Vegetation successional pattern was studied on an extensive, urban brownfield located in
Results/Conclusions
Six main vegetation groups were identified: pioneer forest, tall herbs (with dominant, non-native Artemisia vulgaris and Solidago spp.), shrubland (Rhus spp.), grassland (Festuca rubra, Calamagrostis epigeos, and Panicum virgatum savannah), species rich open herb/grassland assembly. Multivariate data analysis was applied regarding soil characteristics as environmental variables. The CCA ordination diagram revealed that soil characteristics such as soil depth, NPK, Al content as well as heavy metal loads (Pb, Cr, Cd) were the strongest factors affecting the spatial distribution of species. Highest marginal effects were showed by Ni (lambda-1=0.62), P (lambda-1=0.53), and K (lambda-1=0.49) content, soil elements which explained a high share of the variance as single variables, but the conditional effects indicated that only P (p=0.002), K (p=0.032) content and soil depth (p=0.05) qualified for the final model to explain species composition. Non-native species contributed markedly to local species diversity and hot spot of species richness were retained largely in areas with low NPK loads and shallow soils. This suggests the importance of environmental stress in maintaining biological diversity during succession.