Modelling non-equilibrium distributions of invasive species: a tale of two modelling paradigms

Invasive species, biological control and climate change are driving demand for tools to estimate species' potential ranges in new environments. Flawed results from some tools are being used to inform policy and management in these fields. Independent validation of models is urgently needed so we compare the performance of the ubiquitous, logistic regression and the CLIMEX model in predicting recent range extensions of the livestock tick, Rhipicephalus (Boophilus) microplus, in Africa. Both models have been applied to the tick so new, independent data can be used to test their ability to model non-equilibrium distributions. Logistical regression described the spatial data well but failed to predict the range extensions. CLIMEX correctly predicted the extensions without fitting the non-equilibrium data accurately. Our results question the validity of using descriptive, statistical models to predict changes in species ranges with translocation and climate change. More test cases that include independent validation are needed.