Invasive species' potential to homogenize biotas contributes to their status as major threats to biodiversity. Though more than a dozen homogenizing and differentiating mechanisms have been proposed, they have not been incorporated into a unifying framework and null expectations are only beginning to be formulated. Empirical studies of homogenization have also given conflicting results, indicating that invaders' effects are context-dependent. Previous studies have demonstrated a connection between occupancy rates and similarity, with ubiquitous species homogenizing landscapes and idiosyncratic ones differentiating them, but this relationship has not yet been put in precise terms and is thus sometimes misinterpreted. Scientists have also noted that changes in similarity depend on initial similarity, but explanations of this pattern have yet to be formalized. Using an occupancy-based null model of community assembly, we incorporate both of these observed patterns into predictions about homogenization and differentiation and extend those predictions as occupancy rates change over time.
Results/Conclusions
Our model predicts that as invasive species proceed from low occupancy rates to high occupancy rates, there will be a transition from a net differentiating effect to a net homogenizing one. The threshold occupancy rate at which this transition occurs is equal to the pooled occupancy rate of the other species in the community. Thus, highly differentiated landscapes with many rare or endemic species will have low thresholds and will be most strongly homogenized by invasion. The model also suggests that more speciose communities (including those defined at a coarse spatial grain) will have more "inertia" against both homogenization and differentiation than more depauperate communities. Though this null model omits important biological complexity, its predictions give a baseline against which real invasions can be measured and may help clarify our thinking about similarity metrics and homogenization while indicating which landscapes are most vulnerable and how similarity might change over time.
