Por que muitos estudos encontram uma alta congruência taxonômica?: consequências para os estudos de comunidades

Abstract

The way species are distributed among the highest taxonomic classifications can cause correlations between the community response with data identified in species or higher taxonomic levels, due to the taxonomic hierarchy, which generates dependence between the taxonomic levels. As the data identified in genus are dependent on the data identified in kind we evaluate whether it is the taxonomic structure or if some other structure presented by the data is able to generate "taxonomic" congruence between the levels of taxonomic resolution. We used data from 25 biological groups to test the degree of congruence between the taxonomic levels using Mantel and Procrustes for the observed data and for the genus data aggregated from the random assignment of species. We also evaluated some underlying mechanisms that could be responsible for generating congruence among the taxonomic levels: a) the weights given to the species; b) the effect of the most common species; c) variance in species abundance; and d) variation in species occurrence. More than half of the groups evaluated have a congruence between species and genus data (r ≥ 0.7). However, the congruence was also satisfactory even when the species were randomly assigned to the genera. The main factor capable of generating high congruence is the effect that the species with greater variance in abundance, and to a lesser extent due to the species considered dominant. That is, regardless of which genus these species are randomly allocated, they end up determining the final pattern of ordinations (Procrustes) and matrices of distance (Mantel). This prevalence few species end up driving the results of the ordinations, making the patterns of community analysis reflect only the strongest pattern that is contained in the distribution of occurrences and / or abundance of a few species.