A salmon hatchery is a building that takes adults, spawns them, rears the young, and puts fish back in the water — by pipe, truck, or net pen. It can keep a lineage on the books. It is not a salmon run.
If you searched salmon hatchery, you are probably writing a budget, a Hatchery and Genetic Management Plan, a recovery docket, or a foundation grant, and you need to know whether the fish that left the building become the run you are paying for. Sometimes they help. The truck can be full. It is still not a riverscape.
what a salmon hatchery actually is
A salmon hatchery is artificial propagation with a street address. In the manager's verbs: collect broodstock, spawn, incubate, rear, mark, release, monitor. NOAA Fisheries reviews West Coast programs through a Hatchery and Genetic Management Plan (HGMP) — broodstock source, release numbers, marking, and genetic and ecological risk controls — not a press release.
People collapse three different jobs into the same word.
- Harvest and mitigation production. Put fish into a fishery, or into a mitigation ledger for hydropower and lost habitat. Columbia River Mitchell Act hatcheries are the famous cousin. The count that matters is catch, or a mitigation credit. It is a real delivery. It is not proof the gravel still makes a run.
- Supplementation. Use the hatchery to increase naturally spawning adults in a target population, usually with local or integrated broodstock, so more fish reach the gravel than the wild return would have supplied. The intent, in the manager's words, is to boost natural production — not only the harvest.
- Conservation broodstock. Hold and spawn a lineage in captivity because natural abundance is too low to carry itself. The near-term job is: do not go extinct this year. Conserve the genetic resources of that stock. A conservation hatchery is trying to produce fish that can still return and reproduce in their native habitat — not a generic smolt for a net.
Those three share plumbing. They do not share an object. Supplementation is still a hatchery release in the ledger; the table below is what you hold, not the program label. A hatchery can be the right tool for a lineage and still leave you holding a number instead of a run.
The fish and the riverscape exist whether or not anyone stocks a hatchery, fills a truck, or buys a certificate. The hatchery is a tool. The salmon run is a living fish moving through a living riverscape — cold, slow, connected water from gravel to ocean and back.
three things people call the same thing
Recovery memos, hatchery budgets, and foundation dashboards use these as if they were interchangeable. They are not the same object. Only the last row is the run.
| what it is | what you hold | what it proves | what it does not do | |
|---|---|---|---|---|
| hatchery release | Fry, parr, or smolts from a production, mitigation, or supplementation program, piped or trucked to water | A delivery of fish and a release count | That fish left the building, marked or unmarked, into a named reach | Make a self-sustaining run, or prove the water is still cold, slow, and complex enough to rear the next year |
| conservation broodstock | Captive spawning to keep a named lineage from blinking out when wild returns cannot carry it | A genetic insurance policy — a stock still on the books | That this lineage did not go extinct this year | Substitute for gravel, cold water, connected floodplain, or treaty harvest in a living riverscape |
| wild-origin spawners in a living riverscape | Fish born in gravel, reared in slow cold water, ocean, return to spawn where the rooms of the house still work | The run | That the place still makes salmon | Issue from a truck. You cannot hatchery a riverscape into existence |
Read the last row first. A hatchery release can be how you get more fish onto the second row, or how you keep the second row from going to zero. It is never, by itself, the third.
A hatchery fish is not a salmon run. The count can be real. It is still not the riverscape.
when the hatchery is the right tool
Hold both.
A hatchery can keep a lineage from blinking out. That is not a slogan. In May 1991, Idaho Fish and Game, NOAA, and the Shoshone-Bannock Tribes started a captive broodstock program for Snake River sockeye (Oncorhynchus nerka) — before NOAA listed the evolutionarily significant unit endangered that November. The near-term job was to conserve the unique genetics of the Redfish Lake stock in Idaho's Sawtooth Valley. Restoration work now focuses on Redfish, Alturas, and Pettit lakes. Oregon and Bonneville Power Administration sit in the cooperator list. The tribes' habitat and lake work is part of the program, not scenery around it.
The program's own long-term language still points past the building: rebuild abundance, and provide sport and treaty harvest. NOAA's delisting frame still asks for natural-origin adults in those lakes — not a hatchery production total booked as recovery. The people running Eagle, Springfield, and the Washington captive facilities already know the difference. Honor the work. The hatchery kept a lineage on the planet. That is still not the run.
In basins where these fish are already a treaty right, a First Food, and a relative, a lineage that blinks out is not a spreadsheet miss. It is a broken obligation. A captive program that keeps Snake River sockeye extant is doing a real job. So is a tribal or state hatchery that puts fish in a fishery while the riverscape is still being rebuilt. Do not sneer at the truck that prevented a zero.
You might be thinking this is an argument that hatcheries ruined salmon, or that hatchery managers should find other work. It is not. NOAA, state agencies, tribal hatchery programs, and Mitchell Act crews are cousins. Critique the object held — a release count versus a living run — not the person who spawned the pair.
when the truck is the trap
The trap is calling a truck a run — in a recovery spreadsheet, a season announcement, a mitigation report, or a foundation dashboard that closes when the last smolt hits the water.
The other trap is stocking while the water is still too warm, too fast, and too straight. Juveniles need slow, cold, complex water. A full truck into a channelized ditch is a competent delivery into a missing riverscape. The hatchery did its job. The house still has no rooms.
A third trap is stopping the sentence at supplementation. Integrated broodstock, local origin, and a high share of natural-origin spawners in the hatchery are how serious programs try not to replace the wild population with a cultured one. Those are real practices. They still do not make the floodplain. They still do not make the side channel. They still do not make the estuary room of the house — that coast has its own post.
Carcasses on the gravel still move marine nutrients into the forest. That is a different invoice, already written: the fish that fertilizes forests. This post is the living run, not the nutrient after the spawn.
genetic and fitness claims, labeled
Do not turn this into a hatchery war. Name what is supported, what is contested, and what is not a single cause.
| claim | confidence | what it means on the ground |
|---|---|---|
| Hatchery-origin salmonids often have lower reproductive success in the wild than natural-origin fish of the same stock | Supported for relative reproductive success, strongest in steelhead pedigree studies (Araki and colleagues on Hood River; later reviews across salmonids) | A release count is not a spawning count. Fish that left the building are not automatically the next generation in the gravel |
| Captivity can select for traits that work in the raceway and fail in the creek (domestication) | Supported as a mechanism | The building is a different selective environment than gravel. That is why conservation programs talk about local broodstock, natural-origin parents, and rearing that is less like a factory |
| How large the fitness penalty is, how fast it accumulates, and whether a well-integrated conservation program avoids dragging the wild population | Contested / program-dependent | HSRG metrics (pHOS, pNOB, PNI) exist because gene flow is treated as managed risk, not a closed case. Results vary by species, broodstock, and how the fish are released |
| Hatcheries are why salmon declined | Not supported as a single cause | Habitat, harvest, hydropower, ocean conditions, and hatcheries interact. Blaming the hatchery manager for a missing riverscape is the wrong ledger |
Steelhead (Oncorhynchus mykiss) carries some of the clearest pedigree evidence. Do not paste those percentages onto every Chinook or sockeye program and call it settled. Do not ignore them either. A recovery biologist already lives in that tension. This post is not here to relitigate it. The operational point is smaller: even when the genetics are well managed, the truck is still not the riverscape.
the same honesty, other living systems
A planted tree is not a forest — the bag can be full; it is still not a stand.
A nursery fragment is not a reef — the tree can be full; it is still not an animal city.
A rented hive is not pollination — the box can be real; it is still not the living service.
Same shape. Different stock. Keep the tool if it becomes the system. Hold the living system if that is the thing you actually want.
keep the hatchery. hold the riverscape.
If you already fund or run a hatchery, keep it — and change what you report.
- Report the run, not only the release. Smolts out the pipe is a hatchery metric. Wild-origin spawners in water that can still rear juveniles is a run metric.
- Keep two ledgers. Conservation broodstock that prevented a zero is one line. A living riverscape that still makes the next year is a different line. The second line is the one NOAA delisting language already points at.
- Do not stock a ditch and call it recovery. If the water is still too warm and too straight, the truck is a delivery into a missing riverscape. Fund the rooms of the house — gravel, slow water, connection — on a named place.
- If the run is the point, fund the riverscape as a present-tense asset. ensurance is how beneficiaries fund that condition now: a certificate is a 1:1 hold on one agent for a named place, people, or purpose. It is not a hatchery invoice and not a salmon token. Live doors are small:
salmon.syndicate,rivers-lakes.ensurance,habitat.ensurance. There is no$SILVERand nosalmon.ensurance. Price is a bridge. It is not the worth of the fish.
You already pay for the failure either way — hatchery budgets, trucked fish, closed seasons, lost treaty harvest, emergency trap-and-haul. The hatchery can stay. The object you are trying to hold is the run.
Keep the hatchery if it becomes a run. Hold the riverscape if you want the fish.
you don't have a salmon problem. you have a missing riverscape →