Exploring the potential of metabarcoding to disentangle macroinvertebrate community dynamics in intermittent streams

Taxonomic sufficiency represents the level of taxonomic detail needed to detect ecological patterns to a level that match the requirement of a study. Most bioassessments apply the taxonomic sufficiency concept and assign specimens to the family or genus level given time constraints and the difficulty to correctly identify species. This holds particularly true for stream invertebrates because small and morphologically similar larvae are hard to distinguish. Low taxonomic resolution may hinder detecting true community dynamics, which thus leads to incorrect inferences about community assembly processes. DNA metabarcoding is a new, affordable and cost-effective tool for the identification of multiple species from bulk samples of organisms. As it provides high taxonomic resolution, it can be used to compare results obtained from different identification levels. Measuring the effect of taxonomic resolution on the detection of community dynamics is especially interesting in extreme ecosystems like intermittent streams to test if species at intermittent sites are subsets of those from perennial sources or if independently recruiting taxa exist. Here we aimed to compare the performance of morphological identification and metabarcoding to detect macroinvertebrate community dynamics in the Trebbia River (Italy). Macroinvertebrates were collected from four perennial and two intermittent sites two months after flow resumption and before the next dry phase. The identification level ranged from family to haplotype. Metabarcoding and morphological identifications found similar alpha diversity patterns when looking at family and mixed taxonomic levels. Increasing taxonomic resolution with metabarcoding revealed a strong partitioning of beta diversity in nestedness and turnover components. At flow resumption, beta diversity at intermittent sites was dominated by nestedness when family-level information was employed, while turnover was evidenced as the most important component when using Operational Taxonomic Units (OTUs) or haplotypes. The increased taxonomic resolution with metabarcoding allowed us to detect species adapted to deal with intermittency, like the chironomid Cricotopus bicinctus and the ephemeropteran Cloeon dipterum. Our study thus shows that family and mixed taxonomic level are not sufficient to detect all aspects of macroinvertebrate community dynamics. High taxonomic resolution is especially important for intermittent streams where accurate information about species-specific habitat preference is needed to interpret diversity patterns induced by drying and the nestedness/ turnover components of beta diversity are of interest to understand community assembly processes.