eDNA as a tool for non-invasive monitoring of the fauna of a turbid, well-mixed system, the Elbe estuary in Germany.

The Elbe is one of the longest European rivers and features a large, turbid and well-mixed estuary, which runs through the inner city of Hamburg. The Elbe has been closely monitored using classical catch techniques in the past. Here we tested a COI-based eDNA approach for assessing the biodiversity within the Elbe. We sampled three stations in the Elbe, included low and high tide events, as well as two adjoining lakes to compare the recovered faunas. To analyze the data, we employed two different pipelines: the automated mBRAVE pipeline utilizing the BOLD database and one including NCBI BLAST. The number of OTUs with species or higher-level identifications were similar between both approaches with 352 OTUs and 355 OTUs for BLAST and mBRAVE, respectively, however, BLAST searches recovered another 942 unidentified metazoan OTUs. Many taxa were well represented; however, fish species were poorly represented, especially in the Elbe estuary samples. This could be a result of the universal COI primers, which also yielded high read numbers for non-metazoan OTUs, and small-bodies taxa like Rotifera, which might have been sampled together with the eDNA. Our results show a strong tidal influence on the recovered taxa. During low tide, downstream stations resembled sites further upstream, but the former showed a very different OTU composition during high tide and early tide. Such differences might be due to varying impacts of upstream-originating eDNA during tide cycles. Such factors need to be considered when routinely employing eDNA for monitoring programs.