The effect of excluding taxa with low abundances or taxa with small distribution ranges on ecological assessment.

The present study aims to investigate whether taxa with a small distribution range or taxa with low abundances indicate specific habitats or a high ecological quality and what the effect is if these taxa are excluded from ecological assessment. We compared autecological features between stream dwelling taxa with a mean abundance >5 individuals per sample and a mean abundance \le5 individuals per sample as well as between taxa with a small distribution range and taxa with a large distribution range. The number of rare taxa (either with a small distribution range or with low abundances) in a sample was related to the ecological quality classes. To test the effect of exclusion of rare taxa we constructed 8 data sets all including 142 samples of Dutch lowland streams. From each data set we stepwise excluded taxa that had low abundances or taxa that were known to be restricted in their distribution range. With help of the AQEM assessment software we calculated the final ecological quality classes and the metrics that were included in the multimetric for the original data and the 8 selected data sets. Autecological features of the taxa within the different selections showed that taxa with small distribution ranges were often running water taxa, living on stones and gravel and indicating oligosaprobic water conditions in contrast to taxa that had a large distribution range. There were only small differences between taxa with low and high abundances. However, current velocity preference was lower for taxa with abundance \le5 individuals per sample, saprobic values were higher and scores for typical stream habitats, such as lithal, psammal and akal were lower compared to high abundant taxa. If taxa with low abundances were excluded a higher ecological quality class was achieved in most cases, while excluding taxa with a small distribution range resulted in lower ecological quality classes. In conclusion, excluding taxa with a small distribution range led to worse ecological quality classes because these taxa have special autecological features that often indicate natural streams. On the other hand, excluding taxa with low abundances resulted in higher ecological quality classes because these taxa indicate more disturbed situations and because the number of taxa per sample was strongly reduced. Although the documentation of rare taxa (either with low abundances or with small distribution ranges) is often time and cost-intensive regarding field work, laboratory work, data processing, and analyses, the indicative power of these taxa for natural circumstances is essential and therefore rare taxa should be included in ecological assessment studies. [ABSTRACT FROM AUTHOR]