay are not fully understood. To unravel the interactions between different processes and to allow for more comprehensive analyses, we aim to compile all previously published proxy data and convert them into the Linked open Paleo Data (LiPD) format, overall increasing interoperability, reusability and impact. The PAGES C-SIDE working group recently highlighted substantial open-ocean sea-ice extent changes during the mid-glacial period (Marine Isotope Stage 4, ~72 to 60 ka), coinciding with a significant drop in atmospheric CO2. However, sea-ice changes are notably absent during the early glaciation (Marine Isotope Stage 5d, ~115 to 105 ka), suggesting they cannot account for the early CO2 decrease (Chadwick et al., 2022). As an initial step, we present our compilation of sea surface temperature (SST) reconstructions from marine sediment records across the Southern Ocean (Latitude > 30S) for the last glacial cycle. This compilation assesses SST changes in different zones and basins, evaluates SST gradients, and explores the interplay between SST and sea ice. Our findings reveal a consistent glacial-interglacial amplitude of 5-10C across all basins and zones, with uniform timing. SST gradients from the Antarctic to Subantarctic Zone remain unchanged over time, eliminating them as a mechanism for early CO2 decrease. Additionally, we observe that 50% of the total MIS 5e-to-LGM cooling in Southern Ocean SST occurred from MIS 5e to MIS 5d, with a second drop from MIS 5a to MIS 4, essentially reaching LGM cooling. A distinct decoupling of SST cooling and sea ice expansion over MIS 5 suggests that circulation and/or subsurface temperatures may exert a stronger influence than SST on sea ice extent. This emphasizes the necessity for additional proxy compilations to further disentangle these complex relationships.