It’s Hard to Know When a Species Is Extinct. A Mathematical Model Could Help.

There weren’t supposed to be any old-growth forests left on Cebu, a lush island in the heart of the Philippines—and with the trees’ absence, no Cebu Flowerpeckers, either. For decades, biologists had written off the stocky songbirds as casualties of rampant deforestation after a 1949 survey found no habitat remaining. But when a biologist scoured the landmass in 1992, he found a pair of the elusive birds, whose males’ black-and-white bodies splotched with primary colors recall a Mondrian painting.

The bird’s story proved an instructive, if painful, lesson. “If we didn’t make that assumption in the 1950s [that the Cebu Flowerpecker was extinct], there would be much more forest than there is now,” says Stuart Butchart, chief scientist at BirdLife International. Scientists named the blunder the ‘Romeo error,’ for just as Shakespeare’s Montague took Juliet for dead, ornithologists assumed the bird was a goner, and in doing so failed to protect its shrinking habitat.  

The mistake was almost fatal for the island bird, but understandable—it’s near impossible to prove a species is truly gone and difficult to know when to deem it officially extinct.

New research aims to make this judgment easier for biologists by using a mathematical model to calculate the likelihood that a given species is extinct. A first test of the model, published this month in Biological Conservation, accurately predicted the endangerment status of dozens of bird species—and classified eight as extinct or possibly extinct for the first time.

“Even though it is going to be more work to apply it, it looks like the results are going to be good enough to make it worth the extra effort,” says ornithologist Kenn Kaufman, Audubon’s field editor.

Biologists currently determine a species’ conservation status using a series of guidelines developed by the International Union for Conservation of Nature (IUCN), the global conservation authority. They consider a variety of information, including population trends, major threats, conservation work, and search effort. But drawing a concrete conclusion isn’t always as straightforward as it sounds.

“It all boils down to the challenge of definitively proving something does not exist,” Gary Langham, chief scientist for the National Audubon Society, wrote in an email. “It’s much easier to prove something does exist.”

Using a standardized model, one that could be applied to any species, could help conservationists make better, more objective decisions about where to spend limited resources. To that end, Butchart and other researchers at BirdLife International tested a mathematical approach developed last year on a real group of species: birds. The model accounts for two central factors—threats posed to species, and just how widely and reliably people have searched for their presence—to assess the conservation status of any species on the brink.

The researchers tested the model on 61 species, all potentially or confirmed extinct, weighing the fates of birds like the Eskimo Curlew, not seen since 1963, the Guanacaste Hummingbird, missing for over a century, and the Ivory-billed Woodpecker, the subject of ongoing, feverish speculation.

For the most part—80 percent of the time—BirdLife’s results aligned with current IUCN listings. The model confirmed it’s possible the Eskimo Curlew is extinct, though not definitively because people haven’t sufficiently surveyed for the bird across its continent-spanning territory. The Guanacaste Hummingbird hasn’t been seen since 1895, but because scientists only recognized the bird as its own species in the last two years, it remains critically endangered, according to the researchers. And the possibility of recent Ivory-billed Woodpecker sightings counter significant threats to its habitat, so it also remains critically endangered.

Looking back, Butchart speculates that had scientists applied the model to the Cebu Flowerpecker decades ago, before its rediscovery, they would have likely considered it critically endangered but alive.

Where the study results differ from birds’ current conservation status, the researchers recommend the IUCN reclassify species. They suggest eight birds should be listed as extinct or possibly extinct, including Hawaii’s chunky Po'o-uli, the only relict of an ancient line of honeycreepers; Brazil’s Pernambuco Pygmy-owl, a predator the weight of a golf ball; and the Spix’s Macaw, a charismatic blue parrot reclassified as extinct in the wild. (The Brazilian government plans to introduce around 50 captive-bred Spix's Macaws to the wild next year.)

The IUCN expects to accept BirdLife’s recommendations and update their Red List in the next two years after some review, a representative told Audubon in an email. For future decisions, the model would not replace biologists' assessments, but rather be an additional tool for them to consider.

“This first test is very promising,” says Resit Akcakaya, a conservation biologist at Stony Brook University not involved in the research. “If widely used, these methods will resolve a lot of uncertainties that people have about how to decide whether to list a species as extinct or not.”

The model could also be used to evaluate the conservation status of species within a country’s boundaries, as well as on a global scale, says Natalia Ocampo Peñuela, a tropical forest ecologist at the Swiss Federal Institute of Technology in Zurich who was not involved in the new study. “It’s a great approach; anyone can apply it,” she says.

The downside of the new system is that organizing data for the model is more arduous than simply sending it to an expert for review, but that’s a good thing for species, says Butchart. “It’s easy to make snap judgements based on reading a few emails and paragraphs. It’s considerably more work to force yourself to really break down each component of that decision,” he says. “But it does allow you to be more open and transparent about the assumptions you’ve made.”

In a world with limited resources devoted to conservation, that objectivity is vital. “We want to halt extinctions,” Ocampo Peñuela says. “The first thing we need is to see where the ambulance needs to go.”