The role of dispersal limitation in the post-glacial forest expansion of southern and central Europe

The global vegetation cover underwent strong changes during the past glacial cycle. These have been driven by climatic fluctuations but also by spatiotemporal vegetation dynamics, including migration to new climatologically suitable areas and interactions with other species. However, how much migration lag contributed to the vegetation change after the Last Glacial Maximum (LGM) is often not clear. We used the newly-implemented model LPJ-GM 2.0 to simulate the vegetation change of southern and central Europe from the end of the LGM (18.5 ka) to the preindustrial era (1.5 ka). The model couples a migration module to the dynamic global vegetation model LPJ-GUESS, thus allowing species to migrate simultaneously while interacting with each other. We compared two dispersal settings (free dispersal and dispersal limitation) against pollen data to test the reliability of the migration module to provide realistic paleo-vegetation reconstructions for biome and species distributions. Furthermore, we calculated range shifts of the leading edges and centroids to detect potential species-specific migration lags and range filling delays across simulation time. Our results show that the setting with dispersal limitation is better at capturing the initial post-glacial expansion of non-boreal forests in southern and central Europe than the scenario assuming free dispersal. Range shift analysis shows significant migration lags for most tree species at times of sudden temperature rise (start of the Bølling–Allerød warming event and following the Younger Dryas). Overall, our study suggests that it is necessary to include migration processes when simulating vegetation range expansion under rapid climate change, with implications for future vegetation projections.