Four Dams on the Snake River Make Clean Power, So Why Tear Them Down

Here is a policy puzzle for you. A state passes a law demanding that every kilowatt-hour it sells be carbon-free by 2045. Then a movement in the same corner of the country proposes to blow up four of the largest carbon-free power plants in the region. Both of these are called environmentalism, and both are supposed to be good for the planet. One of them has to be wrong.

I am talking about the four dams on the lower Snake River in Washington State, and the long campaign to breach them for the sake of the salmon. I will give the salmon case its strongest statement, and I will do it first, because it is a serious case and the people who make it are serious people. Then I will answer it, and I will tell you where I think it goes wrong.

The Four Dams

The lower Snake River runs from the Idaho border through southeastern Washington to meet the Columbia River near the Tri-Cities. Four federal dams sit on that stretch, all built by the Army Corps of Engineers. From downstream to upstream they are:

• Ice Harbor, completed in 1961.
• Lower Monumental, completed in 1969.
• Little Goose, completed in 1970.
• Lower Granite, completed in 1975.

Together they do two big jobs. They generate electricity, and they create a slack-water channel deep enough for barges to travel inland from the Columbia to the port of Lewiston, Idaho, a river port hundreds of miles from the ocean. They also provide some irrigation water, though that is a smaller part of the story.

The power is substantial. The dams are part of the federal Columbia River power system, whose electricity is marketed by the Bonneville Power Administration. The four together produce on the order of a thousand average megawatts, enough for a large city, and they are carbon-free and dispatchable. Dispatchable is the word that matters: unlike the sun and the wind, a hydroelectric dam can raise or lower its output on demand, which is what makes it useful for balancing a grid that has more and more intermittent sources. I give you that figure as an approximation. You can confirm the current numbers on the Bonneville Power Administration's website.

The barging matters too. The river carries a large amount of grain, especially wheat grown in the Palouse and across the inland Northwest, down to export terminals on the lower Columbia. By the Port of Lewiston's figures, one barge carries about as much as 35 rail hopper cars or 134 semitrucks, and barge advocates say it moves that cargo with less fuel per ton than either. Take the river out of the system, and that grain has to move another way.

The Salmon Case, at Its Strongest

Now the other side. I want to give it as its defenders would.

Salmon are an icon of the Northwest and a foundation of tribal culture. The tribes of the region fished the rivers for thousands of years, and in the treaties of the 1850s, tribes such as the Nez Perce reserved the right to take fish at their usual and accustomed places. Those treaties are the law of the land under Article VI of the Constitution, which makes treaties "the supreme Law of the Land." That is not a small point, and anyone who brushes it aside is not taking the Constitution seriously.

Snake River salmon and steelhead have declined badly. Several runs are listed under the Endangered Species Act, among them Snake River sockeye, which was listed as endangered in 1991, and Snake River spring and summer chinook, which were listed as threatened in 1992. Young salmon heading to the ocean must pass the dams, and adults returning to spawn must climb past them. The reservoirs warm the water in summer, slow the migration and create habitat for predators. The advocates of breaching argue that nothing short of removing the four lower dams, and restoring the free-flowing river, will bring the runs back to healthy levels. They point to the distance the fish must travel to the high-elevation spawning grounds in Idaho and note that cold water upstream may become more valuable as the climate warms.

That is a real argument, made by biologists, tribal leaders and fishermen who have devoted their lives to the question. I do not wave it away. A free society owes its treaty commitments, and it owes the salmon something too.

One: Clean Power Is the Whole Point

Here is the first answer. The dams produce a very large amount of carbon-free electricity, and the states that would lose it have laws demanding more of exactly that. Washington's Clean Energy Transformation Act, passed in 2019, requires the state's electricity to be carbon-neutral by 2030 and entirely free of greenhouse gas emissions from fossil sources by 2045. Oregon has its own clean-electricity targets. Both states want power that is clean and also reliable.

So how do you replace a thousand average megawatts of dispatchable, emission-free power? The wind does not blow on demand, and the sun sets before the evening peak. You can build batteries, but they are expensive and short-lived at the scale required. You can build more transmission, but see We Raised the Empire State Building in About a Year, So Why Not a Power Line for how long that takes. In practice, the plausible replacement for a lost source of firm power is natural gas. If the region tears down a carbon-free plant and replaces part of its output with a gas plant, the net effect on emissions is upward.

You cannot hold both positions. Either carbon matters, in which case you do not remove the carbon-free source, or it does not, in which case the climate argument for the clean-energy mandates is a sham. The conflict is not mine. It belongs to the people who want both.

Two: The Fish Have Been Getting Help

The second answer is that the dams are not the static, hostile obstacles the picture suggests. Over decades, a great deal has been done to help fish pass them.

Each dam has fish ladders, a series of stepped pools that let adult salmon climb past. Each has juvenile bypass systems that guide young fish away from the turbines. Operators release water over the spillways at certain times to carry smolts past the dams without going through the machinery, and engineers have installed spillway weirs and other devices that make that passage gentler. A fish transportation program has carried young salmon by barge past the dams. The Lower Snake River Compensation Plan, authorised by Congress in 1976, funds hatcheries to offset the fish lost to the dams' construction. The federal government has spent billions across the Columbia system on this kind of mitigation, and survival through the dams for juvenile fish has improved substantially compared with the early years.

None of that proves the runs are healthy. It shows that the question is not "dams or no dams" but "what is the net effect of all the changes?" and that the answer is a matter of degree.

Three: The Ocean Is Half the Life Cycle

The third answer is the one that gets the least attention. A salmon spends part of its life in a river and a good deal of it in the Pacific, and whatever happens at sea has an enormous effect on how many return.

Ocean conditions swing. The Pacific has long cycles of warm and cool phases that affect the food supply for young salmon, and some years the survival rate at sea is far better than in others. Warm-water events, such as the marine heat wave in the northeast Pacific around 2014 and 2015, have been tied to poor survival for a number of fish stocks. Predators at the river mouth and along the way, including sea lions, terns and cormorants, take a significant share. Harvest and hatchery practices play roles. So does what happens in tributaries and habitat across the whole basin.

All of this means that a single intervention, such as removing four dams, is not a guarantee. In good ocean years, returns can surge, and in bad years they can crash, regardless of the dams. A claim that breaching would restore the fish has to deal with the fact that the ocean is a major variable no dam can control. That does not mean the dams are harmless. It means that betting a regional power and transport system on one factor, out of many, is a risky wager.

Four: The Klamath Comparison

Advocates of breaching often point to the Klamath River on the California-Oregon border. Four dams there, Iron Gate, Copco 1, Copco 2 and J.C. Boyle, were removed in what was described as the largest dam removal project in American history. The reservoirs were drained in early 2024, and by the end of that year the dams had been taken out. The river now flows freely for the first time in more than a century.

I think that is wonderful news, and I will say so plainly. It is a fine outcome and the people who worked on it deserve credit.

But look at the differences, because the comparison cuts the other way. The Klamath dams were privately owned, built for power, and produced a modest amount of electricity, with a combined capacity of about 163 megawatts by the removal project's own figures, against roughly 3,000 megawatts of nameplate capacity at the four Snake dams. They were old and would have needed expensive upgrades to be relicensed, and their owner, a utility, agreed to removal as part of a negotiated settlement with tribes, states and others. They had no navigation role, and no wheat barges depended on them.

The Snake River dams are a different animal. They are federal projects, they generate far more power, they support a working inland waterway, and removal requires an act of Congress, since Congress authorised them. The White House Council on Environmental Quality said in 2023 that breaching would need approval from Congress. The Klamath succeeded because the costs were small, the owner was willing and the parties agreed. None of those conditions applies on the Snake. To say "it worked there, so do it here" is to compare a pickup truck to a freight train.

The Argument in Numbered Form

1. The dams are major carbon-free, dispatchable power sources in a region that has legislated an end to fossil-fuel electricity.
2. They support a working waterway that moves grain more cheaply and with less fuel than the alternatives.
3. A great deal has been done to protect the fish, from ladders to bypass systems to hatcheries, and passage survival has improved.
4. Ocean conditions and predation are major factors that breaching would not touch.
5. The Klamath is not a precedent, since those dams were small, private, and removed by agreement, while the Snake dams are large federal assets whose removal needs Congress.

Taken together, these points say that removal is a very expensive, very uncertain gamble with benefits that cannot be guaranteed, and costs that are certain.

The Constitutional Anchor

There is a reason a tear-down is not something an agency can do by itself. Article I, Section 8 gives Congress the power "To regulate Commerce with foreign Nations, and among the several States," and navigation on the nation's rivers is a classic exercise of it. The lower Snake projects were authorised by Congress, and what Congress built, Congress can unbuild. That is how the Framers meant it to work. A matter that affects power for millions of ratepayers, the livelihoods of farmers and shippers, and the treaty rights of tribes is the kind of matter a legislature should decide, in public, after a vote.

The trouble with the present approach is that the decision keeps being pursued by indirection: through lawsuits over fish plans, through agency studies and through negotiated agreements that never quite call themselves a decision. A free people has the right to see the choice plainly and to know who made it. For other cases of the federal government managing its large holdings, see Why Washington Owns Most of Nevada and What the Property Clause Allows.

What Fairness Demands

Let me be fair to the other side before I close. The salmon matter. The tribal treaty rights matter. A policy that kept every dam and did nothing else would deserve criticism. The question is how to honour both the commitment to the fish and the need for clean power.

There are options short of breaching: continued improvement in passage, more spill where it helps, better predator control, habitat work in the tributaries, investments in hatcheries and, yes, a serious conversation with the tribes about what a fair package would include. A deal that paid for improvements and carried real commitments on both sides would be better than a fight in court. The test is simple: will the proposed measure produce more fish, at a cost we can bear, without sacrificing the clean power and the grain trade that people depend on? If the answer is yes, adopt it. If the answer is a guess, ask for evidence.

This is the same standard I apply everywhere. Climate policy and conservation both ought to be judged by results, and a project that reduces a clean power source while claiming to help the environment owes the public an explanation. For a wider look at what a clean-power mix requires, see Count Deaths per Unit of Power and Nuclear Beats Nearly Everything Else, and for how federal land and species laws can set people against their own goals, see Why the Endangered Species Act Punishes the Landowner Who Helps a Bird.

The Case in Brief

1. The four lower Snake River dams, Ice Harbor, Lower Monumental, Little Goose and Lower Granite, supply carbon-free, dispatchable power and support grain barging to Lewiston.
2. The salmon case is serious, resting on treaty rights, endangered listings and the effects of reservoirs on fish passage.
3. But the dams already carry ladders, bypass systems and hatcheries, ocean conditions heavily influence returns, and removal offers no guarantee.
4. The Klamath dam removals completed in 2024 are not a fair precedent, since those dams were small, privately owned and removed by agreement.
5. Removal would need an act of Congress, and the public should demand that the choice be made openly, with the cost to clean power counted.

What you can do: Write to your senators and representative asking whether they support or oppose breaching the lower Snake River dams, and ask what carbon-free source would replace the lost power.