Challenges and insights in comparing simulated tree cover changes over the last 20,000 years with reconstructions for the Northern Hemisphere

Pollen records are the most widespread archive for past climate and vegetation changes, offering valuable insights into Earth’s environmental history. These records provide a unique opportunity to evaluate Earth System Models. In recent years, the availability of quantitative plant cover reconstructions on a continental scale has increased, exemplified by the consistent dataset of REVEALS-based reconstructions provided by Schild et al. (2024) for the entire Northern Hemisphere.We use this data set for comparison with the changes in tree cover simulated by the Max Planck Institute Earth System Model (MPI-ESM) for the last 20,000 years. While the overall agreement between model and data is promising, there are significant regional discrepancies. Notable differences emerge in boreal regions such as Alaska/Western Canada and Siberia, where the model predicts a delayed and weaker tree cover increase during the deglaciation. Conversely, in temperate forest-steppe transition zones, the model shows an earlier and stronger tree cover expansion, balancing out the Northern Hemispheric mean change.However, systematic biases complicate the interpretation of this comparison. For instance, the model tends to simulate excessively cold conditions in boreal latitudes, while the reconstructions likely overestimate tree cover in these regions. As a result, the agreement in vegetation history remains uncertain leaving the comparison of absolute values between reconstructions and model results questionable. An EOF analysis highlights common modes of vegetation changes over the last 20,000 years in MPI-ESM and reconstructions, deepening our understanding despite these uncertainties.References:  Schild, L., Ewald, P., Li, C., Hébert, R., Laepple, T., and Herzschuh, U.: LegacyVegetation 1.0: Global reconstruction of vegetation composition and forest cover from pollen archives of the last 50 ka, Earth Syst. Sci. Data Discuss. [preprint], https://doi.org/10.5194/essd-2023-486, in review, 2024