The Next Mass Extinction?

Martin Bernetti/AFP/Getty Images

An employee of Chile’s Agricultural and Livestock Service removing a dead Guanay cormorant from a beach in Coquimbo, May 29, 2023

In August 2023 dead elephant seals washed ashore on beaches in Argentina. First a handful appeared outside Rio Grande, a coastal city on the eastern side of the Tierra del Fuego archipelago. The next day there were more. Then the sightings moved up the seaboard, from Rio Gallegos to the Chubut province to Rio Negro to Buenos Aires. Park rangers saw pregnant seals head toward the shore, seize up in the water, spasm, and drown.

Later that month, an Argentine veterinarian named Marcela Uhart traveled out to Península Valdés, a flat mushroom-shaped peninsula jutting into the Atlantic that serves as a popular seal breeding ground. Raised on a ranch in the pampas, Uhart has studied fauna in the region for nearly thirty years; she currently works in Patagonia for the University of California, Davis. Arriving at the peninsula, she found over a thousand dead seal pups, their bodies rotting into the sand. “It was surreal,” she told me last month. “The beach [was] filled with carcasses, scattered individuals, sick and dying, pups that were caught by the tide.”

It was quickly ascertained that the seals were dying after contracting the strain of the H5N1 virus—a highly pathogenic, deadly version of the avian flu. (The “H” in the name corresponds to the hemagglutinin antigens that bind the virus to the surface of cells, the “N” to the neuraminidase proteins that help release the virus from infected cells.) The previous fall Peru’s Ministry of Health had reported pelicans sick with bird flu off the western coast of South America; in the winter researchers had found sea lions wasted on the Chilean shore. No one knew how the flu had traveled from one side of the continent to the other. Scientists theorized that it was passed through predation or waste, or physical proximity to sick animals, but the specific mechanisms of promulgation were unclear. On Punta León, a sharp cape below Península Valdés, the beach was strewn with dead terns, even as seagulls picked at the bodies, seemingly unaffected. Some had pulled lifeless creatures into their nests to use as cushions.

By summer, when Uhart first heard about the seals in Argentina, H5N1 had spread to every continent except Antarctica and Australia, appearing in both mammals and birds. Scientists were monitoring the situation closely. More than a hundred million birds breed on the Antarctic mainland and its surrounding islands every year, including boobies, cormorants, petrels, gulls, and sandpipers. Some species, like the snowy albatross and the emperor penguin, nest in the region exclusively. In a closely connected polar ecosystem—where fur seals hunt birds, leopard seals eat fur seal pups, and organisms slip in and out of the water from icy shores—a new disease can be catastrophic. “If the virus does start to cause mass mortality events across penguin colonies, it could signal one of the largest ecological disasters of modern times,” a group of British researchers wrote last year in a paper documenting the flu’s southward proliferation.

Since they neither had a solid understanding of how the virus spread nor knew of a way to stop it, all scientists could do was wait for news to arrive. In the fall, it did. Members of the British Antarctic Survey discovered a pair of infected skuas on the Antarctic mainland, making the current avian influenza epidemic the deadliest and most wide-ranging outbreak of its kind in history.

By the time I spoke to Uhart, the breeding season in Patagonia had ended. Over 17,000 baby elephant seals—96 percent or more of the juveniles in the region—were estimated to have died, as well as more than 500,000 birds. In some areas there were no longer any organisms to infect. Still, Uhart told me, she saw sick and dead animals on each visit to the beach: a sea lion, a duck, a tern. “My suspicion is that the virus will linger on,” she said. “We just don’t know whether it will continue to cause epidemic outbreaks, or whether it will just trickle in like it is now.”

The story was different in Antarctica, where Uhart was heading on an expedition as part of an effort to document the farthest reaches to which the flu had spread. There had been reports of infected leopard seals and Emperor penguins, but tissue samples would have to be gathered and tested to confirm them. In a frozen, sparsely populated area, access would be difficult, updates would lag behind actual infections and deaths, and much of the impact would need to be extrapolated using predictive models. All the same, the potential for calamity was clear. “There’s never been such an event in Antarctica before,” Uhart told me. “This could be a mass extinction.”


H5N1 was first discovered in 1996 in a farmed goose in Guangdong, China. Over the next year it spread through animal markets and farms, eventually making its way to Hong Kong, where it spilled over into humans. In 1997 eighteen people were reported to have caught it; six died. For the next two decades, the “Goose Guangdong virus,” as it came to be known, spread around Asia, Europe and Africa, first among poultry, then among swans, parrots, guineafowl, flamingos, ducks, grebes, and more.


Like many viruses, bird flu moves like a rumor: it changes as organisms contract it. This makes poultry farms, where tens of thousands of birds are kept near each other, particularly crucial inflection points in the evolution of the disease. Often spreading through wild birds in a mild form, the virus turns lethal in captivity. These highly pathogenic versions of bird flu infect an animal’s entire body: heads swell, coordination dissolves, extremities turn purple.

As bird flu spread from farms to the wild and back, it mutated several times, often reassorting, or combining genetic material, with other viruses, and then reassorting with strains of the original virus that had evolved in parallel. By 2020 the phylogenetic tree of the earliest Goose Guangdong virus had gone global many times over, fragmenting into hundreds of slightly modified versions. That fall the disease broke out in Europe, first in wild birds, then in poultry. It quickly spread to Africa and Asia, and in January 2022 the first case of H5N1 in six years was discovered in the United States, in a wigeon in South Carolina. Within a month the virus was reported on turkey farms. Then it popped up in South America.

The current outbreak has attracted such concern not only because of the disastrous threat it poses to wildlife and farmed birds but also because deadly diseases in humans often start, in some form, in animals. Influenza, HIV, Ebola, and the Bubonic plague all likely have roots in human-animal interactions. Sometimes wild animals act as reservoirs for a disease, and a single encounter can set the virus loose among humans; in other cases, like that of Covid-19, the disease gestates and evolves in animals under human care. The effect that these diseases have on humans depends on several factors, most importantly their ease of transmission. Covid-19’s infectiousness and long incubation period combine to make its spread rapid and difficult to track. Ebola, on the other hand, is passed from person to person more slowly, through bodily fluids, and people often die before it can spread widely.

Among humans, avian flu is deadly but so far not particularly infectious. Since 2003 there have been over 860 human cases of H5N1, 52 percent of them fatal; confirmed cases of human-to-human proliferation remain sparse. But the difference between a virus that infects chickens and one that infects humans can be a handful of nucleotides passed between broiler hens in adjacent cages. Factory farms are in this sense essentially labs for malignant pathogens, with hundreds of thousands of immobilized test subjects. Last month bird flu was detected in American dairy cows; a week later, in Texas, it was diagnosed in a human.

The method of mass poultry disease control recommended by the American Veterinary Medical Association under “constrained circumstances” is called “Ventilation Shutdown plus.” It consists of killing any birds potentially exposed to the disease by cutting off the airflow to their sheds and jacking up temperatures inside or flooding the space with carbon dioxide. VSD+ is cheap and efficient, which is significant because livestock deaths from avian flu cost American farmers—and the government, by virtue of farming subsidies—billions of dollars. (Since the 1990s, estimates place the global cost of outbreaks at around $80 billion.) But animals subjected to VSD+ tend to suffocate slowly, over the course of hours, as a result of which a number of animal welfare organizations have protested the method. “It can take hours of profound suffering for animals to finally die,” Vicky Bond, a veterinarian and president of The Humane League, said in 2022. In April of that year, a farm in Iowa killed 5.3 million chickens with VSD+ within twenty-four hours of detecting bird flu. “They cooked those birds alive,” one employee told The Guardian.

Chaideer Mahyuddin/AFP/Getty Images

An employee of the Indonesian government examining chicks for signs of bird flu infection at a farm in Darul Imarah, Aceh, March 2, 2023

Nonetheless, the United States Department of Agriculture requires every farm where avian flu is found to cull its entire population, in part to reduce the risk of spillover. The H5N1 vaccine, which has been proven to protect animals from sickness and death, doesn’t prevent the contraction and promulgation of the disease. “If you vaccinate, then you are sweeping the issue under the rug,” Thomas Peacock, a virologist at the Pirbright Institute in England, told me. More humane ways of euthanizing poultry than VSD+ also exist, but they are slower and more costly. By 2023 more than a hundred million farmed animals had been killed worldwide in attempts to halt the spread of bird flu.



Once they have spread in the wild, zoonotic diseases are nearly impossible to control. Every new version of bird flu is another point in the disease’s expanding history. When scientists study how different kinds spread, which animals they can infect, and their danger to humans, they are effectively trying to bind a book as it’s being written. By the time Alaska’s state veterinarian reported that the first polar bear had succumbed to H5N1, it had been dead for months; a number of wandering albatrosses had died in South Georgia before bird flu was confirmed on the island; the state of the disease in Antarctica has thus far been largely inferred. This is partly why the current epidemic of avian flu is so perplexing, and why governments place such stress on managing infections among domesticated animals.

Now that the disease has spread across the planet, scientists have expressed both urgency and fatalism. Surveys and research efforts are underway, but it can be difficult to know what to do with the information one has. “We’re not going to stand a group of soldiers on the dock and shoot every bird flying overhead,” Peacock told me. Though a number of California condors have been vaccinated, only a few hundred are left in the wild, and over the past decades tens of millions of dollars have been spent on conserving them. Other species are much less valued, accessible, trackable, and understood. It would be futile to attempt to vaccinate a critical mass of any major species in Antarctica, much less the rest of the world. “There’s not a huge amount we can do in terms of ecology, except sit tight and see what happens,” Peacock said.

No one can be sure, in other words, whether the Antarctic extinction that Uhart fears will come to pass. The uncertainty may persist even as scientists collect further data, expedition by expedition. “There’s humungous underreporting, and there are huge challenges in testing,” Uhart told me. “The true impact of this panzootic on wild birds may not be recognized for years to come,” the Australian biologists Marcel Klaassen and Michelle Wille wrote in a 2023 paper.

From what we understand about avian flu’s etiology, Uhart said, the virus might have been quelled twenty years ago, when it was first discovered in a Guangdong goose, or later stopped in poultry flocks, or still later contained on farms in Europe. But such preemptive culling, based at best on epidemiological predictions, would have required multinational cooperation, across industries and governments. The chances of that were about as great as the chances of our catching every infected bird in the sky and every infected seal in the ocean. “We know that the only time and place and opportunity to control the disease is in poultry,” Uhart said. “But we let it get into wild birds. And it just did what viruses are meant to do.”

Subscribe and save 50%!

Get immediate access to the current issue and over 25,000 articles from the archives, plus the NYR App.

Already a subscriber? Sign in