JSNH&B home • 2015 • vol. 6 no. 1

Time to Give a Dam: O’Shaughnessy Dam is No Longer Needed

by Sarah E. Null

Professor, Watershed Sciences Department, Utah State University


In 1923, the City of San Francisco completed O’Shaughnessy Dam, impounding Hetch Hetchy Reservoir in Yosemite National Park (CA). That water project was a classic early 20th century system using the remote, high-elevation, Tuolumne watershed to deliver pristine water requiring minimal treatment—with hydropower as a side-benefit. New York City, Seattle, Portland, and Los Angeles’ Owens Valley and Mono Lake water supplies are other examples of such systems. From an engineering perspective, O’Shaughnessy Dam remains a functional success, so much so that the dam was largely ignored for decades. It worked as it was designed—and still does, providing reliable, inexpensive, high-quality water to about 2.4 million people the San Francisco Bay area (Figure 1).Figure 1
Figure 1. Hetch Hetchy System and service area. Source: Null and Lund, Journal of American Water Resources Association, 2006

So, why can O’Shaughnessy Dam be removed? Quite simply, it is no longer needed for water supply. Over the past century, San Francisco and California have changed dramatically: new, large reservoirs have been built in the Tuolumne watershed, water treatment technology has improved, and California’s population has increased tenfold, as have the tourists who visit Yosemite National Park each year.

Removing O’Shaughnessy Dam is a popular topic for newspaper and magazine articles and the idea of removing O’Shaughnessy Dam to restore Hetch Hetchy Valley (or not to dam it to begin with) has been discussed for a century. It is an old idea, yet it has gained momentum and publicity in the past decade. In 2004, Governor Schwarzenegger asked California’s Department of Water Resources to review existing studies on O’Shaughnessy Dam and the Hetch Hetchy System, and determined the idea is feasible, but potentially costly. Also in 2004, a newspaper series by Tom Philp for the Sacramento Bee detailed the recent Hetch Hetchy debate, pushed for decisions on Hetch Hetchy to be revisited, and won the Pulitzer Prize for editorial journalism. In 2012, a San Francisco ballot measure gauged support for additional research on removing O’Shaughnessy Dam, but was soundly defeated.

Independent and peer-reviewed research from myself and others has provided the science and rationale for the studies and events listed above. I became interested in water resources in general and O’Shaughnessy Dam in particular before all these events, after working four summers at Mono Lake in Eastern California. Mono Lake and Hetch Hetchy Valley are similar because both are water supplies for large California cities: Los Angeles and San Francisco, respectively. For that reason, I’ll briefly describe Mono Lake conditions and events before returning focus to O’Shaughnessy Dam.

Society’s Changing Water Values

Beginning in 1941, tributaries to Mono Lake were diverted to Los Angeles for water supply and hydropower generation (Figure 2). With little fresh water to the lake, evaporation caused Mono Lake to shrink, making the naturally saline lake even saltier. This threatened the brine shrimp and alkali flies at the base of the food web and the migratory birds higher up. The battle to save Mono Lake was waged in the early 1990s and was supported by most residents of Los Angeles—who were willing to pay marginally higher water bills in exchange for saving a unique and fragile ecosystem. Eventually, a 1994 court case ensured that water would return to Mono Lake, but the decision was a compromise because the lake will be restored to an intermediate level—Los Angeles can use some of the water. Today Mono Lake is rising, the ecosystem at Mono Lake is being improved while maintaining water uses for people.Figure 2
Figure 2. Los Angeles Aqueduct transporting water to Southern California. Source: Sarah Null

When I left Mono Lake and the Eastern Sierra Nevada for graduate school, I was interested in water resources management in the arid American west, how values concerning water change over time, and how water is allocated for different uses. I returned to graduate school because I wanted to learn the technical skills to address such questions. Mono Lake provided a nice example of changing societal values, and also a unique perspective from which to view similar problems of environmental needs competing with traditional water supply needs.

Given my background, Jay Lund, my graduate advisor, suggested I study the feasibility of removing O’Shaughnessy Dam. My task was to quantify water scarcity to the San Francisco Bay Area and to agricultural water users in the Central Valley with and without O’Shaughnessy Dam. At that time, I was most familiar with John Muir’s writings on Hetch Hetchy Valley—which give the feeling of a majestic and awe-inspiring place. I also knew in a general sense that there had been a national debate over damming Hetch Hetchy Valley when O’Shaughnessy Dam was originally proposed. I learned that despite considerable interest, attention, and many, many opinions, nobody had analyzed the feasibility of removing O’Shaughnessy Dam to understand the water storage and delivery implications. Despite this lack of research, my feeling at that time (2001) was that the conventional wisdom for removing O’Shaughnessy Dam said that it could not be done.

When I began my research, I believed that water transfers across the state (say, from neighboring Sierra Nevada rivers, like the Stanislaus or Merced Rivers) would likely be necessary to maintain water deliveries without O’Shaughnessy Dam. However, in the end, my results were blissfully simple. O’Shaughnessy Dam is no longer needed because additional, large reservoirs have already been built in the Tuolumne River watershed.

Further Reading: Restoring Mono Lake
Hart, J. 1996. Storm Over Mono: The Mono Lake Battle and the California Water Future. University of California Press. Los Angeles and Berkeley, CA.

The Historical Debate to Dam Hetch Hetchy Valley

O’Shaughnessy Dam was highly controversial at the time it was proposed in 1906. It was famously fought by John Muir, who questioned whether a private reservoir for San Francisco belonged in a National Park. However, his efforts were ultimately unsuccessful.

San Francisco Mayor James Phelan and city engineer Michael O’Shaughnessy believed Hetch Hetchy Valley could be used to its greatest potential by damming it to ensure a stable water supply for San Francisco. San Francisco voters approved construction of a dam in Hetch Hetchy Valley by an 86% majority vote in 1908. This was shortly after the 1906 earthquake—when city water pipes broke and subsequent fires burned for days—and it became clear that the city had poor water delivery. Despite this, the Taft administration suspended the decision, which needed federal approval because Yosemite was in a National Park. It was not until 1913 under the Wilson administration, that the Raker Act was passed in U.S. Congress, allowing a municipal dam to be built in a National Park. O’Shaughnessy Dam was completed in 1923, raised in 1938, and has been used for water supply and hydropower generation for the past eighty years.

In 1987, Donald P. Hodel, then Secretary of the Interior under Ronald Reagan, renewed interest in Hetch Hetchy Valley by suggesting that restoration of Hetch Hetchy Valley may be possible. The U.S. Bureau of Reclamation produced a report for the National Park Service on possible water replacement scenarios to enable dam removal. Soon after, the Department of Energy issued a report discrediting the proposed replacement scenarios. Neither report was a true feasibility analysis, but rather conjecture of how water may be replaced if O’Shaughnessy Dam were to be removed.

Further Reading: The Historical Hetch Hetchy Debate
The Sierra Club’s timeline of the ongoing battle over Hetch Hetchy.

Hetch Hetchy Valley Revisited

Today, O’Shaughnessy Dam and eight other reservoirs, make up San Francisco’s Hetch Hetchy water system. Additionally, Don Pedro Reservoir, owned and operated by Turlock and Modesto Irrigation Districts, was enlarged from 289,000 acre-feet to 2,030,000 acre-feet in 1971. Reservoir space to store water—and water to fill that space—are not the same, and the Tuolumne Watershed has plenty of the former (Figure 3, Table 1). Reservoirs in the Tuolumne watershed can store over 2.7 million acre-feet of water; however the mean annual flow of the Tuolumne River is less than 2 million acre-feet/year.Figure 3
Figure 3. Ratio of surface water storage capacity to mean annual flow by watershed. Red hues indicate watersheds with more surface storage than mean annual streamflow and blue hues indicate watersheds with less surface water storage than mean annual streamflow. Source: Null et al., Journal of Environmental Management, 2014

Table 1. Mean annual flow, total reservoir storage, and ratio of storage to mean annual flow by watershed.
Data sources: California Data Exchange Center and California Jurisdictional Dams
Watershed Mean Annual Flow (taf) Storage (taf) Ratio of watershed storage to mean annual flow
American 2,729 1,797 0.66
Bear 399 182 0.46
Cache 833 628 0.75
Calaveras 268 320 1.20
Chowchilla/Fresno 309 246 0.80
Cosumnes 369 41 0.11
Cottonwood 188 4 0.02
Elder/Thomes 189 0 0.00
Feather 4,473 5,406 1.21
Kaweah 430 143 0.33
Kern 733 568 0.77
Kings 1,678 1,245 0.74
Merced 994 1,042 1.05
Mokelumne 749 851 1.14
Putah 347 1,643 4.74
Sacramento 5,666 4,754 0.84
San Joaquin 1,808 1,270 0.70
Stanislaus 1,171 2,841 2.43
Stony 471 246 0.52
Trinity 1,278 2,463 1.93
Tule 144 82 0.57
Tuolumne 1,905 2,717 1.43
Upper Clear 318 241 0.76
Yuba 2,364 1,430 0.61

Thus, it takes exceptionally wet years for reservoirs to fill. Most years reservoirs sit partially empty. Water managers expect this trend to be exacerbated with climate change. With the warmer and drier climate anticipated for California, the whole region will be short of water, but multiple, large reservoirs will remain empty for much of most years. Anecdotally, we observed this trend during summer 2013 through winter 2014, when most reservoirs were nearly empty following a prolonged drought in California (Figure 4).Figure 4
Figure 4
Figure 4. Folsom Reservoir (Sacramento, CA) on July 20, 2011 at about 97% capacity (top) and on January 16, 2014 at about 17% capacity (bottom). Source: California Department of Water Resources

When O’Shaughnessy Dam was built, John Muir and others valued the beauty of Hetch Hetchy Valley in Yosemite National Park, but such wilderness resources were relatively abundant compared with California’s 1920 population of 3.4 million. Today, California’s population exceeds 38 million and Yosemite National Park alone receives about 4 million visitors each year. Most visit Yosemite Valley, while a flooded Hetch Hetchy receives less than 2% of the park’s visitors. Thus, recreational space in Yosemite Valley is scarce—as anyone could tell you who has tried to reserve a campsite. Opening another valley for tourism and to preserve wildness would offset some of that pressure.

Additional reservoirs in the Tuolumne watershed and additional tourists in Yosemite Valley are but two changes that have occurred in the 90 years since O’Shaughnessy Dam was built. California has grown and its economy has prospered. The other dams that have been built in the Tuolumne watershed and the Hetch Hetchy water system allow for diverse water supply and treatment options. The trade off of a unique recreational valley for one reservoir of a classical water supply system might now be revisited.

Can It Be Done? Modeling San Francisco’s Hetch Hetchy System

I used a computer model named CALVIN, an economic-engineering optimization model, to analyze the water supply and hydropower implications of removing O’Shaughnessy Dam. This allowed me to evaluate the importance of O’Shaughnessy Dam in the context of the Hetch Hetchy water system and California’s other water supplies and demands. Mathematical modeling provides thorough and systematic analysis of water supply systems with changing water uses and priorities. Another simulation model, TREWSSIM, was used by water scientists and economists at Environmental Defense.

O’Shaughnessy Dam has a storage capacity of 360,360 acre-feet. It is considered a multipurpose reservoir because current uses include water storage, hydropower generation, and to a lesser extent flood reduction (most flood management is provided by downstream New Don Pedro, although about 30,000 acre-feet of winter flood storage is provided by O’Shaughnessy). O’Shaughnessy Dam is only one component of nine reservoirs and connecting pipelines that make up the Hetch Hetchy water system, and provides 25% of Hetch Hetchy water system storage (Figure 1) and 14% of reservoir storage in the Tuolumne watershed. A hypothetical pipeline was added in my model to link downstream Don Pedro Reservoir with the Hetch Hetchy Aqueduct. Currently, the Hetch Hetchy Aqueduct crosses Don Pedro Reservoir, but the two are not connected.

Water Supply and Deliveries

Model results suggest that a pipeline linking the Hetch Hetchy Aqueduct with the much larger Don Pedro Reservoir could allow almost all water currently captured at O’Shaughnessy Dam to be captured and used downstream with Don Pedro Reservoir (Table 2). Removing O’Shaughnessy Dam would eliminate reservoir storage, but would have no effect on urban water deliveries if accompanied by re-operation of the larger downstream reservoir and water market reductions in some Central Valley agricultural deliveries, averaging less than 600 acre-feet/year (average agricultural use exceeds 5,000,000 acre-feet/year).

However, Don Pedro Reservoir is not a part of the Hetch Hetchy System. It is owned and operated by Turlock and Modesto Irrigation Districts (they allow San Francisco storage space in Don Pedro Reservoir under a water banking arrangement). For removal of O’Shaughnessy Dam to be considered, new institutional agreements must be reached between San Francisco Public Utilities District and Turlock and Modesto Irrigation Districts.

Water Supply and Deliveries with Climate Warming

Population growth and the possibility of climate warming—which is likely to reduce the natural storage of winter snowpack—are often used to justify reservoirs or to rationalize constructing new reservoirs. Model results suggest this is unwarranted. When model runs were made using a warm and dry hydrology (worst-case scenario for water supply) and urban populations three times that of California’s current population, O’Shaughnessy Dam did not reduce water scarcity or scarcity costs (Table 2). Under these conditions, the region faced severe water scarcity so streamflow was immediately delivered to urban and agricultural water users, while reservoirs remained nearly empty.

Table 2. Average annual historical and climate change water deliveries and scarcity.
Source: Null and Lund, Journal of American Water Resources Association, 2006
Historical Hydrology & Year 2020 Demands Climate Change Hydrology & Year 2100 Demands
With O’Shaughnessy Without O’Shaughnessy With O’Shaughnessy Without O’Shaughnessy
Urban Regions*
Average Deliveries (taf/yr) 1440 1440 2314 2314
Average Scarcity (taf/yr) 0 0 52 52
Agricultural Regions**
Average Deliveries (taf/yr) 5258 5258 2537 2536
Average Scarcity (taf/yr) 1 1.5 2722 2723

* Urban Regions include: the City and County of San Francisco, San Mateo County, Santa Clara Valley, Alameda County and Alameda Zone 7 Water Districts, and parts of Madera, Merced, San Joaquin and Stanislaus Counties
** Agricultural Regions Include: San Joaquin Valley west of San Joaquin River and San Joaquin Valley east of the San Joaquin River between the Tuolumne and Merced Rivers

Water Treatment

The water impounded by O’Shaughnessy Dam has a rare exemption from municipal filtration requirements. The watershed is protected because the upstream portion is entirely within Yosemite National Park. Recreation is prohibited in the reservoir to preserve water quality. This exemption would be lost if O’Shaughnessy Dam were removed, requiring additional water treatment facilities. If Hetch Hetchy did not have this exceptional grandfathered status as an unfiltered supply, it would require filtration (costing roughly $1-2 billion), like nearly all cities in California and the nation. The filtration avoidance status of O’Shaughnessy Dam makes the dam valuable to the Hetch Hetchy system. Water systems constructed since the 1940s typically filter water rather than protect entire watersheds from recreation.

Hydropower Generation and Revenue

If Hetch Hetchy Valley were to be restored, hydropower generation would be reduced at O’Shaughnessy Dam and surrounding high-elevation reservoirs, costing perhaps $12 million annually (see Paradise Regained by Environmental Defense for more sophisticated hydropower reduction and replacement estimates). While hydropower facilities downstream of O’Shaughnessy Dam contribute less than 1% of California’s statewide power supply, they are carbon-free, locally important sources of energy.


Although not a focus of my research, recreation in Hetch Hetchy Valley would increase if O’Shaughnessy Dam were to be removed. As stated above, recreation is prohibited in Hetch Hetchy Reservoir and in the Tuolumne River above the reservoir to preserve filtration avoidance status (although recreation in Hetch Hetchy Reservoir was listed as a benefit when the dam was originally proposed). If Hetch Hetchy Valley were restored, recreation and tourism in Hetch Hetchy Valley would likely be substantial, providing revenue and benefits to Yosemite National Park and nearby towns.

Flood Protection

Reservoirs typically reserve space for flood water during California’s wet winter to control floods, and protect downstream residents and property. Flood protection also was not explicitly considered in my research, although flood storage space from O’Shaughnessy was applied to Don Pedro—the large downstream dam. Primary flood control benefits are provided downstream of O’Shaughnessy Dam, by Don Pedro Reservoir. However, if O’Shaughnessy Dam were to be removed, Don Pedro Reservoir would be operated differently to account for uncontrolled inflow. Additional research is needed on flood protection without O’Shaughnessy Dam.

Further Reading: Removing Hetch Hetchy Feasibility Research
Environmental Defense. 2004. Paradise Regained: Solutions for restoring Yosemite’s Hetch Hetchy Valley.
Null, SE, JR Lund. 2006. Re-assembling Hetch Hetchy: Water supply without O’Shaughnessy Dam. Journal of the American Water Resources Association, 42(4): 395-408. doi: 10.1111/j.1752-1688.2006.tb03846.x.
Null, SE, JR Lund. 2006. Restoring Hetch Hetchy Valley: The role of modeling in policy. EOS Transactions AGU, 87(42): 449-451.
Stokstad, E. 2006. Restoring Yosemite’s Twin. Science News Focus, 314(5799), 582-584. doi: 10.1126/science.314.5799.582

A Restored Hetch Hetchy Valley

Removing O’Shaughnessy Dam would not benefit fish runs that migrate between freshwater streams and the ocean because the high elevation excludes runs like Chinook salmon and steelhead trout. Also fish passage is already blocked by downstream dams. Removing other dams in California or along Pacific Coast rivers would be more beneficial to migratory salmon and trout fisheries. However, removing O’Shaughnessy Dam would allow Hetch Hetchy Valley to once again be managed for wilderness preservation—and perhaps improve connectivity for wildlife between Hetch Hetchy Valley and the rest of Yosemite National Park.

If O’Shaughnessy Dam were to be removed, restoration efforts would likely be intensive since Hetch Hetchy Valley is in a National Park. Restoration could include removal of the concrete face of the dam, which would be more thorough for restoration, but also entails operating heavy machinery in a restoration site (common with dam removals). Or the reservoir could be drained, but the dam left in place as a historical monument, with restoration focusing on the valley behind it. For either option the lower 118 ft of the dam, the portion that was excavated into bedrock, would most likely be left to make the longitudinal stream profile of the Tuolumne River function normally.

O’Shaughnessy Dam is unique because the upstream watershed is mostly granite bedrock so sedimentation is negligible. Rates of sedimentation in natural Sierra Nevada lakes typically vary based on lake size. The smallest lakes can receive 2 ft/1000 yrs of sediment, whereas larger natural lakes such as Tenaya Lake may receive 6 in/1000 yrs. Hence, the reservoir behind O’Shaughnessy Dam probably receives no more than 6 in/1000 yrs of sediment. Similarly, dams typically increase nutrient retention in reservoirs. This too is most likely low because there is little pollution above the reservoir. The snow-fed water in the Tuolumne River is cold, and there are relatively few aquatic organisms in the river. Were O’Shaughnessy Dam to be removed, no dredging or removal of silt would be necessary. It is assumed the Tuolumne River would return to its natural channel without human assistance. During 1977, a critically dry year, the river was in its original channel in the upper four miles of Hetch Hetchy Valley that were exposed from low reservoir levels.

Herbaceous vegetation could return to Hetch Hetchy Valley within a year or two following dam removal. Woody shrubs and tree saplings could follow over the next decade. Thus, it would not take long for Hetch Hetchy Valley to become a pleasant recreation and wilderness site. Very large trees could take 50-100 years. The bathtub ring left by the reservoir would be noticeable long into the future. The bathtub ring occurs from the absence of lichen, as well as the bleaching of natural water stains from submersion of the granite walls. Lichen could grow within 75-120 years. The staining of the granite walls would not return on a human timescale.

Further Reading: Restoring Hetch Hetchy Valley
Alternatives for Restoration of Hetch Hetchy Valley Following Removal of the Dam and Reservoir. Riegelhuth, Botti, and Keay. Yosemite National Park.

The debate over O’Shaughnessy Dam and Hetch Hetchy Valley

Have we made progress in the debate over removing O’Shaughnessy Dam from Hetch Hetchy Valley? The answer to that question is a resounding yes. Prior to the turn of the 21st century, a thorough analysis on the feasibility of removing O’Shaughnessy Dam was lacking. But in the past 10 years, independent, peer-reviewed research by myself and others have clearly shown that removing O’Shaughnessy Dam is feasible, but is potentially costly. The question of whether California and San Francisco needs this dam for water supply has been answered. They do not. The true debate now is whether the City of San Francisco, state of California, or the federal government will fund removing O’Shaughnessy Dam. And if so, how should it be funded?

It turns out that societal values concerning water are more ephemeral than the water infrastructure we build. O’Shaughnessy Dam was so well engineered that water managers and residents gave it little thought for decades. It simply functioned as it was designed and delivered clean, gravity-fed water to San Francisco and portions of the Bay Area. But now that numerous, large reservoirs have been built in the Tuolumne River watershed and reservoir storage is abundant, the population of California exceeds 38 million, 4 million visitors travel to Yosemite National Park each year, and we have improved water treatment technology since the 1920s, the dam has outlived its utility.

It is time to once again to turn our thoughts to O’Shaughnessy Dam, not to fund maintenance on the dam, but to address the fact that it is redundant. It takes space in Yosemite National Park—space that is probably better used for tourism and land preservation.

Restoring Hetch Hetchy Valley could open a valley about half the size of Yosemite Valley and nearly identical to Yosemite Valley in terms of beauty to wildlife and the public. Removing O’Shaughnessy Dam can be done—if supported by the public, politicians, and the institutions that manage water in the Tuolumne watershed and San Francisco Bay Area.

In November 2012, a San Francisco ballot measure asked residents to fund a study on whether to remove O’Shaughnessy Dam. It collected only 23% of the vote. Yet, it was a crucial step for San Franciscans to start thinking about their water supply differently. A reservoir in Hetch Hetchy Valley is not San Francisco’s birthright. To think of Hetch Hetchy Valley and water management in those terms discounts the wonderful advances we have made improving water treatment technology, setting up water markets to buy and sell water within the state, and balancing priorities between managing water for human and environmental water uses over the past century. It takes time for mindsets to shift concerning water in the arid American west, including realizing when dams still function but are no longer needed.

The City of Los Angeles reprioritized their water supply choices by supporting water allocation to Mono Lake. Residents of the San Francisco Bay Area are now in a similar position. Coordinated use of the Hetch Hetchy System and Don Pedro Reservoir could provide surprisingly robust water deliveries without O’Shaughnessy Dam. More dams do not mean more water and Hetch Hetchy Valley’s best use is no longer a municipal reservoir. Difficult and controversial decisions have been the hallmark of California water management, and will continue to be into the future. This means that we must continually reassess water demands, components of California’s water supply system that provide benefits, those that need improvement, and those that have outlived their utility. Sometimes a fresh perspective can be an advantage to highlight promising changes and bring decision-making up to speed with current realities. The current debate over removing O’Shaughnessy Dam has little to do with water supply, but rather revolves around economic costs, public support, and institutional agreements. It is in those arenas that this decision lays. A restored Hetch Hetchy Valley could be a new birthright for the 21st century San Francisco.

Images Credits

- Courtesy of Sarah Null