Introduction bias: Imbalance in species introductions may obscure the identification of traits associated with invasiveness

Abstract

1The introduction stage is usually overlooked in trait-based studies of invasiveness, implicitly assuming that species introductions are random. However, human activities promote the movement of specific types of species. Thus, species deliberately introduced for distinct purposes (e.g. gardening, forestry) or as contaminants of human commodities (e.g. stowaway) will likely show particular traits. If species with certain traits have been preferentially introduced (i.e. introduction bias), some traits may have been mistakenly linked to species invasion abilities due to their influence on introduction probability. In this work, we propose a theoretical framework with different scenarios of introduction bias. The introduction scenarios are: (1) Random introduction, independent from traits; (2) Biologically biased introduction, following the worldwide distribution of the trait; and (3) Human biased introduction, following a theoretical introduction pathway that favours the introduction of species with high values of the trait. We evaluate how the introduced trait distributions in these scenarios may affect trait distributions in naturalized and invasive species pools under different hypothesized associations between traits and the probabilities of naturalization and invasion. The aim of this work is to identify situations where ignoring introduction bias may lead to spurious correlations being found between species traits and species ability to become naturalized or invasive. Our framework strongly points to the need to evaluate the traits of species that have become naturalized or invasive along with the traits of species that have failed to do so in order to unravel any existing introduction bias that may confound the correlation between species traits and invasion success. Overlooking a possible introduction bias may lead to the overestimation of the correlation between the trait and the species invasion ability, especially in cases when the pool of introduced species shows extreme values of the trait distribution (as compared to a random introduction). Trait-based studies that deserve special attention to avoid undesired effects of introduction bias on their findings are: those that investigate naturalization using only the pool of naturalized species, and those studies that examine invasiveness by comparing invasive species with native species.