Forecasting alien species establishment and source regions: Quantitative assessment of potential ant invasions in Japan.

Due to the costs and difficulties of mitigating past biological invasions, there is a critical need for improved predictions of establishment risk for alien species and their source regions to guide the deployment of preventive measures. Here, focusing on a global pool of ant species known to be spread by humans, we develop a computational workflow to predict threats for a country or region of interest. Specifically, the workflow (1) predicts which alien species are most likely to be established based on climatic suitability with species distribution models, (2) clusters areas threatened by similar assemblages of alien species, and (3) identifies global regions that can act as important sources for these species. We apply this workflow to estimate which ants with human-assisted invasion histories around the globe may establish in Japan, an island country with broad climatic and topographic diversity. To reduce forecast uncertainty, we exclude models that we assess to result in dubious transfers based on evaluations of species already established in Japan and avoid making model extrapolations. To better account for the full invasion process, we also estimate introduction risk and spread within Japan and integrate these with our establishment risk and potential sources estimates. Our results indicate that all prefectures of Japan have potential risks of new alien ant establishments, though lower latitudes and small archipelagoes have the highest predicted vulnerability. When combined with the likelihood of spread, we expect shifts in vulnerability toward highly populated areas and in proximity to international ports. Interestingly, the source regions with the most alien species presenting establishment threats are in southern Europe and the subtropical Americas rather than in Asia, in part because many Asian species have already been introduced to Japan. When considering introduction risk based on global trade patterns, the United States was most likely to be a source of future introductions. We discuss the implications of these results for global management policies and cargo surveillance. The workflow described here can be deployed worldwide for different taxa to predict the establishment potential of alien invasions and their sources, and also to design more practical and efficient preventive strategies.