Direct fishing and eDNA metabarcoding for biomonitoring during a 3-year survey significantly improves number of fish detected around a South East Asian reservoir

Biodiversity has to be accurately evaluated to assess more precisely possible dam effects on fish populations, in particular on the most biodiverse rivers such as the Mekong River. To improve tools for fish biodiversity assessment, a methodological survey was performed in the surroundings of a recent hydropower dam in the Mekong basin, the Nam Theun 2 project. Results of two different approaches, experimental surface gillnets capture and environmental DNA metabarcoding assays based on 12S ribosomal RNA and cytochrome b, were compared during 3 years (2014–2016). Pitfalls and benefits were identified for each method but the combined use of both approaches indisputably allows describing more accurately fish diversity around the reservoir. Importantly, striking convergent results were observed for biodiversity reports. 75% of the fish species caught by gillnets (62/82) were shown by the metabarcoding study performed on DNA extracted from water samples. eDNA approach also revealed to be sensitive by detecting 30 supplementary species known as present before the dam construction but never caught by gillnets during 3 years. Furthermore, potential of the marker-genes study might be underestimated since it was not possible to assign some sequences at lower taxonomic levels. Although 121 sequences were generated for this study, a third of species in the area, that exhibits high endemism, are still unknown in DNA databases. Efforts to complete local reference libraries must continue to improve the taxonomic assignment quality when using the non-invasive and promising eDNA approach. These results are of broader interest because of increasing number of hydropower projects in the Mekong Basin. They reveal the crucial importance to sample tissues/DNA of species before dam projects, i.e. before the species could become endangered and difficult to catch, to obtain more precise biomonitoring in the future as we believe eDNA metabarcoding will rapidly be integrated as a standard tool in such studies.