ABSTRACT The ability of environmental DNA (eDNA) to provide rapid assessments of mammal taxa composition at the watershed scale can make it an efficient survey method on large‐scale landscapes, complementing camera traps. Due to the rugged and inaccessible terrain of many areas in Bhutan, camera trapping is associated with logistical challenges, increasing the cost of sampling considerably. In this study, conducted in the Upper Punatsangchhu catchment basin of Bhutan, we investigated the ability of eDNA water samples to capture the diversity of terrestrial mammals in comparison with camera trapping, using six watersheds within the basin as a baseline sampling frame. Combined, the two methods detected a total of 72 mammalian species: eDNA metabarcoding identified 60 species, while camera trapping detected 33 species, with an overlap of 21 species between the two methods. In addition, eDNA metabarcoding detected 90% of the IUCN Red List species detected by the camera traps. Small mammals were frequently detected using eDNA metabarcoding, while camera trapping more often detected large mammals. The mean detection probabilities recorded from eDNA were higher for all species grouped by orders and size categories compared with camera trapping. Biodiversity models based on eDNA metabarcoding and camera trapping both retrieved dominant effects of temperature and isolation in structuring the mammal assemblage. We conclude that eDNA sampling based on watersheds accurately represents the spatial distribution of species across each watershed in our study area in Bhutan to provide a rapid assessment of mammals from river water.
Sherub et al. (2026) studied this question.