Predicting Fish Invaders Using Trait-Based Models: Lessons from the Laurentian Great Lakes

Trait-based models have been used to predict the establishment, spread, and impact of potential invaders based on the assumption that successful invaders have distinct ecological and life history traits. The fish assemblage of the Laurentian Great Lakes is well suited for trait-based model development as a robust test dataset comprised of known successful and failed fish introductions is available. Since 1989, a series of trait-based models have been developed to predict fish invasions in the Laurentian Great Lakes. A review of these models provides insight into the influence of method and trait selection, and the influence of the composition of the invaded assemblage (i.e. context dependence). Models based on DFA and CART had comparable results, a variety of traits were important in the different models, and lake-specific models indicated that results were context dependent.