Reassessment and applications of the Mg/Ca - δ18Oc proxy system recorded in shells of the Arctic planktonic foraminifera Neogloboquadrina pachyderma

Observation-based reconstructions of Arctic sea surface temperatures in response to changing climate boundary conditions are critical to constrain climate sensitivity and evaluate the uncertainties of model simulations. On long and pre-instrumental timescales, this is only possible by employing climate proxies. Yet, most proxies of essential climate variables, such as sea surface temperatures (SST), suffer from limitations when applied to cold temperatures that characterize Arctic environments. These limitations prevent us from constraining uncertainties for some of the most sensitive climate tipping points that can trigger rapid and dramatic global climate change such as Polar Amplification, the disruption of AMOC, sea ice loss, and permafrost melting that are intrinsic to the polar regions. Here, we present a new approach to reconstructing sea surface temperatures (SST) using paired Mg/Ca - δ¹⁸O_c recorded in shells of the Arctic planktonic foraminifera Neogloboquadrina pachyderma. We show that in this proxy system, the Mg/Ca – palaeothermometry is affected by variations in seawater carbonate chemistry, which can be successfully quantified and removed from paleotemperature reconstructions allowing a reassessment of the absolute temperature and the magnitude of marine polar amplification to climate forcing on glacial-interglacial timescales. By applying this novel approach to existing records, we show that the magnitude of high latitude SST cooling during glacial periods has been underestimated and that the new estimate of SST change between the Late Holocene and the LGM exceeds model-based estimates of marine polar amplification by up to 3.0 ±1.0˚ C. Our findings open up opportunities to better constrain the oceanic carbonate system enabling a quantification of high-latitude ocean-atmosphere carbon exchange as well as to benchmark the performance of CMIP6 and future generations of climate models.