Prey partitioning between sympatric canid species revealed by DNA metabarcoding

The recovery of apex predators relies on restoring the full suite of trophic interactions within the ecosystem. Diet analysis with DNA metabarcoding technology can help deliver insights into these trophic interactions with fine-grained resolution. The recovery of wolves in Washington state offers an excellent case to study the trophic cascade impacts of the apex predators on the ecosystem and explore prey partitioning between sympatric canid species. We used DNA metabarcoding technology on scats to characterize the diet composition and its spatiotemporal variations of wolves and coyotes and quantified the diet niche overlap between these two canid species in northeastern Washington. In total, 19 different prey taxa were detected. Frequency of occurrence data showed that wolves primarily preyed upon deer ( Odocoileus sp. ) (47.47%) and moose ( Alces alces ) (42.42%). Coyotes also consumed moose (30.10%) and deer (21.36%), but snowshoe hares ( Lepus americanus ) were the most common prey (61.17%) in their diet. There were significant spatial variations in the wolf diet composition ( p = 0.001) with wolves in the Dirty Shirt pack range consuming more moose (71.43%). Coyotes showed significant spatial and temporal dietary variations (season: p = 0.037; pack: p = 0.003; pack:season p = 0.043). Our data suggested that coyotes use ungulate carrion subsidies from wolves as food resources. DNA metabarcoding with fecal DNA provides an excellent noninvasive tool to characterize diet profile at the fine-grained level and can be applied to other carnivore species to help understand the impacts of recovery of apex predators on the local ecosystems.