Predicting the spatio-temporal pattern of range expansion in an invasive species under lack of equilibrium with climate

In a context of rapid global change, there is growing need for tools that facilitate biodiversity conservation under non-equilibrium conditions. Biological invasions are one of the main drivers of biodiversity loss. Their impact can be mitigated with early detection and rapid action, which normally requires the ability to anticipate which places are most likely to be invaded next. Predictive species distribution models are a key tool towards this aim, but they rely on the assumption that the species is in equilibrium with the environment. This assumption is clearly violated by invasive species during range expansion. This study shows that, with a rescaling and transformation procedure, a true presence-only species distribution model, using presences from both native and invaded areas, can mitigate the violation of environmental equilibrium and provide reliable spatio-temporal predictions of invasion risk. Using the Asian hornet Vespa velutina in Europe as a model species, we compared models built with presence-only, presence-background, and presence-absence data. We obtained the best spatio-temporal predictions with presence-only models using all available presences and environmental variables, if they were rescaled for a more sensible prediction of environmental suitability. Using a retrospective study of the invasion in order to validate the models, we show that rescaled presence-only models efficiently predict the temporality of the invasion. Thus, presence-only models can be a valuable tool for invasive species management, as they predict where and when the species may first become established, even under non-equilibrium conditions.