Wedged between Arizona and Utah, less than 20 miles up river from the Grand Canyon, a soaring concrete wall nearly the height of two football fields blocks the flow of the Colorado River. There, at Glen Canyon Dam, the river is turned back on itself, drowning more than 200 miles of plasma-red gorges and replacing the Colorado’s free-spirited rapids with an immense lake of flat, still water called Lake Powell, the nation’s second largest reserve.

When Glen Canyon Dam was built — in the middle of the last century — giant dams were championed as a silver bullet promising to elevate the American West above its greatest handicap — a perennial shortage of water. These monolithic wonders of engineering would bring wild rivers to heel, produce cheap, clean power, and stockpile water necessary to grow a thriving economy in the middle of the desert. And because they were often remotely located they were rarely questioned.

We built the Hoover Dam, creating Lake Mead, Glen Canyon Dam and more than 300 other dams and reservoirs at a cost of more than $100 billion. Such was the nation’s enthusiasm for capturing its water that even the lower part of the Grand Canyon seemed, for a time, worth flooding. Two more towering walls of concrete were proposed there, and would have backed up water well into the nation’s most famous national park.

But today, there are signs that the promise of the great dam has run its course.

Climate change is fundamentally altering the environment, making the West hotter and drier. There is less water to store, and few remaining good sites for new dams.

Many of the existing dams, meanwhile, have proven far less efficient — and less effective — than their champions had hoped. They have altered ecosystems and disrupted fisheries. They have left taxpayers saddled with debt.

And, in what is perhaps the most egregious failure for a system intended to conserve water, many of them lose hundreds of billions of gallons of precious water each year to evaporation and, sometimes, to leakage underground. These losses increasingly undercut the longstanding benefits of damming big rivers like the Colorado, and may now be making the West’s water crisis worse.

In no place is this lesson more acute than at Glen Canyon.

And yet even as these consequences come into focus, four states on the Colorado River are developing plans to build new dams and river diversions in an effort to seize a larger share of dwindling water supplies for themselves before that water flows downstream.

The projects, coupled with perhaps the most severe water shortages the region has ever seen, have reignited a debate about whether 20th century solutions can address the challenges of an epochal 21st century drought, with a growing chorus of prominent former officials saying the plans fly in the face of a new climate reality.

“The Colorado River system is changing rapidly,” says Daniel Beard, a former commissioner of the U.S. Bureau of Reclamation, which oversees all of the federal government’s dams in the West. “We have a responsibility to reassess the fundamental precepts of how we have managed the river.”

That reassessment — Beard and others say — demands that even as new projects are debated, it’s time to decommission one of the grandest dams of them all, Glen Canyon.

---

Glen Canyon Dam was erected as a political and environmental compromise, an evolution of the earliest water wars on the Colorado River.

In 1922 seven states — California, Arizona, Nevada, New Mexico, Utah, Colorado and Wyoming — signed a compact agreement dividing the Colorado between them, and, later, with Mexico. The northern states agreed to send an annual quota of water downstream to California, Nevada and Arizona, and the dam building began. But the faster those states grew the more water they used, and by mid-century Colorado, Wyoming and the rest of the upper basin feared it was only a matter of time before the south laid claim to the entire river. The northern states sought their own mega-dam — one which could give them control over the flow of the river and provide a gate through which they could mete out exactly how much of it was sent downstream. Their political jockeying in Congress eventually won the promise that the federal government would build more dams in the north too.

In 1956 the Colorado River Storage Project Act paved the way for the construction of four more large power-generating dams in the upper basin of the Colorado River. In its planning, the federal Bureau of Reclamation had zeroed in on a dam site on a tributary in northwestern Colorado called the Green River. But the reservoir it proposed to create would submerge a tract of treasured, fossil-laden parkland called Dinosaur National Monument. Environmentalists, led by legendary Sierra Club executive director David Brower in one of the nation’s early epic conservation battles, fought passionately to preserve the monument.

All sides agreed instead to proceed at a remote spot in southern Utah called Glen Canyon, in a region far from highways, about 200 miles northeast of Las Vegas. Glen Canyon Dam would help normalize the erratic flows of the Colorado, and flood a no-man’s land of barren sandstone domes and inaccessible dendritic canyons — transforming them into a surreal oasis called Lake Powell.

Damming rivers this way is as old as civilization itself, stretching back some 8,000 years to the foothills of Mesopotamia. From the start, the idea was to stem the risk of devastating floods by creating a catchment for the unpredictable torrents that rattle down from upstream. Once a river was restrained, its domesticated waters could be guided through ditches and canals to irrigate land for agriculture, and to use the force of gravity to power its delivery over great distances. For desert regions that got little rain, but watch seasonal snowmelt sluice by during the spring rush, dams became a way to capture that water and hold it until the time of year it was needed most.

With time, dams were developed especially for their ability to generate power. And since force is not just a function of the amount of water flowing, but of the mass of water held behind the dam, the dams grew broader. In 1942 the Bureau of Reclamation completed the low-slung Grand Coulee dam, a butchy mass of concrete more massive than the Pyramid at Giza, stretching for nearly a mile across Washington’s Columbia River. It is still today the single largest hydropower producer in the country.

Glen Canyon Dam is of a different sort — a tall, elegant, sweeping structure engineered in an arch bowing against the pressure of the water, enabling a relatively thin sheet of concrete to withstand unfathomable forces behind it. Arch dams like these were perfectly suited for the Colorado’s narrow chasms, and Glen Canyon — like the Hoover Dam — created a reservoir so deep that the sheer height of the water behind it promised to generate enormous currents of power. By all measures, its completion was a feat.

But it took 17 years for the reservoir to fill, and just 19 years after that, it began a steady decline. Today its potential has been severely undercut by its own inefficiencies. Thanks to the steady overuse of the Colorado River system — which provides water to one in eight Americans and supports one seventh of the nation’s crops — Lake Powell has been drained to less than half of its capacity as less water has flowed into it than has been routinely taken out. That relative puddle is no longer capable of generating the amount of power the dam’s builders originally planned, and so the power has become more expensive for the government to deliver, with the burden increasingly falling on the nation’s taxpayers. Since the dam’s power sales are relied on to pay for the operations of other smaller dams and reservoirs used for irrigation in the West, as Glen Canyon financially crumbles, so might the system that depends on it.

But it’s not just the reservoir’s overuse that is causing it to drain, it’s the very site and concept chosen for Lake Powell itself: the reservoir loses an extraordinary amount of its precious water. When a dam is built in the desert, its water is spread over a wide area under hot sun and wind, leading to massive evaporation. More than 160 billion gallons of water evaporate off of Lake Powell’s surface every year, enough to lower the reservoir by four inches each month. Another 120 billion gallons are believed to leak out of the bottom of the canyon into fissures in the earth, — a loss that if tallied up over the life of the dam amounts to more than a year’s flow of the entire Colorado River. According to the environmental group Colorado Riverkeeper, if the lost water were sold, it would generate some $350 million each year.

Cumulatively these debits, Beard says, amount to “the largest loss of water on the Colorado River,” — an amount equal to six percent of its total flow and enough to supply some nine million people each year.

Not every dam site shares Glen Canyon’s problems. Each dam serves a unique purpose — whether it’s power generation or water storage — and every region has different needs.

But Glen Canyon is far from the only project to fall out of favor — major projects are being decommissioned or reevaluated across the country. In some places there isn’t enough water to justify the environmental and economic costs of blocking a river. In others the dams have turned out to block the flow of sediment, stop fish migration, or threaten endangered species in ways that weren’t anticipated in the middle of the last century.

The Hoover Dam’s Lake Mead, which Wednesday fell to its lowest level ever, 145 feet below capacity, also loses hundreds of billions of gallons of water to evaporation and is just 37 percent full. The reservoir behind Arizona’s Coolidge Dam, one of the first major projects in Western water development and one of Arizona’s largest reservoirs, is virtually empty.

And dams are coming down. Six Western dams were deconstructed in 2015 alone. Just last month California and Oregon agreed to dismantle four more power-generating dams on the Klamath River, having realized that the facilities were crippling native salmon fisheries, which also have enormous economic value. “This is a good exercise of humankind correcting some of the mistakes that it’s made in the past,” California Gov. Jerry Brown said when announcing the plan. And in early May a federal judge in Oregon ruled that — because of extensive ecological damage — the system of dams on the Snake and Columbia Rivers “cries out for a major overhaul.”

Still, on the Colorado, today’s water managers refute the notion that it’s time for a change.

Glen Canyon Dam may be past its prime, says Michael Connor, the deputy secretary of the Interior and a former Commissioner of Reclamation, but its not past its usefulness. Though he calls the amount of water lost to evaporation and leakage “incredibly significant,” Connor credits Glen Canyon with numbing the pain of the recent drought. “Look at the last 15 years,” he says. “It’s the lowest inflow in history and there’s been no shortages on the Colorado River and that’s because of Glen Canyon Dam and Lake Powell and Hoover Dam and Lake Mead.”

There is also a political tide to be reckoned with; the delicate peace struck between the seven competing states and Mexico — and the fear that they’d never again be able to reach an agreement the likes of which they all signed in 1922. “Getting rid of Lake Powell… it would basically make the compact stand on its head,” said Bronson Mack, a spokesman for the Southern Nevada Water Authority in Las Vegas, pointing to the role that Lake Powell plays in guaranteeing the northern states have enough water to deliver to Nevada and the south each year. “We’ve always said cracking open the compact is going to land seven states in decades of litigation, so there has not been an appetite for it.”

---

Decommissioning Glen Canyon Dam, however, could offer a solution that politicians cannot afford to ignore — a cheap, immediate, and significant new source of water where it is most desperately needed.

The idea is this: Since two of the nation’s largest reservoirs — just more than 300 miles apart — depend on the same dwindling water source but are each less than half full, they should be combined into one. Lake Mead would be deeper, and its evaporative losses would increase. But the surface area of Lake Powell would be dramatically reduced, and the evaporated water from there would be saved. Furthermore, moving the water out of Glen Canyon would move it from a valley that leaks like a sieve, into one that is watertight. The ongoing losses in Mead — according to proponents of the plan — would be more than offset by the immense savings at Lake Powell.

In all, according to Tom Myers, a hydrologist who was commissioned to research the implications of the plan for the Glen Canyon Institute, an environmental group advocating for combining the two reservoirs, about 179 billion gallons of water would be saved each year — more than enough to supply the population of the city of Los Angeles.

The argument has logical weight because both reservoirs have been struggling to remain half full, and may not ever refill. Researchers at the Scripps Institution of Oceanography at the University of California San Diego gave Lake Mead a 50 percent chance of running dry by 2021. The federal Bureau of Reclamation itself forecasts that the amount of water runoff will decrease another 9 percent by 2050 in the Colorado River basin, as temperatures increase. And last year a group of academic researchers declared that the Colorado River basin — already in its 16th year of drought, may be headed toward the worst water crisis in 1,000 years.

Meanwhile the Bureau predicts that demand will continue to increase on the river so much that by 2060 the region will run short by a trillion gallons each year.

“They all show a huge deficit and they all show the reservoirs will likely never fill again,” said Eric Balken, executive director of the Glen Canyon Institute. “So let’s rethink the game plan here.”

According to the proposal the Glen Canyon Dam itself would not be removed. Rather, its gates would be opened, and the water behind it allowed to pass through, restoring the natural flows through into the Grand Canyon just below it, draining the Lake Powell reservoir, and allowing the legendarily scenic landscape of Glen Canyon to be resurrected.

The water would not be lost. It would simply flow down through the Grand Canyon and be recaptured behind the Hoover Dam in Lake Mead.

“To me it is a no brainer,” said David Wegner, who studied Glen Canyon as a scientist with the Department of Interior for more than 20 years, and later advised the U.S. House of Representatives Committee on Natural Resources on management of Western water issues. “You’ve got very few options.”

According to Balken, the process could unfold in stages, and it wouldn’t take much. For a while, the water from the lake would simply drain through the dam, as if the plug were pulled from a giant bathtub. The lake levels — already nearly 100 feet below their peak now — would be allowed to drop another 100 feet, until they reach the intakes for the dam’s generators. One option would be to maintain the lake levels there, allowing minimal power production. Just that incremental shift would allow vast tracts of land now submerged to be restored.

A second stage of drawdown could lower the levels further, to a set of release pipes about 200 feet above the foot of the dam. The dam’s power plant would be shut down, saving tens of millions of dollars in operating costs. It would also dramatically alter water flows through the Grand Canyon, and would have to be carefully coordinated with water levels in Lake Mead, downriver. For the Colorado to be truly restored to its natural riverbed, a third stage, in which bypass tunnels would have to be drilled to allow the water to circumvent the concrete footing of the dam, would have to be pursued.

But proponents of the plan — which they call “fill Mead first” — say the lake may not need to be completely drained. Every vertical foot the waters drop reveals whole stretches of long-drowned desert wilderness. Much of the Glen Canyon valley has already been resurrected, as the lake levels have receded, and several of Powell’s recreational boat launches now hang above the shoreline, dry.

As the waters fall further, broad swaths of river pinned between vertical canyon walls would be transformed into remote wilderness valleys, their floors once again inviting to explore on horseback or on foot. Dozens of archeological sites, their walls covered in petroglyphs, would be revealed. For a while parts of these lands would be covered in a heavy silt, the result of decades of fine mud settling at the bottom of the lake. But vegetation would quickly root in the fertile soils, and heavy storms would flush the mud out the bottom of the canyon, scouring the sandstone clean. Eventually, perhaps after a few decades, even the white watermarks painted across the sides of the valley — telltale signs of the manmade flood — would begin to disappear.

As the silt gets washed downstream, a cloudy stew would course through the Grand Canyon, temporarily disrupting the gentle ecology of that part of the river. But that silt has been sorely missed, and it would soon settle, restoring beaches long ago deprived of sediment. The flow of the river through the Grand Canyon would once again be defined mainly by the amount of precipitation gathered by the mountains upstream. Four native endangered species of fish would presumably thrive in their restored waters.

Restoring the land of Glen Canyon this way has long been the campaign of ardent environmentalists. Brower — who mounted an intense effort to save Glen Canyon almost as soon as he’d agreed to allow it to be drowned — called the reservoir his greatest regret, and the Glen Canyon Dam has been a potent symbol of the desecration of wild places ever since. Now the shortages on the river, and the promise that climate change are certain to make them worse, have breathed new, pragmatic life into their arguments.

Whether the “Fill Mead First” argument pencils out for its benefit to the water supply depends on whether Myers’ assumptions about the amount of water that leaks out of the bottom of Lake Powell are accurate. The U.S. Bureau of Reclamation has long said that water that seeps into the ground eventually returns to the river. Myers studied the fault patterns in the rock bed, and the direction of groundwater flow, and concluded that much of the water seeps away never to come back — an amount that adds up to about 124 billion gallons each year. His research was published in the Journal of The American Water Resources Association in 2013.

Still, the river’s water managers believe his models are flawed. “We don’t agree with the fundamental assumptions that this water is being lost permanently,” said Colby Pellegrino, the Colorado River Programs Manager for the Southern Nevada Water Authority. Pellegrino’s office ran its own numbers to see if combining the reservoirs would save water, and said the math worked out as a wash.

“This is an attempt to find a water supply rationale which supports their recreational focus and narrow view of what the river should look like,” she said.

---

From a viewpoint above Glen Canyon Dam, a placid lake stretches in one direction, and a tangle of electrical wires in the other. The dam’s power station is a nest of infrastructure locked behind a razor wire fence at the top of a vertiginous cliff, and it’s capable of sending 1,320 megawatts of electric power shooting out across the moon-like landscape of southern Utah, toward millions of homes.

Glen Canyon Dam’s staunchest defenders say its power output is clean, affordable, and vital. Moreover, they say those who think climate change has doomed the traditional ways of managing the Colorado’s water underestimate the political complexities of doing away with the power dams generate.

“The questions that are being asked are very legitimate questions,” Connor says. “But you’ve got to have some political support… and that just doesn’t exist.”

The government relies on Glen Canyon Dam to produce power — and not only to produce it, but to sell it, in order to repay the U.S. Treasury for a big chunk of the cost of Glen Canyon’s construction and infrastructure upgrades, and to fund water conservation initiatives across Western states.

From the start, Glen Canyon was conceived as a giant hydroelectric generating plant, one so large its planners boasted that the sale of its power would not only pay for the dam’s construction, but finance smaller dams across the West. Today Glen Canyon is the largest facility run by the Western Area Power Administration, the federal energy agency that sells wholesale, federally generated hydropower onto the Western electrical grid.

The dam’s power reaches 3.2 million customers from California to New Mexico, according to a recent analysis commissioned by the Glen Canyon Institute. And while much of the power is resold at retail rates, its greatest dependents are Native American tribes and the U.S. Department of Defense — which are accustomed to getting their power wholesale, at a government-subsidized price that Glen Canyon Institute’s consultants calculated was a little more than one third market rates.

Were Glen Canyon to stop operating, its defenders say, those who rely on its power would face astronomical price hikes — or perhaps lose their access to electrical entirely.

That may eventually happen, however, whether the river’s managers agree to dismantle Glen Canyon or not.

The lower the lake level drops, the less power the dam can generate, and the less power WAPA has to sell. Researchers in the northern Colorado River basin states have even begun considering when Lake Powell’s lake levels might reach “dead pool,” or the level at which the turbines can no longer produce any power at all.

Already, the lake’s levels have dropped more than 90 feet since 1999. If they drop another 100 feet or so, the dam’s turbines begin sucking air. Glen Canyon has been generating at roughly 43 percent of its capacity. Some experts predict in the future that nest of wires will convey, on average, about 600 megawatts of power.

That has left WAPA — and the Department of Interior — in a bind. The agency has long-term contracts to deliver power, and it can’t simply come up short. When there’s not enough water WAPA has to purchase the power it can’t make itself in order to meet its obligations. WAPA spent $62 million on extra power to fulfill its contracts in its fiscal 2014, and, after managing its water to bring the water levels in Lake Powell back up, $22 million for its 2015 operations, the vast majority of which was to make up for shortfalls at Glen Canyon.

The shortfalls — like falling dominoes — could eventually leave the region’s broader dam and water infrastructure system in the lurch, as smaller dams and reservoirs that rely on money from Glen Canyon’s power sales may have to do without.

For example the Dolores Water Conservancy District, a small utility in southern Colorado which operates the McPhee Dam, recently used $7 million in power-revenues from the Colorado River Storage Project to pay for pump upgrades at its facilities. If Glen Canyon Dam wasn’t generating power, that upgrade might never have happened. “Without the basin power funds the viability of our project becomes a lot more challenging,” says the district’s general manager, Mike Preston. “I don’t know how we would do it frankly.”

Meanwhile nature is weaning the West off its power subsidies all on its own. As WAPA’s costs increase, the economic benefit of the dam has already decreased. If production drops further, the conservation programs paid for by the dam’s slush fund will go unfunded.

Power consultants hired by the Glen Canyon Institute to analyze the impact of shutting down the dam found that western power consumers have already reduced their reliance on Glen Canyon power, and that power constitutes less than one percent of western supply. There are end users that rely on Glen Canyon for all of their power — the Navajo Tribe, and the U.S. Department of Energy’s labs in Albuquerque, New Mexico — and these few groups would expect large cost increases as their subsidies disappear. But the vast majority of residential consumers would see their bills jump nominally, by just eight cents per month, should Glen Canyon shut down.

Still — as the Glen Canyon Institute report points out — if Lake Mead and Lake Powell were combined, that worst-case blow would likely be blunted. With the added mass of water behind the Hoover Dam — another federal hydropower facility — generators there would produce more power than they do now, and that surplus, according to the Glen Canyon Institute, could replace about 17 percent of what Glen Canyon produces today.

---

The amount of water in Lake Powell, of course, depends on the amount of water that flows into it. But even as the region scrambles to save its reservoir — and its river — projects are being planned which would take large additional amounts of water out of the Colorado before it can even fill Powell or fuel the Glen Canyon Dam.

Colorado is planning to build a new reservoir at a place called Windy Gap, and to more than triple the capacity of its Gross Reservoir by raising the height of the dam there. Wyoming is considering an expansion of its Fontenelle Dam on the Green River. Utah just filed an application to build a 140-mile pipeline to divert water out of Lake Powell. The planned new projects would divert another 83 billion gallons of water away from Glen Canyon Dam each year

Downriver, New Mexico is preparing to spend close to one billion dollars to build a new dam and reservoir on the Gila River — a river mostly spoken for. The project promises to capture water only in years when there is so much extra that Arizona right-holders downstream can’t use it all, and to do so despite science which forecasts the Gila will have less and less water in the future. Furthermore, like Lake Powell, the reservoir site in New Mexico also promises to leak, so much so that builders now want to line the entire reservoir valley with rubber before it’s filled, an undertaking that will cost more than construction of the dam itself.

“Every supplier in the basin is trying to build anything they can to get as much water as they can,” said Gary Wockner, the executive director of the environmental organization Save the Colorado, “trying to get the last legally allowed drop that they can before the red flag goes up.”

That every drop of water on the Colorado counts is undisputed. Last month the lower basin states announced they were close to a historic, and difficult, agreement to voluntarily cede a small share of their water, just to keep the reservoirs functioning. An ambitious, multi-million effort to buy even small amounts of water off of Colorado farmers and send it to fill Lake Powell has, after two years, failed to garner more than a few million gallons. It appears that stakeholders are ready to do whatever it takes, including re-envisioning the use of dams.

“I think it’s questionable whether you can build a lot more reservoirs in hot, arid environments,” the Interior department’s Connor told ProPublica.

Still, the proposal to “Fill Mead First” appears to be a step too far.

Ultimately the decision to drain Lake Powell — or perhaps to forgo any of the other new dam and water projects now in the works on the river — comes down to a question of whether the seven states and Mexico that share the Colorado river really need the water badly enough.

When they do, abandoning parochial concerns about how the river is supposed to work, and changing the status quo, however uncomfortable or complicated, will begin to seem worth it.

"There’s just a lot that’s built on this scheme of management and the existence of Glen Canyon Dam,” said Jim Lochhead, the chief executive of Denver’s water utility and a former Colorado state negotiator on Colorado River matters. Lochhead says the change would likely require an act of Congress, plus an agreement between seven state legislatures and a revised treaty with Mexico, and a multi-year federal environmental impact analysis.

“A half a million acre feet sounds like a lot of water,” he said, referring to the amount that would be saved by combining the Powell and Mead reservoirs, “but I don’t think it’s significant enough, frankly, to justify going through all of that.”