CHANGES IN HABITAT CONDITIONS OF INVADED FOREST COMMUNITIES IN PODUNAJSKá Ní?INA AND THE IMPACT OF NON-NATIVE SPECIES ON BIODIVERSITY (SW SLOVAKIA)

摘要

With the phytocenological research of riparian mixed forests and pannonic woods in the Podunajská ní?ina in 2018 and 2019, we obtained 113 phytocenological relevés of invaded and non-invaded stands. Using the Modified Permutation Test, we derived modified Ellenberg indication values (EIVs) in order to determine statistically significant differences and correlation relationships between diversity indices and EIVs. We recorded 15 invasive species in the stands, for example, Ailanthus altissima, Ambrosia artemisiifolia, Asclepias syriaca, Aster lanceolatus, Helianthus tuberosus, Impatiens parviflora, Negundo aceroides, Solidago canadensis, S. gigantea and Robinia pseudoacacia. We also confirmed the occurrence of species important for conservation: Epipactis helleborine agg., E. voethii, Cephalanthera longifolia, Clematis integrifolia and the like. We tested the statistical significance of EIVs as explanatory variables by redundancy analysis and Monte Carlo permutation test (stepwise selection, number of permutations 499, p < 0.05). In addition to the EIV continentality, the contribution of light, moisture, nutrients, soil reaction and temperature was higher than random. The model explains 20.21% of the variability pursuant to the coefficient of determination (R~2), with its adjusted (more accurate) variant (R~2_(adj)) capturing 15.70% of the variability of the model. Using the Kruskal–Wallis test, we confirmed the statistically significant differences (p*) in mean values between count of species, Shannon–Wiener index, Simpson dominance index and taxonomic diversity of invaded and non-invaded vegetation. For EIVs, we found significant differences in the mean values for moisture and soil reaction factors. In our model, count of species, Shannon–Wiener index, Simpson dominance index and taxonomic diversity index were positively correlated with invaded stands. Non-invaded stands showed statistically significant negative correlation with the EIVs moisture, soil reacti