Global and regional nested patterns of non-native invasive floras on tropical islands
Non-native species are being distributed globally as a result of human actions, but we still know little about emerging biogeographical patterns. We tested whether the distribution of plant invaders across tropical oceanic islands has a nested structure, and identified mechanisms to explain nestedness among invaders and islands.
Tropical islands world-wide.
We analysed two datasets: a global one (350 spermatophyte species invading natural areas within 25 archipelagos) and a regional one (145 species within 12 Pacific archipelagos). We quantified island and species nestedness using the NODF metric and evaluated the contributions of each island and species to nestedness.
Globally, the distribution of invaders across islands showed a nested pattern related to island area, elevation (a proxy of habitat diversity) and invasive species richness; the pattern was weakly associated with human population density and independent of isolation from the nearest continent. Invader prevalence among islands was the best predictor of species nestedness. Nestedness was more pronounced at a regional than a global scale.
We found novel biogeographical patterns interconnecting non-native invasive floras at a global scale. Both localized and widespread species are important components of island invasive floras. Invader-rich islands host many rare invaders, and many species are invaders in only one island group, suggesting that prevention efforts should pay attention to rare invaders. We have developed a conceptual model to facilitate understanding of nestedness in island invasion. Both habitat and dispersal filtering are potential mechanisms underlying nestedness, whereas idiosyncratic factors of particular islands (e.g. habitat diversity and socio-economic history) or time-lags may explain ‘invader endemicity'. Nested regional patterns may be explained by ‘hub' islands that serve as early sites of introduction for many invaders, some of which subsequently spread across the region.