ASKED IN 1995 what the Bureau of Reclamation plans to do when sediment threatens to fill Lake Powell, the 186-mile-long reservoir on the Colorado River, former reclamation commissioner Floyd Dominy replied, “We will let people in the future worry about it.” Lake Powell and Glen Canyon Dam are less than 50 years old, yet already we can see that those who will bear their true costs will not be some generation in a distant future, but our children and grandchildren, and even ourselves.
Glen Canyon was the second of two high dams on the Colorado. In 1936 Franklin D. Roosevelt dedicated the earlier of the pair, Hoover Dam, the first of nearly 50,000 large dams built worldwide, 90 percent of them since 1950. To justify this frenzy of construction, proponents and funders routinely exaggerated benefits and longevity while underestimating and ignoring costs. The most obvious cost of Glen Canyon Dam is the chasm that it drowned, western explorer John Wesley Powell’s “ensemble of wonderful features — carved walls, royal arches, glens, alcove gulches, mounds, and monuments.” These the Bureau of Reclamation sacrificed to fuel the juggernaut of development in Arizona, California, and Nevada. But there are other costs.
The dam not only drowned Glen Canyon, it radically changed the Colorado River downstream in Grand Canyon. Before the dam, the river was so muddy that the pioneer explorers joked (as they did about many free-running Western rivers) that it was “too thick to drink and too thin to plow.” The Colorado delivers enough sediment to Lake Powell to fill 1,400 ship cargo containers each day. The dam traps it all, leaving the water below clear as air and starving plant and animal species in Grand Canyon of the sediment they need for habitat and spawning.
Spring snowmelt once swelled the unimpeded river to 100,000 cubic feet per second (cfs) or more, equal to the summertime flow of Niagara Falls. In midwinter the flow dropped to only a few thousand cfs. To meet the demand for power in the Southwest, dam operators smoothed out the seasonal variations while they drastically increased the hourly ones, the released water surging through Grand Canyon like mini-tsunamis, washing away what sediment there is.
We justify dams for the hydropower, flood protection, irrigation water, and recreation that they provide. Yet over time, accumulating silt reduces and then eliminates each benefit. As long as the laws of physics hold, large reservoirs must fill with mud. After Lake Powell fills, the Colorado River will meander across a mud flat and plunge down the face of Glen Canyon Dam in a waterfall that will undercut and eventually collapse the dam. In time, the river will remove the dam debris and lake sediments as though they never existed. Dams imprison rivers, but eventually they annihilate their jailers and escape. Like the truth, a river will out.
Scores of the world’s smaller reservoirs have already filled with mud, or nearly so. Matilija and San Clemente Reservoirs in California, China’s Sanmenxia, and others are full and overflowing. The U.S. Army Corps of Engineers estimated the life of the Cerron Grande Dam in El Salvador at 30 years, less than one-tenth the preconstruction prediction. A mid-1980s study for the World Bank found that world reservoirs contained 1,100 cubic kilometers of sediment, nearly a billion acre-feet and almost one-fifth of worldwide storage capacity.
Lake Powell is so large that it will take longer to fill with sediment than most reservoirs — but how long? Assuming that sediment continues to arrive at the rate calculated from sonar surveys in the 1980s, to fill Lake Powell’s 27 million acre-feet would take 700 years (an acre-foot equals about 326,000 gallons). But that time span is certain to be a drastic overestimate. Many scientists predict that within two decades, rising demand for Colorado River water and falling supply will drop the surface of the reservoir to its lowest level, known as “dead pool.” Because the lowest exit from the dam (the river outlet works) is 237 feet above the original riverbed, at dead pool Lake Powell will still hold 2 million acre-feet of water, one-thirteenth of capacity.
All things being equal, to fill a Lake Powell of 2 million acre-feet with sediment will take not 700 years but about 55 years. And all things won’t be equal.
Lake Powell is destined for dead pool for two reasons: First, because the Southwest already consumes the entire flow of the Colorado River, and demand still rises inexorably. Some of the driest states — Nevada, Arizona, and Utah — are among the fastest growing and none plans to slow down. Take a desert, add water, stir in money — that will continue to be the Southwest’s definition of success until it fails, until whole subdivisions stand empty because they have no water.
Second, while rising population drives demand for water in the Colorado River Basin, climate change shrinks supply. Since 1950, the average temperature in the West has risen over 2°F, nearly twice the global temperature rise during the entire twentieth century. Higher temperatures increase evaporation, which reduces runoff. But calculations show that runoff falls by an even greater percentage than temperature rises. Suppose that evaporation in the Colorado River Basin increases by 2 percent, while precipitation remains the same. Then runoff declines by 14 percent. If precipitation also declines by, say, 1 percent, then runoff drops by 22 percent. How much water is 14 percent of the average flow of the Colorado River? Two million acre-feet, nearly half of the water California takes from the Colorado every year.
A 14 percent decrease in the flow of the river is in the middle of the range that climate models project. Not a single climate study projects an increase in flow. The more pessimistic projections may already be coming true: from 2000 through 2009, inflow to Lake Powell was down by one-third.
Lake Powell’s decline has created a vicious cycle that will reduce its life span even further. You can see the effect as you raft through Cataract Canyon above Lake Powell, where the lowering lake level has exposed thin strata of silt from the reservoir’s formerly submerged delta. Before your eyes, the sandy layers crumble and fall into the river, which washes them down into the lake and deposits them anew. Lake Powell thus gains not only the new sediment that the Colorado River brings to it, but the recycled, older sediment that the river deposited years and decades before. The reservoir also gains recycled sediment from its muddy tributaries: the San Juan, the Dirty Devil, and others. A recent sediment survey found that in five years of declining water levels, Lake Powell received 22 years’ worth of recycled sediment. Thus even the 55-year estimate of the reservoir’s longevity might be too high.
With demand rising and supply falling, Lake Powell could reach dead pool in the 2020s. Impossible? Hardly. Between 1999 and 2005, the reservoir lost two-thirds of its volume and neared the elevation of its generator intakes, 333 feet above the riverbed. Another two or three years at the same rate of decline and Lake Powell would already be at dead pool. Fortunately, a wet 2005 intervened to offer a temporary reprieve.
Another threat from rising sediment is invisible to our eyes but may be the more dangerous. The delta at the reservoir’s north end is one of two sediment deposits on the floor of Lake Powell. A second, smaller wedge has piled up against the dam’s upstream face, the result of turbidity flows racing along the bottom of the reservoir. The sediment wedge at the dam face will rise gradually until it threatens to block the outlet works. Former Reclamation scientist Dave Wegner of the Glen Canyon Institute and former Commissioner Dominy (who died in April) proposed drilling new tunnels below the dam, through the sandstone bedrock, and up through the floor of the reservoir, allowing water and sediment to bypass the outlet works. Would it work? No one knows, for no one has studied the possibility. But since the sediment will never cease to arrive, the tunnels, or any other solution, could never take a holiday.
Moreover, instead of accumulating steadily from small turbidity flows, a large volume of silt could arrive at the dam face all at once. A vast quantity waits suspended not only in Lake Powell’s delta, but in deposits along the river’s upstream tributaries. An earthquake, or an unusually large storm or series of storms, could set this silt moving along the lake floor and up the dam face.
If the surface of the reservoir were below the openings to the generators when silt rose to block the outlet works, no water would be able to leave Glen Canyon Dam and enter Grand Canyon. The tributaries in Grand Canyon add too little water to make up for the loss of the mainstem, so sections of the river would run dry and riverine life would die. These conditions would persist until water rose high enough in Lake Powell to exit either through the disused generator intakes or the spillways near the crest of the dam.
As long as Glen Canyon Dam prevented mainstem water from reaching the Grand Canyon, only a trickle from the tributaries in the canyon would arrive downstream at Lake Mead. To meet downstream water contracts, the Bureau of Reclamation would have to drain Lake Mead to dead pool within a few years. Until the Colorado River flowed again, evaporation would drop Lake Mead even lower and no water would be able to leave Hoover Dam. As long as neither dam could release water, the lower Colorado River would cease to exist.
Though every dam will fill with silt eventually, reclamation agencies design them to hold water, not mud. Wet mud weighs roughly twice as much as an equal volume of water. The area behind Glen Canyon Dam has room for 70 billion tons of sediment. Would the dam hold that much? No one knows. Will the weight trigger an earthquake that precipitates a sediment flow or collapses the dam? No one can rule out the possibility.
The epicenter of the May 2008 earthquake that killed 70,000 people in China’s Sichuan Province was located 3.5 miles from Zipingpu Dam, which itself stands 550 yards from a fault line. Completed in 2006, Zipingpu had not had time to fill with silt, but did hold 315 million tons of water. The earthquake cracked the dam, requiring the Chinese authorities to drain the reservoir. Some seismologists believe that though the weight of the water did not cause the earthquake, it likely advanced its timing and increased its severity. Geologists have long known that reservoirs trigger tremors. In the decade after Hoover Dam went up, some 600 earthquakes struck the area.
If Glen Canyon Dam were to fail when Lake Powell was full of mud, or even half full, one of the great disasters in United States history would unfold rapidly. Sludge would destroy the Colorado River ecosystem in Grand Canyon. The arrival of the silt from Lake Powell might overload Lake Mead and topple Hoover Dam. As the silt continued its way downstream, it could collapse the other Colorado River dams and even reach the Sea of Cortéz. Repair would be so monumental and costly as to be impossible. The Colorado River dams provide water and power to 30 million people; without them, the economy of the Southwest would collapse.
The benefits of large hydropower dams are fated to fade and disappear, in some cases in spectacular fashion. The hundreds of dams that have destroyed rivers across the West represent a kind of national debt that Dominy’s “people in the future” must someday pay. In this century of climate change, that debt is coming due far sooner than anyone imagined.