Closeness is not a problem: New ways to handle spatial dependencies in species distributions

Species distribution models are often cited as being important for the design of nature reserves, assessing impacts of climate change and informing managers on management options for land use. Most present-day analyses suffer from the statistical problem of spatial dependency, meaning that points closer together in space are more likely to have similar characteristics than those further apart, a fact which may critically bias model predictions. A new study in the journal Ecography by an international team of researchers now offers a landmark review of the approaches available to handle spatial dependencies in species’ distributions. The authors evaluate six fundamentally different statistical approaches to identify factors that correlate with species distributions and provide detailed guidelines for their application. By providing the technical means to perform the relevant spatial analyses, this work will prove invaluable for anyone wanting to take spatial dependencies into account.