Invasion success of a global avian invader is explained by within-taxon niche structure and association with humans in the native range

Aim To mitigate the threat invasive species pose to ecosystem functioning, reli- able risk assessment is paramount. Spatially explicit predictions of invasion risk obtained through bioclimatic envelope models calibrated with native species distribution data can play a critical role in invasive species management. Fore- casts of invasion risk to novel environments, however, remain controversial. Here, we assess how species’ association with human-modified habitats in the native range and within-taxon niche structure shape the distribution of invasive populations at biogeographical scales and influence the reliability of predictions of invasion risk. Location Africa, Asia and Europe. Methods We use ~1200 native and invasive ring-necked parakeet (Psittacula krameri) occurrences and associated data on establishment success in combi- nation with mtDNA-based phylogeographic structure to assess niche dynam- ics during biological invasion and to generate predictions of invasion risk. Niche dynamics were quantified in a gridded environmental space while bioclimatic models were created using the biomod2 ensemble modelling framework. Results Ring-necked parakeets show considerable niche expansion into climates colder than their native range. Only when incorporating a measure of human modification of habitats within the native range do bioclimatic envelope mod- els yield credible predictions of invasion risk for parakeets across Europe. Inva- sion risk derived from models that account for differing niche requirements of phylogeographic lineages and those that do not achieve similar statistical accu- racy, but there are pronounced differences in areas predicted to be susceptible for invasion. Main conclusions Information on within-taxon niche structure and especially association with humans in the native range can substantially improve predic- tive models of invasion risk. To provide policymakers with robust predictions of invasion risk, including these factors into bioclimatic envelope models is recommended