Validation of the Brown Marmorated Stink Bug survey in Washington, Utah, North Carolina, and California

Invasive species are particularly problematic because it is often difficult to eradicate once established, it is crucial to economically and effectively allocate resources for early monitoring and eradication programs. Given that the cost to manage an invasive species after it is established is typically 10-15 times higher (or more) than the cost to eradicate the species early during an invasion, the failure to effectively allocate resources for monitoring remains a major impediment to effective invasive species management. Ecological niche models are often used to predict the distribution of invasive species before or after they have been detected in new regions. Such models should also be used to guide surveys to promote the early detection and eradication of invasive species. Here we propose a practical framework that seamlessly uses ecological niche models to develop sampling routes that promote detection of invasive species. Our framework uses habitat suitability predictions and occurrence data on incursion populations to generate potential survey sites, which are then prioritized for sampling based on their size and suitability. The generated survey route is then displayed on an open street map platform. Our framework was developed into the “enmRoute” R package and a user-friendly website to facilitate its application, and we validated our framework with a case study. We show that integrating ecological niche models with human transport routes promotes the identification of survey sites that are predicted to collect more individuals and have a greater potential for species detection than traditional sampling approaches. The data provided here is intended to validate our platform and to optimize the Brown Marmorated Stink Bug survey in Washington, Utah, North Carolina, and California. The habitat suitability predictions for this stinkbug and detection in these four states were provided, together with the scripts that were used to generate survey routes.

Methods
In our framework, candidate survey sites are assembled from occurrence data on incursion populations and habitat suitability models. The logistics of surveying these sites are then determined by linking habitat suitability predictions with data on human transport pathways (i.e., roads) to generate a survey route. Our framework requires that a habitat suitability model has been created for a species in question, whereas the inclusion of incursion populations is optional and depends on whether the invasive species has been established in survey areas. The most common method to create a habitat suitability model is to relate occurrence records to environmental variables. Ecological niche models are often used in invasion risk assessment because model development packages have been created that can be readily used to identify areas of potential establishment.

We used H. halys to demonstrate and validate that our framework could be employed for promoting the selection of sites to collect a maximum number of individuals for a given number of survey sites compared to routes designed without ecological niche models. To monitor the spread of this species, a standardized monitoring network was established in 17 states from 2017 to 2020. We used data from states that had extensive sampling in the western US: Washington, California, and Utah, for validating survey routes from our heuristic algorithm. In each state, H. halys were sampled with sticky panel traps baited with lures spaced at 50 m intervals and collected every two weeks. We standardized the counts for modeling analyses as average H. halys adults per trap per week.