Hotspots of invasive plant abundance are geographically distinct from hotspots of establishment

Abstract

Proactively addressing future impacts of invasive plants depends critically on modeling invasive plant abundance. Yet, the vast majority of spatial models of invasion risk focus on occurrence alone. Expanding spatial analyses to consider not only potential for establishment, but also potential for high abundance is an important next step in ecological modeling. We assembled spatial reports of occurrence and abundance across the coterminous United States for 70 terrestrial invasive plants. Using records of occurrence, we used maximum entropy to model suitable climate and land cover conditions for each species’ potential range (‘establishment range’). Combining abundance information into high, medium, and low ranks, we used ordinal regression to model suitable conditions for high potential abundance of each species (‘abundance range’). We compared establishment versus abundance ranges to assess overlap for individual species as well as geographic hotspots of invasion risk. Across all species, the abundance range encompassed a mean of only 16% of the establishment range. Hotspots where many species’ ranges overlapped were consistent for establishment and abundance ranges in 47% of land area. But, modeling establishment alone missed many abundance hotspots, particularly in western U.S. ecoregions. Expanding ecological modeling to focus on invasive plant abundance provides a considerably different view of geographic invasion risk. Our results suggest that spatial models using occurrence data alone are likely to overestimate areas of high impact in their assessment of ‘invasion risk’. Considering abundance in spatial models of invasion risk is a critical next step for more targeted monitoring and management.