Reading the morning snail mail used to be the go-to way for conservationists to learn which birds could soon be extinct. Stuart Butchart remembers how scientists in remote enclaves would post their dispatches to BirdLife International, the nearly century-old consortium of avian-conservation groups (including the National Audubon Society). The field summaries and expedition journals catalogued the health of birds and their habitat and helped inform endangered-species lists as recently as the 1980s. But the massive library had its gaps, says Butchart, BirdLife’s chief scientist. And so in the past decade, researchers have been turning to artificial, off-the-ground intelligence to plug those holes.
One means of judging a species’ extinction risk is to look at how its range transforms over time. But with landscapes constantly shifting due to climate change and other human threats, experts in the field struggle to stay up-to-date. Finite resources, such as money, hours, and bodies, make exhaustive data collection a challenge. As a result, global conservation profiles for bird species are sometimes left incomplete, with missing population numbers or indefinite ranges.
That’s where remote sensing and geospatial data can help. By tapping satellites and other aerial technologies, scientists are able to obtain high-resolution images of complex habitats, along with real-time updates on how they’re changing, and how birds are moving around in them.
Late last year, Colombian conservation ecologist Natalia Ocampo-Peñuela took these tools a step further by trying to use them to expose blind spots in global endangered-species lists. The results, published in concert with Duke University’s Stuart Pimm and other researchers in the November issue of Science Advances, suggest that conservation groups can fine-tune their understanding of where suitable bird habitat really exists by stacking elevation and tree-cover maps over information from existing field studies.
For example, the team found that the range of the Grey-winged Cotinga, a rare, obscure passerine in Brazil, encompasses just 14 square miles of Atlantic coast forest—far less than the 870 square miles estimated by the International Union for Conservation of Nature (IUCN). Based on these calculations, the bird should be listed as “critically endangered,” rather than “vulnerable,” on the IUCN’s Red List of Threatened Species. The same goes for the Velvet-purple Coronet hummingbird, which should be uplisted from a status of least concern to vulnerable, Ocampo-Peñuela says. Taking such actions could push South American governments to prioritize both birds for conservation programs.
To that end, the team believes that their models should gradually be incorporated into major conservation ranking systems like the Red List. Such upgrades could result in wide-scale reshuffling of species that were once thought to be safe.
But the group’s methods have drawn sharp criticism, even prompting a formal rebuttal from the IUCN. The organization states that it’s already using geospatial data points such as tree cover to guide the Red List. Moreover, it says that the narrow range measurements that Ocampo-Peñuela and her peers applied to their models don’t accurately represent all the ground the birds inhabit, causing them to overestimate some extinction risks.
It’s a valid argument, says Catherine Graham, professor of ecology and evolution at Stony Brook University, who isn’t involved with the IUCN or the Science Advances study. She adds that Ocampo-Peñuela’s limited parameters wouldn’t work across thousands of species in a variety of habitats because there isn’t enough geospatial data to back them all.
What’s more, not all remote technologies churn out quality intel. In April 2016, a team of scientists, including Butchart, rated nearly 300 geospatial-data sources. They found that only five percent met the conservation “gold standard” of being consistent, affordable, accurate, and high in caliber. Plus, there’s a big gap satellites can’t fill—basic facts like species’ birth or death rates, diet, and other vital natural-history knowledge, Graham says.
Both camps do agree on one practice: The only way to use big data is by pairing it with old-school science. Ocampo-Peñuela can attest. In 2014 she spent eight months living out of a tent while count ing and mist-netting birds in the Colombian Andes, nearly 10,000 feet above sea level. Her hand-drawn censuses from the field went on to shape the models in her study. As for Butchart, last year he had the BirdLife letters digitized. It was a way to preserve past labors, even as the rest of the data races to catch up.
Urgent: the Endangered Species Act is under attack! Take action today to keep it from getting dismantled.