Comparison of camera traps, eDNA, and visual encounter surveys for threatened species detection

Accurate, cost-effective monitoring remains a major limitation to conservation efforts for many wildlife species. The USA federally threatened eastern indigo snake (Drymarchon couperi) exemplifies this sampling challenge. They occur at low densities, are cryptic, and highly mobile, which has historically made them challenging and expensive to monitor. However, emerging technologies and methodologies may provide new pathways to augment, complement, or replace conventional sampling to yield improved monitoring programs, both for eastern indigo snakes and for other imperiled and cryptic taxa. Here, we compare environmental DNA (eDNA) sampling, camera trapping, and visual encounter surveys as monitoring methods for the eastern indigo snake. We conducted an in situ study of eDNA decay rate in soil. Subsequently, we paired active searches across four sites in the southeast United States, eDNA sampling of gopher tortoise burrows and drift fences, and camera trapping of gopher tortoise burrows and drift fences and used these data to (1) assess eDNA detectability over time and (2) relative cost-effectiveness of the different survey methods. We found patterns of eDNA detectability that were concordant with previous experimental trials. Of 120 samples collected from locations with confirmed snake presence via visual encounter surveys, 66 amplified (55%) and eDNA was detectable up to six days after snake presence. Using multi-scale, multi-method occupancy modeling we estimated that the probability of eastern indigo snake presence in gopher tortoise burrows (0.29) was higher than at drift fences (0.24) and that cameras had a higher rate of detection (0.50) than eDNA sampling (0.38), although image processing time made camera trapping prohibitively expensive relative to other approaches. To reach a 95% likelihood of detection, the most cost-effective sampling method is visual encounter surveys augmented with eDNA sampling. Our results illustrate that visual encounter surveys remain an effective monitoring method, but supplementing with eDNA may decrease costs and increase detection probability.