Potential distribution and predator-prey interactions with terrestrial vertebrates of four pet commercialized exotic snakes in Mexico.

The anthropogenic introduction of alien species is one of the main causes of biota homogenization worldwide. Part of these introductions provokes biological invasions, which cause serious ecological and economic impacts from local to global scales. In order to prevent this and give proper management to exotic species, it is necessary to use predictive tools to prioritize sites with a higher probability of their establishment, thus preventing impacts in local ecosystems and communities. Here, we used ecological niche and interaction models to estimate the invasive potential of four species of exotic snakes identified as high risk and currently sold as pets in Mexico (Python brongersmai , P. molurus , P. regius , and Malayopython reticulatus), and simulate their impact as top predators (according to the average length of the prey that each snake species consumes) of species of native terrestrial vertebrates to this country. Results indicate that all four exotic snake species could generate severe impact on the community of native vertebrates within the country, while P. molurus could get to cause the largest impact given that it would get to interact and potentially consume a greater number of species. The organisms that would be mostly preyed by these snakes would be birds, followed by mammals, reptiles and finally amphibians. This pattern results from a combination of factors including prey body size preferences for predation in the exotic snakes' native ranges, and the amplitude of the potential distribution areas and the number of species from each vertebrate group in Mexico. This paper shows the utility of combining methodological approaches for risk prevention and risk analysis for invasive alien species. • Exotic pet snakes represent a risk for native vertebrates in Mexico. • Highest negative interactions would be on birds, followed by mammals, reptiles and amphibians. • Risk assessment could improve from combining potential distributions and network analyses. [ABSTRACT FROM AUTHOR]