Understanding the biogeography of invasive plants is integral to predicting their threat across spatial scales. Invasive plant biogeography is most often studied using occurrence-based species distribution models (SDM). Occurrence-based SDMs estimate the ranges of invasive plant species, which determine where these species can impact native ecosystems. Abundance is another component of invasive distributions that influences impact, but occurrence-based SDMs poorly predict abundance. Estimates of abundance allow prioritization of high-risk areas for management. In this study, we compare invasion risk estimates based on abundance-based and occurrence-based models. Using a three-rank classification scheme, we assembled an ordinal abundance dataset for the contiguous United States based on data from the Early Detection and Distribution Mapping System. Abundance and occurrence were modeled as responses to climate and landcover using ordinal regression and maximum entropy, respectively. Within species, we compared the size of the establishment range (where occurrence is possible) to the impact range (where high abundance is likely). Across species, we mapped establishment risk by calculating potential species richness, and impact risk by calculating the potential number of high-abundance species. Areas with high richness or numbers of high-abundance species were identified as richness or abundance hotspots, respectively.
Results/Conclusions
We found sufficient abundance data for 155 terrestrial invasive plant species. The models for 64 of these species (41%) were included in this study. Species were removed due to poor occurrence-based model performance (9 species), abundance datasets that insufficiently sample environmental conditions across geographic ranges (61 species), and poor abundance-based model performance (21 species). The final species list consisted mainly of perennials (55 species) and represented a variety of habits including forbs (20 species), shrubs (10 species), and vines (8 species). The most likely abundance class varied within ranges for all species, and impact ranges were generally small portions of establishment ranges (median: 9%). Infilling of the impact range was low (<4%) for all species, indicating substantial risk for high-abundance infestation of unoccupied areas within potential ranges. Species richness and abundance hotspots overlapped by 50% and the Eastern Temperate Forests ecoregion had the greatest species richness and abundance. Abundance hotspots highlighted some high-risk areas (e.g., the northwest and northern Great Plains) not otherwise identified. Occurrence-based models may overestimate geographic invasion risk in areas like the northeast, where many plants can establish but relatively few are likely to be abundant. Given these insights, more spatial analyses of invasive plants should include existing abundance information moving forward.