As global extinction rates rise, conservation biologists must focus on how best to allocate their limited time and money to safeguard biodiversity. Increasingly, conservation effort is being prioritized on the basis not just of species’ endangerment, but also the evolutionary history that these species embody. Decision-making criteria such as the EDGE (Evolutionary Distinct and Globally Endangered) metric have been successfully used to help set conservation priorities for mammals, birds, corals, and sharks. EDGE’s success proves that phylogenetic conservation prioritization metrics can be used by conservation biologists and policy makers, and that they are popular with the public. Nonetheless, almost every application of an EDGE-like approach has had to deal with the uncertainty presented by missing species data. We are unaware of any systematic assessment of the impact of both missing species and imputation on EDGE metrics. We present such a systematic assessment, simulating the random loss of species from a phylogeny and subsequent imputation of their phylogenetic position. We report the impact that each of these processes has on the data underlying EDGE scores.
Results/Conclusions
We find that ED ranking is remarkably robust to missing species. In our simulations, 30% of species can be missing from a phylogeny and the ED scores of the remaining species have a correlation coefficient of 0.8 with their true values. Surprisingly, phylogenetically-biased missing species can impact ED scores less than their random loss. We also find that phylogenetic imputation, while unbiased, is not accurate in reconstructing species’ true evolutionary distinctiveness. For example, members of an imputed clade of 25 species within a phylogeny of 850 species are, on average, imputed to have ED scores 250 ranks away from their true rank. While our results are derived from relatively simplistic simulations whose limitations we discuss, we suggest that the generality of our simulations makes their implications easier to interpret. On the basis of these results, we provide clear guidance for EDGE scoring in the face of phylogenetic uncertainty. Instead of using imputation to account for the relatively minor impact of missing species or to estimate scores for missing species, we suggest conservation biologists should focus on accounting for phylogenetic uncertainty in the species for which they have data.