The Nile monitor (Varanus niloticus), native to Sub-Saharan Africa, is a popular commodity in the pet trade industry. Unfortunately, this popularity has led to the spread of the exotic predator into new regions across the globe, with subsequent releases leading to establishment in North America (Florida, 1990). Once introduced, this generalist carnivore is a potential invasive threat to native wildlife, requiring proactive measures to effectively prevent the monitor's spread into other global regions. To this end, we create and compare 23 alternative ensemble species distribution models (SDMs) using a model selection approach (with 10 possible modeling algorithms) to develop consensus models that best predict potential Nile monitor distribution on a global scale. Whereas we have created Nile monitor projections in the past, those attempts were only based on native (Africa) range data. Here, we use a total of three sets of presence data based on both native and nonnative (Florida) ranges [1) Native only; 2) Nonnative only; 3) Native + Nonnative] along with combinations of elevation, climate, and vegetation predictor variables to inform the models and compare resulting performances. Ultimately, this allowed for analyses of both ensemble SDMs and reciprocal distribution models (RDMs) all within the Biomod2 package in R.
Results/Conclusions
The most predictive consensus SDMs for our Native (TSS = 0.87), Nonnative (TSS = 0.99), and Native + Nonnative (TSS = 0.86) presence data sets were based on the Classification Tree Analysis, Boosted Regression Tree, and Random Forest algorithms using elevation, climate, and vegetation predictor variables combined. For SDM research, results here show that a model selection approach should be applied, and alternative models should include both native and nonnative range predictors when available for a more inclusive representation of species’ suitable habitats and potential distributions. Overall, our three global Nile monitor projections predict strong habitat suitability in tropical and subtropical regions in the Americas, the Caribbean, Madagascar, Southeast Asia, and Australia. In order to prevent any potential introduction to these areas, it is important for vulnerable regions (highlighted in our SDMs) to actively prohibit/regulate Nile monitors as pets, enforce those restrictions, and promote exotic pet amnesty programs before the monitor has a chance to repeat what it has already done in Florida. These predictive tools offer the best line of defense we have in preventing the spread of exotic species, and our highly accurate Nile monitor ensemble SDMs are no exception.