LAUSR

Abstract Cacti invasion of natural ecosystems and grazing lands threaten native biodiversity and reduces economic output from infested grazing lands. Yet, few studies exist about potential hotspots of cacti invasion on a global scale. We used maximum entropy, boosted regression model and generalised additive models, calibrated using geo-referenced data from both the native and introduced ranges to construct ecological niche models for four invasive alien cacti species: Opuntia ficus-indica (the sweet prickly pear), Opuntia stricta (the sour prickly pear), Cylindropuntia imbricata (the tree cholla) and Cylindropuntia fulgida (the jumping cholla). The models were combined, using the average weighted method approach and projected onto the geographic space to predict terrestrial biomes, as well as areas of special conservation concern at risk of cacti invasion. The results indicate that the Mediterranean, tropical savanna and desert and xeric shrubland biomes are the most susceptible to cacti invasion. Eleven global biodiversity hotspots including the Mediterranean basin, Cape floristic region and Southwest Australia were associated with high risk of cacti invasion. The global maps of potential cacti distribution presented in this work have the potential to serve as an important contribution towards the implementation of a global policy to avoid the negative consequences of cacti invasion.