Summary Accurate knowledge of species occurrence is fundamental to a wide variety of ecological, evolutionary and conservation applications. Assessing the presence or absence of species at sites is often complicated by imperfect detection, with different mechanisms potentially contributing to false‐negative and/or false‐positive errors at different sampling stages. Ambiguities in the data mean that estimation of relevant parameters might be confounded unless additional information is available to resolve those uncertainties. Here, we consider the analysis of species detection data with false‐positive and false‐negative errors at multiple levels. We develop and examine a two‐stage occupancy‐detection model for this purpose. We use profile likelihoods for identifiability analysis and estimation, and study the types of additional data required for reliable estimation. We test the model with simulated data, and then analyse data from environmental DNA (eDNA) surveys of four Australian frog species. In our case study, we consider that false positives may arise due to contamination at the water sample and quantitative PCR‐sample levels, whereas false negatives may arise due to eDNA not being captured in a field sample, or due to the sensitivity of laboratory tests. We augment our eDNA survey data with data from aural surveys and laboratory calibration experiments. We demonstrate that the two‐stage model with false‐positive and false‐negative errors is not identifiable if only survey data prone to false positives are available. At least two sources of extra information are required for reliable estimation (e.g. records from a survey method with unambiguous detections, and a calibration experiment). Alternatively, identifiability can be achieved by setting plausible bounds on false detection rates as prior information in a Bayesian setting. The results of our case study matched our simulations with respect to data requirements, and revealed false‐positive rates greater than zero for all species. We provide statistical modelling tools to account for uncertainties in species occurrence survey data when false negatives and false positives could occur at multiple sampling stages. Such data are often needed to support management and policy decisions. Dealing with these uncertainties is relevant for traditional survey methods, but also for promising new techniques, such as eDNA sampling.

Effective management of alien species requires detecting populations in the early stages of invasion. Environmental DNA (eDNA) sampling can detect aquatic species at relatively low densities, but few studies have directly compared detection probabilities of eDNA sampling with those of traditional sampling methods. We compare the ability of a traditional sampling technique (bottle trapping) and eDNA to detect a recently established invader, the smooth newt Lissotriton vulgaris vulgaris, at seven field sites in Melbourne, Australia. Over a four-month period, per-trap detection probabilities ranged from 0.01 to 0.26 among sites where L. v. vulgaris was detected, whereas per-sample eDNA estimates were much higher (0.29-1.0). Detection probabilities of both methods varied temporally (across days and months), but temporal variation appeared to be uncorrelated between methods. Only estimates of spatial variation were strongly correlated across the two sampling techniques. Environmental variables (water depth, rainfall, ambient temperature) were not clearly correlated with detection probabilities estimated via trapping, whereas eDNA detection probabilities were negatively correlated with water depth, possibly reflecting higher eDNA concentrations at lower water levels. Our findings demonstrate that eDNA sampling can be an order of magnitude more sensitive than traditional methods, and illustrate that traditional- and eDNA-based surveys can provide independent information on species distributions when occupancy surveys are conducted over short timescales.

1. Environmental DNA (eDNA) sampling can be a highly sensitive method for detecting aquatic taxa; however, the cost-efficiency of this technique relative to traditional methods has not been rigorously assessed. 2. We show how methods that account for imperfect and stochastic detection can be used to: (i) determine the optimal allocation of survey effort with eDNA sampling for a fixed budget (i.e., identify the optimal combination of water samples vs. site visits); and (ii) assess the cost-efficiency of eDNA sampling relative to traditional survey techniques. We illustrate this approach by comparing eDNA sampling and bottle-trapping for an exotic newt species (Lissotriton v. vulgaris) recently detected in Melbourne, Australia. 3. Bottle traps produced much lower detection rates than eDNA sampling, but the cost-efficiency of the two methods can be similar because bottle-trapping is cheaper per sample. The relative costefficiency of the two sampling methods was sensitive to the available survey budget, the costs of eDNA primer/probe development and sample processing, and the number of positive quantitative PCR assays (qPCRs) used to designate a water sample as positive for newt DNA. Environmental DNA sampling was more cost-efficient than bottle-trapping for small-intermediate budgets when primer/probe development and sample processing costs were low, and 1/4 or 2/4 positive qPCRs was used to label a water sample as positive for newt eDNA. However, bottle traps were generally more cost-efficient than eDNA sampling when primer/probe development and sample processing costs were high, regardless of qPCR threshold or survey budget. Accepted ArticleThis article is protected by copyright. All rights reserved. 4. Traditional sampling methods may achieve lower detection probabilities compared to eDNA sampling, but the totality of costs can make eDNA sampling less efficient than traditional techniques in some circumstances. Our approach provides a quantitative framework for determining how many water samples and site visits are required to maximize detection probabilities with eDNA sampling, and can calculate the cost-efficiency of any sampling method.

Biological invasions are increasing globally in number and extent despite efforts to restrict their spread. Knowledge of incursion pathways is necessary to prevent new invasions and to design effective biosecurity protocols at source and recipient locations. This study uses genome‐wide single nucleotide polymorphisms (SNPs) to determine the origin of 115 incursive Aedes aegypti(yellow fever mosquito) detected at international ports in Australia and New Zealand. We also genotyped mosquitoes at three point mutations in the voltage‐sensitive sodium channel (Vssc) gene: V1016G, F1534C and S989P. These mutations confer knockdown resistance to synthetic pyrethroid insecticides, widely used for controlling invertebrate pests. We first delineated reference populations using Ae. aegypti sampled from 15 locations in Asia, South America, Australia and the Pacific Islands. Incursives were assigned to these populations using discriminant analysis of principal components (DAPC) and an assignment test with a support vector machine predictive model. Bali, Indonesia, was the most common origin of Ae. aegypti detected in Australia, while Ae. aegypti detected in New Zealand originated from Pacific Islands such as Fiji. Most incursives had the same allelic genotype across the three Vsscgene point mutations, which confers strong resistance to synthetic pyrethroids, the only insecticide class used in current, widely implemented aircraft disinsection protocols endorsed by the World Health Organization (WHO). Additionally, all internationally assigned Ae. aegypti had Vssc point mutations linked to pyrethroid resistance that are not found in Australian populations. These findings demonstrate that protocols for preventing introductions of invertebrates must consider insecticide resistance, and highlight the usefulness of genomic data sets for managing global biosecurity objectives.

Environmental DNA (eDNA) sampling is a promising tool for monitoring cryptic species. Numerous studies have demonstrated that eDNA sampling can achieve higher detection rates than traditional monitoring techniques, such as trapping; however, the consequences of that sensitivity for survey design requirements and resulting survey costs have not been investigated. We demonstrate how site occupancy detection models and optimal survey design methods can be used to evaluate the cost‐efficiency of eDNA sampling vs. traditional survey methods. We apply these approaches to two datasets—one in which eDNA sampling and trapping were conducted simultaneously (paired dataset), and another in which sampling methods were independently deployed (unpaired dataset)—to assess the cost‐efficiency of eDNA sampling for detecting a freshwater mammal: the platypus Ornithorhynchus anatinus. Conditional probabilities of platypus eDNA being captured in a single water sample (paired dataset: 0.838, unpaired: 0.879), and detected in a single water sample by qPCR (paired: 0.892, unpaired: 0.858), were higher than the conditional probability of detecting a platypus with a single trapping visit (paired: 0.470, unpaired: 0.219). eDNA sampling was more cost‐efficient than trapping, regardless of whether the management objective was to (1) minimize the survey budget needed to achieve a particular asymptotic variance of the occupancy estimator, or (2) minimize the survey budget needed to detect a change in occupancy over time. Site occupancy detection models coupled with optimal survey design methods provide a powerful means with which to compare the sensitivity and cost‐efficiency of eDNA sampling vs. traditional sampling methods.