Effect of intra‐seasonal variability on vegetation data

Abstract

Questions It is generally acknowledged that the date of sampling partly determines the floristic composition and layer covers of vegetation plots. However, the effect of vegetation seasonality on the results of vegetation analyses is still insufficiently explored. Here, we investigated two data sets to examine how intra‐seasonal variability of vegetation influences the pattern of: (1) species composition; (2) species richness, number of unrecorded species, plant life‐form spectra; and (3) quantitative changes in cover of individual species and layers. Location Deciduous forests and dry grasslands in SE Czech Republic. Methods We established 40 forest and 46 dry grassland permanent plots, each sampled in three periods (spring, summer and autumn). We compared species composition and β‐diversity patterns in multi‐dimensional space between different periods of the growing season using permutational multivariate analyses (PERMANOVA, PERMDISP). We used ANOVA for repeated measures and Tukey post‐hoc tests to assess differences in species richness, proportions of unrecorded species, changes in layer cover and plant life‐form spectra. Results In the Forest data set, spring plots significantly differed in their species composition from summer and autumn plots, while we found significant differences between different periods of the growing season in the Dry grasslands data set. We confirmed, that the forest species richness was highest in summer plots, while the α‐diversity in spring and summer dry grassland plots were comparable and notably declined in autumn. Although most plant life forms were best recorded in summer, both forest geophytes and dry grassland therophytes were best recorded in spring. Conclusions We suggest that before analysing vegetation plot data, large databases should be carefully stratified according to date of sampling. Based on our results from temperate broad‐leaved forests and dry grasslands, we recommend excluding plots sampled before June, as they might be a potential source of misinterpretation of species composition analyses (with the exception of studies focused on vernal species). The use of autumn plots does not have any strong effect on the results, except in α‐diversity analyses in forest plots, while dry grassland plots are significantly different from the rest of the growing season.

How does vegetation sampling in different parts of the growing season influence classification results and analyses of beta diversity?

“…; Vymazalová et al. ). Summer‐recorded plots (June to August) constitute the majority of plots in the Czech National Phytosociological Database (Chytrý & Rafajová ), where they exceed 70% in most vegetation units.…”

“…In the Forest data set, this was probably caused by lower species richness than in summer (Vymazalová et al. ), and consequently, lower representation of the faithful species with high fidelity (Bruelheide ; Chytrý et al. ).…”

“…The occurrence of many vernal species across a wide range of vegetation units (Chytrý ; Vymazalová et al. ) might have led to more homogeneous records, which resulted in the lowest beta diversity in the Dry Grassland data set being recorded in spring (Figs ). This agrees with findings of Otýpková & Chytrý () that different plot sizes had a larger effect on the position of plots in ordination space in cases where the data set consisted of more homogeneous vegetation (in comparison with the minor effect of the heterogeneous data set).…”

“…) or different sampling dates within the growing season (Magurran ; Vymazalová et al. ). However, the last issue still has to be properly addressed, especially the seasonal variability of vegetation.…”

“…; Vymazalová et al. ) systematically change during the growing season. The basic patterns of alpha and beta diversity among vegetation data recorded in different parts of the growing season have been studied for some vegetation types (Lososová et al.…”