Comparison of three strategies for local bioassessments in streams using environmental DNA and RNA metabarcoding of macroinvertebrates

Environmental DNA (eDNA) metabarcoding of streams can deliver information on biological communities with minimally invasive efforts and at low costs. However, the captured eDNA may not necessarily originate only from organisms physically present at a local site but also come via drift from sources upstream. Therefore, data on community composition can be inherently different to specimens occurring at a certain site, limiting direct comparisons to site-based assessment data. To increase the local signal, two possible ways were recently proposed: First, targeting eRNA collected from water samples, which degrades faster than eDNA, and second, using eDNA from water, in which organisms from the stream were shortly incubated in a water container. Based on these two options, we here derived and tested three strategies to obtain a more localized signal of biodiversity: 1. eRNA metabarcoding of a direct stream water sample, 2. eDNA and eRNA metabarcoding of water, in which animals collected via kicknet multi-habitat sampling were shortly incubated, and 3. eDNA and eRNA metabarcoding of water, in which natural substrate exposures (NSEs) that were actively colonised by invertebrates in the stream for four weeks were shortly incubated. For comparison, we used eDNA directly isolated from the stream as the more regional signal. As the local point-based sample reference, we used the animal bulk sample that was incubated. While our results support that eRNA collected from the stream (strategy 1) does not improve the local signal, both incubation strategies increased the proportion of species identified by both metabarcoding of eDNA/eRNA and the locally collected bulk sample by about 20%. Interestingly, we detected 20% more species in the NSEs (strategy 3) than in the kicknet samples (strategy 2), highlighting the potential of this less invasive invertebrate sampling approach. In summary, our study demonstrates that incubating collected animals in stream water before filtering the water increases the local community signal and provides an animal-friendly strategy for biomonitoring.