Testing floristic and environmental differentiation of rich fens on the Bohemian Massif.

The south-eastern part of the Bohemian Massif (the Bohemian-Moravian Highlands, the Třeboň basin, Czech Republic) is an important hotspot of fen biodiversity. Especially rich fens with calcium- tolerant peat mosses (the Sphagno warnstorfii-Tomentypnion alliance) currently harbour highly endangered organisms. In this study we gathered phytosociological and environmental (water chemistry, water table depth) data from 57 unique and well-preserved fens. The ISOPAM algorithm reproduced the expert-based classification at the alliance level presented in the Vegetation of the Czech Republic monograph. Particular types of vegetation were nearly completely differentiated in the PCA of environmental data and all their pairs differed significantly with respect to pH, which together with calcium was correlated with the major vegetation gradient. The secondary gradient coincided with the concentration of nitrate and potassium, but was not apparent in the bryophyte subset. When only data for vascular plants were analyzed, the major gradient reflected increasing number of species from poor to extremely-rich fens, including ubiquitous grassland species, and only partially coincided with pH and calcium. Contrary to expectations, neither the extremely rich or rich fens were associated with low concentration phosphorus in the water. In addition, particular vegetation types did not differ in the N:P ratio of bryophyte biomass. Species composition of extremely rich fens thus seemed to be determined predominantly by a high pH/calcium level and waterlogging, low iron concentration and absence of sphagna that would hamper regeneration of some competitively weak vascular plants. We demonstrated that the delimitation of the major vegetation types (alliances) along the poor-rich gradient makes great floristic and ecological sense also in the Hercynian Mountains and that pH and calcium rather than nutrient availability differentiate causally major vegetation types by determining structure of the moss layer.