The costs of invasive species in the United States alone are estimated to exceed US$100 billion per year, so a critical tactic in minimizing the costs of invasive species is the development of effective, early-detection systems. To this end, we evaluated the efficacy of adding environmental (e)DNA surveillance to the U.S. Geological Survey (USGS) streamgage network, which consists of >8200 streamgages nationwide systemically visited by USGS hydrologic technicians. Incorporating strategic eDNA sample collection during routine streamgage visits could provide early-detection surveillance of aquatic invasive species with minimal additional cost. For this evaluation, USGS hydrologic technicians collected monthly eDNA water samples, May-September 2018, from streamgages downstream of reservoirs in the Columbia River Basin thought to be vulnerable to invasive dreissenid mussel (Dreissenidae spp.) establishment. We tested water samples for dreissenid mussel DNA and also for kokanee (Oncorhynchus nerka) and yellow perch (Perca flavescens) DNA; the two fishes were used to assess if streamgages are adequately located to provide early-detection eDNA surveillance of taxa known to be present in upstream reservoirs. No Columbia River Basin streamgage samples met our criteria for being scored as positive for dreissenid DNA. We did detect kokanee and yellow perch DNA at all streamgages downstream of reservoirs where these species are known to occur. Field collection, laboratory analyses, and personnel time required for collection of four eDNA samples at a streamgage site cost US$500-US$600 (net). Given these results, incorporating eDNA biosurveillance into routine streamgage visits might decrease costs associated with an invasion since early detection maximizes the potential for eradication, containment, and mitigation.

Environmental DNA (eDNA) sampling is a promising tool for the detection of rare and cryptic taxa, such as aquatic pathogens, parasites and invasive species. Environmental DNA sampling workflows commonly rely on multi‐stage hierarchical sampling designs that induce complicated dependencies within the data. This complex dependence structure can be intuitively modelled with Bayesian multi‐scale occupancy models. However, current software for such models are computationally demanding, impeding their use. We present an r package, msocc, that implements a data augmentation strategy to fit fully Bayesian, computationally efficient multi‐scale occupancy models. The msocc package allows users to fit multi‐scale occupancy models, to estimate and visualize posterior summaries of site, sample and replicate‐level occupancy, and to compare different models using Bayesian information criterion. Additionally, we provide a supplemental web application that allows users to investigate study design for multi‐scale occupancy models and acts as a graphical user interface to the msocc package. The utility of the msocc package is illustrated on a published dataset and the functions in msocc are compared to the primary Bayesian toolkit for multi‐scale occupancy modelling, eDNAoccupancy, using various computational benchmarks. These benchmarks indicate that msocc is capable of fitting models 50 times faster than eDNAoccupancy. We hope that access to software that efficiently fits, analyses and conducts study design investigations for multi‐scale occupancy models facilitates their implementation by the research and wildlife management communities.

Autonomous, robotic environmental (e)DNA samplers now make it possible for biological observations to match the scale and quality of abiotic measurements collected by automated sensor networks. Merging these automated data streams may allow for improved insight into biotic responses to environmental change and stressors. Here, we merged eDNA data collected by robotic samplers installed at three U.S. Geological Survey (USGS) streamgages with gridded daily weather data, and daily water quality and quantity data into a cloud-hosted database. The eDNA targets were a rare fish parasite and a more common salmonid fish. We then used computationally expedient Bayesian hierarchical occupancy models to evaluate associations between abiotic conditions and eDNA detections and to simulate how uncertainty in result interpretation changes with the frequency of autonomous robotic eDNA sample collection. We developed scripts to automate data merging, cleaning and analysis steps into a chained-step, workflow. We found that inclusion of abiotic covariates only provided improved insight for the more common salmonid fish since its DNA was more frequently detected. Rare fish parasite DNA was infrequently detected, which caused occupancy parameter estimates and covariate associations to have high uncertainty. Our simulations found that collecting samples at least once per day resulted in more detections and less parameter uncertainty than less frequent sampling. Our occupancy and simulation results together demonstrate the advantages of robotic eDNA samplers and how these samples can be combined with easy to acquire, publicly available data to foster real-time biosurveillance and forecasting.

The increasing complexity and pace of ecological change requires natural resource managers to consider entire species assemblages. Acoustic recording units (ARUs) require minimal cost and effort to deploy and inform relative activity, or encounter rates, for multiple species simultaneously. ARU‐based surveys require post‐processing of the recordings via software algorithms that assign a species label to each recording. The automated classification process can result in cross‐species misidentifications that should be accounted for when employing statistical modelling for conservation decision‐making. Using simulation and ARU‐based detection counts from 17 bat species in British Columbia, Canada, we investigate three strategies for adjusting statistical inference for species misclassification: (a) ‘coupling’ ambiguous and unambiguous detections by validating a subset of survey events post‐hoc, (b) using a calibration dataset on the software algorithm's (in)accuracy for species identification or (c) specifying informative Bayesian priors on classification probabilities. We explore the impact of different Bayesian prior specifications for the classification probabilities on posterior estimation. We then consider how the quantity of data validated post‐hoc impacts model convergence and resulting inferences for bat species relative activity as related to nightly conditions and yearly site occupancy after accounting for site‐level environmental variables. Coupled methods resulted in less bias and uncertainty when estimating relative activity and species classification probabilities relative to calibration approaches. We found that species that were difficult‐to‐detect and those that were often inaccurately identified by the software required more validation effort than more easily detected and/or identified species. Our results suggest that, when possible, acoustic surveys should rely on coupled validated detection information to account for false‐positive detections, rather than uncoupled calibration datasets. However, if the assemblage of interest contains a large number of rarely detected or less prevalent species, an intractable amount of effort may be required, suggesting there are benefits to curating a calibration dataset that is representative of the observation process. Our findings provide insights into the practical challenges associated with statistical analyses of ARU data and possible analytical solutions to support reliable and cost‐effective decision‐making for wildlife conservation/management in the face of known sources of observation errors.

No abstract