Trouble at sea

It’s late July and I’m standing with Daniel Schindler at the mouth of Sam Creek, a small tributary of western Alaska’s Bristol Bay, home to the largest wild sockeye salmon fishery in the world. The mouth of the creek—barely 20 feet wide—boils with fish. Schindler, a renowned salmon biologist, estimates that there are some 500 sockeye at our feet, their bodies gone cherry red and heads a copper-ore green because it’s spawning time.

“Geez, that’s a big run,” he chuckles.

Schindler is a lead researcher in the University of Washington’s Alaska Salmon Program, which has been studying Bristol Bay salmon since the 1940s and has produced what is likely the most robust dataset of its type in the world.  Months earlier, fisheries managers had forecast an enormous run. But the scene in front of us is beyond staggering. Sam Creek is half water, half fish, and another ball of salmon is thrumming in the lake below, waiting to barge upstream.

While much of the Pacific salmon hatchery infrastructure was developed in response to failing wild runs decades ago, the fry produced in recent years are swimming into an ocean that holds more salmon than it has for the last century. For Schindler, the need to address the impacts of hatchery production is urgent. “The data show really clearly that there are limits to how many hungry salmon the North Pacific can support,” he said. If Pacific countries continue to invest in pink and chum salmon hatcheries, “it’s going to be very difficult to turn back the dial” on production in the years ahead, Schindler warned.

In chest waders and with hoods pulled tight to defend against biting flies, Schindler and I walk up the middle of the creek. “Heyyyy, bear!” he calls out over the rush of water. Schindler has been spending summers here—living in a log cabin at the edge of Lake Nerka, which Sam Creek flows into—for 26 years, and he knows we’re likely not the only ones looking for fish. As each of our footsteps parts a solid mass of salmon, I’m keenly aware that Bristol Bay’s huge returns are an outlier. For a complex set of reasons that likely includes climate change, habitat destruction, and massive influxes of hatchery-raised fish, Pacific salmon runs elsewhere have been crashing.

Yet here in this little corner of Bristol Bay, the scene is joyous, absurd. We are knee-deep in the largest Bristol Bay sockeye return on record. Initial estimates put the run at over 80 million fish, nearly double the most recent 20-year average. But as Schindler and I maneuver up the fish-filled creek, there is one troubling addendum to this historic season: The fish are some of the smallest on record for their age.

The phenomenon Schindler and his team are documenting here piles on to years of data showing that Pacific salmon returning to waterways up and down North America are shrinking. The fish are growing more slowly at sea and, in many cases, returning to spawn younger and smaller than ever before. In some places, the biggest, oldest salmon have completely disappeared.

Smaller salmon mean big uncertainties for fishermen, processors, and the communities and ecosystems that depend on these fish. For Schindler and other scientists who have been following this trend, the drop in salmon size sends a clear message that there’s trouble at sea for these fish, and it raises urgent questions about the role of salmon hatcheries in pushing marine ecosystems to their limits. As climate change drives marine ecosystems into uncharted waters, shrinking fish size could spell a grim future for salmon runs across the North Pacific, perhaps even threatening the bounty throbbing and writhing and fighting to spawn here at our feet.

Five species of Pacific salmon return to rivers and streams in North America: Chinook salmon (also known as king salmon), known for their size and the high oil content of their flesh; chum; coho; sockeye; and the smallest among them—pink.  All of these fish are anadromous, spending parts of their lives in freshwater and parts at sea. With the exception of pinks, Pacific salmon have flexible life histories shaped by both genes and the environment. This means that the fish can mature at varying ages and linger differing lengths of time in fresh- and saltwater.

Nearly all Bristol Bay sockeye spend one or two years in freshwater and two or three in the ocean before they spawn. A Chinook from Southeast Alaska, in contrast, might be two or up to seven years old when it is ready to reproduce. These variations in life history ensure that salmon populations avoid putting all their eggs in one basket, which helps to mitigate the impacts of catastrophic events like floods and droughts. The variability also means that multiple factors can influence the size of salmon when they spawn. While that measure regularly shifts from one year to the next, on average it’s been falling in many populations over the past several decades.

Schindler and a team of researchers will soon release a report showing that the weight of Bristol Bay sockeye has declined by 10 percent since the 1960s. Those findings are consistent with observations of spawning salmon elsewhere in North America. A recent analysis of 60 years of data from salmon populations across Alaska found a decrease in average size among Chinook, sockeye, coho, and chum populations beginning around 1990, and the study’s authors noted that the rate of shrinkage has accelerated since 2000.

The research findings are bolstered by observations from local fishermen. In British Columbia, for example, Chief Harry Nyce of the Nisga’a Nation, whose family typically puts up 15 cases of canned sockeye in half-pint jars each year, said that it takes six or seven fish to fill a case now, compared to just three in the past. And the famed “June hogs,” the massive Chinook that swam up the Columbia River in the spring and could weigh up to 100 pounds, are gone.

In Bristol Bay, which produces nearly half the global sockeye supply, commercial harvesters are forced to adapt. Tim Sands, a biologist with the Alaska Department of Fish and Game, is charged with managing commercial fishing in the western part of the bay. At his office in Dillingham, one of the two hub communities for the fishery, Sands keeps a double-humped graph on the front desk to show the fishermen who come and go throughout the season how using smaller mesh for their nets catches more fish. “I’m not ramming it down their throats,” he told me. “There’s not a whole lot else they can do.” Nick Lee, a gillnetter who has been fishing in the region since 1989, agrees. “If you don’t drop your mesh size to match the fish, there are fish just swimming through.”

Circumstances are more complicated for Bristol Bay seafood processors. Here, production is measured by how many fish pass through the processing line each minute, and smaller fish mean fewer pounds overall, according to Travis Roenfanz, plant manager with Peter Pan Seafood, one of the largest processors in Bristol Bay. At the edge of the Dillingham harbor, the company’s facility operates day and night to maximize production during the peak of the sockeye run, from late June to mid-July. But in 2021, when the fish came in at about a pound lighter than the previous 20-year average, around-the-clock processing wasn’t enough to make up the difference. Roenfanz calculated a loss in productivity of about 120,000 pounds each day, a more than 10 percent drop below what the facility would have seen with average weight fish. And larger salmon can command a heftier premium—their fillets often sold fresh to markets and restaurants—while smaller fish are typically headed, gutted, frozen, and sent to international markets for further processing into lower valued products.

The impacts of shrinking fish extend far upstream, into the very ecosystems where salmon spawn and that support future generations. Salmon are essentially torpedoes of fertilizer for rivers and terrestrial environments. After gorging on ocean organisms—including plankton, crustaceans, and forage fish—salmon head inland to reproduce and die. “They’re bringing with them all of these marine-derived nutrients,” explained Krista Oke, a fisheries ecologist at the University of Alaska, Fairbanks.

The nutrients carried by salmon and their decomposing bodies, including nitrogen and phosphorous, are picked up by and help to sustain a wide variety of other organisms—from streamside grasses and enormous spruce trees to flies and bears. Ultimately, they support not only the health of that ecosystem but also the growth and survival of the next generation of salmon. A study published by Oke and colleagues in 2020 calculated that an 8 percent drop in Chinook length since 1990 had equated to a 28 percent decrease in nutrient delivery to the upstream environment. In places where salmon numbers are also falling, a decline in fish size will likely only compound the nutrient loss for these important inland ecosystems.

Salmon size—particularly the size of spawning females—also influences the fish’s reproductive potential. The smaller salmon are when they return, the fewer eggs they produce and the smaller each egg is likely to be. “That’s a big impact to future productivity,” said Greg Ruggerone, a Seattle-based scientist who has studied Pacific salmon for more than four decades.

Research on the effects of smaller size on salmon reproduction has largely focused on Chinook salmon, and the data are alarming. While a large Chinook female might produce some 17,000 eggs, a smaller individual could produce just 3,000. Schindler and a team of researchers have studied these effects on the Yukon River, where Chinook runs have plummeted in recent years.  The fish that do manage to return are both smaller and younger than in the past, and Schindler and his colleagues estimate that the reproductive potential of female Chinook salmon has plummeted by up to 35 percent since 1970. And since the offspring of large fish generally end up being larger fish, the drop in fish size sets off a genetic feedback loop that favors smaller fish.

Back at the mouth of Sam Creek, we lever ourselves over the gunnel of an 18-foot aluminum skiff anchored on the gravel beach and head to Berm Creek, where there seem to be even more fish than in Sam. Using analog counters, Schindler estimates 2,400 live salmon in a single 200-meter stretch. On either side of the creek, bears have left evidence of the bounty: dead male salmon with just their massive humps chomped off, the carcasses of females torn apart for their fatty eggs. Some carcasses writhe with maggots that Schindler tells me will be devoured by the decomposers in a matter of days. The whole ecosystem, it seems, is trained on this brief rush of nutrients fresh from the North Pacific Ocean.

Schindler collects a tool kit from the skiff and wades up the creek, where he grabs 10 dead fish and hucks them up onto the bank. The lifeless salmon exhibit a rainbow of rot on their skin, with hues ranging from petal pink to butter yellow to deep maroon. Squatting in the mud, Schindler uses a pair of sliding calipers to measure the length of each fish. He then grabs the first fish in the line-up across the back with a gloved hand, and in one deft blow with a white-handled kitchen knife cleaves off the head. With a pair of angle-tipped forceps, Schindler reaches into the cut end of the body to a tiny pocket near the fish’s backbone and pulls out a white, gnat-sized structure that’s nearly transparent in the light. “The otolith sits right at the bottom of the brain case,” he says. He drops the otolith into a manila envelop that’s been marked with the fish’s sex and length.

With growth rings like those in the cross-section of a tree, a fish’s otolith reveals important details about its life, including its age and the number of years the fish spent in freshwater and at sea. Over the course of the summer, Schindler and his team will collect some 6,000 otoliths to analyze over the winter. The resulting data will paint a picture of the Bristol Bay sockeye population’s age and growth rate.

While slower growth at sea is the proximal cause for why Bristol Bay sockeye are getting smaller, the ultimate causes are more complicated. “At first sight, it’s pretty straightforward,” said Jan Ohlberger, a Washington State Department of Fish and Wildlife research scientist and the lead author of the paper that he, Schindler, and their team will soon release, but “pretty quickly it becomes really complex.” Using data from 2.8 million individual Bristol Bay sockeye, measures of salmon abundance over the decades, and data on climate and environmental conditions over time, the team created complex models that allowed them to test hypotheses to explain the drop in fish size over time. As a result of this work, the team concluded that the main reason the fish are growing more slowly is because they are being forced to compete for meals—of zooplankton, small fish, and squid—at sea.  

Scientists and fishermen alike are familiar with the concept that the more salmon there are, the smaller they’re likely to be. This phenomenon, called density-dependent growth, typically refers to salmon competing with other salmon in the same population. “Big run, small fish” is a common adage here in Bristol Bay, management biologist Sands told me. The research by Ohlberger, Schindler, and their team supports this. One of the main drivers in the drop in growth rate of Bristol Bay sockeye are fish in the same population fighting amongst themselves for food. “There are just more hungry mouths to feed,” Schindler said. 

But Bristol Bay sockeye—like all salmon—travel widely in the ocean, sharing waters with fish from across the North Pacific. A Sam Creek sockeye, then, could swim past a pink salmon from a Prince William Sound hatchery nearly 400 miles to the east, a Chinook from Southeast Alaska, or a Japanese hatchery chum salmon from thousands of miles to the west. And trouble mounts when the numbers of these fish are too high for the amount of food available. Despite the fact that numerous individual salmon runs are struggling, Ruggerone has shown that pink, chum, and sockeye salmon have overall become more abundant in recent years than at any other time in the past century.

An explosion of pink salmon is driving the rise of North Pacific salmon populations overall. In recent years, pinks have made up some 65 percent of all Pacific salmon, and more pinks continue to be poured into the ocean. North Pacific hatcheries produced some 1.2 billion paper clip-sized pink fry just last year. Known as ocean ranched salmon, or “wild caught” to marketers, the hungry fish disperse in the sea, bulking up their weight more than 3,000 times in saltwater.  Hatchery-reared fish—primarily pinks and chums–now make up some 40 percent of the total salmon biomass in the North Pacific, with releases more than tripling over the past 50 years.

Here in Bristol Bay, there are no salmon hatcheries for hundreds of miles, in part because wild runs have been so successful. And yet Ruggerone’s research has shown that in years with high pink abundance across the North Pacific, Bristol Bay sockeye grow more slowly. Data from Chinook, coho, chum, and sockeye runs across Alaska also show that competition is correlated to all of these fish getting smaller over the years. “You have this super abundant competitor species,” Ruggerone said about pinks, and they’re vying for the same meals as the other salmon.

Ruggerone and other scientists have been able to assess the impact of pink salmon across North Pacific ecosystems by taking advantage of a unique aspect of the ecology of the species: the dramatic annual fluctuations in their population size. In some parts of their range, pink salmon are up to 40 times more numerous during odd years than in even years. By comparing these ebbs and flows of pink salmon abundance, the researchers have observed some striking correlations with overall ecosystem health. In years of high pink salmon abundance, for example, they’ve seen dramatic declines in nutrient-rich plankton populations, mass die-offs of seabirds, and high mortality rates in some orca populations. Ruggerone and others think the pink salmon are gobbling up prey resources to the detriment of other species.

Warming seas are also reshuffling marine food webs and altering ecological relationships in ways scientists are only beginning to understand. At the same time, rising temperatures also throttle up a salmon’s caloric demands. “As the water temperature goes up, their metabolic needs go up,” Schindler explained. This sets up a potential double whammy for fish as seas continue to warm when there is less food to go around.

There’s evidence that commercial and recreational harvests are also playing a role in shrinking fish. Larger fish are more valuable and highly coveted by commercial and recreational fishermen, whether they’re using commercial gill nets, trolling hook and line gear, or rod and reel. And size-selective gear like gill nets have been combing fish out of the water for generations. Salmon research scientist Neala Kendall has for years investigated the role of commercial salmon harvest in Bristol Bay in driving sockeye to smaller sizes by selectively removing larger fish from the gene pool, a phenomenon known as fisheries-induced evolution. 

Most years, Kendall said, the fish that gill net fishermen in Bristol Bay catch are larger than average. This means the salmon that don’t get caught and manage to reach their spawning grounds are smaller than the population as a whole. And gill nets are more size-selective on female salmon than males because large males are often too big to be trapped in the mesh and instead bounce off the nets and continue upstream. When it comes to smaller sockeye, “we think size-selective fishing is a factor,” Kendall said. But, she added, parsing out the role of commercial harvest in shrinking these fish over generations among the myriad pressures facing salmon is tricky, if not impossible. And research shows that both heavily-fished salmon populations as well as runs that are hardly fished at all are likewise shrinking, making the role of fish harvest more difficult to assess.

The decline in salmon size has complicated efforts to ensure that commercial fisheries are sustainably managed over the long term. Across Alaska, fisheries are managed around escapement goals—target numbers of salmon that evade fishermen and that can head upstream to spawn. In Bristol Bay, seasonal workers stationed at remote towers along clear-running rivers count the fish headed to the spawning grounds, and managers open and close fishing downstream to meet escapement targets. But workers track only the numbers of fish—not what size they are. This means that, in most cases, differences in reproductive potential aren’t taken into consideration.

Given Bristol Bay’s huge returns in recent years, no one has been particularly worried that there aren’t enough fish making it to spawning grounds. But no one expects the record-busting salmon runs here to continue, either. The coming year’s sockeye run is forecast to be 35 percent smaller than in 2022—still a strong run, but 10 percent less than the average of the last decade. In the year ahead, if Bristol Bay salmon returns dwindle and the fish continue to mature at smaller sizes, managers here will be forced to consider the variable of fish size in their management decisions to ensure enough fish in future generations.

In Southeast Alaska, where local Chinook are fewer and smaller relative to past decades, some people are already pushing for management to adapt. “We need to put more fish on the spawning grounds to get the same number of eggs in the gravel,” said Ed Jones, who directs Chinook research in the region for the state’s department of fish and game. “I’ve been chirping about this for a few years now.” But people are resistant to changing escapement goals—higher goals could mean lower harvest limits—and to adding new variables to the management process—namely how many eggs each fish contributes. Jones said that in the near future, he and his team will have data from Chinook runs in the region that could prove that the preponderance of younger, smaller fish is already having a negative effect on production. Hard data, Jones explained, could change the tone of these conversations. 

For some communities, the size of the fish isn’t as important as whether they return at all. “It’s going on three years with no salmon,” said Serena Fitka, executive director of the Yukon River Drainage Fisheries Association, at the beginning of the 2022 season. For three years in a row, fisheries managers have closed the Yukon to all salmon fishing—even for subsistence, which has historically filled freezers and dinner plates year round in remote villages, where whatever food can’t be harvested locally has to be flown or shipped in over hundreds or thousands of miles for a hefty price.

Chinook in the Yukon River have seen the steepest decline in size among Alaska salmon populations. But salmon-dependent, largely Indigenous communities along the river just want to see fish—of any size. A halt to fishing not only means critical food shortages, it also means a halt to cultural traditions that go back generations, and the chance for Elders to pass on heritage—and life skills—to young people. Late May is usually a busy time in St. Mary’s, a small Yup’ik village on the Yukon, Fitka explained, when people are readying their fishing nets and drying racks and gathering firewood for smokehouses. This year, she said, “it was so quiet and eerie.”

Two hundred and fifty miles south, nothing was quiet during salmon processing season at Wassiliisia “DeeDee” Bennis’s house in Dillingham, not far from the Nushagak River, one of the key tributaries of the Bristol Bay fishery. Bennis is of Yupik, Aleut, Finn, and Japanese descent, and she grew up in a small village downriver. She is a former commercial fisherman and tribal administrator, and she remains an avid subsistence fisherman. During the weeks each summer that she spends catching and processing salmon, her children and grandchildren crowd around the plywood cleaning table, a fish oil- and smoke-tarnished boombox playing twenty-four seven in the smokehouse to ward off bears. Everyone poses for snapshots holding the largest fish.

No matter the size of the fish, it is the time of year DeeDee loves the best, when she can pass along cultural traditions and fish cutting, smoking, and canning techniques to her children and their children, traditions that have shaped their lives. “They know where they come from. They know who they are. They know the resource,” she said. But she and her husband JD know that not everyone shares in the bounty, and they send canned salmon around the state to friends who don’t have access to fish.

“There’s a little bit of guilt feeling like you’re in the land of plenty,” JD said. Fishermen’s groups, seafood processors, and NGOs have likewise been donating enormous quantities of Bristol Bay salmon to regions hard hit by failed runs.

The future holds big questions for salmon in Bristol Bay and across the North Pacific. Ocean temperatures are expected to continue to rise, with marine heatwaves becoming more frequent. And with rising temperatures, competition at sea among salmon is likely to increase. Climate change is eating up salmon habitat in the south, reshuffling food webs, and making fish hungrier all the while. And pinks are likely to continue to boom in the warmer conditions as other salmon struggle.

Rising temperatures are already reshaping the life histories of Bristol Bay sockeye, causing fish to head out to sea a year earlier—and smaller—than they did in the past. Up until about 10 years ago, the fish compensated by spending more time in the ocean. But in recent years, more sockeye are returning that have spent only two years at sea, rather than three. While Schindler thinks it’s too early to tell whether this trend will stick, if it does, sockeye will continue to return smaller than ever.

On top of this, there is no limit to hatchery production on the horizon and no effort to even coordinate releases among North Pacific rim countries. Japan, Russia, and the U.S. have made enormous investments in hatcheries over the past 50 years. In Alaska, which leads the United States in salmon fry production, the industry is embedded in the highest levels of the Alaska Department of Fish and Game, the agency in charge of managing wild salmon stocks and overseeing hatchery operations, which are mostly managed by private, fishermen-run organizations. 

Schindler hesitates to advocate, seeing his role as illuminating the trade-offs between the hatchery industry and wild salmon runs instead. But Ruggerone and others have called for international discussions about hatchery production through the North Pacific Anadromous Fish Commission, an international inter-governmental organization dedicated to the conservation of anadromous fish stocks in the North Pacific high seas. However, the organization has no regulatory authority in coastal waters where hatcheries operate. “The NPAFC is the place to discuss these issues,” Ruggerone said. But getting Pacific countries—including the U.S.—to agree on hatchery production limits would be difficult, he explained, in part because some countries, such as Japan, rely almost exclusively on hatcheries to support commercial fisheries where there are no wild salmon runs left.

As sockeye crowded Sam Creek and Bristol Bay’s other narrow streams this past summer, Schindler knew what he was looking at: more fish vying for a finite set of resources at sea. At the Bennis’s kitchen table, however, where DeeDee brought plate after plate of salmon that she and her family had prepared while J.D. told stories about commercial fishing as a kid, all I could think about was bounty. There were gloriously red smoked-sockeye strips, brined and smoked trimmings from chinook fillets, and canned chinook made into salmon salad sandwiches. All delicious and satisfying in a hundred different ways, and at least temporarily, enough to distract us from the uncertainty that the future holds.

Note: This story has been updated to describe more precisely changes in Chinook length and sockeye weight.