The Illusion of “Wild” Salmon Purity
Bycatch, parasites and hatchery genetics: what the word “wild” on a salmon label leaves out, and the numbers behind it.
I still remember the first time a fishmonger tried to talk me out of buying a farmed trout. “Take the wild one,” he said, tapping a pale, tired-looking chum with the back of his knife. “This one lived a real life.” I bought the chum. I was younger, and I liked the story.
However, university courses on fish diseases, ichthyotoxicology, and specialized feed studies — along with numerous summer internships at fish farms and hatcheries, and my subsequent work at such facilities, research institutes, and government regulatory agencies — led me to view that story differently. Not because wild salmon is bad food — far from it. Rather, it is because the word “wild” on a price tag has become a marketing ploy that obscures more than it reveals: exactly where the fish was caught, what other creatures perished alongside it, what substances accumulated in its muscle tissue, and the damage inflicted on its home river by the fishery that harvested it.
This is a long piece. Salmon deserve it.
What a salmon carries upstream
Salmon, in the shop sense of the word, covers an entire family: Atlantic salmon, five or six Pacific species, and a good deal of rainbow trout in disguise. Nearly all of them are anadromous. They hatch in fresh water, fatten in the ocean, and come back to die in the gravel they were born in. That last trip matters more than most people realize. Each returning adult is a parcel of marine nitrogen and phosphorus, delivered free of charge into headwater forests that would otherwise starve for it.
Helfield and Naiman measured this in southeastern Alaska. Trees and shrubs along spawning reaches drew 22 to 24 percent of their foliar nitrogen from salmon carcasses, and Sitka spruce next to spawning streams grew markedly faster than spruce a few hundred meters away (Helfield & Naiman 2001, Ecology). Bears, eagles, gulls, mink, stream invertebrates: the whole riparian food web takes a cut. Pull the adults out of the estuary before they spawn and you are not just harvesting protein. You are cutting the fertilizer line for a watershed.
And the parcels are shrinking. Oke and colleagues analyzed 12.5 million fish measured by the Alaska Department of Fish and Game between 1957 and 2018. Chinook had lost 8 percent of body length on average, coho 3.3 percent, chum 2.4, sockeye 2.1. For a Chinook, being smaller and younger means, per individual, about 16 percent fewer eggs, 28 percent less nutrient transport into the river, 21 percent less fishery value and 26 percent fewer meals for the rural households that depend on it (Oke et al. 2020, Nature Communications). The main drivers appear to be climate and competition at sea, but size-selective fishing pushes in the same direction. Kendall, Quinn and co-authors showed that a century of gillnetting in Bristol Bay selects against the largest sockeye, and that age and length at maturity are heritable traits (Kendall et al. 2014, Evolutionary Applications). Fishing does not only remove fish. It edits them.
The arithmetic of “less than one percent”
The Bering Sea pollock fishery is the largest single-species fishery in the United States and one of the largest anywhere. In 2025 it landed more than 3.1 billion pounds of pollock. Its managers point, correctly, to a bycatch rate below one percent. It is a good number. It is also a number that does what percentages do best: it makes a large thing look small. One percent of 3.1 billion pounds is over 30 million pounds of animals that nobody was fishing for (Alaska Marine Conservation Council, “Less than one percent”).
And those animals are not spread evenly across the ocean’s biodiversity. To a disturbing extent, they are salmon on their way home.
Here are the figures the North Pacific Fishery Management Council itself works with. In the 2021 B season the pollock fleet caught 545,901 chum salmon as bycatch, more than double the 2011–2020 average of 258,009. Bycatch then fell to 242,309 in 2022, 111,843 in 2023 and 35,125 in 2024 as vessels adopted avoidance measures (NMFS/NPFMC, Chum Salmon Bycatch DEIS). In 2025 it bounced back to roughly 151,000 fish, and in February 2026 the Council adopted the first hard cap for chum: 45,000 fish of western Alaska origin per year (Alaska Fish News, Feb 2026). Chinook bycatch went from 8,049 fish in 2024 to 19,797 in 2025, and genetic analysis attributed nearly two-thirds of those Chinook to Kuskokwim and Bristol Bay stocks (National Fisherman, “Beyond bycatch”). The all-time Chinook peak was around 122,000 fish in 2007 (NOAA Fisheries, Chinook Bycatch FEIS).
On average about 18.6 percent of that chum bycatch comes from western Alaska rivers. The rest is mostly Asian fish, much of it hatchery-released, and I will come back to what that means. But hold on to the western Alaska share for a moment, because those are the rivers that are collapsing. Between 2020 and 2023, Yukon summer and fall chum abundance ran 63 to 94 percent below the historical average. Yukon Chinook are in worse shape: fewer than 15,000 crossed into Canada in 2023 against a treaty target of 71,000, and in April 2024 Alaska and Canada agreed to a seven-year moratorium on all Yukon Chinook fishing, one full life cycle, in the hope that the run can rebuild (CBC News; Fisheries and Oceans Canada).
A trawl does not check passports. A Chinook that dies in a pollock net in February is a Chinook that will not reach the Kuskokwim in June, will not spawn, and will not feed the spruce.
Nor is the pollock fleet exceptional. It is, if anything, one of the best-monitored fisheries on the planet. The FAO’s most recent global assessment puts discards from marine capture fisheries at about 9.1 million tonnes a year, 10.8 percent of everything caught, with bottom trawls responsible for close to half of it (Pérez Roda et al. 2019, FAO). Gillnets, the standard gear of coastal salmon fisheries, drown an estimated 400,000 seabirds a year worldwide (Žydelis et al. 2013, Biological Conservation), and fisheries as a whole kill something like 650,000 marine mammals annually (Read et al. 2006, Conservation Biology). None of it shows up on a label that says “wild-caught”.
The species-level view is not more comforting. In December 2023 the IUCN moved Atlantic salmon from Least Concern to Near Threatened after its global population fell 23 percent between 2006 and 2020 (IUCN Red List press release). Chronic overharvest does something quieter than extinction. It shrinks effective population size, narrows the range of life-history strategies in a river, and leaves the stock with less genetic slack to absorb a warming ocean.
What is actually in a wild fillet
Now the part the fishmonger never mentions.
A wild salmon spends two to five years feeding in the open ocean near the top of its food web, and its fat is where lipophilic pollutants go to rest. A Norwegian team compared fillets of wild, escaped and farmed Atlantic salmon in 2020. Wild fish carried about three times as much mercury as farmed fish (56 versus 18 micrograms per kilogram) and roughly three times the dioxins and dioxin-like PCBs (1.48 versus 0.51 picograms TEQ per gram) (Jensen et al. 2020, Foods). Everything was comfortably below EU maximum levels, which is the honest caveat, but the direction is not what the “pristine wild” narrative predicts. An earlier study by Lundebye and colleagues found the same pattern across a longer list of persistent organic pollutants (Lundebye et al. 2017, Environmental Research). The reason is not mysterious. Modern feed contains far less fish oil than it did twenty years ago, and fish oil was the main carrier of those compounds into farmed fish.
Parasites are the more immediate problem, at least if you like your salmon raw. Wild salmonids are a normal host for anisakid nematodes, for the broad tapeworms of the genus Diphyllobothrium (now Dibothriocephalus), and in the Pacific Northwest for the intestinal fluke Nanophyetus salmincola. The U.S. FDA’s Fish and Fishery Products Hazards and Controls Guidance devotes an entire chapter to these organisms and prescribes freezing, typically minus 20 °C for seven days or minus 35 °C for 15 hours, for any fish that will be eaten raw (FDA, Hazards and Controls Guidance, Chapter 5). A 2017 report in Emerging Infectious Diseases found larvae of D. nihonkaiense, the “Japanese broad tapeworm” that everyone assumed was an Asian problem, in the muscle of wild pink salmon from Alaska (Kuchta et al. 2017, EID). Anisakis larvae, swallowed alive, burrow into the stomach or intestinal wall and cause anisakiasis, which regularly ends in an endoscopy suite in countries where raw fish is everyday food. Nanophyetus is less dramatic in people, but it has caused a well-documented series of gastrointestinal infections in Oregon after meals of raw or under-smoked steelhead and salmon (Western College of Veterinary Medicine, parasite fact sheet).
Farmed Atlantic salmon are, in this one specific respect, the safer raw product, and not because of marketing. Anisakis reaches a fish through its prey: live crustaceans and small fish. A salmon that has eaten heat-extruded pellets its entire life has no route of infection. Levsen and Maage examined 4,184 harvest-size farmed salmon from Norwegian farms and found zero anisakid larvae in the fillets (Levsen & Maage 2016, Food Control). That result, together with the European Food Safety Authority’s assessment that the anisakid risk in farmed Atlantic salmon fed compound feed is negligible (EFSA 2010), is why EU law exempts such fish from the freezing requirement that applies to wild fish destined for sushi (Regulation (EU) No 1276/2011).
What a farmed salmon eats, and who checks
Here is where my own field has something concrete to offer, and I would rather describe it than praise it.
A salmon in a hatchery does not eat one feed. It eats a sequence of them, each matched to a life stage with different metabolic demands: starter crumbles at first feeding, smolt feeds during the physiological switch to seawater, grower feeds through the sea phase, finishing diets before harvest. The baseline requirements come from the U.S. National Research Council’s Nutrient Requirements of Fish and Shrimp (NRC 2011), and the industry layers vitamin and mineral premixes on top of them.
The premix is not decoration. Vitamin E, the main antioxidant in fish tissue, is a good example. Hamre’s group established a minimum requirement of about 60 milligrams per kilogram of feed for Atlantic salmon at first feeding and found the practical optimum to be double that (Hamre & Lie 1995, Aquaculture Research). Fish short of vitamin E become anemic and accumulate lipid oxidation products in their tissues (Hamre et al. 1994, Fish Physiology and Biochemistry), while adequate vitamin E measurably protects the fillet from oxidation after slaughter, which is a technical way of saying the fish you buy tastes and keeps like fresh fish (Hamre et al. 2004, Aquaculture Nutrition). B vitamins are the other headline item. As feeds shifted from fish meal to plant proteins, the Institute of Marine Research showed that salmon parr on plant-based diets need B-vitamin levels above the old NRC recommendations to hold protein retention and normal liver size (Hemre et al. 2016, PeerJ), and starter and smolt premixes were reformulated accordingly.
This is verified, not assumed. The Norwegian Food Safety Authority commissions an annual feed-monitoring program run by the Institute of Marine Research. The 2025 round analyzed 88 complete feeds (14 starter, 16 smolt, 58 grower), plus fish meals, fish oils, plant meals, insect meals, eight mineral premixes and seven vitamin premixes, for contaminants, nutrients and additives. No complete feed exceeded a legal maximum for any regulated undesirable substance (IMR, Monitoring Programme for Fish Feed 2025). The fish themselves are sampled too. The parallel program for pharmaceuticals, illegal substances and contaminants in farmed fish found no exceedances of EU maximum levels for dioxins, PCBs, mercury, lead, cadmium, inorganic arsenic or PFAS in 2025 (IMR, farmed fish monitoring 2025).
Health is checked at the population level as well. Every farmed salmon in Norway is counted monthly, dead and alive, and the Veterinary Institute publishes the result. It is not flattering: sea-phase mortality was 14.2 percent in 2025, down from 15.4 in 2024 and 16.7 in 2023, with delousing injuries and infectious disease the leading causes (Norwegian Veterinary Institute, Fish Health Report 2025). I mention this because I would not trust an industry that only publishes its good numbers. The point is that someone counts. Nobody counts the wild fish that die in the net and go over the side.
Sea lice deserve the same honesty. Lepeophtheirus salmonis is a native parasite of wild salmonids, and open net pens sitting in the path of migrating juveniles can amplify it. Krkošek’s 2007 analysis of pink salmon runs in British Columbia’s Broughton Archipelago tied farm-associated louse infestations to steep declines in wild juveniles (Krkošek et al. 2007, Science). The industry’s answer, and in my view the only good one, is to break the link. Land-based recirculating systems and closed or semi-closed sea pens physically separate the farmed population from wild migration routes, and an outbreak inside a closed system can be handled as a veterinary problem instead of an ecological one.
Hatcheries: a numbers game paid for in genes
The last illusion is the one that looks most like conservation.
Around the North Pacific, hatcheries release billions of juvenile salmon every year, most of them chum and pink. In Russia much of this is done as “compensatory” stocking, releases meant to offset damage from dams, ports and pipelines. The intention is to keep catches up and rivers full. The genetic bill arrives later.
Araki, Cooper and Blouin tracked steelhead in Oregon’s Hood River across generations using DNA pedigrees and found that each generation of hatchery rearing cut a fish’s reproductive success in the wild by roughly 40 percent (Araki et al. 2007, Science). Christie’s group later showed why it happens so fast: a single generation in captivity changed the expression of more than 700 genes in the offspring, mostly genes involved in immunity, wound healing and metabolism (Christie et al. 2016, Nature Communications). A hatchery is a strong selective environment, and what it selects for is hatchery fish.
Laikre, Schwartz, Waples and Ryman summarized four kinds of damage that large-scale releases inflict on wild populations: loss of genetic variation, loss of local adaptation, change in population composition and change in population structure (Laikre et al. 2010, Trends in Ecology & Evolution). One mechanism has its own name, the Ryman–Laikre effect: when a handful of captive parents produce a flood of offspring, the effective population size of the combined wild-plus-hatchery stock collapses even as the census number grows (Waples et al. 2016, Current Zoology). Kitada’s systematic review of marine stock enhancement across more than 180 species concluded that most programs are either economically unprofitable or were never evaluated at all, while the genetic risks are well established (Kitada 2018, Fish and Fisheries).
Where the hatchery broodstock already carries farm-escapee ancestry, the problem compounds. In Norway’s River Eira, Hagen and colleagues found that hatchery conditions favored introgressed broodstock so strongly that they produced up to four times as many adult offspring as pure wild parents. The “supplementation” program ended up actively selecting for domesticated genotypes in the wild population (Hagen et al. 2019, Nature Communications).
Russian data point the same way. Alekseev and Zubchenko analyzed nearly four decades of releases from the hatchery on the Kola River in the Murmansk region and documented a truncated age structure, a shifted spawning migration and a lower share of females among hatchery-origin Atlantic salmon. They attribute the differences to a single year of rearing in the hatchery, and recommend cutting releases to 150,000–200,000 fish and selecting broodstock according to the river’s historical sub-population structure (Alekseev & Zubchenko 2021, Voprosy Ikhtiologii).
Now close the loop with the Bering Sea. Between 2011 and 2023, roughly 53 percent of the chum bycatch in the pollock fishery came from Northeast and Southeast Asian river systems, and the Southeast Asia group consists mainly of hatchery-released fish. The billions of hatchery chum poured into the Pacific are competing with wild fish for the same ocean food (Oke’s team lists competition at sea among the causes of shrinking salmon), dying in the same trawls, and spawning with the same wild fish. Inflated catch statistics, genetically diluted rivers and a bycatch problem, all flowing from the same well-intentioned pipe.
So what should be on the plate?
I am not going to tell you to stop eating wild salmon. Cooked, and from a well-managed fishery, it is superb food, and the villages on the Yukon and the Kamchatka coast have far more right to it than a sushi bar in Berlin. What I will tell you is that the halo around the word “wild” is unearned. A responsibly farmed salmon, raised on monitored feed with a stage-specific vitamin program, checked by veterinarians and ideally grown in a closed system, is at least as clean a piece of protein as the fish that came out of an estuary net. And it did not take a Chinook, a seabird or a forest with it.
Get Ekaterina Shubnikova’s stories in your inbox
Join Medium for free to get updates from this writer.
The illusion is not that wild salmon exist. It is that buying one costs nothing but money.
If you work in fish nutrition, hatchery genetics or fisheries management and think I have got something wrong, tell me in the comments. This field moves, and so should the article.
Sources
Bycatch, overfishing and ecosystem damage
- Alaska Marine Conservation Council. “Less than one percent”: how bycatch statistics distort reality.
- NMFS / North Pacific Fishery Management Council (2025). Bering Sea Chum Salmon Bycatch Management, Draft EIS, Executive Summary.
- Alaska Fish News (Feb 2026). Chum cap on Bering Sea pollock trawlers.
- National Fisherman (2026). Beyond bycatch: the bigger story of Alaska pollock.
- NOAA Fisheries. Bering Sea Chinook Salmon Bycatch Management, Final EIS.
- Helfield J.M., Naiman R.J. (2001). Effects of salmon-derived nitrogen on riparian forest growth and implications for stream productivity. Ecology 82:2403–2409.
- Oke K.B. et al. (2020). Recent declines in salmon body size impact ecosystems and fisheries. Nature Communications 11:4155.
- Kendall N.W. et al. (2014). Evolution of age and length at maturation of Alaskan salmon under size-selective harvest. Evolutionary Applications 7:313–322.
- Pérez Roda M.A. et al. (2019). A third assessment of global marine fisheries discards. FAO Fisheries and Aquaculture Technical Paper 633.
- Žydelis R., Small C., French G. (2013). The incidental catch of seabirds in gillnet fisheries: a global review. Biological Conservation 162:76–88. Summary at Mongabay.
- Read A.J., Drinker P., Northridge S. (2006). Bycatch of marine mammals in U.S. and global fisheries. Conservation Biology 20:163–169.
- IUCN (Dec 2023). Freshwater fish highlight escalating climate impacts on species.
- CBC News (2024). Canada, Alaska suspend fishing of Yukon River chinook salmon for 7 years.
- Fisheries and Oceans Canada. Agreement: Yukon River chinook salmon 2024 to 2030.
Contaminants, parasites and consumer safety
- Jensen I.-J. et al. (2020). An update on the content of fatty acids, dioxins, PCBs and heavy metals in farmed, escaped and wild Atlantic salmon in Norway. Foods 9(12):1901.
- Lundebye A.-K. et al. (2017). Lower levels of persistent organic pollutants, metals and the marine omega-3 fatty acid DHA in farmed compared to wild Atlantic salmon. Environmental Research 155:49–59.
- U.S. FDA. Fish and Fishery Products Hazards and Controls Guidance, 4th ed., Chapter 5: Parasites.
- Kuchta R. et al. (2017). Diphyllobothrium nihonkaiense tapeworm larvae in salmon from North America. Emerging Infectious Diseases 23(2).
- Western College of Veterinary Medicine. Nanophyetus salmincola.
- Levsen A., Maage A. (2016). Absence of parasitic nematodes in farmed, harvest quality Atlantic salmon in Norway: results from a large scale survey. Food Control 68:25–29.
- EFSA (2010). Scientific opinion on risk assessment of parasites in fishery products. EFSA Journal 8(4):1543.
- Commission Regulation (EU) No 1276/2011, the freezing exemption for farmed Atlantic salmon.
Feed, vitamin premixes and health control in aquaculture
- NRC (2011). Nutrient Requirements of Fish and Shrimp. National Academies Press.
- Hamre K., Lie Ø. (1995). Minimum requirement of vitamin E for Atlantic salmon at first feeding. Aquaculture Research 26:175–184.
- Hamre K. et al. (1994). Decreased concentration of hemoglobin, accumulation of lipid oxidation products and unchanged skeletal muscle in Atlantic salmon fed low dietary vitamin E. Fish Physiology and Biochemistry 12:421–429.
- Hamre K. et al. (2004). Antioxidant vitamins, minerals and lipid levels in diets for Atlantic salmon: effects on growth performance and fillet quality. Aquaculture Nutrition 10:113–123.
- Hemre G.-I. et al. (2016). Atlantic salmon require increased dietary levels of B-vitamins when fed diets with high inclusion of plant based ingredients. PeerJ 4:e2493.
- Institute of Marine Research (2026). Monitoring Programme for Fish Feed, annual report for 2025. Rapport fra havforskningen 2026–41.
- Institute of Marine Research (2026). Monitoring program for pharmaceuticals, illegal substances and contaminants in farmed fish, annual report for 2025. Rapport fra havforskningen 2026–45.
- Norwegian Veterinary Institute (2026). Fish Health Report 2025.
- Krkošek M. et al. (2007). Declining wild salmon populations in relation to parasites from farm salmon. Science 318:1772–1775.
Hatchery and compensatory releases: effects on the gene pool
- Araki H., Cooper B., Blouin M.S. (2007). Genetic effects of captive breeding cause a rapid, cumulative fitness decline in the wild. Science 318:100–103.
- Christie M.R. et al. (2016). A single generation of domestication heritably alters the expression of hundreds of genes. Nature Communications 7:10676.
- Laikre L., Schwartz M.K., Waples R.S., Ryman N. (2010). Compromising genetic diversity in the wild: unmonitored large-scale release of plants and animals. Trends in Ecology & Evolution 25:520–529.
- Waples R.S., Hindar K., Karlsson S., Hard J.J. (2016). Evaluating the Ryman–Laikre effect for marine stock enhancement and aquaculture. Current Zoology 62:617–627.
- Kitada S. (2018). Economic, ecological and genetic impacts of marine stock enhancement and sea ranching: a systematic review. Fish and Fisheries 19:511–532.
- Hagen I.J. et al. (2019). Supplementary stocking selects for domesticated genotypes. Nature Communications 10:199.
- Alekseev M.Yu., Zubchenko A.V. (2021). Efficiency of artificial reproduction of Atlantic salmon with release of juveniles at different ages: the Kola River case. Voprosy Ikhtiologii 61(1):109–118.
