Modelling European tree species distribution change over the Holocene

The postglacial re-colonization of Europe by temperate tree species from a few glacial refugia during the Holocene (10,000 BP) is a very interesting case to study the mechanisms of the vegetation dynamics. The relative roles that played the climate conditions, the species dispersal capacities and the competition between species in the re-colonization rates remain controversial. We investigate these different aspects with the CARAIB dynamic vegetation model (CARAIB DVM). Transient runs were performed over the Holocene using the HadCM3 AOGCM-reconstructed climate. European-wide species migration at 0.5° x 0.5° is represented in the model using migration rates derived from a high resolution cellular automaton, CATS-UPSCALE. Individual tree species migration rates were pre-calculated with CATS-UPSCALE every 1000 years over each grid cell used by the DVM in the climatic conditions reconstructed by the AOGCM. The impacts of competition on plant dispersal are not taken into account by the automaton. Thus, in CARAIB, a function has been constructed to reduce the potential CATS migration rates in competition conditions. It is based on the species dispersal kernel and on the net primary productivity of the different species present on the grid cell. The migration of one species, from its 10,000 BP refugia, is studied within a landscape defined by a set of other species for which no dispersal limitations are assumed. Here, we illustrate the results obtained for two wind-dispersed (Abies alba and Picea abies) and for a no wind-dispersed (Fagus sylvatica) tree species. The speeds and the paths of the postglacial spread obtained with the DVM are compared to the past distributions of the three species reconstructed from pollen and macrofossil data. The Holocene climate conditions simulated by the HadCM3 AOGCM do not constrain the European re-colonization of the studied species, except in Scandinavia at the beginning of the period for Picea abies. We observe that, during the past 10,000 years, species occupied regions where climate conditions were different from present observed species climate requirements, notably in the 10k species refugia. This result may imply that at present the species do not occupy their potential distribution area and thus that the postglacial re-colonization is not completed yet. We also show that species dispersal capacities cannot explain the observed species migration over the Holocene and that competition has played an important role. Indeed, when we use the potential migration rates (no competition), species spread too fast.