Refine
Document Type
- Article (2) (remove)
Language
- English (2)
Has Fulltext
- yes (2)
Is part of the Bibliography
- no (2)
Keywords
- CCA (2) (remove)
A long-term systematic survey of grassland communities was performed in the Biele Karpaty Mts. in Slovakia. The main aims of the research were i) syntaxonomical classification of meso- and subxerophilous grassland vegetation, ii) analysis of the main gradients in species composition, iii) evaluation of the effect of environmental factors on species composition of grasslands. The data set included 342 phytosociological relevés of grasslands recorded between 1991 and 1999. For the classification of relevés to associations, the expert system for identification of grassland vegetation of Slovakia was used. The main environmental gradients of species composition were analysed by detrended correspondence analysis (DCA). For the ecological interpretation of ordination axes Ellenberg indicator values were used. The relationship between species composition and environmental factors (geology, pedology, climate, topography, management) was analysed by canonical correspondence analysis (CCA). The expert system identified (according to association definitions) 220 phytosociological relevés (64% of the whole data set). Grassland communities were classified within seven associations belonging to four alliances and three classes: Festuco-Brometea: Bromion erecti and Cirsio-Brachypodion pinnati; Molinio- Arrhenatheretea: Arrhenatherion; Nardetea strictae: Violion caninae. The results of the DCA support our assumption that the main environmental gradient in species compositions of grasslands is related to moisture and soil reaction (content of CaCO3 in the soil). The results of the direct gradient analysis (CCA) show that all 23 environmental variables explained 16.15% of the variability of the species data. The most important factors affecting the data variation were precipitation and geological bedrock.
A systematic survey of grassland communities in central Slovakian sub-montane and montane regions (including the Kremnické vrchy Mts., Starohorské vrchy Mts., Veľká Fatra Mts., and Zvolenská kotlina Basin) was performed between 1996 and 2007. The main aim was to identify main environmental gradients in the studied vegetation and to estimate the most important individual variables responsible for the variation of their species composition. Along with the floristic composition, the environmental variables were either recorded in the field (altitude, slope, aspect), calculated (solar radiation, climatic data, and phytochorological affinity), or derived from available maps or GIS digital data layers (type of bedrock, soil parameters). These environmental variables were used as supplementary in the detrended correspondence analysis (DCA) or explanatory in the canonical correspondence analysis (CCA). The affiliation of individual phytosociological relevés to associations was estimated by an electronic expert system for Slovak grassland communities. Altogether, 15 xero-, sub-xero- and mesophilous grassland associations were distinguished. Wet and fen meadows were analysed at the level of alliances. Unconstrained ordination revealed moisture and nutrient gradients as most important for the data set. By means of constrained ordination, the variability of the studied vegetation could be explained by a set of geological, topographic, phytochorological and derived climatic variables, although the percentage of explained variance was rather low and did not exceed 12% for all significant factors combined. Among individual variables, the geological bedrock type, climatic water balance, solar radiation, and slope played the most important role in determining the distribution and variability of individual grassland communities. Affinity to phytochorions determined according to local air temperature gradients was also significant. Soil properties played only a subordinate role in our analyses. The analysis of a more homogeneous subset of the data without wetland relevés gave similar results as the analysis of the complete data set. The differences in results of constrained and unconstrained ordinations are discussed together with the potential reasons for extremely high proportion of unexplained variance revealed by the variation partitioning methods.