Abstract: The Early Eocene Climatic Optimum (EECO, 53.26-49.14 Ma) represents the warmest period of the past 66 million years, with global temperatures reaching their highest levels on a million-year scale. This period serves as a geological reference for future Earth temperatures under the most extreme Shared Socioeconomic Pathways (SSP8.5) scenario in the Sixth Assessment Report of the Intergovernmental Panel on Climate Change (IPCC). Previous studies have reconstructed the global sea surface temperature (SST) state of the EECO using Earth system models and paleoclimate proxies. However, Earth system models often lack constraints on boundary conditions, making it difficult to accurately reconstruct the global SST state of the EECO. Additionally, there are discrepancies between different proxy reconstruction results. This study utilizes an emerging paleoclimate data assimilation approach, conducting simulations with the Earth system model of intermediate complexity, cGENIE. Our 100-member ensemble considers three most important uncertainties: atmospheric carbon dioxide concentration, alkalinity, and rain ratio. We assimilate three types of paleotemperature proxies (TEX86, Mg/Ca and δ¹⁸O of planktonic foraminifera) from 35 ocean drilling sites to reconstruct high-precision, globally coupled SSTs during the EECO. The data assimilation yielded a global SST of 30.7 ℃ (95% confidence interval: 28.8-33.0 ℃ ), characterized by significantly improved accuracy compared to prior estimates from Earth system model simulations. Sensitivity experiments confirmed that different types of proxies have a significant impact on the data assimilation reconstruction results: SSTs reconstructed without TEX86 proxy data were lower than those assimilating all proxies, while those without δ¹⁸O data were higher. This study provides a high-precision reconstruction of globally coupled SSTs during the EECO hothouse period, offering an accurate geological reference for future climate change.

Key words: paleoclimate data assimilation, Early Eocene Climate Optimum, sea surface temperature, paleoclimate proxies, Earth
system models