Summary
A bioRxiv preprint comparing environmental DNA sampling with camera trapping found that water samples detected established coypu populations more reliably with less field effort in southern France. The researchers propose using eDNA for rapid screening and camera traps for information about animal activity and habitat use.
A comparison of two wildlife-monitoring methods found that environmental DNA (eDNA) detected established populations of invasive coypu more reliably than camera trapping in Mediterranean river systems in southern France. The researchers propose using rapid eDNA screening alongside targeted camera trapping rather than treating the methods as interchangeable.
The findings come from a bioRxiv preprint posted on September 15, 2026. The study used replicated observations and Bayesian generalised linear mixed models to estimate the probability of detecting coypu when the animals were present.
eDNA produced the higher detection probability
Environmental DNA monitoring looks for genetic material that animals leave in their surroundings. In this study, researchers collected water samples and compared their detection performance with camera traps, a non-invasive method that records animals passing through selected locations.
A single eDNA water sample had an estimated detection probability of 93.0%, with a 95% credible interval of 72.1% to 98.9%. The corresponding estimate for 30 camera-trap days was 52.8%, with a 95% credible interval of 22.6% to 91.1%.
The model assigned a 96.9% posterior probability to eDNA having the higher detection probability. These estimates were conditional on coypu being present, so they describe how reliably each method detected an established population under the study conditions.
Repeated sampling increased the chance of detection. Two eDNA samples were sufficient to produce a median cumulative detection probability above 95%. Reaching the same median level with camera trapping required four 30-day camera-trapping periods.
The researchers also found that detectability varied by spatial scale. Camera-trap performance differed substantially among local camera locations, while eDNA detection varied more among river catchments. That pattern indicates that the choice and placement of sampling sites can influence results for both methods.
Camera traps add information that eDNA cannot provide
The comparison does not make camera trapping redundant. A water sample can provide evidence that coypu genetic material is present in a catchment, but camera traps can record activity, behaviour and habitat use. Those observations can help wildlife managers understand how animals are using an area after eDNA has identified a population.
This gives the two methods different roles in a surveillance programme. eDNA could be used to screen river systems quickly and identify locations requiring closer investigation. Camera traps could then be concentrated at selected sites to study activity patterns, behaviour and habitat use.
The study examined established coypu populations in Mediterranean river systems in southern France, and the abstract describes replicated data but does not report the number of samples or sampling locations. The findings therefore provide a comparison for the stated setting rather than a universal performance estimate for every habitat or for newly arriving populations. As a bioRxiv preprint, the work is presented before formal journal publication.
Coypu (Myocastor coypus) is an invasive semi-aquatic rodent in the study system. For invasive mammals that are difficult to observe directly, a method requiring fewer field deployments could help prioritise surveillance effort while retaining camera trapping for ecological information that genetic detection alone cannot supply.