Potential of eDNA metabarcoding for the inventory of alpine plants and mammals in the Vallorcine catchment area

Environmental DNA (eDNA) is increasingly used in ecology to detect biodiversity, at relatively low cost. In this study, we evaluate how effectively eDNA reflects mammal and plant diversity in a mountain watershed, using the Nant de Loriaz catchment as a case study. We compared eDNA results with traditional methods such as vegetation plots and camera traps for mammals, tested differences between water and sediment samples, and inspected how eDNA signals change along an upstream-downstream gradient. Our results show that eDNA reflects only part of the diversity detected by traditional surveys, with some species detected only by eDNA and others only by traditional methods. This comparison highlights the complementarity of both approaches, with eDNA improving detection of elusive or hard-to-observe species, while traditional surveys provide more reliable species-level identification for mammals. Sediment samples mostly reflect local communities and may retain DNA for longer periods, providing a snapshot of species that persist in a given location. In contrast, water samples integrate eDNA from both local aquatic sources and upstream inputs, combining signals from nearby and more distant communities. Along the upstream-downstream gradient, we found no clear nestedness pattern; instead, high species turnover indicates that eDNA primarily reflects local communities shaped by environmental conditions rather than passive downstream transport. Our results show that eDNA can improve the precision and resolution of biodiversity identification, but its performance depends on environmental conditions and context. These findings highlight both the potential and the limitations of eDNA for informing conservation decisions.