Different aggregation approaches in the chironomid community and the threshold of acceptable information loss
MetadataShow full item record
Due to the problem of identification, Chironomidae larvae, although very abundant, are often avoided or not properly used in bioassessment programs. The aim of this work was to test how different aggregation processes-taxonomic resolution and the random aggregation approach (best practicable aggregation of species-BestAgg) affect the analysis of chironomid communities regarding any information loss. The self-organizing map method, together with classification strength analysis and Spearman's rank correlation, revealed that the genus-level and BestAgg-abundance matrix most accurately approximated the species-level community pattern. The subfamily-level dataset was ineffective at presenting the chironomid community structure, with a substantially lower concordance with the species-level dataset. The biologic environmental gradients analyses presented the same set of important environmental variables for the species-level, genus-level, and BestAgg-abundance matrix. The indicator values analysis showed that indicator genera provide information very close to that gained from species indicators. According to our results, the numeric relationship between species and higher taxa influences taxonomic scaling, limiting Chironomidae family aggregation, with acceptable information loss only up to genus level. In addition, the BestAgg approach, with the maximum level of aggregation, properly assesses the community structure and consequently describes environmental conditions.