CONICET | Buscador de Institutos y Recursos Humanos

The accurate anticipation of potential biological invasions is a crucial step toward the control of invasive species. The method used most commonly to identify areas suitable for biological invasion is the construction of Ecological Niche Models (ENMs), although the potential accuracy of this approach may be grossly overestimated. In the present study, we examine how biogeographical, biological, and methodological factors may affect our capacity to identify areas suitable for biological invasion. We created virtual species to investigate the incongruences between the fundamental and available niche in a natural environment. Firstly, we verified how differences in species characteristics (environmental tolerance and dispersal capacity) may hinder our ability to predict invasions using ENMs. We also evaluated how different algorithms behave in the context of these differences. We also measured how prediction accuracy varies in different regions of the world, by evaluating the degree of niche mismatch found in each zoogeographic region. In general, the predictions of the ENMs varied according to species tolerance, dispersal capacity, and the algorithm used to fit the model, although the principal source of variation was the degree to which the algorithms under- or over-estimated the fundamental niche. Some zoogeographic regions did indeed prove to be more error-prone than others, due to the variation in the levels of climatic incompleteness and the representation of the fundamental niche within a species? distribution. We demonstrate that the prediction of potential biological invasions using ENMs may incur errors in niche estimation, which may result in suitable locations being overlooked. This reinforces the need for caution in the prediction of biological invasions, given that the fundamental niche may not be expressed adequately within the native range of the species, as determined fundamentally by its biological characteristics.