Non-indigenous species (NIS) establishments are a growing concern. Current quantitative methods for NIS risk assessment generally focus on only one species or only one of the main drivers of establishment [propagule pressure (PP), environmental suitability, or species' traits]. There is a need for quantitative models that estimate establishment probability for species at the pathway level; models would be particularly relevant if they could utilize available data, combine multiple predictors and were made accessible to managers. We present and evaluate methodology that uses establishment data, PP proxy data and any available trait data for the suite of species present in an introduction pathway to generate a joint pathway-level establishment model where species' establishment probabilities are influenced by their traits and environmental tolerances. The consequence of using our joint model is that if traits are predictive and species differ in the number of propagules needed to establish, a family of species-specific PP-establishment curves is estimated. Theoretical results revealed that our model performs well and makes accurate predictions even when trait data are incomplete and/or extraneous data are used in fitting. An empirical analysis of freshwater fish introductions to the United States identified species that have a high chance of establishment. The inclusion of species' trait and environmental data significantly improved upon predictions made with a PP-only model. Further, by considering both PP and species traits, we were able to predict which species have been observed in the U.S. and which species were more likely to persist.Synthesis and applications. Managers can use the methodology presented herein to generate quantitative pathway-level models of establishment probability. This methodology is especially appealing because it gives managers the ability to make quantitative estimates of how proposed management actions will affect establishment probabilities, allows managers to control risk without completely restricting trade, can generate predictions for new species in a pathway given knowledge of the species' trait values and informs on the consequences associated with substitutions for species with trade restrictions. Managers can use the methodology presented herein to generate quantitative pathway-level models of establishment probability. This methodology is especially appealing because it gives managers the ability to make quantitative estimates of how proposed management actions will affect establishment probabilities, allows managers to control risk without completely restricting trade, can generate predictions for new species in a pathway given knowledge of the species' trait values and informs on the consequences associated with substitutions for species with trade restrictions.
