Probabilistic Framework for Aquatic Invasive Species Environmental DNA Monitoring & Inference

Environmental DNA(eDNA) monitoring is a novel detection methodology that is being rapidly adopted for aquatic invasive species(AIS) surveillance due to its high sensitivity and cost-effectiveness. However, the real-world application of eDNA monitoring, such as for invasive Asian carp management, has been contentious due to the uncertainties in the relationship between species’ presence and the presence of the species’ eDNA. In this paper, a Bayesian Network(BN) framework is developed to integrate these uncertainties and to evaluate Asian carp eDNA monitoring performance over a range of laboratory detection methods. An eDNA fate and transport simulation model is developed to estimate the spatial and temporal relationship between eDNA concentration, species’ density and alternative sources of eDNA, such as fish carcasses. These simulation results are entered into the BN to determine the conditional probabilities of eDNA monitoring given Asian carp presence or absence for each laboratory detection threshold. The BN finds that highly sensitive laboratory methods greatly improve detectability, but are more prone to false positives due to detecting eDNA from alternative sources. This model serves as a framework to draw inferences from eDNA monitoring results with greater confidence and will assist in the future design of effective AIS monitoring strategies.