Modeling invasion risk: Combining environmental suitability and introduction likelihood

Background/Question/Methods

Invasive species are of key concern to researchers, managers, and policy makers and can be ecologically and economically costly. Assessing invasion risk requires understanding both where a species can exist and the likelihood of that species arriving. This research aims to develop a model to assess invasion risk for marine and estuarine species that combines habitat suitability modeling and likelihood of introduction using New Zealand as a case study. Invasion risk for the European green crab, Carcinus maenas, is modeled as it is a species of concern, though not yet established, in New Zealand. This model combines the use of Maxent for species habitat modeling and the generation of an introduction likelihood landscape based on patterns of commercial shipping; ballast water and hull fouling are significant vectors for novel species entering New Zealand. Several methods were considered for modeling habitat suitability, introduction likelihood, and invasion risk. Techniques for selecting environmental layers (A prioriand forward-stepwise performance) and background data in Maxent were evaluated. Introduction likelihood was derived from port presence, proximity to ports, and number of commercial vessels entering ports. Finally, invasion risk was determined based on different interaction types/weights between habitat suitability and introduction likelihood.

Results/Conclusions

Modeled invasion risk for C. maenas showed significant spatial differences within New Zealand, though this varied considerably on which methods were employed for modeling habitat suitability, introduction likelihood, and invasion risk. As the base model, A priori environmental layers were considered in Maxent, introduction likelihood was based on number of vessels, and invasion risk was assessed using equal weighting of habitat suitability and introduction likelihood. This model resulted in a lower overall invasion risk arising from little overlap between areas of highest habitat suitability (New Zealand’s south island) and greatest introduction likelihood (near Auckland). However, less constrained habitat suitability and changing the weighting to reflect the importance of either habitat suitability or introduction likelihood in calculating invasion risk resulted in an increased modeled risk across New Zealand. Implications for method choice and the impact it has on how invasion risk is assessed are presented with respect to the underlying ecology and potential for application in management situations.