Influence of taxonomic resolution and data transformation on biotic matrix concordance and assemblage-environment relationships in stream macroinvertebrates

I examined patterns in stream macroinvertebrate assemblage structure and assemblageenvironment relationships at four taxonomic levels (i.e.species, genus, family, and order) and based on four data types (i.e.presence/absence, logarithmic, square-root, and raw abundance) in a boreal drainage basin.Tests of resemblance matrix concordance within taxonomic levels showed that not all matrices were strongly correlated.Presence/absence matrices showed poor correlations with raw abundance matrices, possibly reflecting the fact that a few dominant species were responsible for assemblage patterns in the latter.By contrast, logarithmic and square-root data generally showed strong matrix correlations, and this pattern existed at the species, genus, and family levels.Assemblage-environment relationships were rather similar between species-, genus-, and family-level data sets, given that the same key environmental variables were included in the final environmental dissimilarity matrices in BIO-ENV analysis.In conclusion, researchers should pay considerable attention to data transformations when interpreting assemblage patterns at different taxonomic levels and comparing different studies, as different data transformations may provide differing information and lead to highly differing conclusions.