In late 2014, conservationist Ian Gynther lost hope. After days spent crawling into rock crevices, scouring through camera-trap footage, and carefully laying bait around Bramble Cay—a tiny island at the northern end of Australia’s Great Barrier Reef—there was little room for doubt. The Bramble Cay melomys (Melomys rubicola), a furry little rodent endemic to the island, had gone extinct. “My colleagues and I were devastated,” Gynther, a senior conservation officer at Queensland’s Department of Environment and Heritage Protection, later told The Guardian. “As each day of our comprehensive survey passed without revealing any trace of the animal, we became more and more depressed.”

The disappearance of the Bramble Cay melomys became a grim milestone in the history of conservation biology. Its extinction report, published in 2016, determined the cause of death to be anthropogenic climate change, the first such attribution for a mammalian species.The rodents’ home...

The melomys will not be the last species to meet this fate. As global temperatures rise, more and more of the Earth’s millions of species are experiencing environmental change at a rate that may well be unprecedented in our planet’s history. A recent meta-analysis of research on more than 2,000 species suggested that nearly 50 percent of threatened, nonflying terrestrial mammals and 23 percent of threatened birds had already been negatively affected by climate change in at least part of their ranges.2 And with climate change accelerating many deleterious global dynamics, such as ice melt and ocean acidification, the damage is likely to continue.

Faced with this sobering reality, conservation biologists are increasingly shifting their focus from documenting the effects of climate change on the world’s wildlife to trying to forecast the risk that individual taxa or ecosystems will be lost—and do so early enough to intervene. Due to varying exposure and differences in biology, not all organisms are equally likely to suffer. So, to best allocate limited resources, “what we really want to be able to do is pinpoint those species that are most at risk,” says the University of Connecticut’s Mark Urban, an ecologist who works in Arctic Alaska. “We need to identify the winners and the losers. And then try to help the losers.”

Over the last decade, more and more researchers have been working to measure this risk for taxa across the tree of life. The tools that provide the foundations for these so-called climate change vulnerability assessments (CCVAs) range from models of future habitat availability to analyses based on the physiological effects of projected temperature increases, and can be conducted for one or multiple species at a time. The results, which can be expressed as the estimated extinction risk of a species or its categorization into broad classes such as “highly vulnerable,” are cited by conservation planning organizations such as the International Union for Conservation of Nature (IUCN) and the Intergovernmental Panel on Climate Change (IPCC).

But conservation biologists are still working to improve the methodology behind these assessments. Bruce Stein, chief scientist at the National Wildlife Federation, notes that researchers are increasingly addressing questions such as, “What does climate vulnerability mean?” and, “What are some of the different techniques for assessing it?” Yet, as in any discipline that deals with prediction, researchers working on CCVAs are grappling with a suite of challenges. Data on many species are limited, and the need to juggle multiple sources of uncertainty—including those inherent to the climate forecasts these models employ and the biological assumptions they rely on—have raised questions about how best to approach vulnerability assessments, and how researchers should make use of their results.3

“We’re making some headway in thinking more about becoming forecasters rather than just descriptors,” says Urban. “But I think we still, as a discipline, have a long way to go.”

Predicting the harmful effects of climate change

In the Namib Desert of southwestern Africa, quiver trees (Aloidendron dichotomum) stand out against the vast, rocky backdrop. These towering succulents grow more than 8 meters tall and live for around 200 years on average. In addition to being culturally significant—they’ve been used for centuries by local San people to make bow quivers, and their image appears on Namibia’s 50-cent coin—the trees provide critical habitat and food for many insects and birds. But by the turn of the millennium, the quiver tree populations of Namibia and South Africa were in serious trouble: large swaths were dead or dying.

Wendy Foden, then a master’s student at the University of Cape Town, was one of the researchers who stepped in to investigate. “We went to [almost] every population across its entire range, which is about 2,000 kilometers long,” says Foden, now a researcher at South Africa’s Stellenbosch University and a climate change specialist for the IUCN. “We looked at how many were alive, how many were dead, how many babies.”

More than 50 sites and some 6,000 quiver trees later, the team concluded that the species was being battered in populations at the northern edge of its range as temperatures crept upwards and the desert became drier.4 Although populations in cooler southern regions were faring better, projected temperature increases suggested that they would soon experience similar conditions. Short of migrating 25 miles south within 15 years—a tall order for a plant species with a juvenile phase of around half a century—the quiver trees were soon going to be well outside their comfort zone.

These sorts of geographical shifts in species’ tolerable environmental ranges, or climate envelopes, have become one of the most often-cited consequences of global warming, and a major starting point for biologists to calculate climate-related extinction risk. “The typical approach is to take global climate change models, downscale, and look at how the magnitude and rate of climate change is going to potentially shrink or constrain a species’ climate envelope,” says Lindsey Thurman, an ecologist at the United States Geological Survey (USGS). An early, influential example, published by the University of Leeds’s Chris Thomas (now at the University of York) and colleagues in 2004, applied a version of this approach to endemic species of plants and animals occupying around 20 percent of the world’s landmass. By 2050, the team calculated, between 15 percent and 37 percent of species could be “committed to extinction,” with especially high losses for species in scrubland and temperate forest.5

R.I.P., RAT: The Bramble Cay melomys (Melomys rubicola; above) was a rodent endemic to the Great Barrier Reef. In 2016, following a survey of the roughly four-hectare island with which it shares its name (below), researchers declared the species extinct as a result of climate change.

Natalie Waller

But these sorts of correlative, distribution-based models have come under fire from ecologists in recent years. For a start, they often equate a species’ current range to the range in which it experiences its preferred environmental conditions—an assumption that overlooks other distribution-influencing factors such as food availability. What’s more, such models generally fail to capture biological consequences of climate change that are not reflected in distribution, says Urban. Indeed, an analysis he conducted in 2015 found that, compared to other approaches, distribution-based models paint a more optimistic picture of extinction risk.6 “One of the big issues with these types of models is that they’re probably missing many of the key mechanisms that really determine how species respond to climate change,” he says.

To address these issues, some researchers use an organism’s biological traits to predict its responses under different climate scenarios. This approach often takes the form of thinking, “What can go wrong?” says Foden. “That’s the right question to start with.” Such trait-based models incorporate data on characteristics that can make a species particularly sensitive to the effects of climate change, such as dependence on certain temperatures for survival, or slow population growth rates. (See “Estimating Vulnerability” here.) More-mechanistic models attempt to incorporate relationships between environmental conditions and a species’ developmental or reproductive biology.

Vulnerability assessments built on such foundations are constantly evolving, as biologists turn up ever more answers to the question, “What can go wrong?” For instance, species don’t live in a vacuum, notes Urban, so “species interactions are a key point.” Thanks to these interactions, species may experience indirect effects of climate change that range from the relatively obvious—if one organism faces climate-driven extinction, its predators may also be at risk—to the more nuanced. Last fall, for example, researchers reported that clownfish living in anemones that were bleached as a consequence of ocean warming showed higher signs of stress and reduced fecundity compared with fish living in unbleached anemones.7 So-called cascading effects open up a dimension of species’ vulnerability that modelers have yet to fully explore.

However, not all of the effort to understand responses to climate change concerns various routes to harm. A growing branch of species vulnerability research takes a more positive view, by investigating how species compensate for their changing environments. Although the phenomenon has long been fundamental to biological theory, researchers are only just beginning to come to grips with it in a climate-change context.

Incorporating adaptation into assessments of vulnerability

In 2014, a medium-size butterfly inhabiting the west coast of North America received a surge of media attention mainly because it was still alive. By the mid-1990s, the Quino checkerspot (Euphydryas editha quino) had been pegged as a climate-change victim in waiting. With its range eroded from the south by rising temperatures and from the north by urbanization, the butterfly—along with local populations of its host plant, the dwarf plantain (Plantago erecta)—was running out of space. “It seemed to me that it was really heading for extinction,” says Camille Parmesan, a biologist at the University of Plymouth in the U.K. “This species had such a high number of populations completely gone that I thought were irretrievable, I didn’t give it much hope.”

But as Parmesan worked on plans to raise captive insects in the lab for future reintroductions, the checkerspot did something unexpected. “It went up the mountain,” Parmesan says. In 2015, she and her colleagues reported that the butterfly had relocated east to elevations unheard of for its subspecies, and had switched to a different host plant in the process.8 “Those populations at the high elevations are really healthy,” Parmesan says. “It felt so good to see them really dance, bumping into each other, having a great time.”

Source: Nat Clim Change, 3:919–25, 2013

The case of the Quino checkerspot and other species like it offer a reminder of a fundamental biological response to environmental stress: adaptation (a term that, in CCVAs, refers both to evolutionary changes in a population’s genetic makeup and to individual plasticity, or acclimatization). “We’re often seemingly surprised by how adaptable species end up being,” says Stein. “It’s this notion that nature can find a way—which is sometimes true and sometimes isn’t.”

The potential for organisms to escape doom via adaptation is termed “adaptive capacity” by the National Wildlife Federation and several other organizations. Well-recognized adaptation-favoring traits include superior dispersal ability, phenotypic plasticity, and high genetic diversity. “Even with a lot of uncertainty, you know that high levels of genetic variation are certainly going to make those populations more robust in the future, even if you don’t know what the drivers are going to be,” says Ary Hoffmann, an evolutionary biologist at the University of Melbourne in Australia.

New factors contributing to adaptive capacity are being reported all the time: a study published earlier this year, for example, suggested that warm-blooded vertebrates hold an advantage over their cold-blooded counterparts by being able to tolerate a wider range of climatic conditions and consequently to adapt more quickly.9 Some researchers are beginning to pin down species’ adaptive capacity to variation in specific regions of the genome, too. Scientists in California recently reported that in North American populations of the yellow warbler (Setophaga petechia), successful adaptation to climatic changes was associated with genes involved in exploratory and migratory behavior.10 A study published earlier this summer identified more than 200 regions in the epigenome of the spiny chromis damselfish (Acanthochromis polyacanthus) that were tied to increased tolerance of rising ocean temperatures.11

Adaptive capacity remains a relatively rare feature of vulnerability assessments, however; a recent USGS review found that, of 124 assessments carried out by the US Department of the Interior as tools for conservation planning, the concept appeared in just a third.12 There’s growing momentum to change that. Ignoring adaptation can lead to overly pessimistic predictions that could result in allocating resources where they’re less needed, notes USGS’s Thurman. “We feel very strongly that a better understanding of adaptive capacity can improve the cost-effectiveness of conservation plans.” The IUCN’s current guidelines on vulnerability assessments promote the inclusion of adaptive capacity alongside measures of an organism’s sensitivity and exposure to climate change, and Stein and colleagues have published multiple explainers about how to measure adaptive capacity appropriately.

To help get the message across, some researchers are demonstrating the effect of taking adaptive capacity into account for vulnerability assessments of specific taxa. In October, Benjamin Ofori of the University of Ghana, along with Linda Beaumont and collaborators at Australia’s Macquarie University, ranked 17 Australian lizard species on vulnerability and found that, based on sensitivity and exposure, seven would be classified as “highly vulnerable” to a warmer and drier climate. However, when they incorporated terms to describe each species’ adaptive capacity, that number dropped to two.13 “You have some species that were previously classified as highly vulnerable being reclassified as moderate,” Ofori says. “You even have some species jumping [rank]”—potentially reshaping conservation priorities.

To calculate the likelihood that organisms will be harmed by climate change,
researchers estimate their exposure, sensitivity, and ability to adapt to new conditions.
See full infographic: WEB | PDF

A better understanding of adaptive capacity will not only improve CCVAs, but could also inform new approaches to conservation, says Hoffmann. His group has promoted strategies to increase genetic diversity—and thus, theoretically, adaptive capacity—in the mountain pygmy possum, an endangered marsupial inhabiting the rapidly warming alpine zone in southeastern Australia. According to a report from the team, introducing males from one relatively large population to a smaller one appears to have helped double the size of the latter in just three years.14

But while adaptation can help rescue species from immediate harm, it can only go so far. “With climate change, you can’t stabilize it—certainly not locally, and even globally,” explains Parmesan. Climate change creates “a moving target,” and an organism that adapts under stress might only have reduced its vulnerability for a while.

Unfortunately, that appears to be true for the Quino checkerspot. Despite its recent escape from extinction, Parmesan says, the butterfly has reached a new dead end: the tops of the mountains. While the high-elevation habitat provides a cool home for now, Parmesan and her colleagues predict that even this region will become uninhabitable in coming decades. “That’s when the depression sets in,” she says. Moving the insects to more-favorable habitat—a strategy known as assisted migration—is problematic under current environmental regulations. “It’s wonderful that these wild critters can surprise us and have more adaptability than we think they have,” says Parmesan. “But at the same time, it’s just a little bit of buffering. It gives us a few more years to figure something else out.”

The value of climate change vulnerability assessments

As the methodology behind climate change vulnerability assessments evolves, some conservation biologists are concerned that different efforts are out of sync with one another. Reviews of the literature have shown substantial variation in the way models are applied. The Sapienza University of Rome’s Michela Pacifici, with Foden, Hoffmann, and others, reported that birds and mammals were by far the most often assessed taxa in CCVAs carried out between 1997 and 2014—even though they constitute a fraction of a percent of the world’s biodiversity—and only a handful of studies assessed vulnerability on a global scale.15 While North American and Australian studies frequently made use of trait-based approaches, Europe overwhelmingly favored those based on distribution.  

Such differences can often be traced to the information available on particular taxa or regions, says Thurman. “For many species, we pretty much have no data,” she says. “Even understanding basic natural history can be challenging.” There’s often a difficult choice to be made, therefore, about how much complexity to include. On one end of the spectrum are approaches that require little in the way of biological data such as environmental tolerances and life history traits, but potentially miss important pathways. On the other are more-detailed models that require researchers to make assumptions to fill in the gaps, adding considerable uncertainty to already taxing calculations.

CHANGING THE WORLD: In 2015, unusually high temperatures in Arctic Alaska triggered the arrival of spring one month earlier than normal at ecologist Mark Urban’s research site (top). Though temperatures followed a more normal schedule the following year (bottom), researchers observed the physical effects of the previous summer’s heat, such as the collapse of land on the far left side of the second year’s image. Both photos were taken in the first week of June.
mark urban

Opinions vary as to how much the choice of approach matters. Last year, the University of York’s Thomas and colleagues published an “assessment of assessments” for 12 recent methodologies.16 The aim, he tells The Scientist, was to find out whether different approaches—ranging from distribution-based models to trait-focused assessments—would reach roughly the same conclusions. Using historical data on population distributions and abundances of Great Britain’s birds and butterflies, as well as simulated data, the team pitted the models against each other to predict the present from the past. The results, Thomas says, were disappointing. Just two of the models—both of which relied more heavily on distribution than on biological traits—achieved better-than-random accuracy. And “the methods simply disagreed with one another on how they classified species. By definition, that means at least some of them have got to be wrong.”

Source: Science, 344:1246752, 2014

Thomas says he hopes that one effect of comparative studies such as his will be to motivate more data collection. “We can’t just say the data doesn’t exist. We have to think, how are we going to put in place the monitoring of things so that a few decades in the future, people aren’t wringing their hands still saying, ‘We haven’t got the data.’” In the meantime, he adds, while models may lack accuracy in predicting exactly when species will go extinct, they can identify taxa likely to be vulnerable at some point.

At the same time, the predictions generated by vulnerability assessments are just one step on a road to the much larger goal of finding practical solutions to conserve at-risk taxa, says Stein. As such, CCVAs shouldn’t be seen as “the end product,” he says. “That’s the beginning. Assessing climate vulnerability allows you then to begin applying that to better conserve these things in light of climatic changes.” He and others have advocated for biology-based management strategies that build on CCVAs by helping species realize their adaptive capacity, for example, through assisted migration or breeding programs.

Another crucial consideration is that extinction risk is a product not just of climate change, but of multiple interacting stressors including habitat fragmentation and species invasions. And conservation priorities are based on more than just a species’ risk of dying out. Societally influenced considerations such as a species’ economic value, cultural significance, or perceived charisma affect how conservation dollars are spent; scientists may also consider a species’ importance to the entire ecosystem before making recommendations to decision makers. And some conservationists want to drop the species-focused view altogether in favor of a more holistic, biodiversity-centered approach, Parmesan says. “If we focus on individual species, we’re going to be very upset, because there will be a lot of extinctions.”

With the clock ticking for many species around the world, it’s imperative that biologists, policy makers, and the public decide what really matters, says Stein. To make effective use of concepts such as climate change vulnerability, “we have to be clear about what our values are,” he says. “To think we’re going to be able to keep things as they are today, or go back to some version it was in the past, just is no longer realistic.”  


  1. I. Gynther et al., “Confirmation of the extinction of the Bramble Cay melomys Melomys rubicola on Bramble Cay, Torres Strait: results and conclusions from a comprehensive survey in August-September 2014,” Unpublished report to the Department of Environment and Heritage Protection, Queensland Government, Brisbane, 2016.
  2. M. Pacifici et al., “Species’ traits influenced their response to recent climate change,” Nat Clim Chang, 7:205–208, 2017.
  3. A.A. Wade et al., “Assessments of species vulnerability to climate change: From pseudo to science,” Biodiversity Conserv, 26:223–29, 2017.
  4. W. Foden et al., “A changing climate is eroding the geographical range of the Namib Desert tree Aloe through population declines and dispersal lags,” Divers Distributions, 13:645–53, 2007.
  5. C.D. Thomas et al., “Extinction risk from climate change,” Nature, 427:145–48, 2004.
  6. M.C. Urban, “Accelerating extinction risk from climate change,” Science, 348:571–73, 2015.
  7. R. Beldade et al., “Cascading effects of thermally-induced anemone bleaching on associated anemonefish hormonal stress response and production,” Nat Commun, 8:716, 2017.
  8.  C. Parmesan et al., “Endangered Quino checkerspot butterfly and climate change: Short-term success but long-term vulnerability?” J Insect Conserv, 19:185–204, 2015.
  9. J. Rolland et al., “The impact of endothermy on the climatic niche evolution and the distribution of vertebrate diversity,” Nat Ecol Evol, 2:459–64, 2018.
  10. R.A. Bay et al., “Genomic signals of selection predict climate-driven population declines in a migratory bird,” Science, 359:83–86, 2018.
  11. T. Ryu et al., “The epigenetic landscape of transgenerational acclimation to ocean warming,” Nat Clim Change, 8:504–509, 2018.
  12. L.M. Thompson et al., “Summarizing components of U.S. Department of the Interior vulnerability assessments to focus climate adaptation planning,” USGS Open-File Report 2015–1110, doi:10.3133/ofr20151110, 2015.
  13. B.Y. Ofori et al., “Influence of adaptive capacity on the outcome of climate change vulnerability assessment,” Sci Rep, 7:12979, 2017.
  14. A.R. Weeks et al., “Genetic rescue increases fitness and aids rapid recovery of an endangered marsupial population,” Nat Commun, 8:1071, 2017.
  15. M. Pacifici et al., “Assessing species vulnerability to climate change,” Nat Clim Change, 5:215–24, 2015.
  16. C.J. Wheatley et al., “Climate change vulnerability for species—Assessing the assessments,” Glob Change Biol, 23:3704–15, 2017.

Interested in reading more?

Climate Change

Become a Member of

Receive full access to digital editions of The Scientist, as well as TS Digest, feature stories, more than 35 years of archives, and much more!
Already a member?