Ecoregions and benthic macroinvertebrate assemblages of Swedish streams

A total of 694 streams was sampled for benthic macroinvertebrates in the autumn of 1995 as part of the Swedish national stream survey. After removal of sites considered to be impacted, 428 streams were available to test a geographical classification of benthic macroinvertebrates using ecoregions. Cluster analysis, semi-strong hybrid multidimensional scaling, and a permutation test procedure were used to investigate if the 6 main ecoregions of Sweden could be a useful tool to partition the species variance of benthic assemblages. Taxon richness, abundance, and diversity differed between ecoregions, with the main difference occurring in the northern part of the country where the arctic-alpine region had lower taxon richness, number of individuals, and Shannon-Wiener diversity than the southern parts of the country. The permutation test of assemblage composition and the comparison of the flexible beta unweighted pair-group method using arithmetic averages (UPGMA) cluster groups with the ecoregion classification also showed significant differences between the arctic-alpine and the boreal, boreonemoral, and nemoral regions. Despite this result, the benefit of an ecoregion classification for biomonitoring of streams using macroinvertebrates is not convincing because there is a gradual change in macroinvertebrate assemblage composition from north to south. Benthic macroinvertebrate stream assemblages are structured both by large-scale factors (i.e., on a geographical scale), and by small-scale factors (i.e., on a local scale). Ecoregion classifications alone, therefore, may not sufficiently partition variance in assemblage composition and a nested approach, including other factors such as altitude stream size, and catchment characteristics, is probably needed to improve ecoregion classifications and biological assessments that use stream benthic macroinvertebrates.