Paper detail

Predicting distributions of invasive species

This chapter aims to inform a practitioner about current methods for predicting potential distributions of invasive species. It mostly addresses single species models, covering the conceptual bases, touching on mechanistic models, and then focusing on methods using species distribution records and environmental data to predict distributions. The commentary in this last section is oriented towards key issues that arise in fitting, and predicting with, these models (which include CLIMEX, MaxEnt and other regression methods). In other words, it is more about the process of thinking about the data and the modelling problem (which is a challenging one) than it is about one technique versus another. The discussion helps clarify the necessary steps and expertise for predicting distributions. Some researchers are optimistic that correlative models will predict with high precision; while that may be true for some species at some scales of evaluation, I believe that the issues discussed in this chapter show that substantial errors are reasonably likely. I am hopeful that ongoing developments will produce models better suited to the task and tools to help practitioners to better understand predictions and their uncertainties.