Fig 1.
Biplot of DCA with samples classified (and outlined by envelopes) according to plot assignment (A, B or C) and year of sampling, and species with the highest weight (layers of woody species were combined into a single one). Percentage cover values were transformed with a logarithmic function. Eigenvalues of the first two axes were 0.727 and 0.211; total inertia 4.789. Vectors of supplementary variables are shown in the upper right corner: P–Plain; LD–Low dunes; HD–High dunes; PD–Plateau dunes; SR–Species richness; H’–Species diversity of vascular plants; Age–successional age; Year–year of sampling.
Fig 2.
Different successional stages at the youngest plot A in 2015.
Photo by K. Prach.
Fig 3.
A rather monotonous pine forest at plot B in 2010.
Photo by K. Ujházy.
Fig 4.
Part of plot C with the oldest (ca. 200 years) pines and deciduous trees, mostly beech (Fagus sylvatica) in 2015.
Photo by K. Ujházy.
Fig 5.
Relationship between vascular plant species diversity (H’: Shannon-Wiener index) and successional age in 100 m2 plots.
Fitted by 3rd polynomial function (black solid line: r = 0.719; p < 0.001) with 95% confidence interval (black dashed lines). Symbol colors correspond to those used in Fig 1 for plots and years of sampling.
Fig 6.
Response curves of woody species in the combined tree and shrub layer (a; up) and for woody species in the herb layer (b; down). Percentage cover values estimated in 100 m2 plots were used as a response variable and the Poisson distribution was set using a GAM model. Names of species are abbreviated as follows: Bet. p.–Betula pendula; Fag. s.–Fagus sylvatica; Ile. a.–Ilex aquifolium; Pin. s.–Pinus sylvestris; Pru. s.–Prunus serotina; Que. r.–Quercus robur; Rha. f.–Rhamnus frangula.
Fig 7.
Repeated vegetation maps of the successional pine forest in plot C (200 × 200 m).
Vegetation types were delimited according to the dominant species of the herb layer. The sampling grid of 10 × 10 m subplots is shown.
Table 1.
Percentages of vegetation types (mapping units) in plots (A, B, C) in sampling years (1988, 2003, 2018).
–missing data; 0.0 negligible values.