Model-data synthesis of benthic isotopes suggests a warmer Miocene Climatic Optimum

The global warmth of the Miocene Climatic Optimum (MCO, ~15 Ma) offers valuable insights into warm climate dynamics and future climate change, providing a real-world testbed for validating climate models. However, the magnitude of MCO warmth remains uncertain due to the scarcity and inherent uncertainties of surface temperature proxy records. Existing estimates based on benthic isotopes, which leverage the relatively homogeneous nature of the deep ocean and therefore require fewer records, suggest that the MCO global mean surface temperature (GMST) was up to 5.8±0.7 ◦C warmer than preindustrial (PI). However, this benthic-based estimate is subject to uncertainties in seawater chemistry and is challenging to compare directly with surface temperature records. In this study, we synthesize benthic foraminiferal δ18O records with climate model simulations that feature an equilibrated deep ocean and include water isotope capability to better constrain the MCO warmth. Using a probabilistic inference framework, we estimate that the MCO GMST was most likely 8.0±0.8 ◦C warmer than PI, which is warmer than previous estimates and comparable to warming levels projected under the Shared Socioeconomic Pathway 5 (SSP8.5). In addition, our model-data-based framework provides a maximum likelihood estimation of the surface temperature field, which agrees well with independent surface proxies, thus offering a way to direct compare surface- and deep-ocean-based estimates. This study highlights the importance of equilibrated deep ocean conditions for more accurately estimating both deep ocean and surface temperatures, and our method can be applied to improve the reconstruction of global climate during other time intervals.