Does taxonomic resolution affect the multivariate description of the structure of freshwater benthic macroinvertebrate communities?

Abstract

Ten data sets from published studies of freshwater benthic macroinvertebrates were used to deduce the necessary level of taxonomic resolution (from genus to phylum) for assessing variation in community structure. Quantitative data sets describing communities of benthic invertebrates identified to genus were aggregated to family, order, class, and phylum. Bray-Curtis distances between each pair of communities in a given study were calculated at each level of taxonomic aggregation. Also, quantitative data were converted to qualitative (presence-absence) data and aggregated in the same manner. Jaccard's distances between communities were calculated for these data matrices. Correlations between the resulting distance matrices, calculated at different taxonomic resolutions, were quantified using Mantel's test (N. Mantel. 1967. Cancer Res. 27: 209-220). Distance matrices for different taxonomic resolutions were highly correlated (r >> 0.64) when based on quantitative data for all except 1 of the 10 studies. Correlations were lower and more variable for qualitative data sets. For the data sets investigated, genus-level identification did not usually provide a strikingly different description of community patterns than higher levels (e.g., family, order) of taxonomic identification.