Part Two — When Enhancement Becomes Replacement
Are Hatcheries Saving Wild Salmon,
or Replacing Them?
Part One followed the money.
Part Two follows the fish.
And the question is much simpler than fisheries science sometimes makes it sound:
After decades of producing salmon in hatcheries, are we rebuilding wild salmon populations that can eventually sustain themselves — or are we permanently replacing natural production with artificial production?
That does not mean hatcheries are bad.
Some hatcheries may be absolutely necessary.
If a salmon population is approaching extinction, a carefully managed conservation hatchery can preserve its genetic line and buy time while the problems that caused the decline are addressed.
But that is very different from operating a hatchery for 40, 50 or 60 years with no clear biological point at which the river should begin taking over again.
A legitimate recovery hatchery should contain the seeds of its own retirement.
Not Every Hatchery Has the Same Job
We first have to stop using the word hatchery as though every hatchery exists for the same reason.
A conservation hatchery is essentially a biological lifeboat. Local fish are bred carefully because the population is dangerously small. The objective is to preserve enough genetic diversity for a self-sustaining wild population to become possible again.
An integrated recovery hatchery can temporarily increase numbers while natural productivity and habitat are being repaired.
A mitigation hatchery replaces fish production lost because of something such as a dam or other permanent development.
And a harvest-production hatchery deliberately produces large numbers of fish to support fisheries.
Those purposes are not interchangeable.
DFO's own 2019 Salmonid Enhancement Program report makes that distinction particularly clear. About 335 million juvenile salmon were released that year. About 306 million were associated with a harvest objective, approximately 14.6 million with rebuilding, and roughly 1.1 million with conservation. DFO now says its Pacific hatcheries and spawning channels release close to 400 million juvenile salmon each year.[1][2] (PAC)
So when we hear that hatcheries are needed to “save wild salmon,” the first question should be:
Which hatchery, producing which fish, for what purpose?
What Does Recovery Actually Mean?
Imagine a river that once produced 10,000 wild Chinook.
The population collapses.
A hatchery is built.
Years later 20,000 Chinook are returning.
That sounds like success.
But what if 18,000 of those fish originated in the hatchery and only 2,000 were naturally produced?
We have produced more salmon.
We have not necessarily restored the wild population.
That distinction is central to Part Two.
A recovery program should therefore be judged by questions such as:
Are natural-origin adults increasing?
Are naturally produced adults successfully replacing themselves?
Are more wild juveniles surviving to become smolts?
Is natural reproductive success improving?
Is hatchery influence declining?
Has the watershed become more productive?
Can hatchery production eventually be reduced?
If success is defined simply as more hatchery fish returning, a hatchery can prove itself successful forever.
If success is defined as a self-sustaining wild population, eventually the hatchery has to become less necessary.
A major 2024 review of B.C. hatchery science reached much the same problem from another direction. Supplementation often produced an immediate increase in total spawners, but the evidence for sustained improvement in wild or natural-origin abundance was much weaker. Of 31 rebuilding populations assessed, only nine had sufficient data for a third-generation evaluation of the wild component, and only four of those nine showed improved wild-spawner abundance. The authors also stressed the limited data and possible habitat and environmental effects.[5] (Wiley Online Library)
That is exactly why we need to separate fish production from wild recovery.
What Do pHOS and PNI Actually Mean?
Here is where we need to translate the fisheries language into English.
pHOS means the proportion of hatchery-origin fish spawning naturally.
So when DFO reports a pHOS of 0.91, forget the decimal for a moment.
It means:
About 91 out of every 100 salmon spawning naturally were hatchery-origin fish.
They may now be swimming in a river and spawning in gravel, but they began life in a hatchery.
Then there is PNI — proportionate natural influence.
PNI is an index used to estimate whether natural selection in the river or selection associated with hatchery breeding is exerting greater influence on an integrated population.
A PNI of 0.01 does not mean one percent wild DNA.
It means the calculated natural influence is extremely low relative to hatchery influence.
U.S. hatchery programs use the same kinds of measurements. For example, one NOAA-reviewed recovery program uses a future PNI target above 0.67 because the intention is for the natural environment rather than the hatchery environment to drive local adaptation.[11] (NOAA Fisheries)
Those American targets are not Canadian laws.
But they are useful scientific warning lights.
Now Look at Five Major B.C. Systems
DFO's analysis gives us a remarkable picture.
At Robertson Creek, the average pHOS was approximately 0.91 and the PNI approximately 0.01.
Plain English:
About 91 of every 100 Chinook spawning naturally were hatchery-origin fish.
At Nitinat, pHOS was about 0.89 and PNI about 0.09.
About 89 of every 100 natural spawners were hatchery-origin.
At Conuma, pHOS was approximately 0.97 and PNI approximately 0.03.
About 97 of every 100 natural spawners were hatchery-origin.
On the east side of Vancouver Island, Big Qualicum was approximately 0.67 pHOS and 0.28 PNI, while Campbell/Quinsam was approximately 0.68 pHOS and 0.16 PNI.
DFO classified all five as Integrated-Hatchery populations.[3] (Publications.gc.ca)
These numbers do not prove hatcheries caused the original salmon declines.
They do tell us something else.
Several major B.C. systems are no longer simply wild populations receiving a small helping hand from hatcheries.
Artificially produced salmon now make up a very large — and in some cases overwhelming — portion of the spawning population.
That deserves examination.
West Coast Vancouver Island Should Stop Us in Our Tracks
DFO's recent West Coast Vancouver Island Chinook assessment is probably one of the most important documents in this entire discussion.
DFO reports approximately 15 million hatchery Chinook smolts released annually into about 20 WCVI rivers, with approximately 14 million coming from three major operations.
Yet DFO still assesses natural-origin WCVI Chinook as being in the Critical Zone.[4]
DFO estimates only approximately 15 to 20 percent of WCVI Chinook are natural-origin, while hatchery-origin fish have averaged roughly 85 percent of total abundance in recent assessments.[4] (Waves Vagues)
That is the distinction we have been looking for.
We can have a considerable number of salmon in the system while the natural salmon-producing system underneath remains in serious trouble.
DFO also states that large-scale hatchery production has diminished genetic diversity across WCVI populations, primarily through hatchery strays interbreeding with wild fish.[4] Its separate straying assessment found genetic influence from Robertson Creek, Conuma and Nitinat introduced into many other WCVI populations, producing what the scientists described as genetic homogenization.[3] (Waves Vagues)
That does not mean the hatcheries accomplished nothing.
Quite the opposite.
It demonstrates how remarkably effective hatcheries can be at maintaining total abundance.
But total abundance and wild-population recovery are not the same measurement.
A Hatchery Is Also an Evolutionary Environment
A hatchery is not simply a safer river.
It is a different environment.
In nature, salmon experience floods, predators, competition, changing temperatures, variable food, disease and countless other pressures.
Most do not survive.
The survivors reproduce.
That is natural selection.
Inside a hatchery we deliberately remove many of those pressures.
Food is supplied.
Predators are reduced.
Water is controlled.
Humans select broodstock.
Early-life survival can become dramatically higher.
That is exactly what a hatchery is designed to accomplish.
But evolution does not stop when salmon enter a concrete tank.
A broad scientific review found that hatchery salmonids frequently showed lower reproductive fitness in natural environments, with local wild-origin broodstock generally performing better than non-local or long-domesticated stocks.[6] More recent genomic research found evidence of genetic divergence associated with domestication selection in Chinook hatchery populations after only several generations.[7] (Wiley Online Library)
That does not mean every hatchery fish is somehow defective.
It means the environment in which animals live and reproduce matters.
And Good Hatchery Design Can Matter Too
There is also positive evidence.
Hess and colleagues studied a Chinook supplementation program using local wild-origin broodstock and found a substantial demographic increase without a detectable first-generation loss of reproductive fitness in that case.[8] (PubMed Central (PMC))
That is important.
The scientific conclusion is not:
Hatcheries are bad.
It is:
Purpose and design matter.
Local broodstock matters.
The number of generations in captivity matters.
Family representation matters.
Straying matters.
Natural reproductive success matters.
And the definition of success matters most of all.
Canada's East Coast Shows a Very Different Model
We do not even have to leave Canada to see another philosophy.
Inner Bay of Fundy Atlantic salmon became so endangered that Canada established a Live Gene Bank.
The program exists to prevent extinction and preserve the genetic foundation of the remaining populations.
It uses river-specific, locally adapted fish.
Genetics are managed.
Adaptation to captivity is deliberately minimized.
And Canada's recovery objective is explicit:
re-establish wild, self-sustaining populations.
Most importantly, the federal action plan says the present emphasis of the Live Gene Bank is not simply population supplementation. Its purpose is to preserve genetic variation and recovery potential while the factors preventing recovery — particularly extremely poor marine survival — remain unresolved.[9] (Canada)
Atlantic Canada has not solved Atlantic salmon recovery.
That is not the point.
The useful difference is that hatchery dependence is not being mistaken for the final objective.
The hatchery is the lifeboat.
The destination remains a self-sustaining wild population.
Maine Uses the Same Basic Conservation Logic
On the U.S. East Coast, Maine contains the last remaining natural Atlantic salmon populations in the United States.
Craig Brook National Fish Hatchery maintains seven separate river-specific broodstocks and explicitly describes its mission as preserving their genetic integrity as part of recovering the species.[10] (U.S. Fish and Wildlife Service)
Again, Maine has not solved Atlantic salmon recovery.
But the program provides a useful comparison.
The genetic identity of individual river populations is treated as something to preserve — not something to gradually blend together through decades of uncontrolled hatchery influence.
The U.S. West Coast Still Uses Hatcheries — But Tests Them Differently
Washington, Oregon and the Columbia Basin still operate large hatchery programs.
So this is not Canada-versus-America or hatchery-versus-no-hatchery.
The difference is increasingly in how the programs are tested.
NOAA uses Hatchery and Genetic Management Plans that can examine broodstock, hatchery-origin spawners, natural-origin broodstock, genetics, straying, disease, releases, harvest and monitoring.
Some U.S. environmental reviews have gone further and formally compared alternatives including existing production, modified production, 50-percent production reductions, and other production levels.[11] (NOAA Fisheries)
That is important because a heavily hatchery-dependent population cannot necessarily have the hatchery switched off tomorrow without consequences.
But dependence is not proof of recovery either.
The intelligent question becomes:
Can natural production be strengthened enough that hatchery dependence can be deliberately reduced?
British Columbia should be testing that question too.
Europe Gives Us Another Useful Comparison
Europe does not have one single salmon hatchery policy.
But the European Union increasingly measures river health by the functioning of the whole aquatic system.
Under the EU Water Framework Directive, ecological status includes fish populations, flow, groundwater connection, river continuity, riverbed and riparian condition, temperature, oxygen, nutrients and pollutants.[12]
The EU Nature Restoration Regulation now also calls for restoring at least 25,000 kilometres of rivers to free-flowing condition by 2030, including removing barriers to connectivity.[12] (EUR-Lex)
That does not mean Europe has solved salmon restoration.
It means their policy helps illustrate the other half of the hatchery equation:
Instead of asking only how many fish can be manufactured, ask whether the river's biological machinery is being repaired.
Big Qualicum Gives Us a 60-Year Test
Big Qualicum deserves particular attention because DFO says it was built in 1963 and was British Columbia's first modern salmon enhancement project.[13] (PAC)
More than six decades have now passed.
That does not automatically make the program a failure.
It gives us an extraordinary scientific opportunity.
After more than sixty years:
Has natural-origin productivity increased?
Has hatchery influence declined?
Can the river sustain the population without the same level of artificial production?
If not, what biological problem is still preventing that?
And has that limiting problem actually been repaired?
Those are fair questions after sixty years.
The Missing Half of Enhancement Is the Watershed
This is where habitat belongs in the hatchery story.
Not as another subject.
As the other half of the experiment.
A hatchery can bypass a damaged life stage.
If eggs are dying in unstable gravel, collect them and incubate them safely.
If juvenile survival is poor, rear the fish under controlled conditions.
If young salmon are suffering in degraded habitat, release them later or larger.
Those actions may produce more surviving fish.
But they do not necessarily repair why natural fish were dying.
The hatchery can bypass a damaged life stage without repairing why that life stage is failing.
That is why enhancement must include rebuilding the watershed's ability to manufacture salmon habitat naturally.
Restoration Has to Repair Processes, Not Just Pieces
There is a difference between protecting a creek, repairing a damaged site and restoring the processes that actually build salmon habitat.
Real watershed recovery can involve floodplain connection, groundwater exchange, wetlands, natural sediment movement, riparian forests, estuaries, side channels, fish passage and natural flow.
The scientific evidence shows that meaningful restoration can increase natural salmon production — but scale matters.
Across intensively monitored Pacific Northwest watersheds, researchers have documented measurable improvements in habitat and fish production when restoration addressed important limiting processes at meaningful scales.[14] (OUP Academic)
Roni and colleagues demonstrated the other side of that lesson: scattered, small-scale restoration can be too limited to create a detectable population response. In their modelling, considerable watershed restoration was necessary before a roughly 25-percent increase in smolt production could reliably be detected.[15] (Wiley Online Library)
So counting restoration projects is not enough.
We have to measure whether the river became more productive.
We Have Real Examples
On Idaho's Pahsimeroi River, managers removed barriers, improved flows and screened water diversions.
Accessible habitat increased dramatically.
A median of about 42 percent of Chinook redds subsequently occurred in newly accessible habitat.
Most importantly, juvenile productivity rose from a median of approximately 64 smolts per female spawner to 99.[16] (Wiley Online Library)
That is a number the public can understand.
Each spawning female was producing more young salmon.
That is natural productivity.
California's Butte Creek provides an even larger example. Extensive passage and watershed restoration was followed by roughly a 2,000-percent increase in average spring Chinook abundance, although the population remains vulnerable to drought and still lacks some of the life-history diversity needed for long-term resilience.[17] (The Ecological Society of America)
Again, restoration is not magic.
But ecological capacity can be rebuilt.
Urban Systems and Source Control Fit Here
Urban streams are part of exactly the same hatchery question.
A salmon stream running through a city can be affected by stormwater, road runoff, leaking sewers, cross-connections, septic systems, industrial drainage, fine sediment, warmer water and loss of groundwater recharge.
If we put hatchery juveniles into that stream without controlling those sources, we may simply keep replacing fish exposed to the same problem.
Research in western U.S. urban watersheds found untreated highway runoff could be lethal to adult coho, while soil-based biofiltration prevented the lethal effect.[18] More recent Canadian research found very young coho extremely sensitive to the tire-derived contaminant 6PPD-quinone.[19] (PubMed Central (PMC))
That is why source control belongs here.
Not because Part Two is suddenly a stormwater article.
Because a recovery hatchery cannot compensate forever for a watershed we continue damaging.
Then Add the Cumulative Effects
A salmon does not experience government departments.
It experiences a life cycle.
The egg may begin in gravel affected by road sediment.
The juvenile experiences altered flows and warmer water.
Then perhaps contaminated urban runoff.
Then degraded estuary habitat.
Then poor marine feeding conditions.
Then competition, predators and harvest.
If it survives and returns, it may spawn among large numbers of hatchery-origin fish.
Government can place every one of those pressures in a separate department and a separate report.
The salmon cannot.
DFO's WCVI assessment now says essentially the same thing scientifically: natural Chinook productivity is expected to remain poor until cumulative pressures involving freshwater and estuarine degradation, changing ocean conditions and other limiting factors are alleviated, and DFO calls for a holistic, ecosystem-focused rebuilding approach.[4] (Waves Vagues)
Each pressure may have an agency, a permit and a separate report. The salmon experiences all of them at once.
The Feedback Loop We Need to Test
This next part is a hypothesis.
We should test it rather than simply declare it true.
Habitat is damaged.
Natural production falls.
Hatchery production compensates.
Hatchery fish become increasingly important to fisheries and spawning abundance.
Management becomes increasingly dependent on artificial production.
Natural productivity remains weak.
And weak natural productivity then becomes another reason the hatchery cannot be reduced.
Nobody has to plan that outcome.
Nobody has to be corrupt.
It can simply develop because producing another generation of hatchery fish is immediate and measurable, while restoring an entire watershed may take decades.
That is exactly why Robertson Creek, Nitinat, Conuma, Big Qualicum and Quinsam deserve a proper long-term biological audit.
Do Not Shut Everything Down Tomorrow
There is an equally foolish response we should avoid.
If a salmon population is now overwhelmingly hatchery-dependent, suddenly eliminating production could cause abundance to collapse.
The U.S. experience shows why alternatives need to be modelled rather than guessed.
The answer is not:
Close every hatchery.
The answer is:
protect the remaining natural population;
repair limiting habitat;
reduce straying;
improve broodstock management;
measure reproductive success;
measure natural smolts produced per spawner;
establish a natural-origin baseline;
and then reduce artificial production carefully as natural production strengthens.
A conservation hatchery might remain necessary for years.
Some harvest hatcheries may continue because society deliberately chooses to operate them.
Fine.
But call the objective what it is.
Every Recovery Hatchery Needs an Exit Door
For every major B.C. hatchery claiming a recovery role, the public should eventually be able to answer some very basic questions.
What was the original recovery objective?
What is the natural-origin population today?
How many naturally produced smolts come from each natural spawner?
How many hatchery fish are spawning naturally?
How much straying occurs?
Is genetic differentiation being maintained?
What habitat processes are still limiting natural production?
What has actually been repaired?
At what level of natural production could hatchery releases be reduced?
What would a 25-percent reduction do?
What would a 50-percent reduction do?
And what biological conditions would finally allow the recovery program to end?
Those are not anti-hatchery questions.
They are the questions a genuine recovery program should want answered.
The Question We Should Have Asked Years Ago
Hatcheries are remarkable pieces of biological engineering.
They can protect eggs.
Raise juveniles.
Prevent immediate extinction.
Support fisheries.
Preserve genetic lines.
Compensate for lost production.
And buy time.
But those abilities create a danger too.
The better we become at producing salmon artificially, the easier it becomes to confuse producing salmon with recovering salmon.
They are not the same thing.
For a true recovery hatchery, success should eventually mean:
natural-origin adults increase;
natural reproductive success increases;
smolt production increases;
habitat productivity improves;
hatchery influence falls;
and artificial production can begin to decline.
If the opposite happens — decades pass, hatchery fish dominate spawning, natural production remains weak, important limiting habitat processes remain damaged and there is still no biological exit criterion — then we should at least be willing to ask whether we are still conducting wild-salmon recovery.
Or whether we have created a permanent compensatory salmon-production system.
A hatchery can make eggs.
It can make fry.
It can make smolts.
It can produce adult returns.
What it cannot manufacture is a functioning watershed.
Ultimately, only that watershed can produce truly wild salmon generation after generation without us.
So for Robertson Creek, Nitinat, Conuma, Big Qualicum, Quinsam — and eventually every major recovery hatchery in British Columbia — there is one graph we should publish.
Put hatchery-origin adults on one line.
Put natural-origin adults on another.
Run the two lines back as far as the records allow.
And then ask the simplest question in this entire discussion:
If we remove the hatchery fish from the graph, what has happened to the wild population underneath?
That is the test.
Numbered References
Fisheries and Oceans Canada. Salmonid Enhancement Program — 2019 Report. Reports 335 million juvenile salmon released and separates conservation, rebuilding and harvest enhancement objectives. (PAC)
Fisheries and Oceans Canada. Hatcheries and Spawning Channels. Current DFO overview stating Pacific hatcheries and spawning channels release close to 400 million juvenile salmon annually. (PAC)
Weil, J., Luedke, W., Healy, T.M., Withler, R.E., Brown, N.A.W., Bokvist, J., and Porszt, E. The Magnitude and Extent of Chinook Straying from Hatcheries in Southern British Columbia. DFO Canadian Science Advisory Secretariat Research Document 2025/024; associated Science Advisory Report 2024/031. Primary source for pHOS, PNI, hatchery designations, straying and WCVI genetic homogenization. (Publications.gc.ca)
Brown, N., Holt, C., Irvine, J.R., Luedke, W., McHugh, D., and Thom, M. West Coast of Vancouver Island Natural-Origin Chinook Salmon Stock Assessment in 2024. DFO Science Response 2025/021 and Research Document 2026/020. Primary source for Critical Zone status, hatchery proportion, hatchery releases, genetic effects, cumulative pressures and rebuilding advice. (Publications.gc.ca)
Riddell, B.E., Pearsall, I.A., and Rosenberger, A.G. 2024. A Review of Pacific Salmon Hatcheries in British Columbia, Canada, and Interactions With Natural Populations. Fisheries. doi:10.1002/fsh.11091. (Wiley Online Library)
Araki, H., Berejikian, B.A., Ford, M.J., and Blouin, M.S. 2008. Fitness of Hatchery-Reared Salmonids in the Wild. Evolutionary Applications 1:342–355. doi:10.1111/j.1752-4571.2008.00026.x. (Wiley Online Library)
Howe, N.S., Hale, M.C., Waters, C.D., Schaal, S.M., Shedd, K.R., and Larson, W.A. 2024. Genomic Evidence for Domestication Selection in Three Hatchery Populations of Chinook Salmon. Evolutionary Applications 17:e13656. doi:10.1111/eva.13656. (Wiley Online Library)
Hess, M.A., et al. 2012. Supportive Breeding Boosts Natural Population Abundance With Minimal Negative Impacts on Fitness of a Wild Population of Chinook Salmon. Molecular Ecology 21:5236–5250. doi:10.1111/mec.12046. (PubMed Central (PMC))
Government of Canada. Action Plan for the Atlantic Salmon, Inner Bay of Fundy Population. Defines the Live Gene Bank's genetic-conservation role and the objective of re-establishing wild, self-sustaining populations. (Canada)
U.S. Fish and Wildlife Service. Craig Brook National Fish Hatchery. Conservation program preserving the genetic integrity of seven river-specific Atlantic salmon broodstocks in Maine. (U.S. Fish and Wildlife Service)
NOAA Fisheries. Hatchery and Genetic Management Plans and associated environmental reviews. NOAA materials show use of pHOS/PNI objectives and evaluation of alternative hatchery-production levels, including reduced-production scenarios. (NOAA Fisheries)
European Union. Water Framework Directive 2000/60/EC and Nature Restoration Regulation. River ecological assessment includes fish, hydrology, groundwater connection, continuity, morphology and pollutants; current restoration targets include at least 25,000 km of free-flowing rivers by 2030. (EUR-Lex)
Fisheries and Oceans Canada. Big Qualicum Hatchery. DFO identifies the facility as built in 1963 and B.C.'s first modern enhancement project. (PAC)
Bennett, S., et al. 2016. Progress and Challenges of Testing the Effectiveness of Stream Restoration in the Pacific Northwest Using Intensively Monitored Watersheds. Fisheries 41:92–103. doi:10.1080/03632415.2015.1127805. (Wiley Online Library)
Roni, P., et al. 2010. Estimating Changes in Coho Salmon and Steelhead Abundance from Watershed Restoration: How Much Restoration Is Needed to Measurably Increase Smolt Production? North American Journal of Fisheries Management. doi:10.1577/M09-162.1. (Wiley Online Library)
Copeland, T., Blythe, D., Schoby, W., Felts, E., and Murphy, P. 2021. Population Effect of a Large-Scale Stream Restoration Effort on Chinook Salmon in the Pahsimeroi River, Idaho. River Research and Applications 37:100–110. doi:10.1002/rra.3748. (Wiley Online Library)
Cordoleani, F., et al. 2024. Restoring Freshwater Habitat Mosaics to Promote Resilience of Vulnerable Salmon Populations. Ecosphere 15:e4803. doi:10.1002/ecs2.4803. (The Ecological Society of America)
Spromberg, J.A., et al. 2016. Coho Salmon Spawner Mortality in Western U.S. Urban Watersheds: Bioinfiltration Prevents Lethal Storm Water Impacts. Journal of Applied Ecology 53:398–407. doi:10.1111/1365-2664.12534. (PubMed Central (PMC))
Lo, B.P., et al. 2023. Acute Toxicity of 6PPD-Quinone to Early Life Stage Juvenile Chinook and Coho Salmon. Environmental Toxicology and Chemistry 42:815–822. doi:10.1002/etc.5568. (PubMed)
No comments:
Post a Comment