East of Portland, the Sandy River purls downhill from the glaciers of Mount Hood and flows into the Columbia. Along its fifty-five miles, the Sandy has it all: icy cold whitewater, wilderness, old growth, and wild fish—Chinook and coho salmon, steelhead, and coastal cutthroat trout. Over two-thirds of the watershed is national forest, wilderness, or protected conservancy areas.
In the early twentieth century, Portland General Electric (PGE) engineered the Sandy's watershed to power electric lights in Portland. The Bull Run Hydroelectric Project included the Marmot Dam on the Sandy River and a smaller dam on the Little Sandy River. River water was diverted into wooden flumes and eventually into tunnels dug by hand through mountains. Finally it dropped six hundred feet to the powerhouse on Bull Run River. The elaborate hydraulic engineering powered the razzle-dazzle of a growing Portland. It also blocked migrating salmon and steelhead from about 280 miles of high-quality river habitat, mostly above Marmot Dam.
Like all private dams in the United States, the Marmot Dam operated under a license from the Federal Energy Regulatory Commission (FERC). Licenses are issued for thirty- to fifty-year periods and when they expire, dam owners have to apply for renewal. Increasingly, salmon and watershed advocates are looking at FERC relicensing as a once-in-a-career opportunity to overhaul dam operations so that ecological needs are balanced with human needs. Conditions attached to a new FERC license can result in far more watershed restoration than a hundred smaller projects—riverbank plantings, added logjams—could ever accomplish.
By 1999, the FERC license for the Bull Run Project was in its final years. John Esler, currently PGE's licensing project manager, was working on the Sandy relicensing application when he realized it didn't pencil out. The project produced only a fraction of PGE's power supply at a high cost per kilowatt hour and, because the Sandy River's spring and fall Chinook salmon, winter steelhead, and coho salmon were listed as threatened species under the federal Endangered Species Act, he knew PGE would be required to add fish passage and other mitigations as conditions of relicensing.
PGE went to federal and state agencies and proposed dismantling the hydro project instead of relicensing it. The company ran into opposition where it expected support: from the state agency for salmon, the Oregon Department of Fish and Wildlife (ODFW). ODFW didn't want to lose the money PGE gave it for a salmon hatchery, as mitigation for the dam's impacts on wild salmon. Esler says, “Agencies don't hate dams, they hate change. They're addicted to the mitigation money they get from dams.” Eventually, ODFW and all other agencies supported the dam removal.
Elwha Dam ceremony, September 17, 2011. Bottom: Early stages of deconstruction of Glines Canyon Dam. Photos by John Gussman
In July 2007, after years of negotiations and planning, the forty-seven-foot-high Marmot Dam was dynamited. As planned, a temporary earthen cofferdam continued to block the river until fall rains raised the river level. On October 19, rising water breached the cofferdam. US Forest Service research hydrologist Gordon Grant and other scientists were watching. What the river did next amazed these seasoned scientists who study rising rivers for a living: it immediately began to cleanse itself of the one million cubic yards of sediment trapped in the reservoir. A knickpoint—where a river has a sudden change in gradient—formed on the cofferdam, which eroded quickly over the next forty-five minutes. As the eroding current got stronger, the knickpoint moved upstream at hundreds of yards per hour and the 165-foot-wide cofferdam eroded in several hours. Next morning, a new gravel bar downstream was the only sign of the cofferdam.
In one day the river transformed downstream from a single channel to a braided channel. It created new gravel bars and islands and resculpted riverbanks. It flushed reservoir sediments downstream in waves. Grant says that about 20 percent, or two hundred thousand cubic yards, of the stored sediment was exported within the first two days, “exceeding all expectations.”
Three days after the initial breach—and weeks earlier than fish biologists had predicted—coho salmon swam upriver through the breach. They swam up a river still turbid with sediment into habitat where coho hadn't been in nearly a century.
The Marmot Dam is part of a small, select group of large dams in the United States that have been removed in the past decade. Close to eighty thousand large dams (defined as higher than twenty-five feet or storing more than sixteen million gallons) have been built in North America, most in the twentieth century. In the last decade, according to the American Rivers organization, from twenty to fifty dams per year have been removed or decommissioned. The removals total less than 1 percent of dams, but even this small number represents a sea change in thinking: from constructing dams as tools for human convenience to removing dams to restore rivers and fish, including salmon. If dam removal is seen as a quick fix, however, it reflects an arrogance that humans can engineer natural systems and completely control the results. Dam removal can become the latest iteration of the simplistic view that rivers are merely channels with dams as stoppers that can be put in and taken out.
Elwha Dam, before deconstruction. Photo by Valerie Rapp
In the late 1800s, engineering advances made it feasible to build large dams and the invention of alternating current made it possible to transmit electricity long distances: river water could be turned into abundant electricity, replacing coal's foul smoke and dangerous kerosene.
Dams on the Pacific Northwest's rivers solved other problems as well. Winter storms flooded young cities like Eugene and Portland. Untreated sewage, pulp mill wastes, and other industrial effluents were discharged directly into the Willamette and other rivers. In winter, all these foul discharges flushed downstream, but during low summer flows, a stench rose from rivers. Dams in the mountains controlled disastrous winter floods and released water gradually over the summer, diluting the wastes poured into rivers.
Demands for electricity, flood control, irrigation water, and waste dilution propelled the construction of large dams throughout the Pacific Northwest. When the last dams were finished in the 1970s, three-fourths of the Pacific Northwest's power came from hydroelectricity, and electricity rates were the cheapest in the nation. Water from reservoirs irrigated millions of acres in Idaho, Washington, and Oregon.
Since little was known about river ecology, people believed that the year-round stability of river flows benefited fish and rivers as well. But we now know that a dam fixes a river like neutering fixes a male dog. The dam cuts the river's potency, upstream and down. Below the dam, the river is starved of the driftwood logs, gravel, and sand that build riverbeds, gravel bars, and salmon spawning habitat. Not surprisingly, a fixed channel eventually suffers channel shrinkage. Water flow is regulated to meet human needs. Wild salmon, whose life histories are finely adapted to rivers' natural flood pulses and flow changes, encounter water that is low or high at the wrong seasons. Adult salmon can climb fish ladders, but young salmon migrating downstream are chewed up and spit out from dam turbines.
The decline or loss of wild salmon is a loss for the entire ecosystem. In undammed salmon rivers, scientists have found that up to 30 percent of the nitrogen in the upstream food web derives from ocean sources, carried upstream in the bodies of spawning salmon. Salmon's upstream migration brings ocean nutrients to mountain watersheds, a natural fertilization cycle that enriches forests.
The very first day of dam de-construction, September 15, 2011. Photo by John Gussman
Just as the last large dams were completed in the late twentieth century, several factors converged to create a time when dam removal became a real possibility. Many wild salmon stocks were being listed as threatened or endangered, the first dams built were about a hundred years old and needed upgrades for safety, and the regional power grid had enough power sloshing around that the loss of a few small power sources was not catastrophic. Watershed councils were connecting people to their watersheds through salmon festivals and weekend projects such as riverbank plantings. The possibility of a fast explosive cure for damaged rivers was alluring.
But slicing through the snarls of a knotted-up, damaged ecosystem is not the same as thoughtfully teasing apart the tangles. Successful dam removal takes teams of hydrologists, fish biologists, engineers, botanists, and economists to look at the whole watershed and its ability to rebuild itself. Teams—and local communities—can see what costs, benefits, and tradeoffs, ecological and cultural, are possible in that watershed.
The complexity of a well-designed dam removal is evident in the story of the biggest dam removal ever done in the United States, and probably the biggest dam removal in the world: taking out the two dams on the Elwha River on Washington's Olympic Peninsula.
Glacial meltwater from the Olympic Mountains plunges through old-growth forest, deep gorges, and open bottomland, nourishing the Elwha River. Historically, the Elwha had ten runs of anadromous fish—spring and fall Chinook, coho, pink, chum, and sockeye salmon, plus summer and winter steelhead, sea-run cutthroat trout, and sea-run bull trout. Although the Elwha is only forty-five miles long with a hundred miles of tributary streams, about four hundred thousand salmon used to return every year.
The 105-foot-tall Elwha Dam was completed in 1913 and completely blocked salmon from all but the lower five miles of the river. Although state law required fish passage, none was provided. Run after run of salmon bashed themselves against the concrete barrier, trying to find a way upriver.
Dick Goin, longtime local fisherman, recalls that even in the 1940s and 50s, quite a few salmon returned to the lower river. “The operation of the dam was always destructive to the fishery,” Goin says. Dam electricity mainly powered mills in nearby Port Angeles. Operators tended to “overrun it during the week and shut it off on the weekend to let it refill. Killed millions of fish.” He remembers seeing big Chinooks, still gravid with eggs, stranded on suddenly dry riverbed, just flopping.
Upriver, the 210-foot-tall Glines Canyon Dam went online in 1927. The bedload of gravels and sand that eroded from the geologically active Olympic Mountains dropped in the reservoir behind the Glines Canyon Dam. With no new gravel and sand coming downriver, beaches at the river mouth eroded by up to 150 feet. To the east, Ediz Hook, the long curved sand spit protecting the harbor of Port Angeles, eroded.
The Lower Elwha Klallam Tribe dreamed of the Elwha dams being removed some day, but never thought it would happen. However, in the 1980s, FERC relicensing was coming up for the two dams. One night in September 1987, an unknown person painted a giant crack and the words “Elwha Be Free!” on the face of Glines Canyon Dam. Radical group Earth First! claimed responsibility for the prank. Slowly the idea transformed from the tribe's dream and monkeywrenchers' pranks to an idea backed by the Friends of the Earth, Seattle Audubon, Olympic Park Associates, and Sierra Club's Washington State chapter.
Although the dams were privately owned, Olympic National Park surrounded the two dams and included the Elwha backcountry, about 80 percent of the watershed. National Park Service and tribal leadership got deeply involved in the dam relicensing process. By the early 1990s, federal money was appropriated to study the idea.
After years of river studies, negotiations, and meetings with tribes, local residents, and environmental groups, the decision was made to remove both Elwha dams. In 2000, the federal government, which provided most of the $325 million of the ultimate removal cost, bought and began operating the dams. Brian Winter, Elwha River Restoration project manager, says that $75 million of the dam removal budget was spent on a water treatment plant and related work to clear dam-removal turbidity from Port Angeles drinking water: “A reality check of what you have to do if you take a large dam out.” Native trees, shrubs, grasses, and wildflowers were planted or seeded on the exposed slopes and terraces as reservoirs were drawn down, to avoid colonization by nonnative plants after removal.
Finally, after years of planning and preparatory work, dam removal started on September 17, 2011. The dams—much larger than Marmot and more difficult to remove—and powerhouses were deconstructed over three years. Final pieces of the Glines Canyon Dam came out on August 26, 2014. The first wild Chinook salmon returned to the upper river in less than two weeks. A fish biologist spotted the first redd (gravel nest built by spawning salmon) above Glines Canyon on September 29, 2014.
Former Elwha Dam site in January 2014. Photo by John Gussman
Dam removal can certainly be a powerful action to restore rivers. A well-thought-out dam removal is an act of humility. It removes human-made barriers and frees wild rivers and wild salmon to do what they know how to do. Reconnecting streams is highly effective ecologically and can be a cost-effective means of restoration. Big flushes of sediment accumulated behind dams can rebuild downstream river habitat and, in the Elwha's case, renew an estuary.
Dam removal can also be powerful for reconnecting people. Loss of the salmon was a source of sorrow for the Lower Elwha Klallam Tribe, but many other people didn't like losing the dams and the locally generated electricity for nearby Port Angeles. Before removal began, the community went through a long consensus-building process that got most people to support the project.
In cases where upper watersheds are mostly protected wilderness, such as the Elwha and Sandy, dam removal can unlock miles of high-quality habitat. Eighty percent of the Elwha's watershed is within Olympic National Park, never logged, with salmon habitat in excellent condition, and permanently protected as wilderness. The eagerness of wild salmon to return has surprised even salmon biologists. As with the Marmot Dam site, wild Chinook salmon and bull trout swam past the old Glines Canyon Dam site less than two weeks after the last chunks of concrete came out.
Ecologically, dam removal seems like a no-brainer—but that's not always so. Dam benefits, such as power with minimal carbon emissions, flood control, irrigation, and reservoir recreation, have to be balanced against the costs and benefits of removal. Sediment may hold old industrial toxins. In some cases, dams effectively separate wild and hatchery stocks of native fish. Former reservoirs can become hotspots of invasive plants that are inedible for most wildlife. After a dam is in place for decades, river systems and human communities adapt to the dam and reservoir. Relicensing can be a rare opportunity to require changes in dam operations to benefit fish and rivers.
On the east side of the Cascade Range, the Deschutes River runs 250 miles north to the Columbia River, through forests of ponderosa pine, sagebrush rangelands, and canyons of red-brown rimrock. The 204-foot-tall Pelton Dam sits astride the Deschutes River west of Madras, and generates enough electricity to power about 45,000 homes. Besides power, the Pelton Round Butte Hydroelectric Project provides irrigation and supports recreation and tourism, especially in popular Lake Billy Chinook.
PGE also owned the Pelton Round Butte project. In the late 1990s, ahead of FERC relicensing, PGE paid for scientific studies of the river, project, and salmon. Research hydrologist Grant and his team found that the Deschutes was different from other Pacific Northwest rivers. Because its water came mostly from porous volcanic rocks, which acted like a giant rocky sponge, its flow was unusually stable. Unlike the Elwha and Sandy rivers, the Deschutes River had been little changed below the dam. Little sediment had accumulated behind Pelton Dam.
The Confederated Tribes of the Warm Springs Reservation wanted to buy into the project and change its operations to be more salmon-friendly. The tribes purchased a share of the project in 2002, with options to eventually buy up to 50.01 percent ownership. Tribal members worked at the project, and the project powered tribal industries.
In 2005, FERC issued a new license, good for fifty years, to PGE and the tribes. The multiparty agreement signed earlier by twenty-two organizations, state and federal agencies, tribes, and environmental groups became conditions of the new license.
The agreement requires spending about $135 million to make the project more salmon-friendly. The original fish-passage system failed, and a new fish-passage system has been built at considerable expense. That system, plus other projects, should reestablish steelhead and Chinook salmon to the Upper Deschutes, Metolius, and Crooked rivers.
Some people see relicensing as a chance to take out every dam up for renewal—a way to redeem the original sin of damming free-flowing rivers. Others see any dam removal as a sacrilege, an attack on the basic infrastructure of Western civilization. As a scientist, Grant sees relicensing as a chance to honestly evaluate the past, present, and potential future impacts and benefits of dams. “All dams are different,” Grant says, “and our tradeoff space is far more complicated than we give it credit for.”
Each possible dam removal will be unique, as individual as our Pacific Northwest rivers, ecosystems, and communities. As the science of dam removals grows, so-called knowledge bases and decision-support tools will be developed. They will be useful, in the way that any flashlight helps when you're exploring in the dark. But ultimately, dam removal will be as much a social science as a geomorphic science, a cultural upheaval and readjustment as much as a river system digesting and sculpting sediments.
Undammed rivers will never be the same as they were before dams. Terraces of sediments will remain in former reservoirs, and nonnative plant species will have at least a foothold. Because of climate change, some winter precipitation will fall as rain instead of snow, mountain snowpack will melt earlier in the spring, runoff will peak earlier, and late-summer streamflows will be lower. Tree species are already slowly migrating. New ecological mixes of plants will develop. Some of the insect outbreaks plaguing our forests have a distinct climate signal.
The important change we should be considering isn't from “build dams” to “blow up dams,” but from “rivers are simple” to “rivers are complex.” We should find ways to work with natural systems, undo our harm where possible, and let rivers and wild salmon rebuild watersheds.
We need to get it right with our experiments on the Elwha, Sandy, and other rivers before we make decisions about the really big dams of the Columbia Basin hydropower system, especially the four federal dams on the lower Snake River. The completion of these final four dams in the late 1960s and early 1970s seemed to be the critical point at which salmon smolt mortality became so high that runs already in decline crashed to extremely low numbers. Recovery plans for Columbia Basin salmon have repeatedly been ruled inadequate even though hundreds of millions of dollars are spent annually on attempted recovery efforts.
All the costs, benefits, and mitigations on the Columbia and Snake rivers will be many times bigger than they were for the Elwha, Sandy, and Deschutes rivers. The expected gain will be a moderate reduction in extinction risks for salmon. To breach the Snake River dams, find out we were wrong, and rebuild them, would be unthinkable.
No dam lasts forever. Concrete crumbles. Trapped sediment slowly fills reservoirs. Our rivers will take out every dam eventually, the same way they've carved through lava flows and glaciers in the past. We can decide only if we want to free some rivers in our lifetimes.
Our vision of dam removal is essentially cultural. On the day that demolition of the Glines Canyon Dam began, one man said, “It's the end of an era.”
Another man replied, “It's the beginning of an era.”
4 comments have been posted.
What a phenomenal article. I found it thoughtful and informative, an article that delicately considers a complex and sensitive issue. Thank you Ms. Rapp!
Lorena Williams | July 2015 | Durango, CO
Steven Hawley, who has reported extensively on dam removals, raises a good point. I did not have space in my O.Hm. article to discuss all the complexities of dam mitigations on Oregon's Deschutes River, and I may have portrayed the relicensing mitigations in a too-rosy tone. A key question is, did the expensive selective-water-withdrawal tower and other mitigations at Pelton-Round Butte Project work? If not, why not, and what can still be changed? PGE's comments in Feb., 2015, point out that Oregon's Dept. of Environmental Quality has certified that the dam project and new water-withdrawal tower are not the causes of the Deschutes not meeting some water-quality standards. PGE writes that "Management of water temperatures and water quality requires patience as we strive to untangle multiple factors and influences" -- yes, a bureaucratic evasion of the issue, but it also has some truth to it. The new water-withdrawal tower has been working, according to PGE, for salmon smolts to pass downstream. Salmon had to be reintroduced to the upper watershed, and adults only began returning in 2011 - 2012. I don't like the continued use of some adult salmon for hatchery broodstock, but it has been required by fish agencies. Mr. Hawley is correct that we need to pay attention to how restorations and mitigations are actually working, and continue to raise questions.
Valerie Rapp | May 2015 | Blue River
I couldn't help but note that Valerie Rapp's article, "The River Fix" contains two errors, one minor, the other more substantial. As depicted in the film documentary DamNation, Mikal Jubal is the man responsible for painting "Elwha be Free" on the face of Glines Canyon Dam one night in 1987. Ms. Rapp wrongly writes that the act was committed by "an unknown person." More importantly, the depiction of PGE's 2005 relicensing of Round Butte Dam as a paragon of prudent ecological decision-making is way off the mark. Like its ill-conceived predecessor, the new fish passage system installed at Round Butte in 2009 isn't returning fish above the dam in sufficient numbers to restore populations. Worse, the new system has caused a dangerous rise in nutrient levels in water passing through the dam. The result is that PGE is violating quality standards in the lower 100 miles of the Deschutes, jeopardizing existing healthy trout, steelhead and salmon populations, while simultaneously failing to meet restoration goals in the Deschutes above Round Butte Dam. Ms. Rapp cautions we need to "get it right" with our dam removal "experiments" on our rivers before turning our attention to tearing out even bigger dams. But the long record of failed experiments with hatcheries and hauling fish in tanker trucks make the recent success of dam removals all the more compelling. In fostering an an appreciation for the complexities of rivers, as Ms. Rapp writes we should, we may yet find that returning the river to its rich biological compexity is work best left to natural processes that occur when a river flows unimpeded. Steven Hawley Hood River, Oregon
steven hawley | May 2015 | Hood River
I am not so sure removal of dams on the Snake and Columbia rivers will be a good idea. First of all what happens to all of the fish downstream when the silt behind the dams hits them? More than likely they will die. Then what happens to the cost of power? It more that likely will soar sky hight.
Jim Brown | May 2015 | La Grande, Oregon