Testing for differences in beta diversity with asymmetric dissimilarities

In this paper, expanding on a well-known connection between Whittaker's beta diversity and multivariate dissimilarity coefficients, a simple test for differences in beta diversity among different sets of plots is proposed. Being based on any suitable (not necessarily symmetric) plot-to-plot dissimilarity coefficient the test is appropriate for analyzing multivariate data that are usually zero-inflated due to the large number of rare species and for which the number of variables (species) is typically larger than the number of observations (plots). To illustrate the proposed test in practice, we used data from two beech forests with different management history in Abruzzo (central Italy).