2026-06-30

Still Early Days on the Klamath:

Salmon Are Returning, but a Warming River Is Making Them Sick

A Water Policy Series — June 2026

Introduction

Two truths arrived on the Klamath in the same season, and they do not agree.

The first is a triumph. A little more than a year after crews finished pulling four hydroelectric dams out of the Klamath River — the largest dam removal in the history of the United States — Chinook salmon have swum more than 360 river miles inland to the confluence of the Sprague and Williamson rivers in southern Oregon, water no salmon had reached in over a hundred years.¹ More than ten thousand large fish passed the site of the former Iron Gate Dam last fall, and biologists counted roughly sixty-five thousand wild juvenile Chinook in Fall Creek alone.¹ Bald eagles, otters, and beavers moved back into a valley that had been a reservoir. By every measure the engineers set, the river did what a freed river is supposed to do.

The second truth is a warning. This spring, in one of the driest years on record, Oregon State University researchers found Ceratonova shasta — the parasite that kills juvenile salmon — at all six of their monitoring sites, and a federal health report found forty-six percent of the young Chinook it tested in the upper river infected, more than double last year’s rate.² The bulk of the dying traces to hatchery stock, and the conditions driving it — low snowpack, early melt, low flows, warm water — are precisely the conditions a parasite prefers.² The same river that carried salmon to Oregon for the first time in a century is, this year, making their offspring sick.

Both are true, and holding them together is the whole point. “Dam removal gave the Klamath River and our fish a fighting chance,” the Karuk Tribe’s vice chairman, Kenneth Brink, put it. “But no one should confuse dam removal with an instant cure.”² Taking out the concrete was the visible act, the photographed one, the one that felt like an ending. It was the easy part. The river’s real problem was never only the dam, and pulling the dam does not solve it — it hands it back, unblocked, to a warming climate and to the people who must now manage a wild river instead of a plumbed one. That is the harder story, and it is the one the decade of removals now lining up behind the Klamath will each have to live out.

I. The Cure That Wasn’t a Cure

The science, the Karuk’s senior biologist Toz Soto said, “is telling us two things at once.”² Salmon are reoccupying their historic range faster than almost anyone predicted; this year’s drought is punishing the young fish that range produced. The temptation is to grab whichever truth fits the story you already wanted to tell — a vindication for those who fought for removal, an I-told-you-so for those who opposed it. Both readings are wrong, because both mistake a single season for a verdict.

What actually happened is that removal fixed one thing and exposed another. For a century, the dams were a bottleneck: they physically blocked salmon from reaching 400 miles of cold upstream habitat, and they created warm, still reservoirs that bred the very parasite now in the news. After removal, both improved at once — the fish reached the habitat, and, in the first two springs, C. shasta prevalence and the river’s toxic algal blooms both fell below what the dammed river had produced in comparable years.^{1 3} That is the part the triumphant coverage got right.

But a free-flowing river is not a managed one. The parasite has a life cycle that depends on a tiny worm living in the riverbed; the thing that breaks that cycle is high water — winter and spring flows strong enough to scour the worm and flush the spores downstream before they can find a fish.³ A dam could never provide those flushing flows; that was one of the arguments for taking it out. But neither can a drought. Removal restored the river’s ability to flush itself. It could not give the river the snowpack to do the flushing. So in a dry year the restored river behaves like what it now is — a wild system at the mercy of the weather — and the weather this year was merciless. “One good year or one bad year will not define the future of this river,” Soto said. “What we need now are several winters with strong flushing flows.”²

That last point is worth sitting with, because it is easy to hear “recovery” and picture a patient healing on a calendar — a little better every year until the infection rate quietly falls to zero. That is not how this works. The parasite’s cycle does not degrade with the passage of time; it breaks only if that year delivers the flow. A wet winter can scour the riverbed clean and drop infection rates sharply; a dry one, like this one, can undo that progress just as fast, because the worm host and its spores are always ready to rebuild the moment flows go slack. So the trajectory is real — the first two post-removal years were better than comparable dammed years — but it will look like a sawtooth, not a smooth curve, for as long as the Klamath basin keeps handing out drought years at the rate it has lately. The cure was never the dam’s removal, and it is not simply time either. The cure is repeated high water, year after year, and no one controls the snowpack that produces it.

II. Why the Dams Came Down (and It Was Not the Salmon)

Here is the part of the story the salmon photographs obscure: the Klamath dams did not come out because we decided the fish were worth more than the power. They came out because the dams stopped making financial sense, and the fish were the beneficiaries of an accounting decision.

The four dams generated something like 700,000 to 800,000 megawatt-hours a year — enough for roughly seventy thousand homes.⁴ That sounds like a great deal to give up until you read the ledger the way the utility, PacifiCorp, was forced to. To keep operating, the dams needed federal regulatory relicensing, which would have required more than $400 million in mandated fish ladders and water-quality fixes.⁵ An early federal estimate concluded the dams would lose about twenty million dollars a year once those obligations were counted, and both states’ utility commissions found that removal would mean cheaper power for ratepayers than keeping the dams alive.⁵ Decommissioning, by contrast, was capped at $200 million.⁵ The dams had become what advocates bluntly call deadbeat dams: aging structures whose costs and liabilities had outrun their usefulness.

Follow the money and the point sharpens. The removal’s roughly $450 million price was met not by a grand federal appropriation for salmon but by a $200 million surcharge on PacifiCorp’s own customers — who were, in effect, paying the cheaper of two bills — and $250 million from a California water bond voters had passed years earlier.^{5 6} The ratepayers funded it because removal beat relicensing on cost. This is not a cynical reading; it is the load-bearing one. It explains which dams come down and which do not, and it means the salmon’s champions won the Klamath by making an argument about liability, not about beauty. The fish rode in on the accounting.

That accounting logic has two consequences worth separating. The first is about which dams come down at all: removal is available wherever a dam’s costs now exceed its value, as they did on the Klamath, and it is not available — not yet, not cheaply — where a dam still earns its keep. That test, not sentiment about salmon, is what will decide the next round of removals. The second consequence is easier to miss: the same deal that pays to pull out the concrete has nothing to say about what happens to the fish afterward. No line item in that $450 million covered flushing flows, hatchery reform, or a parasite outbreak, because none of those were the problems the accounting was solving. The ledger that freed the river closes the moment the concrete is gone. The biology is just beginning, and it was never on the books.

III. What Removal Undid, and What It Left

It helps to be precise about what the act of removal actually accomplished, because the process itself reveals how much of the problem it never touched. Pulling a dam is mostly not demolition. The concrete comes out in a single season; the hard, year-long work is sediment — the decades of muck trapped behind the dam that, released too fast, would bury the river downstream.⁷ On the Klamath, the engineers drew the reservoirs down slowly through the winter of 2024, timing the release to high flows so the sediment moved through and the water cleared by spring, then spent the following year replanting a reservoir footprint with billions of native seeds.⁷ It was a triumph of managed subtraction: the careful undoing of a single obstruction.

But notice what that undoing addresses and what it does not. It restores *connection. That matters downstream because it tells you the tool’s limit. Removal is available wherever a dam’s costs exceed its value. It is not available — not yet, not cheaply — where a dam still earns its keep. And nothing about the accounting that freed the river does anything to feed the fish once it is free. The books close the moment the concrete is gone. The biology is just beginning. It restores flow — the river runs cold and continuous where it once pooled warm behind concrete. What it cannot restore is a climate. The dam was a barrier; remove it and the barrier is gone. The drought is not a barrier. The warming that lets a parasite thrive is not a barrier. The dense pulse of hatchery fish that carries disease is not a barrier. Removal is the right and necessary answer to the barrier, and it is simply the wrong shape of tool for everything the barrier was hiding — the slow, distributed, weather-driven pressures that were always the more serious threat and that a hundred years of dammed river let everyone ignore.

This is why the 46% infection figure is not a refutation of removal but rather its natural sequel. Behind the dam, the river’s health was somebody’s operational problem — a matter of flows to be released, temperatures to be managed, a machine to be run. In front of the freed river, health becomes an ecological problem that is harder to address because no one manages it. The Klamath did not trade a sick river for a well one. It traded a controlled river for a wild one, and a wild river in a warming basin is going to have bad years. The question removal poses — and cannot answer — is whether the good years will outnumber them.

IV. The Hard Part

If the dam was the easy part, the hard part has a name too: keeping fish alive in a river that is warming, crowded with predators, and prone to disease. And here the frontier of the work is quietly shifting, in a way worth tracing, from getting fish past concrete to keeping fish alive in open water.

For a century the science of salmon-and-dams was a science of passage — fish ladders for adults going up, bypass screens and spillway weirs for juveniles coming down. That work succeeded on its own terms: adult ladders on the Columbia now pass better than ninety-six percent of the fish that reach them.⁸ But passage was always a solution to the barrier, and the Klamath’s lesson is that the barrier was never the whole disease. The newer work aims at the parts removal exposes. Engineers at firms like Natel have built turbines with thick, forward-swept blades that let juvenile fish pass through a working dam at 98 to 100 percent survival — a way to make the dams that stay far less lethal, since most dams will stay.⁹ Biologists are mapping and rebuilding thermal refuges — the cold groundwater seeps and side channels where young fish shelter when the main stem turns lethal — because in a warming river the cold water is the habitat that matters most.¹⁰ And hatchery managers are rethinking the very practice at the center of this year’s Klamath die-off — a different kind of fix than the flushing flows above, one that changes the fish’s odds rather than the parasite’s biology. A traditional hatchery release dumps a large, synchronized batch of juveniles into the river at once, which concentrates a great many susceptible fish together in the same infectious water at the same time; more hosts moving through spore-laden water simultaneously means more infections per unit of exposure, regardless of how much C. shasta is actually out there. The reform — release smaller, staggered batches forced dump¹¹ — lowers that host density and tends to produce bigger, hardier fish that clear the risky stretch of river faster. It is damage control, not a cure: it shrinks how badly a bad year hurts, but only repeated flushing flows shrink the parasite’s presence in the riverbed itself.

None of these is a headline the way a falling dam is a headline. That is exactly the problem. Removal is a discrete, photogenic act with a ribbon-cutting; thermal-refuge restoration, hatchery reform, and predator management are diffuse, unglamorous, and never finished. They are the maintenance the river will need for as long as the climate keeps warming — the work that begins the day the cameras leave. The Karuk have said from the start that the river “is still in the early stages of healing.”² Healing is not an event. It is a practice, and the practice is the hard part.

V. The Next Rivers

None of this is an argument against removal — it is an argument for understanding what removal buys, because the Klamath is not the end of a story but the template for the next one. The pipeline behind it is already forming, and every river in it will meet the same lesson at the same door.

The clearest heir is the Eel River in California, where Pacific Gas & Electric filed in July 2025 to surrender its license and tear out Scott Dam and Cape Horn Dam — obsolete structures that no longer generate power and that block 288 miles of habitat for threatened salmon and steelhead.¹² The Eel is the Klamath’s model made explicit: the dams are deadbeat, the owner wants out, and — the piece that usually kills these projects — a water-diversion agreement signed the same summer arranges to replace the supply the dams once sent to the Russian River, defusing the downstream fight before it starts.¹² Behind it wait Matilija Dam near Ventura, a structure so filled with sediment it stores almost nothing while blocking steelhead, and Enloe Dam in Washington, a power-less concrete wall that has sealed off 1,520 miles of habitat since 1922 and whose removal study was completed this year.¹³ And looming over all of them is the great unresolved case, the four Lower Snake River dams, whose roughly 1,000 megawatts and barge navigation make them the fish-recovery prize no one has figured out how to afford politically — a reminder that where a dam still earns its keep, the Klamath’s accounting does not apply, and the salmon lose.¹⁴

The scale is larger than any single river. In 2025 the country removed a hundred dams and reconnected nearly five thousand miles of stream — the most in a single year in American history — and the advocates’ stated goal is thirty thousand dams by 2050.¹⁵ Most of those will be small, obscure, deadbeat barriers whose removal is a clear and easy good. But each one will teach its watershed the Klamath’s two-truths lesson: that the falling dam is the beginning of the work, not the end of it, and that the river on the far side of removal is a wild thing that will still need tending through every dry year the century has left to send.

Conclusion

Look at the coverage of the Klamath and the shape of the mistake becomes clear. The eye stops at the surface — at the dam removal, the salmon surging north, the ribbon cut and the cameras gone — and reads the story as finished there. But the story isn’t finished — it keeps unfolding long after the hype ends. Below the visible act of removal runs the part no photograph caught: a warming basin, a parasite that thrives in low water, a hatchery practice due for reform, and years of flushing flows the river will need and the sky may not provide.

The people who fought for the Klamath’s dams to come down were right, and the river is better for their work. But the truest thing said about it this year came from the tribe that has watched the river longest: do not confuse removal with a cure. The dam was the barrier, and the barrier is gone, and that was the easy part — a single obstruction, carefully subtracted, paid for because the arithmetic finally favored the fish. What remains is harder and has no ribbon: to keep a wild river’s young alive through the bad years, on purpose, indefinitely, in a climate that is not going to help. The falling dam looked like the finish line. It was the starting line, and the race it started is the one that never ends.

Sources