Oligocene climate signals and forcings in Eurasia revealed by plant macrofossil and modelling results

The Oligocene represents a transitional time period from a warm climate to a cooler climate that is more representative of the modern day; yet, a general view of continental climate pattern and forcings are still lacking. Different proxies and models show striking disparities, especially in mid-high latitudes, requiring validation of Oligocene climate reconstruction in order to understand the large-scale processes that drive the observed climate changes. Here, we compiled 149 macrofossil floras in the mid-high latitudes of Eurasia, then quantitatively reconstructed the Oligocene climate using Coexistence Approach (CA) and combined previously published paleoclimate data. During the Oligocene, Eurasian mid-high latitudes were mainly dominated by a humid subtropical climate. Mean annual temperature (MAT) ranged between 5.4 °C and 25.5 °C with mean annual precipitation (MAP) ranging from 338 to 2453 mm. Three regions (Europe, central Eurasia and eastern Asia) indicate different climatic regimes, with a generally warmer and wetter climate in Europe and a colder and drier climate in central Eurasia when compared to eastern Asia. No significant reorganization of climate was observed between the Early and Late Oligocene. The climate anomalies between the Oligocene and present indicate that geographic changes (e.g. retreat of the Paratethys Sea) played an important role in shaping the climate pattern of Eurasia. By comparing the fossil data to a range of different HadCM3L model simulations of the Oligocene with differing boundary conditions (e.g. CO₂ and topography), we demonstrate similar large-scale climate spatial patterns between models and fossil data, however, models simulated much higher temperature seasonality (lower simulated winter temperatures and higher simulated summer temperatures) in Eurasia. Mean annual temperature analysis indicates that simulations with 560 and 1120 ppmv CO₂ matched better with fossil data when compared to other simulations, depending on the topography. These results provide some constraints that should be considered for future paleoclimate modelling.