The role of sampling effort, taxonomical resolution and abundance weight in multivariate comparison of stream dwelling caddisfly assemblages collected from riffle and pool habitats

We examined how the sample representativeness of a single assemblage and the separation of two assemblages can be improved by finding appropriate combinations of sampling effort, taxonomical resolution and abundance weight for stream dwelling caddisflies as model organisms. We found that these parameters strongly influenced the outcome of multivariate analyses both individually and when considered jointly. This is the first study to show that sample representativeness of an assemblage can decrease with increasing sampling effort. We assume that the turnover rate of the assemblage and the rarity of the species are responsible for this phenomenon. We found that the separation of two assemblages can be improved by increasing sampling effort and applying abundance data. Further, we observed that the effect of taxonomical resolution on the separation of ecological assemblages was highly context-dependent. Decreased taxonomical resolution, i.e. changing from species to genus or to family, did not decrease, or even more increased the separation of assemblages. In sum, this study demonstrates the importance of the careful selection of sampling, laboratory and data processing related factors (i.e. sampling effort, taxonomical resolution, abundance weight) in the multivariate comparison of assemblages. (C) 2010 Elsevier Ltd. All rights reserved.