Sunbury Lock Case Study

Fish communities obtained from monthly eDNA samples on the River Thames were compared with Environment Agency electrofishing records in the nearby area.
eDNA detected more species per unit effort and provided consistent data on a number of important species that are routinely under-sampled in electrofishing.

THE CHALLENGE OF TRADITIONAL SURVEYING METHODS

Electrofishing is the primary method of conducting fish surveys in rivers and is routinely conducted as part of the Environment Agency’s monitoring. This requires expensive equipment and trained operators, which limits the frequency and geographical coverage of surveys. It is also invasive – fish are temporarily stunned and float to the surface where they can be counted and identified. Although the fish recover, this inevitably causes stress to the animals. Moreover, some species are less susceptible to the electric currents and so tend to elude electrofishing surveys – this particularly applies to species like loaches, sticklebacks and bullhead which live at the bottom of the river amongst the stones where they are somewhat shielded.

METHODOLOGY

eDNA samples were collected from our long-term monitoring site on the River Thames at Sunbury Lock. One sample is collected each month by members of the NatureMetrics team and analysed using the Aquatic eDNA Fish assay.
Environment agency electrofishing data was downloaded for the closest locations on the river. These covered 37 different surveys from 9 different sampling locations between 1995 and 2013.

RESULTS

The 8 eDNA samples detected a total of 19 fish species with an average of 13.9 species per sample.
The 37 electrofishing surveys detected a total of 17 species (plus a hybrid) with an average of 8.5 species per survey.
All species consistently detected by electrofishing were also consistently detected by the eDNA survey, including European eel (Anguilla anguilla).
Both methods showed roach and perch to be the most dominant species in terms of relative abundance of sequence counts (eDNA) and individuals (electrofishing).
Several species rarely or never detected by electrofishing were detected in multiple eDNA samples. This includes European bullhead, lamprey, stone loach, ruffe, three-spine stickleback and minnow. While it is possible that this is at least partly due to temporal changes in the fish community (the most recent electrofishing survey is from 2013), several of these species are small in body size and live at the bottom of the river, making them less susceptible to electrofishing.
The only (non-hybrid) species absent from the eDNA data that had been recorded by electrofishing within the previous 25 years was Atlantic salmon (Salmo salar). Two individuals were recorded in a survey in 2002 and a single individual in 2008.

Larger Version of Data

Figure 1:Bubble plot showing the species detected in 8 eDNA samples (blue header) and 37 electrofishing surveys (yellow header) close to Sunbury Lock on the River Thames. The size of the bubble indicates the proportional abundance of species within each sample, based on sequence counts (eDNA) or number of individuals (electrofishing).

CONCLUSIONS

eDNA is a powerful, non-invasive method for fish surveys, which appears to perform at least as well as electrofishing for community assessment and matches the main conclusions in terms of species relative abundances. Although surveys were conducted at different times, our data suggests that eDNA is more effective than electrofishing at detecting the presence of species such as stone loach, sticklebacks and bullheads, and this is a pattern that we have also seen in other datasets.

Our data is consistent with results published in the scientific literature, which show that eDNA outperforms electrofishing for fish surveys in streams and rivers (e.g. Goutte et al., 2020; McColl-Gausden et al., 2020).

REFERENCES:

Goutte, A., Molbert, N., Guerin, S., Richoux, R. & Rocher, V. (2020). Monitoring freshwater fish communities in large rivers using environmental DNA metabarcoding and a long-term electrofishing survey. Journal of Fish Biology, 97(2): 444-452
McColl-Gausden, E.F., Weeks, A.R., Coleman, R.A., Robinson, K.L., Song, S., Raadik, T.A. & Tingley, R. (2020). Multispecies models reveal that eDNA metabarcoding is more sensitive than backpack electrofishing for conducting fish surveys in freshwater streams. Molecular Ecology, 30(13): 3111-3126