MERGE SP: Pollen-based maps of plant migration in Europe during the Holocene: Creating a dataset for the evaluation of migration models in dynamic vegetation models

The correct inclusion of migration in dynamic vegetation models, such as LPJ-GUESS, is important for evaluating how changes in future climate will affect vegetation. Of special interest is how rapidly species will spread northwards in response to increasing temperature, especially with respect to tundra shrubification and tree-line advance, which could adversely affect arctic and subarctic biomes. For instance, it has been shown that LPJ-GUESS does not model the migration of Fagus (beech) and Picea (spruce) properly. Data from the early Holocene describing how species advanced as the ice retreated would be useful to evaluate tree migration models. The VR LandClim I & II projects (PI Marie-José Gaillard) have established databases of Holocene pollen records for sites across Europe for 25 time windows (from 50 to 11500 years before present) and corresponding REVEALS estimates of past plant cover for 30 taxa. Data from five of these time windows have previously been interpolated in to construct gridded land-cover maps for Europe of the fractions of coniferous, deciduous forest and open land at a one degree spatial scale for use in climate modeling. In this project we aim to use pollen data to estimate the presence/absence of key species across Europe for the entire Holocene. The purpose of the project is to: 1) Expand the spatial model to allow for temporal dependencies. 2) Apply the resulting model to pollen-based presence/absence data for key species in Europe, producing space and time continuous estimates of species distributions for the entire Holocene. 3) Produce a data-set of pollen-based plant migration patterns during the Holocene that can be used for evaluation and calibration of dynamic vegetation models and related tree migration models.