Proxy-model comparison of EECO upper ocean temperatures using coccolith clumped isotope thermometry

Inorganic carbonate proxies measured on planktic foraminifera, such as δ¹⁸O and Mg/Ca, suggest that the early Eocene Climatic Optimum (EECO; ~50 Ma) was a period characterised by extremely warm global ocean temperatures (Hollis et al., 2019). However, significant inter-proxy offsets in reconstructed absolute temperatures  exist and stem from differing sensitivities to diagenesis, uncertainties of past seawater composition, and variations in the water depth recorded by different proxy archives. Here we present clumped isotope-derived upper ocean temperature reconstructions measured on coccoliths, from a suite of globally distributed sites for an interval (50.7 Ma–50.4 Ma) representative of peak EECO conditions. Coccolith clumped isotope-derived temperatures are independent of the seawater carbonate chemistry and are species-independent (Clark et al., 2024), which allows for reliable past temperature reconstructions. We compare our results to HADCM3 and GFPL annual time series model simulations with multiple CO₂ levels from the recent DeepMIP model compilation (Steinig et al., 2024). Since we consider variability across “deep” time to be minimal, we identify the depths and months our coccoliths most likely calcified at for each site, to allow for better comparison to the annual time series model simulation. This is especially important since modern coccoliths calcify at different depths across different ocean basins (Clark et al., 2024; Mejia et al., 2023).

Using the well-constrained coccolith calibration (Clark et al., 2024), we find that coccolith clumped isotope-derived temperatures generally agree well with previous surface ocean temperature reconstructions from other inorganic carbonate proxies (Hollis et al., 2019). During the EECO, we also find a similar magnitude of variability in coccolith calcification depths as observed in the modern ocean. This suggests that coccolith calcite records upper ocean temperature signals rather than solely sea surface temperature. Furthermore, the variability in overlap between the model and coccolith clumped temperatures across an annual year, in particular for the mid latitudes, further confirm that coccolith calcite captures upper ocean temperature signals largely during coccolithophore blooming months. Our coccolith clumped isotope-derived temperature data confirms the relatively flat latitudinal gradients and warm high latitudes found by other temperature proxies during the EECO, highlighting the potential of coccolith clumped isotopes as an useful tool for reconstructing past upper ocean temperatures during past warm climates.

Hollis et al., 2019, https://doi.org/10.5194/gmd-12-3149-2019 ; Mejia et al., 2023, https://doi.org/10.1016/j.epsl.2023.118313 ; Clark et al., 2024, https://doi.org/10.22541/essoar.173042174.40363194/v1 ; Steinig et al., 2024, https://doi.org/10.1038/s41597-024-03773-4