Quantitative evaluation of species composition dissimilarity within a community and among communities

Abstract

AbstractIt is often necessary to perform quantitative comparisons and assessments of species composition among plant communities under various land use conditions and landscapes, as well as among plant communities at different stages of restoration or degradation in arid grasslands. In landscapes composed of a number of plant communities, in which many quadrats for a survey are set in each community, the species composition in each quadrat is measured using community variables, such as the presence/absence data (i.e., binary variable), biomass, cover, the number of individuals, or the frequency of occurrences (i.e., quantitative variables) of each species per quadrat. In this study, we defined two types of species composition dissimilarity in vegetation surveys: between different communities and within each community which were calculated based on the species composition of each quadrat, using the Bray–Curtis index. The second type dissimilarity occurs incidentally among quadrats (or micro‐sites) even in a community with a relatively uniform structure. Both dissimilarity measures take values from 0 to 1, where larger values indicate larger dissimilarities, and the quadrat‐to‐quadrat variation in dissimilarity also indicates the degree of heterogeneity in the spatial pattern of species composition. Quantifying dissimilarity between and within communities remains a challenge in ecological applications. We proposed a new model to quantify community dissimilarity and applied this model to a grassland vegetation survey dataset. Quantitative variables resulted in more precise measures of community structure than binary variable. The advantages of our model relative to other broadly used community structure metrics based on the Jaccard and Sørensen indices were discussed.