Fit for purpose? Evaluating benthic invertebrate DNA metabarcoding for ecological status class assessment in streams under the Water Framework Directive

Abstract

The ecological state of aquatic ecosystems is systematically monitored using various bioindicators in many countries worldwide. In the European Union, freshwater biomonitoring is the central component of the EU Water Framework Directive (WFD, 2000/60/EC) and currently based on morpho-taxonomic methods. DNA metabarcoding is a novel approach to assess the ecological state fast and efficiently based on organismal DNA signatures and thereby support and upscale biomonitoring. However, compliance of metabarcoding with existing morpho-taxonomic methods must be ensured prior to official implementation. Thus, this study, co-designed by research institutions and environmental agencies, explored necessary key parameters and performed method intercalibration for the implementation of metabarcoding into WFD assessments of running waters. We focussed on benthic invertebrates as the most commonly used bioindicators. We analysed 170 invertebrate samples collected as part of the German federal state WFD routine stream biomonitoring, first via microscopic determination and then using metabarcoding. Our goals were to quantify overlap in i) taxonomic composition and ii) ecological status derived with both methods. For this purpose, we established data harmonisation measures to integrate invertebrate metabarcoding data into the official national WFD classification modules considering abundance and presence/absence data. Our results revealed a high (ca. 70 %) overlap of bioindicator taxa found with both methods. Metabarcoding identified significantly more small invertebrate taxa and detected similar proportions of the important bioindicator ‘EPT’ taxa (mayflies, stoneflies, caddisflies). Despite deviations in some detected bioindicator taxa, the derived ecological status classes were highly correlated between methods, particularly after intercalibration (R2 = 0.74, Spearman rho = 0.86). Regardless of whether we used abundance or presence/absence data, the resulting stream type classifications showed strong agreement. Thus, our study not only demonstrates the consistency of the methods for the stream types analysed but is also the first to operationalise a path to integration of metabarcoding data into the WFD assessment modules based on formal intercalibration guidelines.