Power to Act: Establishing eDNA Survey Power for Imperiled and Invasive Fishes in Maine.

Kinnison, Michael, T.¹, Vaughn Holmes², Geneva York³, Jacob Aman⁴, Merry Gallagher⁵, ¹Maine Center for Genetics in the Environment and School of Biology and Ecology, University of Maine, Orono, ²School of Biology and Ecology, University of Maine, Orono, ³Environmental DNA CORE Facility, University of Maine, Orono, ⁴Wells National Estuarine Research Reserve, Wells, Maine, ⁵Maine Department of Inland Fisheries and Wildlife, Bangor, Maine

For many species of concern (e.g., threatened or invasive), it is often just as important to be confident they are absent as it is they are present. However, eDNA can miss low-abundance species if insufficient sampling effort is applied, leading to false negatives that can be problematic for management. At the same time, excessive or inefficient eDNA surveys can result in unsustainable field and lab costs. Identifying survey designs that provide high confidence of detection, while controlling costs, are thus essential to effective and sustainable applications of eDNA. Here we review two different studies applying hierarchical occupancy models to establish survey designs with specified power for detecting either invasive centrarchid fishes (Micropterus and Pomoxis spp.) in lakes, or imperiled anadromous smelt (Osmerus mordax) in streams. In both case studies we found that our initial baseline surveys provided good probability of species detection but could be improved through modest adjustments in effort at particular levels of the sampling hierarchy (e.g., number of sites, events, local replicates, or lab replicates). These case studies show how eDNA surveys 1) can achieve very high detection probabilities within the context of challenging real-world settings, 2) be improved through iterative analyses of actual survey data, and 3) be adaptively managed to prioritize eDNA resources to the most challenging cases. Notably, the methods we case-study are widely applicable, often to existing eDNA datasets, and afford a relatively rapid pathway to actionable eDNA inference for fisheries and conservation.