Effects of occurrence data density on conservation prioritization strategies

Abstract

Place-prioritization analyses are a means by which researchers can translate information on the geographic distributions of species into quantitative prioritizations of areas for biodiversity conservation action. Although several robust algorithms are now available to support this sort of analysis, their vulnerability to biases deriving from incomplete and imbalanced distributional information is not well understood. In this contribution, we took a well-sampled group (i.e., Icteridae or New World blackbirds) in an intensively sampled region (the contiguous continental United States), and developed a set of pseudo-experimental manipulations of occurrence data density—in effect, we created situations in which data density was reduced 10- or 100-fold, and situations in which data density varied 100-fold from region to region. The effects were marked: priority areas for conservation shifted, appeared, and disappeared as a function of our manipulations. That is, differences in density of data can affect the position and complexity of areas of high conservation priority that are identified using distributional areas of species derived from ecological niche modeling. The effects of data density on prioritizations become more diffuse when considerations of existing protected areas and costs related to human intervention are taken into account, but changes are still manifested. Appropriate considerations of sampling density when constructing ecological niche models to identify distributional areas of species are key to preventing artifactual biases from entering into and affecting results of analyses of conservation priority.