Modelling species distributions and environmental suitability highlights risk of plant invasions in western United States

Abstract Aim Non‐native invasive plants impact ecosystems globally, and the distributions of many species are expanding. The current and potential distributions of many invaders have not been characterized at the broad scales needed for effective management. We modelled the distributions of 15 non‐native invasive grass and forb species of concern in western North America to define their environmental niches and predict potential invasion risk. Location Arid and semi‐arid western United States. Methods We modelled species distribution using presence/absence data from > 148,000 vegetation survey plots to predict the probability of presence of each species based on associated climate, soil, topography and disturbance records. Boosted regression tree models were trained and tested within a buffer distance from presence observations of each species. Buffers were optimized by adding absence observations from increasingly large areas until prediction of presence was maximized, as defined by the area under the precision‐recall curve (AUPRC). We evaluated model AUPRC and other metrics by cross‐validation within training areas, then projected final models within focal EPA level III ecoregions. Results Focal species were present in 2.7% to 21% of plots within model training areas of 40 to 340 km radius. Models conservatively estimated potential species presence, with typically low false positive rates and moderate sensitivity. The most influential predictors of presence included minimum temperature (especially for grasses), climatic water deficit, precipitation seasonality and fire history. Cold desert ecoregions were impacted by the most species, with predicted range expansion into the prairie ecoregions and infilling for Warm Desert species. Main conclusions Invasive forb and grass species are likely to expand their ranges and continued increases in temperature, aridity and area burned will increase invasion risk. Monitoring species presence and absence and mapping known and potential ranges with a focus on presence detection, as in our methodology, will aid in identifying new invasions and prioritizing prevention and control.