Modelling spread of the invasive macrophyte Cabomba caroliniana

1. Predicting spread of non-indigenous species requires an understanding of where propagules are being transported, and whether these propagules can survive in the novel habitat and successfully integrate into the recipient community. In this study, we model potential spread of invading Cabomba caroliniana in Ontario, Canada, using a combination of passive and active dispersal models coupled with an environmental suitability model, thereby considering the first two stages of the invasion process. 2. Measures of propagule pressure incorporated both human-mediated dispersal via trailered boats, and advective flow from invaded to non-invaded systems, while habitat suitability was forecasted by combining native and global data sets and using boosted regression trees. 3. Risk of invasion differed depending on the combination of approaches used and the time period considered. Three lakes appear to be at greatest risk owing to a combination of high boater and water movement from invaded sources, and high environmental suitability. The best predictors of lake suitability were pH, mean lake temperature and dissolved calcium concentration. Hundreds of lakes in Ontario may be suitable for establishment of Cabomba, highlighting the need for vector management.