Combining eDNA Metabarcoding, Hydrology-Based Modeling and Camera Trap Datasets to Assess the Potential of River eDNA in Monitoring Terrestrial Mammals

Global efforts aimed at safeguarding and restoring biodiversity require methods to monitor progress towards conservation objectives. Such methods should provide a systematic and robust assessment of biodiversity for the lowest cost. River environmental DNA (eDNA) metabarcoding has been successfully applied to measure biodiversity in dendritic riverine habitats and is increasingly used to describe communities of terrestrial vertebrates in ecosystems that are challenging to survey using traditional methods. However, interpreting eDNA surveys in riverine habitats requires an understanding of the influence of eDNA transport, decay, and production on the distribution of eDNA. To this end, the hydrology-based eDITH (eDNA Integrating Transport and Hydrology) model incorporates such factors and can recover reliable spatial biodiversity patterns for aquatic taxa, but its potential to successfully model terrestrial taxa is so far unexplored. Here, we applied eDITH to eDNA metabarcoding data for terrestrial mammals collected over two mountainous catchments (575 and 745 km²) in British Columbia, Canada. We assessed prediction transferability between neighboring catchments and compared model predictions with observations from camera trapping. We found that for 9 out of 15 taxa detected by both eDNA and camera traps, predicted distributions predominantly matched observations from camera trap surveys, illustrating that eDITH can uncover patterns of mammal distribution in mountainous catchments. While lacking knowledge of actual taxon density prevents us from determining whether discrepancies stem from data limitations or complex eDNA production-density relationships, good transferability of predictions to the neighboring catchments suggests that eDNA distribution of some terrestrial and semi-aquatic mammals is partly determined by habitat preference and hydrology. Downstream sampling can recover most biodiversity across the catchment, but the inclusion of upstream samples can aid in detecting elusive species. This study underscores the broader applications of river eDNA beyond aquatic species and illustrates its potential use in addressing terrestrial mammal biodiversity monitoring objectives with tailored sampling approaches.