Deglacial and Holocene sea ice and climate dynamics at the Western Antarctic Peninsula

The reconstruction of past sea-ice distribution in the Southern Ocean is crucial for an improved understanding of ice–ocean–atmosphere feedbacks and the evaluation of Earth system and Antarctic ice sheet models. The Antarctic Peninsula (AP) has been experiencing a warming since the start of regular monitoring of the atmospheric temperature in the 1950s. The associated decrease in sea-ice cover contrasts the trend of growing sea-ice extent in East Antarctica. To reveal the long-term sea-ice history at the northern Antarctic Peninsula (NAP) under changing climate conditions, we examined a marine sediment core from the eastern basin of the Bransfield Strait covering the last Deglacial and the Holocene. For sea-ice reconstructions, we focused on the specific sea-ice biomarker lipid IPSO25, a highly branched isoprenoid (HBI), and sea-ice diatoms, whereas a phytoplankton-derived HBI triene (C25:3) and warmer open-ocean diatom assemblages reflect predominantly ice-free conditions. We further reconstruct ocean temperatures using glycerol dialkyl glycerol tetraethers (GDGTs) and diatom assemblages and compare our sea-ice and temperature records with published marine sediment and ice core data. A maximum ice cover is observed during the Antarctic Cold Reversal 13 800–13 000 years before present (13.8–13 ka), while seasonally ice-free conditions permitting (summer) phytoplankton productivity are reconstructed for the late Deglacial and the Early Holocene from 13 to 8.3 ka. An overall decreasing sea-ice trend throughout the Middle Holocene coincides with summer ocean warming and increasing phytoplankton productivity. The Late Holocene is characterized by highly variable winter sea-ice concentrations and a sustained decline in the duration and/or concentration of spring sea ice. Overall diverging trends in GDGT-based TEX86L and RI-OH' subsurface ocean temperatures (SOTs) are found to be linked to opposing spring and summer insolation trends, respectively.