Species composition drives macroinvertebrate community classification

Community classification enables us to simplify, communicate, track and assess complex distribution patterns. Yet, the distribution of organisms may not coincide with predefined geographical and environmental boundaries, and therefore, biology itself should be leading the classification. In this study, we showed how to arrive at such a biology-based classification by clustering locations based on similarity in species composition. A hierarchical classification structure allowed for the selection of classification levels that suit multiple scales of analysis. We also showed how to objectively identify the number of clusters present in a dataset based on the distribution of specific indicator species, allowing to identify clear boundaries in species composition on multiple scales. The resulting biology-based clusters were identified and characterized by local and regional environmental conditions, showing the limited explanatory power of these environmental conditions and the added value of taking biology itself as a starting point of the classification. By departing community classification from species composition, the unknown environmental, geographical, and biotic drivers influencing species composition are accounted for.