Search

ABSTRACT Worldwide, freshwater vertebrate populations are declining with increasing pressure on rivers due to numerous environmental and climatic threats. Environmental DNA (eDNA) could potentially provide a more efficient and non‐invasive mechanism to monitor freshwater systems, either as a complement or in replacement to traditional methods to accurately assess species' distributions. Here, we utilize a hierarchical multispecies N‐mixture model to compare three fish sampling methods: traditional fyke netting and active and passive environmental DNA sampling along a 30 km stretch of the Canning River in Western Australia. We used the fitted model to compare capture probabilities among sampling methods and reveal the sampling effort required to describe the species assemblage. Results indicated that while all methods could detect fish, combined eDNA methodologies detected one more fish species than those caught by fyke netting. In addition, active eDNA sampling produced the highest capture probabilities and more consistently described the entire fish assemblage at any given site. Fyke netting and passive eDNA did not show significant differences in their average capture probabilities, and both methods had lower abilities to capture individual species than active eDNA. Active eDNA also required fewer replicate samples to detect the expected observed richness, and fyke netting required the most replicates. Additionally, a hierarchical multispecies abundance model showed that active environmental DNA (eDNA) sampling is the most effective method for monitoring freshwater fish populations. This study contributes to our understanding of eDNA in aquatic systems and demonstrates that, at least under current conditions, active sampling is still the preferred method in freshwater systems with low flow compared to both passive sampling and fyke netting.

ABSTRACT Monitoring alien species is critical to their management. However, early detection of invading alien freshwater fish can be challenging due to the difficulty of observing fish in low abundance. Environmental DNA (eDNA) has emerged as a new and potentially more sensitive method for sampling invasive species as compared to conventional methods, but the comparative financial cost is not often assessed. Adoption of eDNA by managers requires studies that showcase its cost‐effectiveness relative to conventional approaches. Here we use eDNA to assist in the management of an aggressive alien fish, the pearl cichlid (Geophagus brasiliensis), that is invading an urban river in south‐western Australia. We applied an occupancy model to survey data collected 6 years apart (2015, 2021) to assess how the species' distribution had changed and to evaluate whether an instream barrier had the potential to limit upstream invasion. To understand the effectiveness of eDNA, we used our model to quantify the relative efficiency (capture probability) of two eDNA sampling methods (active eDNA and passive eDNA) and fyke netting, as well as the number of replicate samples required per site to deliver >95% detection. We coupled the number of replicates needed with the cost per replicate to determine the cost‐efficiency of each method. We found that G. brasiliensis abundance was higher in downstream reaches in both survey years, and there was no evidence that its distribution had changed through time. However, G. brasiliensis was present above the instream barrier. Active eDNA sampling was considerably better at detecting G. brasiliensis than the other methods, making it the most cost‐effective method. Fyke nets came in a close second, and passive eDNA was a very distant third. Our results directly inform management in the study river and broadly highlight the cost‐effectiveness of active eDNA as a freshwater biosecurity tool.