Environmental <scp>DNA</scp> as a Detection and Quantitative Tool for Brown Bullhead Catfish ( <scp> <i>Ameiurus nebulosus</i> </scp> ): A Comparison With Conventional Netting

ABSTRACT Invasive fish are a serious threat to indigenous biodiversity. However, methods and tools to detect and control invasive fish can be poorly developed, costly, and/or labour‐intensive. Environmental DNA (eDNA) methods have significant potential for use in pest fish management. However, different studies asking how eDNA compares with conventional sampling, whether a relationship exists between eDNA data and population density, and how differences in field and laboratory eDNA techniques affect detection rates can return species‐ and context‐specific results. We addressed these questions for lacustrine environments using the brown bullhead catfish ( Ameiurus nebulosus ), an introduced pest in Aotearoa New Zealand. We collected eDNA samples alongside catch and biomass data from fyke nets set in two lake systems in the Bay of Plenty and Waikato regions of New Zealand's North Island. We investigated the effects of filter pore size (1.2 μm, 5 μm, dacron) and laboratory assay type (catfish‐specific metabarcoding, multi‐species metabarcoding) on catfish detection and compared our eDNA data to contemporary and historic catfish abundance records. All three filter types were similarly successful at detecting catfish presence when a minimum of five replicates was taken from each sampling site, while the catfish‐specific metabarcoding assays returned higher catfish compositional data and fewer false negatives compared to the multi‐species assays. We identified mixed results when examining the correlation between eDNA catfish read composition and historic or contemporary netting data, with significant correlations corresponding to random combinations of both filter size and eDNA assay type. Our results highlight the value of eDNA as a tool for pest fish detection that can complement more conventional biomonitoring methods, while limited correlations between eDNA data and field‐based fish netting rates emphasize the need for optimisation before eDNA can be used quantitatively for biosecurity surveillance efforts.