Caution with curves: Caveats for using the species–area relationship in conservation

Conservation biologists use the species–area relationship for a variety of purposes, including upscaling diversity from small plots to regions, predicting species loss, and for identifying biodiversity hotspots and prioritizing actions to protect them. Despite its widespread use, several complications that affect the accuracy of its application are often overlooked. First, interpretation of the species–area relationship is a function of the census design used to construct it. Nested census designs guarantee only that one individual each of a given number of species is within the sampled area, but we are almost always concerned with the loss or protection of more than one individual of each species. Census designs using non-contiguous plots are useful for sampling large regions, but their interpretation is not straightforward because species number is a function of the spatial extent of the landscape and the size of the sample units. Second, power function behavior is often assumed, even though the species–area relationship often displays curvilinearity on log–log plots across scale ranges pertinent to conservation. Finally, applications of the species–area relationship often assume that the area of interest is contiguous while in practice it seldom is, and so calculations using the species–area relationship need to account for beta diversity between disjunctive areas.