Machine learning aids in identifying at-risk plants
Classification method helps prioritise species at risk of extinction. Samantha Page reports.
More than 15,000 additional plant species should be categorised as “at risk” on the Red List of Threatened Species maintained by the International Union for Conservation of Nature (IUCN) says a group of biologists which has developed a machine-learning protocol to identify species under threat.
The researchers say that to date a mere 5% of the known plant species in the world have been evaluated for conservation risk
“Our method isn't meant to replace formal assessments using IUCN protocols,” explains Anahí Espíndola, of the University of Maryland, US.
“It's a tool that can help prioritise the process, by calculating the probability that a given species is at risk.”
Espíndola and her team use a machine-learning classification and prediction method known as Random Forest to evaluate 150,000 uncategorised but known plant species and predict their Red List status. These ranged from “least concern”, in which there is no risk of extinction, to “critically endangered”.
Their results, which appear in a paper in the journal Proceedings of the National Academy of Sciences, suggest that 10% of those species should be categorised as having some sort of risk.
“On a global scale, the level of threat to plants is much higher than expected,” they write.
“Additionally, our results indicate that several geographic regions should receive more attention from conservation biologists and/or decision-makers than they currently do.”
Classifying plants is a time-consuming, and therefore expensive, endeavour, which is one reason the Red List lacks so many species. The researchers hope their analysis will make the process more efficient by identifying priorities.
“Ultimately, we hope it will help governments and resource managers decide where to devote their limited resources for conservation,” Espíndola says. “This could be especially useful in regions that are understudied.”
The research revealed one “important but troubling” result. The geographical regions that are well represented on the Red List are not where many of the major botanical biodiversity hotspots occur, nor where there are clusters of at-risk species.
"When I first started thinking about this project, I suspected that many regions with high diversity would be well-studied and protected. But we found the opposite to be true," Espíndola says.
"Many of the high-diversity areas corresponded to regions with the highest probability of risk. When we saw the maps, we were surprised it was that clear."