REASSEMBLING HETCH HETCHY: WATER SUPPLY WITHOUT O ...

ABSTRACT: The Hetch Hetchy System provides San Franciscowith most of its water supply. O’Shaughnessy Dam is one com-ponent of this system, providing approximately 25 percent ofwater storage for the Hetch Hetchy System and none of its con-veyance. Removing O’Shaughnessy Dam has gained interestfor restoring Hetch Hetchy Valley. The water supply feasibilityof removing O’Shaughnessy Dam is analyzed by examiningalternative water storage and delivery operations for San Fran-cisco using an economic engineering optimization model. Thismodel ignores institutional and political constraints and hasperfect hydrologic foresight to explore water supply possibili-ties through reoperation of other existing reservoirs. The eco-nomic benefits of O’Shaughnessy Dam and its alternatives aremeasured in terms of the quantity of water supplied to SanFrancisco and agricultural water users, water treatment costs,and hydropower generation. Results suggest there could be lit-tle water scarcity if O’Shaughnessy Dam were to be removed,although removal would be costly due to additional water treat-ment costs and lost hydropower generation.(KEY TERMS: water supply; dam removal; Hetch Hetchy; opti-mization; economics; filtration avoidance; restoration.)

Null, Sarah E. and Jay R. Lund, 2006. Reassembling HetchHetchy: Water Supply Implications of Removing O’Shaugh-nessy Dam. Journal of the American Water Resources Associa-tion (JAWRA) 42(2):395-408.

INTRODUCTION

O’Shaughnessy Dam, located in the Hetch HetchyValley of Yosemite National Park, was built by theCity of San Francisco in 1923 as a component of theHetch Hetchy water system. The Hetch Hetchy Sys-tem has 10 other reservoirs, numerous water con-veyance pipelines, and water treatment facilities. Thissystem provides water to 2.4 million people in the SanFrancisco Bay Area, including the City and County ofSan Francisco and 29 wholesale water agencies inSan Mateo, Alameda, and Santa Clara Counties(USBR, 1987).

O’Shaughnessy Dam was controversial when pro-posed and built in the early 1900s. Throughout thepast century, the idea of removing O’ShaughnessyDam to restore Hetch Hetchy Valley has never entire-ly gone away, although today California is much dif-ferent. Yosemite National Park is now one of the mostloved and visited parks in the United States. Califor-nia’s Bay Area is a major urban region, with millionsof residents requiring water delivery; the TuolumneRiver now has several times more storage capacitywith the construction of the New Don Pedro Reser-voir; and water treatment technology and treatmentstandards have vastly improved. For Hetch HetchyValley restoration to be considered, the remainder ofthe Hetch Hetchy System or other alternative sourcesmust be able to supply water without O’ShaughnessyDam. The importance of O’Shaughnessy Dam must beevaluated in the context of the greater Hetch HetchySystem and other water supply and demand manage-ment options.

1Paper No. 04039 of the Journal of the American Water Resources Association (JAWRA) (Copyright © 2006). Discussions are open untilOctober 1, 2006.

2Respectively, Doctoral Student, Geography Graduate Group, University of California-Davis, 152 Walker Hall, One Shields Avenue, Davis,California 95616; and Professor, Department of Civil and Environmental Engineering, University of California-Davis, 3109 Engineering Unit3, One Shields Avenue, Davis, California 95616 (E-Mail/Null: [email protected]).

JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION 395 JAWRA

JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATIONAPRIL AMERICAN WATER RESOURCES ASSOCIATION 2006

REASSEMBLING HETCH HETCHY: WATER SUPPLYWITHOUT O’SHAUGHNESSY DAM1

Sarah E. Null and Jay R. Lund2

This study provides quantitative estimates for thewater supply feasibility of removing O’ShaughnessyDam using a spatially refined economic engineeringoptimization model. It identifies cost minimizingalternatives for San Francisco’s water supply andhighlights how removing O’Shaughnessy Dam couldchange current operations, water supply, deliveries,hydropower generation, water treatment, and eco-nomic water supply costs (Null, 2003). Modeling canprovide quantitative support for discussion or actions.This analysis indicates whether water scarcity wouldincrease substantially without O’Shaughnessy Dam.Such information also might be useful as part of amore comprehensive benefit cost analysis of HetchHetchy Valley restoration, not undertaken here.

Removing O’Shaughnessy Dam could raise manyinstitutional, political, and economic questions. How-ever, this study ignores most institutional and politi-cal implications to focus on optimization of watersupply and economic factors. It is helpful to ignorepolitical and institutional constraints to understandthe physical limitations of a water supply system. Itthen becomes clear whether human constraints orphysical constraints are limiting factors. Recently,much has been written and discussed about removingO’Shaughnessy Dam (Jensen, 2004; Leal, 2004; Philp,2004a-2004l; Rosekrans et al., 2004). Tom Philp of theSacramento Bee recently won the 2005 Pulitzer Prizefor editorial writing based significantly on this study(Stanton, 2005). In addition, Governor Schwarzeneg-ger’s staff authorized California’s Department ofWater Resources (CDWR) to review this and otherstudies relating to O’Shaughnessy Dam and HetchHetchy Valley (CDWR, 2005).

This paper begins with a brief overview of HetchHetchy Valley, the Hetch Hetchy System, and poten-tial O’Shaughnessy Dam removal decisions. CALVIN(CALifornia Value Integrated Network), the computermodel used to evaluate alternatives to O’ShaughnessyDam, is explained. A summary of model parameters,infrastructure modifications, benefits, and model limi-tations follow. Model runs are described, and resultsare discussed in terms of operational and economicperformance. The paper concludes with a short dis-cussion of the implications and water supply costs ofremoving O’Shaughnessy Dam and some larger insti-tutional and economic factors regarding removal ofO’Shaughnessy Dam.

Background

Prior to construction of O’Shaughnessy Dam, HetchHetchy Valley was similar to neighboring YosemiteValley. Both valleys were formed from jointed granitebedrock, and they initially were cut by the Tuolumne

and Merced Rivers, respectively. Glaciers thenscoured the valleys, widening and polishing the sur-rounding granite. Both valleys once had natural lakesthat filled with sediment, forming flat meadows(Huber, 1987).

In 1906, following the San Francisco earthquake,the shortcomings of San Francisco’s water supplybecame clear. City water planners targeted HetchHetchy as a dam site for San Francisco’s water supply(Hundley, 1992). San Francisco’s voters approved theconstruction of a dam in Hetch Hetchy Valley by an86 percent majority vote in 1908 (Sierra Club, 2002).In 1913, the Raker Act was passed in Congress,enabling a large reservoir to be built in a nationalpark (Hundley, 1992). O’Shaughnessy Dam was com-pleted in 1923, was raised in 1938, and has been usedfor the past 80 years.

O’Shaughnessy Dam has a storage capacity of360,360 acre feet (af) (444.5 millions of cubic meters,mcm). Its current uses include water storage,hydropower generation, and some flood reduction(USBR, 1987). The reservoir behind O’ShaughnessyDam is not used for recreation. In terms of total waterstorage in the Hetch Hetchy System, O’ShaughnessyDam is not exceptionally large. Its 360 thousand acrefeet (taf) (444 mcm) of surface water storage areapproximately 25 percent of the surface storage in theHetch Hetchy System and 14 percent of storage onthe Tuolumne River. For this study, removal ofO’Shaughnessy Dam applies only to the dam and itsreservoir. No pipelines, diversion capacity, or con-veyance facilities would be removed from the HetchHetchy System.

As illustrated in Figure 1, the Hetch Hetchy Sys-tem consists of several reservoirs, power plants, andconveyance facilities. Downstream of O’ShaughnessyDam, the Kirkwood Power Plant and Moccasin PowerPlant generate hydropower. Nearby, Cherry Reservoirand Eleanor Reservoir are currently operated primar-ily for hydropower at Holm Powerhouse. Adequatewater supply storage is usually provided byO’Shaughnessy Dam. Another much larger dam, NewDon Pedro Reservoir, is downstream of O’Shaugh-nessy Dam, the Hetch Hetchy Aqueduct intake, andCherry and Eleanor Reservoirs (City and County ofSan Francisco, 1999). New Don Pedro Reservoir isowned and operated by Modesto and Turlock Irriga-tion Districts, with a water banking arrangementallowing the City and County of San Francisco theuse of 570 taf (703 mcm)of water storage.

Together, these reservoirs and water bankingarrangements, with numerous Bay Area reservoirsand the connecting pipelines, make up the HetchHetchy System, operated by the San Francisco PublicUtilities Commission (SFPUC). Total surface storagein the Hetch Hetchy System is approximately 1,500

JAWRA 396 JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION

NULL AND LUND

taf (1,850 mcm) (Table 1). In addition to the HetchHetchy System total, 1,460 taf (1,801 mcm) remainsof 2,030 taf (2,504 mcm) New Don Pedro Reservoirstorage capacity. The Hetch Hetchy System supplieswater to 77 percent of urban uses in the City andCounty of San Francisco, as well as parts of SanMateo, Santa Clara, and Alameda Counties, supply-ing more than 2 million water users (USDOE, 1989).Three powerhouses on the upper Tuolumne Rivertogether provide approximately 2 billion KW hrs/yr ofhydropower (USBR, 1987). This is a clean source ofenergy and an important source of revenue for theSFPUC.

Although water storage is a priority for SFPUCand the Hetch Hetchy System, it is not the 360 taf(444 mcm) of water storage that makes O’Shaugh-nessy Dam valuable; rather water from O’Shaugh-nessy Dam has filtration avoidance status (SFPUC,2005). Typically, filtration of urban surface water sup-plies is required by federal law. Filtration avoidancemeans O’Shaughnessy Dam impounds high qualitywater that meets water quality standards under

the federal Surface Water Treatment Rule. Only minimal water treatment is currently required, suchas addition of lime for corrosion control and chlorineor chloramine as a disinfectant (SFPUC, 2005). Thewatershed above O’Shaughnessy Dam is pristine; itlies within Yosemite National Park. O’ShaughnessyDam is the only reservoir in the Hetch Hetchy Systemthat has filtration avoidance (and one of only about ahalf-dozen sources nationally). Even Cherry andEleanor Reservoirs, less than 10 miles (16 km) fromO’Shaughnessy Dam, do not qualify for filtrationavoidance.

Recently the SFPUC approved seismic retrofits forthe Hetch Hetchy System as part of its CapitalImprovement Program (SFPUC, 2002). Increasedlocal San Francisco water storage would be valuablein case of an earthquake, terrorist act, or other emer-gency repair. For this study, all water storage is val-ued equally, regardless of location. However, becauseO’Shaughnessy Dam is at the upstream end of thesystem, it is of little value in the event of emergencyinterruptions in the conveyance system.


REASSEMBLING HETCH HETCHY: WATER SUPPLY WITHOUT O’SHAUGHNESSY DAM

Figure 1. Map of San Joaquin and San Francisco Bay Regional CALVIN Model.

Potential for Removing O’Shaughnessy Dam

There are valid arguments for both keeping andremoving O’Shaughnessy Dam. Many arguments onboth sides are economic. Hydropower is generatedwhen water is released from O’Shaughnessy Dam. Inaddition, loss of filtration avoidance determinationwould incur considerable costs to the Hetch HetchySystem and water users. Finally, some environmen-talists believe O’Shaughnessy Dam is a poor choicefor removal because its removal benefits no threat-ened or endangered species and would make onlyminor improvements for ecological connectivity on theTuolumne River system. The land under the reservoircould be restored, but this is a small land area to jus-tify removal on ecosystem restoration grounds.

Arguments for removing O’Shaughnessy Dam pri-marily center around ethical concerns and increasingopen space in Yosemite National Park for tourism andrecreation. It has been argued, most famously by JohnMuir, that a reservoir for San Francisco residentssimply does not belong in Yosemite National Park,land that in theory belongs to all Americans (Muir,1912). Furthermore, Yosemite National Park is one ofthe most heavily visited national parks in the nation.Within the park, Yosemite Valley is grossly congested.

Restoring Hetch Hetchy Valley could open a valleythat is half the size of Yosemite Valley and nearlyidentical in terms of beauty. Revenue and increasedlocal spending from tourism could offset some or all ofthe losses from removing O’Shaughnessy Dam. Thisproblem can be thought of as weighing two scarceresources, water and space in Yosemite NationalPark.

In recent years, construction of new dams hasdeclined due to economic factors, environmental con-siderations, and the best locations being already inuse. This has led to the idea of increasing water effi-ciency without new infrastructure (Poff and Hart,2002). Options for increasing the efficiency of watersupplies are wide reaching, including coordinated useof existing water infrastructure, conjunctive usebetween surface water and ground water, water con-servation, water transfers, and generally more sophis-ticated operations (CDWR, 1998). With this shift ofideology, the popularity of dam removal has risen dra-matically. At least 467 dams were removed in theUnited States in the latter part of the 20th Century(Poff and Hart, 2002).

METHODS

Although simulation is the most common modelingapproach, optimization models offer another methodfor exploring solutions to water resource problems.Optimization methods can suggest promising solu-tions with the “best” performance from among manyalternatives. Optimization models require explicitobjectives to be maximized or minimized by modifyingdecisions within system constraints. Until recently,optimization models were too computationally bur-densome to be practical for large systems or problems.Now, more powerful computers and software makeoptimization of large systems feasible (Labadie, 2004).

CALVIN is a network flow based economic engi-neering optimization model of California’s inter-tiedwater management system. It was developed at theUniversity of California-Davis, based on the U.S.Army Corps of Engineers’ HEC-PRM reservoir opti-mization software (Jenkins et al., 2001; Draper et al.,2003).

The objective function for CALVIN is to minimizetotal economic costs, mathematically represented as

where Z is total cost (U.S. dollars), cij is the costcoefficient on arc ij (U.S. dollars), and Xij is flow from


NULL AND LUND

TABLE 1. Storage in the Hetch Hetchy System.

Reservoir Capacity (taf)

Hetch Hetchy System Storage

O'Shaughnessy* 0,360

Eleanor 0,028

Cherry 0,268

New Don Pedro 0,570**

San Antonio 0,050

Calaveras 0,097

Lower Crystal Springs 0,058

Pilarcitos 0,003

San Andreas 0,019

Total Hetch Hetchy System Storage 1,454

Other Tuolumne River Storage

New Don Pedro (MID and TID) 1,460

Total Basin Storage

All Reservoirs 2,914

**Filtration Avoidance Permit.**Banking arrangement with City and County of San Francisco.Notes: Total Storage in New Don Pedro Reservoir = 2,030 taf; MID

= Modesto Irrigation District; TID = Turlock Irrigation Dis-trict

Minimize Z c Xij ijji

= ∑∑ (1)

node i to node j (taf), in space and time. Convex non-linear costs can be represented as piecewise linear.CALVIN also has constraints representing physical oroperational bounds. These include constraints forconservation of mass, upper bounds, and lowerbounds, written mathematically as

for all nodes j

for all arcs

for all arcs

where Xij is the flow from node i to node j, aij is gainsor losses on flows in arc ij (taf), bj is external inflowsto node j (taf), uij is upper bound on arc ij (taf), andlij is lower bound on arc ij (taf) (Jenkins et al., 2001).

To represent hydrologic variability, CALVIN uses72 years of monthly unimpaired historical data (1922through 1993), during which the three worst droughtson record occurred, those of 1929 through 1934, 1976through 1977, and 1987 through 1992. California’sentire interconnected water system has been modeledwith CALVIN (Jenkins et al., 2001; Draper et al.,2003). Additional CALVIN models of California havebeen used for various regional and climate changewater management studies (Lund et al., 2003).

Physical parameters in the model include infras-tructure capacities, environmental requirements, andhydrology. Surface reservoirs and ground waterbasins have upper and lower bounds. Maximum sur-face reservoir capacity is limited by the seasonallyvarying flood storage level, and minimum capacity isset to dead storage. For ground water basins, themaximum capacity is the amount of water that can bestored in the aquifer, and the lower bound is the low-est historical level of that ground water. Pumping andconveyance capacities also are limited. Environmentalrequirements include minimum instream flows andflows to refuges. Due to the controversy inherent ineconomic valuation of environmental uses of water,environmental water requirements are modeled asconstraints.

Economic parameters include economic penalty/demand functions and operating costs. The objectivefunction of CALVIN (which drives the model) is tominimize the sum of water scarcity costs to urban andagricultural water users, lost hydropower benefits(compared with some ideal condition), pumping costs,water treatment operating costs, and water qualitycosts. Water scarcity costs are economic penaltiesimposed if agricultural and urban demands are notfully met. The economic value of water for agricultureis derived from the Statewide Agricultural Production

Model (SWAP) (Howitt et al., 2001). SWAP is a sepa-rate model of agricultural water and land use deci-sions that maximizes farm profits given cropproduction functions and limits on water, land, tech-nology, and capital inputs. SWAP is similar to theCentral Valley Production Model (CVPM) (USBR,1997), except that values of water delivery can varybetween months. For each agricultural region, SWAPis used to generate economic loss functions thatdecrease as water delivery increases. The economicloss represents the reduction in agricultural profitsfrom limited water deliveries (higher marginal valueof water), compared to ideal profits if water were not alimiting factor (marginal value equals zero).

Urban economic cost functions are based ondemand curves for urban water use. These cost func-tions vary by month and are assumed to have con-stant seasonal elasticity (Jenkins et al., 2001).CALVIN uses projected demands for the year 2020 forboth agricultural and urban demand areas.

Operating costs correspond to variable costs, pri-marily for ground water pumping, surface pumping,and water treatment. Hydropower penalty curvesoften are nonlinear and thus are difficult to model.For this study, separate hydropower storage andrelease penalties were chosen to represent variablehead facilities. These independent piecewise linearpenalty functions were fit to the full multivariate non-linear surface using a least squares approach (Ritze-ma, 2002).

CALVIN produces a monthly time series of waterdeliveries for each demand location that can be trans-lated into scarcity costs based on economic value func-tions. For this paper, scarcity is defined as thedifference between the quantity of water deliveredand the maximum quantity demanded. Maximumdemand delivery is derived from economic value func-tions at the point where marginal benefit of addition-al water is zero. Scarcity cost is the economic value ofmaximum water deliveries minus the economic valueof water actually delivered. The marginal willingnessto pay, or shadow value, for additional water is alsooutput for each demand node and time step. This alsois the slope of the economic value function at thedelivered quantity of water.

Model Area and Assumptions

To examine the effects of removing O’ShaughnessyDam, this study focuses on the San Joaquin and SanFrancisco Bay Regional CALVIN model (Figure 1).The model area includes 13 surface reservoirs, exclud-ing O’Shaughnessy Dam, and five ground waterbasins. Major conveyance facilities include: the HetchHetchy Aqueduct, the California Aqueduct, the Delta



X a X b

X u

X l

ij ij ij jii

ij ij

ij ij

= +

≤

≥

∑∑ (2)

(3)

(4)

Mendota Canal, the South Bay Aqueduct, and thePacheco Tunnel. Seven hydropower plants are includ-ed. Minimum instream flows are imposed on theTuolumne River below New Don Pedro Reservoir, onthe San Joaquin River below the confluence with theStanislaus River at Vernalis, and on the StanislausRiver below Goodwin Reservoir (Ritzema and Jenk-ins, 2001).

Six urban demand regions and four agriculturaldemand areas are included in the model area. Project-ed year 2020 demand data was obtained fromCDWR’s (1998) bulletin on per capita urban water useby county and detailed analysis unit (DAU). Demandfor several smaller cities are represented as fixed2020 demands (Ritzema and Jenkins, 2001).

The San Francisco Public Utilities Commissiondemand area combines the City and County of SanFrancisco with most of San Mateo County. The SantaClara Valley (SCV) demand area includes Santa ClaraValley Water District, Alameda County Water Dis-trict, and Alameda County Zone 7. The SFPUC andSCV water urban demand areas and all agriculturaldemand areas are represented using economic valuefunctions (Ritzema and Jenkins, 2001).

Although the regional model is large enough toallow for coordinated use among many storage andconveyance facilities, the focus of this study is theHetch Hetchy System (Figure 2). In CALVIN, the localSan Francisco area reservoirs (Calaveras, LowerCrystal Springs, San Andreas, and San Antonio) havebeen represented as a single, aggregated service areareservoir. Tiny Pilarcitos Reservoir (3 taf or 3.7 mcm)was omitted. Cherry and Eleanor Reservoirs are rep-resented as a single reservoir because of the inabilityto disaggregate inflows into these reservoirs and theexistence of an interconnecting tunnel. Nonstoragereservoirs in the Hetch Hetchy System, such as EarlyIntake and Moccasin Reservoir, are represented asjunction nodes instead of reservoirs. Hydropower isincluded in the Hetch Hetchy portion of the model forKirkwood, Holm, Moccasin, and New Don Pedropower plants. Minimum instream flows of 50 to 125cfs (1.42 to 3.54 m3/s) are imposed on the TuolumneRiver downstream of New Don Pedro Reservoir(USBR, 1987).

Although raising the dam height at CalaverasReservoir has been discussed to increase storage inthe Hetch Hetchy System, Calaveras Reservoir isgiven a maximum capacity of 91 taf (112 mcm). Like-wise, storage at this site has not been lowered to thecurrent restriction of 28 percent of total maximumstorage (SFPUC, 2006). While additional service areawater storage is a priority for SFPUC and the HetchHetchy System in the event of a seismic event or ter-rorist act, this study assumes no new storage.

In CALVIN, storage space in New Don Pedro Reser-voir is not divided between different owners andinterests. Rather, the maximum capacity of New DonPedro Reservoir was set to 2,030 taf (2,504 mcm), andwater is allocated to different urban and agriculturaldemands as needed. Because CALVIN is economicallydriven, water scarcity is allocated to demands withlower economic values for water, usually agriculturalareas. Here, results should be interpreted to indicatethe extent of water scarcity. However, in the realworld, water scarcity often occurs to demand areasbased on water rights and contracts. In the TuolumneRiver area, agricultural users with senior waterrights would be unlikely to face scarcity unless theymarketed water to other users.

For model runs in which O’Shaughnessy Dam hasbeen removed, an inter-tie between New Don PedroReservoir and the Hetch Hetchy Aqueduct has beenadded. Physically, the Hetch Hetchy Aqueduct crossesNew Don Pedro Reservoir. As stated above, the HetchHetchy System owns rights to some storage space inNew Don Pedro Reservoir; however, there currently isno way to route this water to Bay Area users exceptby releasing it through the Tuolumne River to theSan Joaquin River, pumping from the Delta, thenrouting it to the Bay Area via the California Aqueduct(entailing significant quality degradation). This hypo-thetical New Don Pedro-Hetch Hetchy Aqueductinter-tie increases flexibility in the conveyance systemand ensures higher quality water to Bay Area cus-tomers than water pumped from the Delta. Inter-tiecapacity is limited by the downstream Hetch HetchyAqueduct capacity of 465 cfs (13.17 m3/s).

Water Treatment

Water treatment costs typically have high fixedcosts and economies of scale. In CALVIN, treatmentcosts can only be modeled with unit costs on treat-ment links. Operating treatment costs from the LosAngeles Aqueduct System (another high quality sys-tem) were applied to the Hetch Hetchy System to esti-mate variable treatment costs from loss of thefiltration avoidance permit. Additional fixed costs forconstructing water treatment plants require a calcu-lation outside of the model.

Detailed economic analysis of increased watertreatment construction costs are beyond the scope ofboth CALVIN and this study. Removal of O’Shaugh-nessy Dam would prompt loss of filtration avoidancestatus, incurring considerable construction costs fornew treatment facilities. This study could not includea detailed analysis of construction costs of new facili-ties. Although expansion of water treatment facilities


NULL AND LUND

is a long term goal for the SFPUC, new treatmentfacilities are costly, and even deferral of such a largeexpense has considerable financial benefits.Rosekrans et al. (2004) estimates capital costs forexpanding filtration facilities to treat all water sup-plies range from US$134.4 million to US$228 millionabove existing treatment cost plans. Avoiding such acapital expense makes keeping O’Shaughnessy Dam apriority for the SFPUC. More detailed economic stud-ies of water treatment costs might be useful as part ofa more comprehensive benefit-cost analysis (Null,2003).

Year 2100 Demands and Dry Climate Warming

Year 2100 model runs uses a warm and dry hydrol-ogy representing effects of possible climate warmingand roughly three times the current urban popula-tion. For this study, a warm and dry climate scenario(PCM) is used to develop a comprehensive hydrologyfor year 2100 model runs as a worst case scenario interms of water supply (Zhu et al., 2005). Populationestimates for the year 2100 are from Landis and Reil-ly (2003). For a complete discussion see Lund et al.(2003, particularly Appendix A). These extreme modelruns are used to examine the effects of removing



Figure 2. Hetch Hetchy System Model Schematic.

O’Shaughnessy Dam when water demand is much,much greater and climate 26 percent drier in the verydistant future.

Additional changes for the “2100” model runsinclude: coastal demand regions have unlimitedaccess to seawater desalination at a unit cost ofUS$1,000/af (US$811 per thousand cubic meters);urban wastewater recycling is available for up to 50percent of return flows, also at a cost of US$1,000/af(US$811/tcm); some environmental demandsincrease, although no changes are made to environ-mental demands for links for the Tuolumne River;and operation and maintenance (O&M) water treat-ment costs increase to represent the loss of filtrationavoidance by the year 2100. (Variable treatment costswere increased to the same level as 2020 model runswith higher treatment costs.)

Model Runs

Six model runs are compared for this study, oneconstrained to current operating policies with 2020demands, three unconstrained runs from the year2020 modeling set, and two model runs with year2100 demands and a warm, dry hydrology represent-ing possible effects of climate warming (Table 2). Twounconstrained year 2020 runs include O’ShaughnessyDam. One has increased water treatment costs to rep-resent loss of filtration avoidance; the other has nochange to water treatment costs to represent filtra-tion avoidance being maintained. All runs withoutO’Shaughnessy Dam have an inter-tie from New DonPedro Reservoir to Hetch Hetchy Aqueduct and high-er O&M water treatment costs, reflecting an end offiltration avoidance. Results from a base case model-ing set represent 2020 conditions with current operat-ing and allocation policies (Jenkins et al., 2001).Some of these results are included for comparison andare referred to as base case results. Both model runswith year 2100 water demands have increased watertreatment costs. In one run O’Shaughnessy Dam isremoved and an inter-tie linking New Don PedroReservoir with the Hetch Hetchy Aqueduct is added.

Limitations

Consequences of O’Shaughnessy Dam removalcould be substantial. The complexities of the systemand management options require a quantitativeapproach, for which computer modeling is well suited.However, several limitations apply to the approachtaken here. As with all modeling studies, manage-ment and river systems are simplified. Economic benefits from recreation are currently neglected.Recreation and tourism in Hetch Hetchy Valley wouldlikely be substantial, providing revenue and benefitsto Yosemite National Park and nearby towns, but thisis not a comprehensive benefit cost analysis. Anotherimportant limitation with CALVIN is perfect hydro-logic foresight. This allows the model to prepare forand aggressively operate during droughts, reducingwater scarcity and associated costs. This limitationtends to be less important when large amounts ofstorage (including ground water) are available (Drap-er, 2001). Urban and agricultural economic demandsare assumed to be fixed, and ground water basins areextremely simplified. Jenkins et al. (2001) providesmore on the limitations of CALVIN.

All model runs, except the base case, are uncon-strained by current institutional and legal allocationpolicies. This reduces water scarcity since water oper-ations and allocations have perfect institutional flexi-bility. However, removing current policy constraintsdo show what could be possible with existing infras-tructure. In these model runs, operations and waterallocation are only economically driven. Absence ofinstitutional implications and public or political sup-port for the idea is perhaps the greatest limitation.These are not part of this model but are neverthelessdriving factors. Additional examination of these fac-tors would be useful.

RESULTS

Overall, year 2020 model runs show little waterscarcity when O’Shaughnessy Dam is removed and an


NULL AND LUND

TABLE 2. Model Runs by Filtration Condition, Dam Status, and Year.

Keep Filtration Avoidance Lose Filtration Avoidance

O'Shaughnessy Dam Retained 2020 20202020 Base Case 2100

Remove O'Shaughnessy Dam With Scenario modeled, produced no new results 2020New Don Pedro Inter-tie 2100

inter-tie added between New Don Pedro Reservoirand the Hetch Hetchy Aqueduct. Although storage atthe O’Shaughnessy Dam site is eliminated, flow in theTuolumne River does not change. Capture of signifi-cant quantities of water into the upper Hetch HetchyAqueduct remains possible. Adding an inter-tiebetween New Don Pedro Reservoir and the HetchHetchy Aqueduct allows capture of Tuolumne Riverwater at New Don Pedro Reservoir. Thus, the lowerHetch Hetchy Aqueduct can remain full at all timesregardless of the existence of O’Shaughnessy Dam.

Although water deliveries are not greatly affectedby removing O’Shaughnessy Dam, removal could becostly. Hydropower generation suffers without waterstorage at O’Shaughnessy Dam. Also, removing thisreservoir ends SFPUC’s filtration avoidance and cre-ates a need for additional water treatment facilities.Filtration avoidance is important and valuableenough to drive decisions regarding potential removalof O’Shaughnessy Dam.

Water Storage

Total water storage in the Hetch Hetchy Systemfalls in model runs without O’Shaughnessy Dam.Without O ’Shaughnessy Dam, storage drops byapproximately 350 taf (432 mcm), the capacity ofO’Shaughnessy Dam, in all water years. To assesswhether the storage space lost from removingO’Shaughnessy Dam is critical to meet water deliver-ies, storage in the other Hetch Hetchy System reser-voirs is evaluated. Water storage remains about thesame in Cherry/Eleanor Reservoir, New Don PedroReservoir, and local San Francisco reservoirs when

O’Shaughnessy Dam is removed from the model.Model runs indicate that much storage is never usedin local San Francisco reservoirs. Thus, considerablestorage remains in the overall Hetch Hetchy Systemwithout O’Shaughnessy Dam.

Water Deliveries and Scarcity

With and without O’Shaughnessy Dam, full deliv-eries are made to urban demand areas for all opti-mized 2020 model runs (Table 3). Although it ispossible to deliver water to San Francisco and SantaClara Valley urban areas via the Hetch Hetchy Aque-duct, the Pacheco Tunnel, or the South Bay Aqueduct,deliveries via all these pipelines are unaffected byremoving O’Shaughnessy Dam. This indicates thatremoving O’Shaughnessy Dam would change opera-tion of the Hetch Hetchy System but need not affectsurrounding water resources. When model runs areconstrained to current operational constraints (basecase), a small amount of scarcity occurs to urbanwater users (Jenkins et al., 2001; Ritzema and Jenk-ins, 2001).

In all model runs, full deliveries are made for envi-ronmental uses. This includes minimum instreamflows on the lower Tuolumne River and flows towildlife refuges such as the San Joaquin and MendotaRefuges.

A slight decrease in deliveries to agriculturaldemand areas occurs in model runs withoutO’Shaughnessy Dam. Total annual average deliveriesto agricultural areas is 5,257,983 af/yr (6,485.63mcm/yr) with O’Shaughnessy Dam but average 575af/yr (0.71 mcm/yr) less without O’Shaughnessy Dam.



TABLE 3. Average 2020 Deliveries, Scarcity, and Scarcity Cost.

Base Case With With O’Shaughnessy Without O’ShaughnessyO’Shaughnessy* Dam Dam**

Urban Regions

Annual Average Deliveries (taf/yr) 1,424 1,440 1,440

Annual Average Scarcity (taf/yr) 16 0 0

Annual Average Scarcity Cost (US$K/yr) 15,290 0 0

Agricultural Regions

Annual Average Deliveries (taf/yr) 5,259 5,258 5,257

Annual Average Scarcity (taf/yr) 0 1 1.5

Annual Average Scarcity Cost (US$K/yr) 0 5 11

**Constrained to current operating policies.**Results do not change with loss of filtration avoidance.

In the driest year this increases scarcity to 72.5 taf(89.43 mcm). It slightly reduces water delivered dur-ing dry years to Turlock Irrigation District (CVPM 11)and Modesto Irrigation District (CVPM 12) (Figure 3).There is a small transfer of water from agriculturaluses to urban uses during a few very dry years. Addi-tional urban water conservation would reduce watertransfer amounts.

Water rights and the allocation of storage space todistinct operating agencies are not included inCALVIN. Essentially CALVIN assumes that SFPUCpurchases a small amount of water from irrigationdistricts during shortage events, and this amount ofwater purchased increases slightly without O’Shaugh-nessy Dam. These results indicate the minimum levelof water scarcity from removing O’Shaughnessy Dam.

Using results from the base case modeling set, con-strained by 1997 operating policies (includingO’Shaughnessy Dam), water scarcity is observed inSFPUC and Santa Clara Valley residential demandarea in 1921 through 1934, 1977, and 1986 through1993. Over the entire 72-year time span, averageannual water scarcity is 6 taf (7.4 mcm) for SFPUCand 10 taf (12.3 mcm) for Santa Clara Valley (Jenkinset al., 2001).

Conveyance

Without O’Shaughnessy Dam, flows through theupper Hetch Hetchy Aqueduct (above New Don PedroReservoir) are less but remain considerable. Flows inthe Tuolumne River above the O’Shaughnessy Damsite do not change with removal of the reservoir. Onlystorage is eliminated. Thus, considerable quantities ofrunoff could be diverted at the dam site in much of

most years. Seasonal flows in the upper Hetch HetchyAqueduct illustrate the importance of springsnowmelt (Figure 4). The upper aqueduct is always atcapacity in April and May from spring runoff. During other months, flows through the upper aqueduct varyconsiderably with streamflow.

Flows through a New Don Pedro inter-tie to thelower Hetch Hetchy aqueduct are the inverse of flowsthrough the upper Hetch Hetchy Aqueduct (Figure 5).During April and May, flows through the New DonPedro inter-tie are zero because the aqueduct isalready at capacity with diversions at the O’Shaugh-nessy Dam site. During other months, water is divert-ed from New Don Pedro to bring the lower portion ofthe Hetch Hetchy Aqueduct to capacity. Thus, thelower Hetch Hetchy Aqueduct (downstream of NewDon Pedro Reservoir) is always at capacity whenflows through a hypothetical New Don Pedro inter-tieare incorporated. The New Don Pedro inter-tie addsflexibility to the Hetch Hetchy System. WereO’Shaughnessy Dam to be removed, additional flexi-bility and conveyance from New Don Pedro wouldremain of value.


NULL AND LUND

Figure 3. Total Annual Agricultural Scarcity.

Figure 4. Seasonal Flows in the Upper Hetch Hetchy Aqueduct.

Hydropower and Water Treatment

Hydropower generation is reduced substantiallywithout O’Shaughnessy Dam, primarily from elimi-nating hydropower generation at Kirkwood PowerPlant, which requires head from O’Shaughnessy Dam.Generation at Moccasin Power Plant is reduced signif-icantly and is reduced slightly at Holm Power Plant.The loss of hydropower generation at Kirkwood andthe reduction at Moccasin and Holm average 457GWhr/yr. The average annual hydropower revenueloss is approximately US$12 million/yr, assumingmonthly varying wholesale electricity prices (Ritze-ma, 2002).

Filtration treatment O&M costs of US$17/af(US$13.79/tcm) are used based on operating filtrationcosts for other California cities with high qualitysource water. If filtration avoidance were lost, treat-ment operating costs could rise to US$17/af(US$13.79/tcm), or US$6 million/yr. Most likely, thesecosts would be passed on to urban water users, rais-ing monthly water bills to rates comparable to otherCalifornia cities. Overall, water quality would remainhigh because reservoirs such as New Don Pedro(which do not have filtration avoidance) also havegood water quality.

Year 2100 Results

In model runs with projected year 2100 demandand a warm, dry hydrology (having 26 percent lessaverage inflow) representing possible climate warm-ing changes, the entire region is short of water but notshort of storage capacity. In these runs, water scarcityis extensive, especially for agricultural areas. Thisunderscores an important distinction: water and stor-age space are not the same. In year 2100 runs, wateris generally not stored in surface reservoirs forextended periods; it is used promptly to meetincreased demands, reducing evaporative losses towater supply.

The year 2100 model run with O’Shaughnessy Damstores an annual average 180 taf (222 mcm) in HetchHetchy Reservoir. It fills and empties the reservoirnearly each season, filling to capacity in 46 percent ofthe years and emptying to dead storage 74 percent ofthe years. The year 2100 model run withoutO’Shaughnessy Dam stores an annual average of 115taf (142 mcm) more water in New Don Pedro Reser-voir. There is always less surface storage in theremaining Hetch Hetchy System reservoirs in year2100 models than in year 2020 models. Despite con-siderable storage space, there is not enough water tomeet demands. Excess water rarely exists to be storedfor future years; rather it is usually sent to demandareas within a year. More ground water is used fordrought storage in year 2100 models than in year2020 models. As demand increases in the future, con-junctive use (which reduces evaporative losses) maybecome more widespread. Little difference in groundwater storage exists between the year 2100 resultswith and without O’Shaughnessy Dam.

The lower Hetch Hetchy Aqueduct was not atcapacity for one month of one year in the model runwith O’Shaughnessy Dam and without the New DonPedro inter-tie. The Hetch Hetchy Aqueduct belowNew Don Pedro Reservoir remains full despite theremoval of O ’Shaughnessy Dam with year 2100demands. The inter-tie may be desirable even withO’Shaughnessy Dam.

In both year 2100 model runs there is water scarci-ty to urban water users. Residential San Franciscousers average 12.7 taf/yr (15.7 mcm/yr) of waterscarcity, and Santa Clara County water users average25.5 taf/yr (31.5 mcm/yr) of water scarcity (Table 4).Maximum monthly scarcity to these regions is 3 tafand 6.3 taf (3.7 mcm and 7.8 mcm), respectively.Combined average scarcity to other Central Valleyurban demand areas is 13.5 taf/yr (16.7 mcm/yr).Removing O ’Shaughnessy Dam does not changeurban water deliveries or scarcity. Water scarcity toall agricultural demand areas ranges from 429taf/year to 1,265 taf/year(529 mcm/yr to 1,560mcm/yr) for runs with and without O’ShaughnessyDam.

During times of water scarcity, maximum monthlymarginal willingness to pay for additional water var-ied between US$929 to US$969/af (US$754 toUS$786/tcm)for San Francisco and Santa Clara Val-ley urban areas. Maximum marginal willingness topay for more water in Central Valley urban demandareas varied between US$412 to US$522/af (US$334to US$423/tcm). The maximum monthly marginalwillingness to pay for agricultural users was far less,between US$267 to US$352/af (US$217 to US$286/tcm). The large difference in marginal willingness topay between coastal urban water users and other



Figure 5. Monthly Flows Through a New Don Pedro Inter-Tie.

Central Valley urban and agricultural users arisesbecause the lower Hetch Hetchy Aqueduct is at maxi-mum capacity. Despite water scarcity to urbanregions, more water cannot be delivered without addi-tional conveyance capacity.

Both desalination and recycled water was availableat a price of US$1,000/af (US$811/tcm). Additionaldesalination and recycling facilities in CALVIN wereused in only two months in year 2100 runs. The runwith O’Shaughnessy Dam used a total of 28 taf (35mcm) of recycled and desalinated water. The runwithout O’Shaughnessy Dam used a total of 24 taf (30mcm). Overall, scarcity was extensive and the urbanmarginal value of water was high before recycled ordesalinated water was used. Users opt to face somescarcity (reducing use by conservation efforts) ratherthan pay for additional water from these sources(unless costs of desalination decrease significantly).Some scarcity can be optimal.

DISCUSSION

A stable water supply for the San Francisco penin-sula could be maintained after removing O’Shaugh-nessy Dam. The numerous reservoirs and pipelines inthe Hetch Hetchy System provide considerable flexi-bility for delivering water. Adding an inter-tie linkingNew Don Pedro Reservoir with the Hetch HetchyAqueduct makes it possible for little change to occurto water deliveries when O’Shaughnessy Dam is

removed. This inter-tie from New Don Pedro Reser-voir allows for the potential isolation of O’Shaugh-nessy Dam decisions from other parts of the SanJoaquin and Bay Area. Water storage space in NewDon Pedro Reservoir is shared by three entities:SFPUC, the Modesto Irrigation District, and the Tur-lock Irrigation District. This inter-tie would also allowwater transfers, exchanges, or other forms of flexibleoperations among these agencies in the event of along drought or operational interruptions to facilitiesupstream.

When O’Shaughnessy Dam is removed in combina-tion with a warmer drier hydrology and waterdemands increased to projected year 2100 demands,there are surprisingly few effects of dam removal onwater deliveries and operation of the Hetch HetchySystem. Some water scarcity occurs to urban residen-tial demand areas and considerable water scarcity toagricultural demand areas regardless of the existenceof O’Shaughnessy Dam. Scarcity occurs because insuf-ficient water exists in the system despite unused surface water storage and the lower Hetch HetchyAqueduct remaining at capacity at all times.Although water desalination and water recycling areavailable, some water scarcity costs (for water conser-vation) are preferable to higher costs of additionalwater supplies. With increased future demands, waterstorage in ground water basins increases, suggestinggreater conjunctive use in the future.

Removing O’Shaughnessy Dam carries consider-able financial costs. These include lost hydropowerrevenue, construction costs for additional water treat-ment facilities, increased treatment operating costs,and dam removal costs. Expanded opportunities fortourism and recreation in Hetch Hetchy Valley andresulting regional economic development would beneeded to justify dam removal and restoration eco-nomically. If urban, agricultural, and environmentalwater demands can be met without O’ShaughnessyDam, the decision to remove the reservoir and restoreHetch Hetchy Valley becomes economic and political.

Filtration avoidance makes O’Shaughnessy Damextremely valuable for SFPUC and the Hetch HetchySystem, saving the SFPUC operating and deferredcapital costs. It is very possible that this filtrationavoidance status will drive decisions regarding damremoval. However, if filtration avoidance status werelost, O’Shaughnessy Dam would lose most of its valueto the Hetch Hetchy System. In that case, economicvalue and revenues from recreation and tourism inHetch Hetchy Valley might offset lost hydropower rev-enue and increased treatment costs. However, ifHetch Hetchy Valley was opened to recreation, theeconomic benefits would go primarily to the NationalPark Service, concessionaires, and the local economy,while SFPUC would incur most costs. Further


NULL AND LUND

TABLE 4. Average 2100 Deliveries, Scarcity, and Scarcity Cost.

Climate Change HydrologyWith Without

O’Shaughnessy O’ShaughnessyDam Dam

Urban Regions

Average Deliveries (taf/yr) 2,314 2,314

Average Scarcity (taf/yr) 52 52

Average Scarcity Cost 22,748 22,730(US$K/yr)

Agricultural Regions

Average Deliveries (taf/yr) 2,537 2,536

Average Scarcity (taf/yr) 2,722 2,723

Average Scarcity Cost 18,087 18,090(US$K/yr)

thought and research are needed to examine thesepossibilities and changes.

Water management in California is dynamic.Changes in climate, water laws, water markets, andtechnology could alter the way water is moved andvalued. It is beyond the scope of this project to providea comprehensive economic benefit cost analysis or anestimate of public support for removal of O’Shaugh-nessy Dam. A thorough benefit cost analysis of poten-tial dam removal would be useful. Travel cost surveysand contingent valuation surveys could be used toestimate the economic benefits and public support forexpanded recreation potential. Estimates of increasedregional economic development also would be useful.Future ecological studies include creating a restora-tion plan for Hetch Hetchy Valley and measuring theimpacts of dam removal on the Tuolumne River andsurrounding ecosystems. These benefits of restorationcould then be compared with losses from lowerhydropower production and other costs. Additionalresearch on institutional aspects of water transfers orexchanges among SFPUC, the Modesto Irrigation Dis-trict, and the Turlock Irrigation District also would beuseful.

CONCLUSIONS

Removing O’Shaughnessy Dam need not substan-tially increase water scarcity. Without O’ShaughnessyDam, capture of considerable quantities of runoffcould be possible at the dam site for much of mostyears. Only storage is eliminated. Flow in theTuolumne River above the O’Shaughnessy dam sitedoes not change with removal of the reservoir. If NewDon Pedro Reservoir is connected directly with theHetch Hetchy Aqueduct, almost all demands can besatisfied without affecting the larger region outsidethe Tuolumne Basin. This substantially reduces theneed for difficult large-scale coordinated use agree-ments.

Conveyance sometimes can substitute for waterstorage. Connecting New Don Pedro Reservoir withthe Hetch Hetchy Aqueduct allows operators of theSFPUC flexibility to use different reservoirs mosteffectively to meet full water deliveries to demandregions. Even with O’Shaughnessy Dam remaining inthe Hetch Hetchy System, the inter-tie improvesdelivery reliability.

Removing O’Shaughnessy Dam reduces hydropow-er generation and reduces revenues by approximatelyUS$12 million/yr.

The loss of filtration avoidance that would occurwith removal of O’Shaughnessy Dam may be a pre-dominant economic factor in the debate to remove

O’Shaughnessy Dam. Construction costs of additionalfiltration facilities would be high. RemovingO’Shaughnessy Dam could increase Bay Area drink-ing water costs significantly, to levels common formost California cities.

Even with much greater future water demands anda drier warmer future climate, removing O’Shaugh-nessy Dam has little effect on the Hetch Hetchy Sys-tem and water deliveries to demand regions. Althoughthere is unused surface storage space with projectedfuture demands, there are not enough water inflows.Water and surface storage are not interchangeable.

Optimization modeling is useful in identifyingpotentially effective reoperations for water resourcesystems to support restoration. This optimizationapproach found an unexpected level of public accep-tance as a basis for a series of Pulitzer Prize winningeditorials in a major metropolitan newspaper, perhapsindicating a greater potential for such methods forpublic policy purposes.

Water delivery to urban and agricultural users maybe limited by political and institutional constraintsrather than water storage or other physical compo-nents of the Hetch Hetchy System. It is hoped thattechnical studies such as this will help enlightendeliberations and suggest opportunities.

ACKNOWLEDGMENTS

The authors thank Mimi Jenkins and Stacy Tanaka for guid-ance and modeling expertise throughout this project. The authorsalso thank three reviewers for their helpful comments on thismanuscript. This work was partially funded by a fellowship fromthe U.C. Davis John Muir Institute for the Environment.

LITERATURE CITED

CDWR (California Department of Water Resources),1998.The Cali-fornia Water Plan Update. Bulletin 160-98, California Depart-ment of Water Resources, Sacramento, California.

CDWR (California Department of Water Resources), 2005. HetchHetchy Restoration Study. Available at http://hetchhetchy.water.ca.gov/. Accessed in April 2005.

City and County of San Francisco, 1999. San Francisco Water andPower: A History of the Municipal Water Department and HetchHetchy System. City and County of San Francisco, San Francis-co, California.

Draper, A.J. 2001. Implicit Stochastic Optimization With LimitedForesight for Reservoir Systems. Ph.D. Dissertation, Universityof California-Davis. Davis, California.

Draper, A.J., M.W. Jenkins, K.W. Kirby, J.R. Lund, and R.E. Howitt.2003. Economic-Engineering Optimization for California WaterManagement. Journal of Water Resources Planning and Man-agement 129(3):155-164.

Howitt, R.E., K.B. Ward, and S.M. Msangi. 2001. CALVIN WaterManagement Appendix A: Statewide Water and AgricultureProduction Model. Available at http://cee.engr.ucdavis.edu/



faculty/lund/CALVIN/Report1/Appendices/AppendixA.pdf.Accessed in March 2005.

Huber, N.K. 1987. The Geologic Story of Yosemite Valley. U.S. Geo-logical Survey Bulletin 1595, Washington, D.C., 64 pp.

Hundley, N., 1992. The Great Thirst: Californians and Water,1770s-1990s. University of California Press, Berkeley and LosAngeles, California.

Jenkins, M.W., A.J. Draper, J.R. Lund, R.E. Howitt, S. Tanaka, R. Ritzema, G. Marques, S.M. Msangi, B.D. Newlin, B.J. VanLienden, M.D. Davis, and K.B. Ward, 2001. Improving Califor-nia Water Management: Optimizing Value and Flexibility. Cen-ter for Environmental and Water Resources Engineering.Available at http://cee.engr.ucdavis.edu/faculty/lund/CALVIN/Report2/CALVINReport2001.pdf. Accessed on February 7, 2006.

Jensen, A., 2004. Hetch Hetchy Reclaimed: Hetchy Clients NeedEqual Water Supply; 9/7/04. Sacramento Bee, Sacramento, Cali-fornia.

Labadie, J.W., 2004. Optimal Operation of Multireservoir Systems:State-of-the-Art Review. Journal of Water Resources Planningand Management 130(2):93-111.

Landis, J.D. and M. Reilly, 2003. How Will We Grow: Baselines Pro-jections of California’s Urban Footprint Through the Year 2100.Working Paper 2003-2004, Department of City and RegionalPlanning, Institute of Urban and Regional Development, Uni-versity of California-Berkeley, Berkeley, California. Available athttp://www-iurd.ced.berkeley.edu/pub/WP-2003-04-screen.pdf.Accessed on February 7, 2006.

Leal, S., 2004. Hetch Hetchy Reclaimed: SF: Proceed with ‘ExtremeCaution’ at Hetch Hetchy; 9/12/04. Sacramento Bee, Sacramen-to, California.

Lund, Jay R., Richard E. Howitt, Marion W. Jenkins, Tingju Zhu,Stacy K. Tanaka, Manuel Pulido, Melanie Tauber, RandallRitzema, and Inês Ferriera, 2003. Climate Warming and Califor-nia’s Water Future. Available at http://cee.engr.ucdavis.edu/faculty/lund/CALVIN/ReportCEC/CECReport2003.pdf. Accessedin March 2005.

Muir, J., 1912. The Yosemite. The Century Co., New York, NewYork.

Null, S., 2003. Re-Assembling Hetch Hetchy: Water Supply Implica-tions of Removing O’Shaughnessy Dam. Master’s Thesis, Uni-versity of California-Davis, Davis, California.

Philp, T., 2004a. Hetch Hetchy Reclaimed: Looking Again at HetchHetchy. Sacramento Bee, 8/22/04, Sacramento, California.

Philp, T., 2004b. Hetch Hetchy Reclaimed: The Lost Yosemite.Sacramento Bee, 8/22/04, Sacramento, California.

Philp, T., 2004c. Hetch Hetchy Reclaimed: CALVIN Says the DamCan Go. Sacramento Bee, 8/29/04, Sacramento, California.

Philp, T., 2004d. Hetch Hetchy Reclaimed: The Dam Downstream.Sacramento Bee, 8/29/04, Sacramento, California.

Philp, T., 2004e. Hetch Hetchy Reclaimed: In 1987, an Attempt toBring Back the Valley. Sacramento Bee, 8/30/04, Sacramento,California.

Philp, T., 2004f. Hetch Hetchy Reclaimed: Editorial - San Francis-co’s Paradox. Sacramento Bee, 8/30/04, Sacramento, California.

Philp, T., 2004g. Hetch Hetchy Reclaimed: Muir’s Plea. SacramentoBee, 9/5/04, Sacramento, California.

Philp, T., 2004h. Hetch Hetchy Reclaimed: Yosemite on the Cheap.Sacramento Bee, 9/7/04, Sacramento, California.

Philp, T., 2004i. Hetch Hetchy Reclaimed: Addiction Explained.Sacramento Bee, 9/12/04, Sacramento, California.

Philp, T., 2004j. Hetch Hetchy Reclaimed: A River’s ‘Rajahs.’Sacramento Bee, 9/19/04, Sacramento, California.

Philp, T., 2004k. Hetch Hetchy Reclaimed: Drain It, Then What?.Sacramento Bee, 9/19/04, Sacramento, California.

Philp, T., 2004l. Hetch Hetchy Reclaimed: Hetch Hetchy’s Future: Itis Time for New Chapter, New Champions. Sacramento Bee,9/20/04, Sacramento, California.

Poff, L.N. and D.D. Hart. 2002. How Dams Vary and Why it Mattersfor the Emerging Science of Dam Removal. Bioscience 52(8):659-668.

Ritzema, R.S., 2002. CALVIN Climate Warming Appendix D:Hydropower in the CALVIN Model. Available at http://cee.engr.ucdavis.edu/faculty/lund/CALVIN/ReportCEC/AppendixD.pdf. Accessed in March 2005.

Ritzema, R.S. and M.W. Jenkins, 2001. CALVIN Water Manage-ment Appendix 2C. Region 3 Results: San Joaquin and SouthBay Area. Available at http://cee.engr.ucdavis.edu/faculty/lund/CALVIN/Report2/Appendix2C.pdf. Accessed in March 2005.

Rosekrans, S., N.E. Ryan, S.H. Hayden, T.J. Graff, and J.M. Balbus,2004. Paradise Regained: Solutions for Restoring Yosemite’sHetch Hetchy Valley. Environmental Defense, New York, NewYork.

SFPUC (San Francisco Public Utilities Commission), 2002. WaterSystem Improvement Program. Available at http://sfwater.org/main.cfm/MC_ID/7/MSC_ID/6. Accessed in February 2006.

SFPUC (San Francisco Public Utilities Commission), 2005. SFPUCReleases 2004 Water Quality Report. Available at http://sfwa-ter.org/detail.cfm/MSC_ID/51/MTO_ID/63/MC_ID/10/C_ID/2525/holdSession/1. Accessed in February 2006.

SFPUC (San Francisco Public Utilities Commission), 2006. Calav-eras Dam Replacement. Available at http://sfwater.org/main.cfm/PRJ_ID/141/MC_ID/7/MCS_ID/90. Accessed in February2006.

Sierra Club, 2002. Hetch Hetchy Timeline. Available at http://www.sierraclub.org/ca/hetchhetchy/timeline.asp. Accessed inJanuary 2005.

Stanton, S., 2005. Bee Editorial Writer Wins Pulitzer Prize. Sacra-mento Bee, 4/4/2005, Sacramento, California.

USBR (U.S. Bureau of Reclamation), 1987. Hetch Hetchy: A Surveyof Water and Power Replacement Concepts. Sacramento, Cali-fornia.

USBR (U.S. Bureau of Reclamation), 1997. Central Valley ProjectImprovement Act: Draft Programmatic Environmental ImpactStatement. Documents and Model Runs. U.S. Department of theInterior, Bureau of Reclamation, Sacramento, California (2 CD-ROMS).

USDOE (U.S. Department of Energy), 1989. Hetch Hetchy, Strikinga Balance: A Review of the Department of the Interior’s Surveyof Water and Power Replacement Concepts for Hetch Hetchy.U.S. Department of Energy, Washington D.C.

Zhu, T., M.W. Jenkins, and J.R. Lund. 2005. Estimated Impacts ofClimate Warming on California Water Availability UnderTwelve Future Climate Scenarios. Journal of the AmericanWater Resources Association (JAWRA) 41(5):1027-1038.


NULL AND LUND

REASSEMBLING HETCH HETCHY: WATER SUPPLY WITHOUT O ...

Documents

Transcript of REASSEMBLING HETCH HETCHY: WATER SUPPLY WITHOUT O ...