Using publicly available data to conduct rapid assessments of extinction risk

Abstract

The IUCN Red List plays a key role in setting global conservation priorities and is populated via rigorous, time-intensive assessments. Here, we test rapid preliminary assessments of plant extinction risk using one Red List metric: Extent of Occurrence (EOO). We developed REBA (Rapid EOO-Based Assessment) to harvest and clean data from the Global Biodiversity Information Facility, calculate each species' EOO and assign EOO-based Red List categories. We validated REBA classifications against 1671 North American plant species already on the Red List and found ~87% overlap between REBA's classifications and the IUCN's. However, REBA's false-negative rate for species outside the Least Concern category was substantial (~68%). To elucidate factors that might drive such a high rate of under-classification, we used hierarchical Bayesian models to show that certain plant types (e.g., Geophytes) and threats (e.g., Invasive and Other Problematic Species, Genes, and Diseases) increased the probability of under-classification. While REBA requires further refinement, it has yielded valuable insight into how preliminary assessment methodologies may become more effective.