Similarity indices, ordination, and community analysis tests using the software R

Abstract

Similarity indices are frequently used in ecological analyses to compare community structure between ecosystems and experimental treatments. However, many ecological analyses have used classical similarity indices, even though new similarity indices and statistical tests have been proposed. In this review, we introduce representative and useful similarity indices and statistical tests for analyzing ecological communities. We recommend using the statistical software R with the "vegan" package to analyze similarity indices. We synthesized the frequency with which similarity indices have been used for ecological studies and found that classical similarity indices, such as the Jaccard index, have been used by researchers in the past decade. Here, we emphasize that the Chao index is one of the best options among similarity indices when considering the probability of rare species in communities. To test for differences in community similarity, PERMANOVA is one of the best tests for ecological analyses. Ecologists should use adequate similarity indices and tests.