全球变化知识资源中心

Aim The former continental-scale studies modelled coarse-grained plant species-richness patterns (gamma diversity). Here we aim to refine this information for European forests by (a) modelling the number of vascular plant species that co-occur in local communities (alpha diversity) within spatial units of 400 m(2); and (b) assessing the factors likely determining the observed spatial patterns in alpha diversity. Location Europe roughly within 12 degrees W-30 degrees E and 35-60 degrees N. Taxon Vascular plants. Methods The numbers of co-occurring vascular plant species were counted in 73,134 georeferenced vegetation plots. Each plot was classified by an expert system into deciduous broadleaf, coniferous or sclerophyllous forest. Random Forest models were used to map and explain spatial patterns in alpha diversity for each forest type separately using 19 environmental, land-use and historical variables. Results Our models explained from 51.0% to 70.9% of the variation in forest alpha diversity. The modelled alpha-diversity pattern was dominated by a marked gradient from species-poor north-western to species-rich south-eastern Europe. The most prominent richness hotspots were identified in the Calcareous Alps and adjacent north-western Dinarides, the Carpathian foothills in Romania and the Western Carpathians in Slovakia. Energy-related factors, bedrock types and terrain ruggedness were identified as the main variables underlying the observed richness patterns. Alpha diversity increases especially with temperature seasonality in deciduous broadleaf forests, on limestone bedrock in coniferous forests and in areas with low annual actual evapotranspiration in sclerophyllous forests. Main conclusions We provide the first predictive maps and analyses of environmental factors driving the alpha diversity of vascular plants across European forests. Such information is important for the general understanding of European biodiversity. This study also demonstrates a high potential of vegetation-plot databases as sources for robust estimation of the number of vascular plant species that co-occur at fine spatial grains across large areas.