Using environmental DNA to model occupancy of an endangered freshwater fish in a regulated river catchment

Context Detecting rare freshwater species in regulated river systems is challenging, especially when populations are fragmented and occur at low densities. Environmental DNA (eDNA) offers a sensitive, non-invasive alternative to traditional sampling methods. Aims To estimate the distribution and occupancy of an endangered native fish species (Macquaria australasica), compared with a widespread introduced species (Oncorhynchus mykiss), and to identify habitat influences on eDNA detection. Methods Water samples were collected from 21 sites across the Cotter River (Australian Capital Territory, Australia) and screened using species-specific eDNA assays for M. australasica and O. mykiss. Detection data were analysed using a multi-stage occupancy model to account for false-positive and false-negative errors. A generalised linear mixed model was used to assess environmental drivers of detection. Key results O. mykiss was detected at all sites and showed higher occupancy and eDNA concentrations than did M. australasica. Detection of M. australasica was lower in upper catchment sites and more variable across the river system. eDNA detections of M. australasica at translocation and upstream sites suggest possible persistence or expansion. Conclusions eDNA combined with occupancy modelling effectively shows spatial patterns in species detection. Implications These findings demonstrated the value of eDNA and occupancy modelling for monitoring rare species.