Assessment of vegetation species diversity based on the integrated statistical approach for heterogeneous data (a case study for the "Kaluzhskie zaseki" State Nature Reserve)

This paper deals with the problem of correct estimation of species diversity in heterogeneous vegetation data. We used a set of 722 phytosociological relevés from "Kaluzhskie zaseki" Reserve as a data example. Our main task was to correct the previously made estimations of plant diversity in the Reserve using modern statistical methods. The former analysis did not take into account a different number of relevés referenced to plant community types which had been determined earlier. To solve the problem we used the integrated statistical approach proposed by Colwell et al. (2012). The approach allowed us to obtain comparable interpolated or extrapolated estimates of species richness for different sample sizes on the basis of the unified theoretical framework and the species accumulation curves conception. Curves were built for 10 most represented community types. We carried out a comparison of species richness estimates resulted from the curves with the help of confidence intervals (95%) because a proper statistical test is not yet developed. Besides the species richness (gamma diversity), we also assessed alpha and beta diversity of vegetation. Means and medians of numbers of species per plot were calculated for the alpha diversity assessment of the community types. We assessed beta diversity using the following quantitative measures: (1) simple Whittaker’s index, (2) the average within-group Jaccard distance and (3) the index proposed by McCune and Grace (2002). Interpolated and extrapolated species richness was obtained in the EstimateS program; all other calculations were performed with the R statistical software. Analysis of species richness revealed two main groups of community types and two transitional ones. The least obvious result obtained by the statistical analysis was the same level of species richness in all communities dominated by nemoral plants in the ground layer. The earlier calculated higher species richness in broadleaved forests (dominated by Quercus robur, Fraxinus excelsior, Ulmus glabra, Tilia cordata, Acer platanoides and A. campestre) in comparing with forests dominated by Populus tremula, Betula spp. or Pinus sylvestris in the overstorey and by nemoral plants in the ground layer was determined by a different number of relevés referenced to these communities. Communities dominated by meadow and meadow-edge species in the ground layer are in the lead in terms of plant species richness. Our results confirmed that assessment of alpha, beta and gamma diversity of vegetation should be conducted by methods less dependent from sample sizes. We recommend using Colwell et al. (2012) approach for species richness compatible estimations, Jaccard and McCune indices for beta diversity measuring. Acknowledgments: this work was supported by Russian Foundation of Basic Research, grant No. 13-04-02181-а The article contains 2 tables, 5 figures, 22 ref.

Keywords

Authors

References