Using coarse‐scale species distribution data to predict extinction risk in plants

AIM: Less than 6% of the worlds described plant species have been assessed on the IUCN Red List, leaving many species invisible to conservation prioritization. Large‐scale Red List assessment of plant species is a challenge, as most species’ ranges have only been resolved to a coarse scale. As geographic distribution is a key assessment criterion on the IUCN Red List, we evaluate the use of coarse‐scale distribution data in predictive models to assess the global scale and drivers of extinction risk in an economically important plant group, the bulbous monocotyledons. LOCATION: Global. METHODS: Using coarse‐scale species distribution data, we train a machine learning model on biological and environmental variables for 148 species assessed on the IUCN Red List in order to identify correlates of extinction risk. We predict the extinction risk of 6439 ‘bulbous monocot’ species with the best of 13 models and map our predictions to identify potential hotspots of threat. RESULTS: Our model achieved 91% classification accuracy, with 88% of threatened species and 93% of non‐threatened species accurately predicted. The model predicted 35% of bulbous monocots presently ‘Not Evaluated’ under IUCN criteria to be threatened and human impacts were a key correlate of threat. Spatial analysis identified some hotspots of threat where no bulbous monocots are yet on the IUCN Red List, for example central Chile. MAIN CONCLUSIONS: This is the first time a machine learning model has been used to determine extinction risk at a global scale in a species‐rich plant group. As coarse‐scale distribution data exist for many plant groups, our methods can be replicated to provide extinction risk predictions across the plant kingdom. Our approach can be used as a low‐cost prioritization tool for targeting field‐based assessments.