Final Paper - Executive Summary
November 2000

1. The World Commission on Dams Global Case Study Programme

This study is one of eight case studies being undertaken worldwide with a common methodology and approach to inform the World Commission on Dams investigation of the development effectiveness of large dams. This particular case study concerns Grand Coulee Dam (GCD) and the Columbia Basin Project (CBP), which were originally authorised as a single project but today they have come to be known by the two principal parts of the original project. GCD constitutes the dam and powerplant, Lake Roosevelt, and ancillary facilities at the dam. CBP constitutes the project to pump water from Lake Roosevelt to irrigate more than one million acres of land in the semi-arid region in the State of Washington known as the Columbia Plateau.

2. Context And Scope of the Grand Coulee Dam and Columbia Basin Project Case Study

During the early 1930s, the US Army Corps of Engineers ("the Corps") and the Bureau of Reclamation ("Reclamation") each produced similar but separate plans for a dam near an enormous canyon known as the Grand Coulee. Each report contained a plan for pumping water from a reservoir behind the dam to be pumped into an equalising reservoir that would then feed a series of canals, which would then irrigate part of the Columbia Plateau. Each plan recognised that hydroelectric power revenues would have to be used to subsidise irrigation water because without a substantial subsidy, farmers would not be able to earn their livelihood by cultivating the semi-arid lands. The Corps recommended against federal implementation of their plan, in part because it was felt there was insufficient demand for power. In contrast, Reclamation supported federal involvement in carrying out their plan. According to the Commissioner of Reclamation, by the time the project was completed, the demand for power would be sufficient to absorb the electricity generated at GCD, and the increased population from major cities in the region - Spokane, Seattle, Tacoma, and Portland - would provide a local market for the products of irrigated farms.

After the presidential election of Franklin Delano Roosevelt in 1932, members of the Roosevelt administration decided to go ahead with a different version of the plans advocated by the Corps and Reclamation. Roosevelt's principal objective in building a dam at the Grand Coulee was to make good on campaign promises by putting unemployed people to work building the dam. Roosevelt was also an advocate of public power, and he felt that inexpensive power provided by a dam at the Grand Coulee would curb tendencies of private utilities to charge excessively high rates for power. The version of the dam authorised by Roosevelt's administrators was a low dam at the Grand Coulee that would provide hydropower, but not irrigation water. Roosevelt was concerned about the price of the project as proposed (ie, a high dam plus irrigation facilities), particularly since he had already committed to supporting Bonneville Dam and was troubled about investing more of the New Deal's funding in the Northwest than its population warranted. Initially, the project was to be constructed by the State of Washington with federal funds, but GCD and CBP soon became a federal project to be built and operated by Reclamation.

In 1935, Roosevelt's administrators shifted their position by supporting a high dam that would provide both hydropower and irrigation water according to the 1932 plan developed by Reclamation. Lobbying for the high dam had been intense and it came from a number of quarters, including Reclamation, a congressional delegation from Washington, and local project supporters. Moreover, the irrigation components were a means of meeting Roosevelt's goals for a planned project where farmers from other parts of the country could be relocated, and for an ample supply of cheap public power that could help reduce the high rates being charged by private utilities. Following the Roosevelt administration's decision, a Supreme Court decision gave Congress an opportunity to authorise the dam, which it did. The congressional authorisation indicated that the project would be for the purpose of controlling floods, improving navigation, regulating stream flow, providing irrigation water for the reclamation of public lands and Indian reservations, and "for other beneficial uses". The authorisation called for "the generation of electric energy as a means of financially aiding and assisting such undertakings".

GCD, which is a mile wide and 550ft (170m) high, is the largest producer of electricity in the US and the third largest producer of electricity in the world. Construction of GCD and its associated electrical generating facilities was carried out in two stages, separated by nearly three decades. The first stage, which consisted of the dam and two powerhouses, started in 1933 and was completed in 1951. The second stage, consisting of the "Third Powerplant," began in the mid-1960s and was completed in 1975. GCD has a total generating capacity of 6 809MW, and during the 1990s, gross generation in some years was greater then 25 billion kWh.

Construction of CBP began in 1945 and was largely completed in 1955. Detailed planning for settlement and CBP agricultural activities was conducted under the auspices of the Columbia Basin Joint Investigations in the mid-1940s. Settlement of project lands began in 1949, and the last addition of lands to CBP occurred in 1985.

Remaining sections of this report are organised as follows. Section 3 describes the projected and actual impacts, as well as unintended consequences, of GCD and CBP. There are eight major sub-sections contained within Section 3 and they discuss, in consecutive order, the following issues: irrigation, hydropower, flood control, recreation, ecosystem impacts, social effects on non-indigenous peoples, effects on Native Americans in the US, and effects on First Nations in Canada. Section 4 considers system-wide operations and basin-wide impacts; Section 5 examines the distribution of costs and benefits and provides stakeholder perspectives on the project. Section 6 examines the decision-making process for the construction of GCD and CBP, while Section 7 examines the evolution of policies affecting GCD and CBP from the time the projects were first built to the mid-1990s. Finally, Section 8 introduces lessons learned from the case study. Below we summarise the principal findings from each of these sections.

3. Projected and Actual Impacts of the Grand Coulee Dam and Columbia Basin Project

Irrigation

Only about half of the 1 029 000 acres (416 000ha) proposed for CBP (as of 1945) are receiving irrigation water because the "second half " of project lands were never developed. Although the actual acreage that received water is almost half of what was planned, the rate of development of irrigated acreage was faster than originally anticipated.

Gross value of output per acre (in constant dollars) doubled from 1962 to 1992. This resulted from improved crop yields, the shift from traditional field crops to high-value fruits and vegetables, and farm size expansions that took advantage of improved farm technologies and economies of scale. CBP land areas devoted to wheat, potatoes and tree crops were underestimated by Columbia Basin Joint Investigations planners in the 1940s; those crops have been instrumental in generating substantial growth in the gross value of CBP agriculture. If the predicted crops had been grown in the expected proportions, the 530 000 acres (215 000ha) irrigated in 1992 would have produced approximately $338 million in $1998. The actual cropping patterns resulted in a gross value of production of $637 million in 1998 US dollars, nearly twice as much as predicted. Average CBP farm size is about 500 acres (200ha), far greater then early constraints of 160 acres. Government restrictions on maximum farm size and actual farm sizes have both increased as a result of changes in farm technology and economics.

Reclamation estimated the construction cost to bring irrigation water to 560 000 acres (227 000ha) at $1.25 billion, whereas the actual cost was $3.6 billion, nearly three times the predicted value for the same area. The original (1932) goal of having CBP farmers repay 50% of the cost of constructing irrigation works, including drainage facilities, has long since been replaced by estimates closer to between 10% and 15%. The subsidy to irrigators results, in part from payments made by the US Treasury (and in the future, to be made by Bonneville Power Administration (BPA) ratepayers) for a substantial portion of the total construction costs allocated to irrigation. There has been no link between repayment schedules and total irrigation project costs. Instead, irrigators are charged based on their "ability to pay".

In 1932, Reclamation envisaged substantial pumping subsidies for irrigators. Although comparisons depend heavily on assumptions made about rates that BPA charges its customers (which vary by type of customers), the predictions of the power subsidy in 1932 (ie approximately $19 million a year) are a reasonable approximation of the actual subsidy. However, foregone revenues from power sales are only one dimension of the energy-related subsidy to CBP. Another dimension involves revenues forgone because water that could otherwise be used to generate hydropower is diverted to irrigation. A conservative estimate places the annual value of these forgone revenues at $39 million. In addition, withdrawal of water from the Columbia River for irrigation purposes means there is less in-stream flow to support the river's anadromous fisheries. Irrigation has also decreased the quality of return flows and groundwater aquifers.

Prices of CBP irrigation water were set low as a matter of public policy. The justification for subsidies to irrigators was based on meeting a policy objective of fostering the growth of rural communities with many small family farms. However, this policy objective was only partially satisfied because of factors such as changing agricultural technology and economics, and unanticipated drainage costs. Originally estimated at about $5 million (about 60 million in $1998), drainage costs have grown to close to $130 million (in nominal dollars; ie dollars spent in the years in which costs were incurred). Part of the spiralling drainage cost was absorbed by Reclamation, and (in 1962) part of the cost was shifted to irrigators when their repayment obligation was raised from an average of $85 ($775 in $1998) per acre to about $132 ($710 in $1998) per acre.

Multiplier effects predicted qualitatively in 1932 have occurred, evidenced by the economic vitality of the CBP area. While these indirect economic benefits are important from a regional perspective, their significance is diminished when a national accounting stance is adopted.

Hydropower

Reclamation estimated the cost of GCD and the left and right powerhouses to be about $2 billion (in $1998). The actual cost of these facilities, which is represented in Reclamation records in nominal dollars, was $270 million. Based on assumptions about when these expenditures took place, the total cost for this portion of GCD was $2 670 million (in $1998). This was about one-third more than envisaged by the early designers. In 1932, Reclamation's planners estimated that there would be 15 main generating units of 105MW capacity each, yielding a total of 1 575MW. The planners also estimated that the facility would yield 800 000 kW of continuous firm power, which would be available for commercial sale. Project critics believed that demand for this power would not materialise in a timely fashion.

The availability of low-cost power in the US Northwest during the late 1930s (a result of the Corps' Bonneville Dam) and early 1940s (the result of power generated by Bonneville Dam and GCD) made the US Northwest an attractive location for numerous war-related activities, most notably, aluminum plants and other heavy industry, shipyards, and a nuclear weapons facility at Hanford, Washington. Since 1941, a combination of very low electric power rates, high electricity demand from aluminum and other industries, and population growth has provided a ready market for GCD power. Indeed, the existence of low-cost power from GCD and other dams on the Columbia River helped fuel the economic development of the US Northwest and created further demands for low-cost power.

As a consequence of the war-related economic activity generated in the US Northwest that occurred in the late 1930s and early 1940s, power generation for GCD was much higher than anticipated. For example, from 1949 on, actual capacity exceeded predicted capacity; and from the time GCD started generating power in 1941, the amount generated has been greater than predicted values. The Third Powerplant was not anticipated in the 1932 project design. Once the Third Powerplant came online in 1975, the 1932 predictions of generating capacity and generation were exceeded by an enormous margin.

Estimated cost for the Third Powerplant (converted to $1998 from values reported in Reclamation's 1967 planning documents) was $1.9 billion, whereas the actual cost was $2.93 billion, an overrun of approximately 55%. In spite of the large cost overrun, if the 1967 benefit cost calculations for the Third Powerplant had been redone using actual cost instead of estimated cost, the result would have been a benefit-cost ratio of a still favourable 2:1 instead of the actual 3.18:1.

Power from GCD becomes part of the regional supply transmitted and marketed by BPA, and thus it is not possible to identify customers served by GCD power per se. However, power produced by the Federal Columbia River Power System (FCRPS) and marketed by BPA is sold to several types of customers through a variety of power sales agreements and rates. Customers include public utility districts, municipalities, rural co-operatives, federal agencies, a collection of enterprises referred to as direct service industries (DSIs), and private utilities. GCD plays a pivotal role in FCRPS. For example, from 1950 to 1953, power from GCD accounted for about three-quarters of total FCRPS revenues. As more projects came online, the contribution of GCD diminished, but it still remained the centrepiece of hydropower generation, accounting for 20% to 33% of total FCRPS kilowatt hours from the late 1950s to the present. According to BPA, on a cumulative basis, by the early 1990s, GCD had generated revenues equivalent to over $2.9 billion (in nominal dollars).

Ancillary services and dynamic benefits are unanticipated benefits of GCD hydropower production. Using the language of power engineers, examples of these services and benefits include the following: voltage support, spinning and non-spinning reserves, energy imbalance, generator dropping, and station service. The sale of these ancillary services and dynamic benefits will become increasingly important in future years as power markets develop, but it is currently not possible to accurately estimate the market value of these services and benefits. Other benefits of GCD that were not anticipated in 1932 include the atmospheric pollutants avoided by using hydropower instead of the most likely alternative: coal-fired steam electric powerplants. While these benefits of pollutants avoided may be notable, it is not possible to provide defensible quantitative estimates of their value.

Flood Control

Project planners did not provide for flood control as part of GCD. Indeed, flood control only began to be considered in the context of GCD after massive damages accompanied record flooding in 1948. Although the idea of providing upstream storage (in both Canada and the US) on the Columbia River for power generation had been under discussion since the mid 1940s, the flood of 1948 spurred those discussions forward and placed flood protection on the agenda as a potential major function of GCD and other dams upstream of urban population centres such as Portland, Oregon.

Following a 1964 agreement between the US and Canada known as the Columbia River Treaty, four "Columbia River Treaty projects" were built: Duncan, Keenleyside (formerly named Arrow Lakes), and Mica dams in British Columbia (BC), Canada, and Libby Dam in Montana. Once the last of the Columbia River Treaty projects had been built in 1973, key dams in the US and Canada began to be operated jointly to optimise power generation and flood control, while respecting other project purposes, such as irrigation. On the US side, the principal players concerning management of the dams were the Corps, Reclamation, and BPA. For purposes of operating Canadian storage for basin-wide flood control, the Treaty called for management by the Corps, BPA, and BC Hydro.

Because GCD is operated as part of a large system that includes other FCRPS dams and the Columbia River Treaty dams in Canada, it is difficult to completely separate the influence of GCD in controlling floods from the effects of other dams. However, a 1999 Corps study of the annual flood control benefits derived from Grand Coulee Dam storage estimates the amount to be $20 200 000 in $1998.

Although the literature on flood control contains many references to ways in which flood protection works lead to more intensive development of flood plains with consequent future increases in flood damages (when flood control works fail), no specific documentation of this phenomena occurring in the Columbia River Basin could be found.

Recreation

In 1932, Reclamation planners did not anticipate the recreational benefits that are now associated with GCD and CBP. However, the Columbia Basin Joint Investigations of the mid-1940s anticipated the recreational potential of Lake Roosevelt. While no specific projections were made of potential visits to the project area, the investigations concluded that new recreational areas should provide for capacity of at least 40 000 to 80 000 additional people within the two decades following the end of World War II. Locations of major recreational facilities created by GCD and CBP are as follows: Lake Roosevelt National Recreational Area (Lake Roosevelt NRA), the dam itself, and CBP lands. These three categories of facilities are administered separately. Collectively, they currently receive annual visitor days totalling over three million.

Lake Roosevelt NRA houses numerous campgrounds, swimming beaches, boat ramps and boat docks. Six major facilities operated by the Colville Indian tribe offer houseboat and fishing rentals, fuel, and food. A 20-mile (32km) stretch of the shoreline along the Spokane River is also included in NRA. Recreation at GCD consist largely of guided tours and a laser light show featuring animated graphics on the downstream surface of the dam; the show runs every evening during summer and routinely attracts large crowds.

CBP created multiple opportunities for recreation, including major fishing areas such as Banks Lake and Potholes Reservoir. In addition there are six Washington State Parks and 32 other major state and federally managed recreation facilities in the CBP area. Recreation-related business is an economic mainstay for many communities located on our near major project reservoirs. These communities include Coulee Dam, Grand Coulee, and Electric City.

A number of issues related to recreation have become sources of tension. These include project operations that lower reservoir levels during the summer tourist season, and management conflicts between the National Park Service and the Colville and Spokane tribes.

Ecosystem Impacts

At that time GCD was planned, assessing ecological effects of proposed federal projects was neither a requirement nor a priority. The principal ecosystem impact of GCD and CBP has been on anadromous fish that once spawned past the dam, mainly, chinook, sockeye, and steelhead. Annual salmon populations in the Columbia River Basin have fluctuated widely, and estimates of numbers of fish vary significantly. Prior to white settlement in the 1800s, the basin supported a population estimated at between 7 and 30 million salmon and steelhead, with runs to the upper Columbia (ie, the portion of the basin above GCD) of between 500 000 and 1.3 million. Salmon and steelhead runs experienced substantial declines prior to construction of GCD because of factors such as the development of commercial fisheries, over-harvesting, grazing, timber harvesting, mining, dams on tributaries, Bonneville and Rock Island dams, roads, highways, railroads, and destruction of estuarine and freshwater wetlands. In 1938, before GCD cut off the upper Columbia River for migrating anadromous fish, the basin-wide run of salmon and steelhead was 2.2 million and the run to the upper Columbia River was estimated at 25 000.

Reclamation took control of GCD and CBP in 1933, but it was several years before it reached an agreement with the US Department of Fisheries (later, the US Fish and Wildlife Service (USFWS)) and the State of Washington regarding a plan to mitigate adverse effects on anadromous fish. This plan, the Grand Coulee Fish Maintenance Programme (GCFMP), was put forward by a board of consultants to the US Secretary of the Interior in 1937. Except for the Board's recommendation for a hatchery on the Okanogan River, the Board's plan was implemented.

The goal of GCFMP was to maintain a certain number of fish, not necessarily to preserve specific runs. For those in charge of the programme, fish were valued based on their commercial importance, and fisheries managers strove to maintain the commercial catch as measured in pounds. The economic losses associated with sport and commercial fishing were estimated at between $250 000 and $300 000 in 1937 dollars ($2 841 000 to $3 409 000 in $1998). At the time GCFMP was developed, no significant consideration was given to the importance of salmon to indigenous people in either the US or Canada. Indeed, the Minister of Fisheries in Canada indicated little concern over the potential loss of salmon and steelhead because there were no commercial salmon fisheries on the Columbia River in Canada. The state of knowledge of ecosystems at the time was such that virtually no consideration was given to the maintenance of genetic biodiversity.

GCFMP entailed trapping all of the fish runs at the Rock Island Dam downstream of GCD and transporting them to four tributaries - the Wenatchee, Entiat, Methow, and Okanogan rivers - between Rock Island and Grand Coulee dams, for natural propagation; and to three hatcheries - Leavenworth, Entiat, and Winthrop - for purposes of artificial propagation. The hatchery at Leavenworth was completed in 1940, and the other two were completed the following year.

The underlying concept of GCFMP was that juvenile upper Columbia River fish stock reared at the aforementioned hatcheries and tributaries would, as adults, return naturally to the tributaries to spawn. The hatcheries would remain in operation merely to augment the run to populations larger than could be hoped for from natural runs alone. Natural propagation turned out to be far more successful than artificial propagation using hatcheries.

At the time GCFMP was developed and first implemented, there was an extensive commercial fishery and there were only two other main-stem dams (apart from GCD). Upper Columbia River stocks were then estimated at 25 000 in an average year. Although there were numerous problems that persisted over several years because of the experimental nature of the entire enterprise, the GCFMP target outcomes were attained. GCFMP met its own numerical goal: maintaining upper Columbia River stocks of 36 500 salmonids in an average year. In fact, during recent years, the upper Columbia River stocks have averaged approximately 48 700 per year.

A simple comparison of the Board's target with today's population sizes may lead one to believe that GCFMP has been successful since the original target has been exceeded. However, the changes that have occurred over the past 60 years make a meaningful comparison between the present populations and targets set in the late 1930s difficult, if not impossible. The inability to compare present populations to the Board's target of 36 500 salmonids is demonstrated by considering differences in commercial fishery takes. In the early 1930s, over 75 000 wild salmonids were caught in lower Columbia River commercial fisheries, whereas today no in-river commercial harvesting of salmonids is permitted. If a comparison is to be made, the escapement at Rock Island (25 000 wild salmonids) in the 1930s should be added to the 76 300 wild salmonids caught at that time by commercial fisheries to give 101 300 wild salmonids as the comparable 1930s escapement figure at Rock Island. This comparison - 101 300 in the 1930s versus 48 700 in recent years - places the effects of GCD and GCFMP in a different light. Another factor demonstrating why it is not meaningful to compare the 25 000 fish escapement at Rock Island in the 1930s with the current average escapement of 48 700 is that no hatcheries were in operation in the early 1930s, whereas numerous hatcheries are operating today. The salmonids at Rock Island in the 1930s were wild, whereas the majority of the 48 700 salmonids at Rock Island today are from hatcheries.

The aforementioned differences (among many others, including differences in the value attached to biodiversity and changes in regulations on sport fishing) are not the only factors that make it difficult to compare the salmonid runs of 1930s to those of the 1990s. Such a comparison is also difficult because of three significant irreversible effects that GCD and GCFMP have had on the mid- and upper-Columbia River anadromous fish. First, GCD blocks 1 100 miles (1 770km) of riverine habitat, making it a practical impossibility to significantly increase the number of salmonids in the upper-Columbia River Basin. This change is virtually irreversible since the likelihood that GCD will ever be removed is practically nil: the dam is of central importance to the Columbia River hydropower system. Second, the Board's plan makes it a practical impossibility to substantially increase the population sizes of the upper-Columbia River stocks in the mid-Columbia River tributaries because GCFMP was developed around a habitat area that cannot support significant population increases. And third, GCFMP made it impossible to ever recover the original stocks of the four mid-Columbia River tributaries. These stocks, considered insignificant at the time of the Board's analysis, were completely subsumed within the upper-Columbia River stocks through the trapping and transportation programme in the late 1930s and early 1940s.

High concentrations of total dissolved gas (TDG) can be fatal to anadromous fish, and operations at GCD affect dissolved gas in two ways. First, the dam carries TDG generated at upstream dams. Second, spill at GCD increases TDG down river. TDG is a basin-wide problem.

GCD provided the impetus for stocking Lake Roosevelt with native and introduced species of resident (as opposed to anadromous) fish. Although the Spokane and Colville tribes now operate kokanee and trout hatcheries funded by BPA, a number of problems have prevented attainment of target harvest levels. In particular, whenLake Roosevelt is drawn down beginning in the summer or early fall, the reduced volume and surface area limit food supply and volume of optimal water temperatures during periods that are often critical for trout. Also, there are conflicts between releasing flows for outmigration of anadromous smolts and maintaining ideal conditions for resident fish in Lake Roosevelt

GCD and CBP caused numerous changes in terrestrial ecosystems, including a substantial reduction in the original shrub steppe habitat that once existed in the project area. Although CBP-related habitat change has caused a decrease in the populations of species like pygmy rabbits and burrowing owls, CBP has also created large areas of new habitat in the form of wetlands, reservoirs, and riparian corridors. This has led to increases in many wildlife species, particularly waterfowl. However, physical changes in new habitats, such as artificial wetlands that are not regularly maintained, have led to a falloff in species abundance (compared to levels observed during the 1970s).

Social Effects: Non-indigenous Peoples

In keeping with the President Roosevelt's original motivations for supporting the project, GCD and CBP provided jobs for thousands of workers. Estimates of new employment range between

2 000 and 8 000 workers, depending on the year; the size of the local communities in the area around GCD grew to 15 000 during the construction period.

Between 3 000 and 4 000 non-indigenous peoples were displaced as a result of the creation of GCD. Lake Roosevelt itself flooded at least eight towns of non-indigenous peoples; while some towns were moved successfully to new locations, others were never reestablished. Notification of settlers to be displaced by dam-related activities was not conducted following any formal notification procedure. Compensation offered for land and property to be inundated was a major source of contention between Reclamation and some settlers who were displaced. At the time, Reclamation lacked legislative authority to assist in the relocation of people displaced by Lake Roosevelt.

The 1932 Reclamation Report did not say much about the influence of CBP on the development of the Columbia Plateau, but that subject was studied extensively during the 1940s by the Columbia Basin Joint Investigations. Although some early CBP planners had estimated there would be as many as 80 000 families on 10 000 farms within CBP, the actual results were far different. For example, in 1973, there were in 2 290 farms operating in the project area. Notwithstanding the differences between what was planned and what occurred, CBP has fostered development of an economically viable farming community that contributes significantly to the state's economy. In addition to farming and farm-related economic activity, tourism linked to GCD and CBP has made significant contributions to the overall prosperity of the region.

During the past few decades, social tension has developed between project supporters in eastern Washington and critics of FCRPS in Seattle, Portland, and other urban areas west of the Cascade Mountains. This tension, which revolves around the symbolic and ecological significance of salmon in the US Northwest, evolved gradually as the Seattle, Tacoma, and Portland metropolitan areas grew in size, and as the adverse effects of hydropower on salmon runs became more widely known. Another factor contributing to the escalating tension has been a shift in societal values, as many residents of the US Northwest have developed an increased appreciation for rivers in their natural state and as scientists come to better understand the significance of maintaining biodiversity.

Effects on Native Americans in the US

In keeping with 1930s practices concerning "executive order reservations", the US had no formal process for involving the Colville and Spokane tribes in GCD-related decisions concerning the taking of reservation lands or the destruction of tribal fisheries.

About 2 000 members of the Colville tribe and between 100 and 250 of the Spokane tribe were displaced as a result of GCD-related construction. After 1940, the Department of the Interior abandoned its former practice of acquiring Indian consent; thereafter, the government notified Native American landowners by mail concerning lands to be taken and apprised them of how much they would be paid. Appeals processes existed, but tribal members were not encouraged to engage in the process, nor were they supported when they initiated appeals. Although displaced tribal members appear to have received payments at the same level as displaced non-indigenous settlers, tribal members received their funds significantly later. Additionally, the process for assessing land did not value built structures on Native American lands (eg living quarters) the same as those of non-indigenous settlers. Two tribal towns - Keller and Inchelium - were forced to relocate. Inchelium had no water supply at its new site, and its residents did not regain telephone service for 30 years. The rate of filling Lake Roosevelt was such that many tribal graves were inundated; since then, many burial sites often have been exposed because of fluctuations in reservoir levels and increased bank erosion, leaving them open to theft and desecration. In addition, the physical barrier posed by Lake Roosevelt, and the increased density of settlement on the Columbia Plateau, cut off access of Colville and Spokane tribal members to food and medicinal plants and to each other's reservations.

According to tribal members we interviewed, government officials had indicated that the Colville and Spokane reservations would receive free electrical service from GCD, but they only acquired electricity much later and they usually paid a much higher price than typical off -reservation residents paid.

By completely eliminating runs of salmon above GCD, the project severely disrupted the way of life for the Colville and Spokane tribes: important salmon-based cultural and ritual ceremonies were eliminated, parts of language and crafts associated with fishing disappeared, and tribal members' diets changed significantly. For the Spokane and some of the tribes of the Colville Confederation, salmon probably accounted for about 40% to 50% of their daily diet before GCD. As a result of moving to foods high in fat, sugar, and salt, rates of heart disease, diabetes, and other diet-related illnesses have increased significantly on the reservations.

GCFMP further damaged the Colville tribe's fishery below GCD because the capture and transplantation of most adult fish to mid-Columbia River tributaries destroyed much of the remaining wild salmon that spawned between the Okanogan River and GCD, both on the main-stem Columbia River and on reservation streams. Collectively, problems with fish survival in the early days of the GCFMP hatcheries, loss of fish that spawned in the upper reaches of the Columbia River, and genetic weakening through mixing stock made it much more difficult for members of mid-Columbia River Treaty tribes to maintain their traditional livelihood by fishing.

Fish losses and government mitigation efforts associated with GCD have exacerbated tensions between members of treaty reservations and executive order reservations. More generally, opportunities for misunderstandings among tribes have increased because salmon ceremonies, which used to provide an opportunity for enhanced understanding among members of different tribes, no longer take place. Other reasons for increased tensions include less salmon to go around and perceived inequities on how different tribes were compensated for the adverse effects of GCD (and FCRPS in general).

In 1978, acting on a claim by Colville Confederated Tribes, the Indian Claims Commission awarded more than $3 million (in nominal dollars) for the subsistence value of fish lost to the claimant tribes since 1940. In a claims case settled in favour of the Colvilles in the early 1990s, the US government made a $53 million lump sum payment (in nominal dollars) in recognition of the government's unfulfilled promise of 1933 to pay the tribes an annual share of power revenues. This settlement also included future payments by BPA to the tribes of approximately $15 million annually. Despite these monetary payments, many tribal members feel that no monetary compensation can make up for how GCD caused a loss of fisheries, towns, burial sites, cultural traditions, and, more generally, a way of life.

Effects on First Nations in Canada

As in the case of Native Americans, salmon were also central to the economic, cultural, and spiritual life of Canadian First Nations. Salmon were also integral to the daily diet of First Nations people. In contrast to Native American tribes, Canadian First Nations were unaware of plans to construct GCD. The federal government in Canada dismissed the importance of GCD as an impediment to migration of salmon because there was no commercial salmon fishery in the Canadian portion of the Columbia River Basin. GCD blocked all anadromous fish runs to the Ktunaxa, Shuswap, and Lakes-Sinixt territories. As a result of GCD construction, a number of fish populations or stocks were lost including Arrow, Slocan, and Whatshan sockeye, Columbia and Windermere Lake sockeye, and numerous stocks of chinook. GCFMP adversely affected the Okanogan sockeye population since it called for the harvesting of Okanogan sockeye for hatchery brood stock and the genetic mixing of wild and hatchery stocks.

The importance of salmon fishing to Canadian First Nations is demonstrated by the locations of current and former reserves at important fishing locations, including the confluence of the Kootenay and Columbia rivers. As a result of the loss of salmon along the Columbia River in Canada, members of the Arrow Lakes band disbursed to other communities. The Arrow Lakes band no longer has any enrolled members. Salmon provided the foundation for the subsistence economies of Canadian First Nations, and, at the time GCD was built, the subsistence economies were far more important than wage or commercial economies.

In contrast to the situation for Native Americans in the US, there has been no mitigation for the adverse impacts of GCD on Canadian First Nations. Canadian First Nations are actively pursuing mitigation and compensation for their salmon losses, particularly through efforts to achieve their long-term goal of salmon restoration.

4. Basin-Wide Impacts and System Operations

The 1932 Butler Report as a Plan for the Upper Columbia River Basin

In terms of hydroelectric power projects on the Columbia River above the Snake River, with minor modifications, every one of the dams in a 1932 Corps Report dealing with this reach of the river was eventually built. Key recommendations in this report, prepared by Major Butler of the Seattle District Office of the Corps, were overruled by the Chief of the Corps of Engineers. While Butler recommended that the government proceed with GCD and CBP, the Chief recommended against it. Instead, the official Corps position was that hydropower and irrigation projects in the Butler report be undertaken by parties other than the US government.

Influence of GCD on Decisions to Build other Water Projects within the Columbia River Basin

There is a strong linkage between GCD and many other dams on the Columbia River. In the US, this linkage is clearest for Libby Dam (located upstream in Montana) and Chief Joseph Dam (located immediately downstream). Libby was one of the Columbia River Treaty projects, and decisions to build each of those projects were directly affected by GCD. Operations at Chief Joseph Dam are tied closely to those at GCD.

The desire to increase power generation at GCD by having increased storage in Canada motivated discussions leading to the Columbia River Treaty. If the downstream power benefits to Canada resulting from the Treaty had not been made available, Canada was prepared to build coal-fired powerplants (instead of the Canadian treaty dams) to meet power needs in British Columbia. This makes clear the direct linkage between GCD and the Columbia River Treaty dams.

Effects of the Columbia River Treaty Dams on Canadian Resources and People

While there is substantial agreement about the nature of the flood control and power benefits of the Columbia River Treaty dams within Canada, differences of opinion exist about the adverse environmental and social effects of those dams within Canada. In general, the environmental and social effects concern loss of forestry resources and agricultural lands, effects on recreation and recreation related tourism, effects of fluctuations in reservoir levels, and the influence of the Canadian Treaty dams on the tax base of local communities.

Two major programmes have been established to compensate communities adversely affected by Canadian Treaty dams: one involves the Columbia Basin Fish and Wildlife Compensation Programme, a nonprofit entity separate from - but not independent of - its BC Hydro and BC Environment partners; and the second involves the Columbia Basin Trust, an entity with broad powers created to ensure that benefits derived from the Treaty help renew the economy and natural environment of the region impacted by the Treaty dams.

Cumulative Impacts of Projects on Ecosystems

Basin-wide factors affecting anadromous fish have been classified under the so-called "four

H's": harvesting, referring primarily to adverse effects of over-harvesting; hatcheries; hydropower (ie, the influence of the many dams in the basin); and habitat, as affected by a number of activities including timber harvesting, wetland destruction, and irrigation.

Some scientists have suggested that spawning and spawned out salmon in the upper basin comprised an important source of nutrition for animals such as bears and eagles, and that GCD (along with other dams) blocked an important source of nutrient transfer from the Pacific Ocean to the upper Columbia River Basin. The specific effects of GCD have yet to be unraveled.

Cumulative Socioeconomic Impacts

In the context of the Columbia River Basin, the significance of cumulative socioeconomic impacts of dams is undeniable. Indeed, the notion of cumulative effects is implicit in the basin-wide plans laid out in the 1932 Corps Report prepared by Major Butler. In terms of socioeconomic effects of GCD and CBP, the cumulative effects are seen in the economic development of the CBP area, and, more generally, in the US Northwest.

An important set of cumulative effects involves the influence of the CBP's agricultural outputs on farming activities in other parts of the US and Canada. In particular, many potato farmers in Idaho, once one of the leading producers of potatoes in the US, shifted to other crops in the face of very stiff price competition from potato growers receiving CBP irrigation water. Similarly, growers of apples, potatoes and other crops in British Columbia have felt increased competition from CBP farmers who produce the same outputs with the advantage of subsidised irrigation water.

Relationships between cumulative socioeconomic impacts and cumulative ecosystem effects are also notable. For example, GCD promoters minimised the importance of GCD in interfering with anadromous fish runs by arguing that those runs (which decision-makers valued in economic terms) had already been diminished significantly by Rock Island and Bonneville dams prior to the construction of GCD.

System Operations

At the time it was completed in 1941, GCD was operated to optimise power generation. Between 1948, the time of massive flooding in the lower Columbia River, and 1972, GCD was relied upon more heavily for flood control; in addition, power generation shifted from project specific to more system-wide management. After all four Columbia River Treaty projects came online in 1973, power generation and flood control activities at GCD were organised as a basin-wide management system. Under conditions of the Columbia River Treaty, BPA, the Corps, and BC Hydro schedule basin-wide management, and Reclamation works with BPA and the Corps in running FCRPS. Following passage of the Northwest Power Act of 1980, the creation of the Northwest Power Planning Council (NPPC) and its 1983 fish and wildlife programme marked the increased consideration of anadromous fish in decision-making related to FCRPS. In all GCD operations, irrigation water rights have been respected.

In 1990, BPA, the Corps, and Reclamation used the EIS process established by the US National Environmental Policy Act (NEPA) of 1969 to conduct a "System Operation Review". The three organisations developed an FCRPS system-operating strategy to balance the varied and sometimes conflicting needs of water users in the US portion of the basin. During the mid-1990s, the National Marine Fisheries Service (NMFS) listed the following species as endangered under the Endangered Species Act (ESA): Snake River sockeye, Snake River spring/summer chinook, and Snake River fall chinook. The Biological Opinions issued by NMFS in connection with the above-noted listings outlined what was required for species recovery. The federal operating agencies (BPA, the Corps, and Reclamation) have tried to adopt the general recommendations in the Biological Opinions in developing a preferred system operating strategy for FCRPS.

Current GCD operations reflect trade-offs among various project purposes. Reclamation develops operating requirements specific to irrigation at CBP, and BPA requests that reservoir operators (eg, Reclamation and the Corps) maintain reservoir levels to maximise power output given the limits imposed by requirements for irrigation, flood control, and maintenance of anadromous fisheries. Other issues, such as recreation, are also considered, but they have a lower priority.

Evolution of Basin-wide Planning Institutions

Advocates of basin-wide planning and management made a valiant effort during the 1930s and 1940s to create a new institution for comprehensive planning in the Columbia River Basin: the Columbia Valley Authority (CVA). That effort failed. Long-term squabbles between Reclamation and the Corps, and the reluctance of US Northwest residents to surrender the economic future of their region to some unknown and untested body helped kill the CVA idea.

BPA, initially created in 1937 as a temporary entity to market power from Bonneville Dam, became a permanent agency charged with marketing power from all federal dams on the Columbia River, including GCD. Between World War II and the ratification of the Columbia River Treaty in 1964, BPA, Reclamation, and the Corps dominated decisions associated with hydropower. Once the treaty was in force, BC Hydro joined this group of agencies as one of the principal participants in managing basin-wide operations. Currently, numerous federal, state, and local agencies, tribal governments, and others (including NPPC) are all taking a hand in planning and developing resources within the basin.

This ongoing proliferation of agencies has resulted inevitably in duplication of effort and cases where agencies, unable to work together and poorly co-ordinated, have missed or ignored cautions presented by others. The most striking of these ignored warnings were the ones issued by USFWS in the 1940s in which that agency cautioned that without proper planning, dams proposed by the Corps would seriously endanger annual salmon runs.

5. Distribution of Benefits and Costs

Anticipated Beneficiaries and Cost Bearers

The most comprehensive analysis of beneficiaries prior to the decision to build GCD was made in the 1932 Butler Report on the Columbia River and its tributaries. This report identified direct benefits in the form of profits to farmers and businesses; indirect benefits in the form of increased values for farmland and urban land: and public benefits in the form of an increased supply of produce for consumption and processing as well as reduced food prices. The 1932 report also included an elaborate analysis of secondary benefits based on multiplier effects (eg, the increased economic activity for agribusiness, banking, wholesaling, retailing, and transportation).

Major Beneficiaries and Cost Bearers

Major project beneficiaries include irrigators, who receive subsidised CBP water and power (as well as associated agribusiness); BPA ratepayers, who enjoy relatively low electric power rates; residents and businesses protected from floods downstream; recreators and recreation-related commerce; US Northwest residents (since GCD helped develop the region economically during and after World War II); and BC Hydro ratepayers, who enjoy very low electric power rates.

Major cost bearers include Native Americans and First Nation tribes; non-Native Americans who were forced to resettle; commercial and sport fishing interests (for anadromous fish) in the US and Canada; US taxpayers, who have thus far paid the bulk of the irrigation construction costs; individuals concerned with maintaining ecosystem integrity (because of the loss of wild salmon and the Columbia River in its natural state); some US farmers outside the CBP area (because of increased competition from CBP farmers); and, upstream residents and businesses whose resources and livelihoods were affected by GCD or the Columbia River Treaty dams.

Stakeholder Perspectives

We interviewed nine stakeholders in the US who represented tribal entities and environmental groups, and they shared similar perspectives on the benefits and costs of the project. In aggregate, they viewed major beneficiaries as irrigators, electric power users (particularly, public utilities, citizens, and large aluminum companies), and residents of the Northwest (who gained from the overall socioeconomic development of the region). Tribal members and environmental groups saw the Native American, sport and commercial fishing interests, and US citizens generally (from the perspective of ecosystems lost and modified) as the primary cost bearers. Tribal members felt that the adverse impacts on many Native Americans were adverse, severe, and irreversible. Additional stakeholder input from the Canadian perspective was captured during a WCD meeting held in Castlegar, British Columbia. Attendees at this meeting voiced the opinion that First Nations tribes that had traditionally depended on upstream salmon spawning also suffered many of the same adverse impacts as Native American tribes in the US.

The other 12 stakeholders we interviewed represented governments of communities near GCD, a public utility district, and CBP farmers. In aggregate, they viewed the principal beneficiaries as irrigators, electric power users, residents of the US Northwest, local towns, and recreators. These 12 individuals saw the Native American tribes as the primary cost bearers, but they felt that the tribes also benefited from the project (eg, via tourism) and that the net positive impacts of GCD and CBP far outweighed the cost borne by Native Americans.

Cost Allocation and Repayment

Repayment obligations for BPA include the cost of constructing hydropower-related facilities and interest during construction, interest during the repayment period, and what is known as "irrigation assistance". The latter is the difference between the irrigators' allocated share of construction costs (including interest) and the amount that irrigators actually re-pay. The US Treasury pays for non-reimbursable capital costs (eg, cost for flood control).

As of 30 September 1998 the total construction cost for GCD and CBP was $1.93 billion (in nominal dollars), and nearly 88% of that total cost was returnable from BPA's power revenues. As of that date, BPA had paid the full cost allocated to commercial power ($1.1 billion) but it had paid almost none of the costs associated with irrigation. BPA is not required to begin those payments until 2009, and those payments are scheduled to end in 2045.

Thus far, about $445 million (nominal) dollars (24% of the total project cost) has been paid to the US government. As of 1998, irrigators had paid in about $51 million (less than 3% of total project cost, about 12% of total repayment) of the total project cost. Revenues collected by BPA from power sales had repaid about $389 million (approximately 20% of total project costs and about 88% of total repayment). These figures are in nominal dollars, and the repayment obligations are specified in nominal dollars (ie, dollars that are not adjusted for inflation).

Irrigation assistance, as of 1997, was to be approximately $585 million, which is 87% of the project costs allocated to irrigation. This means that irrigators will eventually pay 13% of their allocated project cost. Since there is no interest on these payments, the revenues will be deposited lump sum in the US Treasury according to the contract for each irrigation block.

All FCRPS projects are grouped together as part of a basin-wide accounting system in which revenues and costs generated by projects in the system are pooled into a common fund. Using this accounting scheme, Reclamation derives income from various revenue-producing power facilities to fund works that would not otherwise be economically feasible under Reclamation law. The basin-wide accounting system has been criticised for obscuring projects that are not economically feasible by lumping them together with more economically viable projects. The practice of basin-wide accounting is widespread in the US and is used, for example, for Reclamation projects in the Colorado, Missouri, and the Bighorn river basins.

6. Options Assessments and Decision-Making Processes

Decision to Build a Dam at Grand Coulee

The 1932 planning studies prepared by the Corps and Reclamation and the associated congressional debates were not the key factors in the decision to build GCD. Instead, it was President Franklin Roosevelt, newly elected in 1932, who pushed the project forward. The mechanism used was legislation establishing the Public Works Administration, which allowed Roosevelt and his advisers to circumvent the congressional authorisation process to bring major infrastructure projects online during the Great Depression. At first, the project was to be a low dam for hydropower to be built by the State of Washington. Within months, the project became a federal project to be constructed and operated by Reclamation. By 1935, the Roosevelt administration amended its original plan and directed Reclamation to proceed with a high dam that would include facilities for irrigating the Columbia Plateau. Later in 1935, a Supreme Court decision forced a change in plans by requiring the administration to submit GCD and CBP, along with many other projects, to the congressional authorisation process. Although construction of the project was well underway, the congressional debates were contentious. In the end, GCD and CBP received congressional authorisation.

Early Attempts to Compensate Native Americans for Expected Losses

While the project was still under the jurisdiction of the State of Washington, tribes that would be affected by GCD used the Federal Power Act of 1920 to ensure that fish passage facilities would be provided at GCD and that annual payments would be made to Colville and Spokane tribes for tribal lands flooded by the reservoir. When the project became federalised, provisions of the Federal Power Act of 1920 requiring fish passage and annual payments no longer applied. However, Secretary of the Interior Harold Ickes endorsed correspondence between the Commissioners of Indian Affairs and Reclamation indicating that the government would pay the tribes for a share of power revenues generated from water on tribal lands. Until court actions in the 1970s, the federal government ignored these provisions.

Columbia Basin Joint Investigations

The Reclamation Act of 1939, which applied to CBP, called for the kind of orderly and systematic planning advocated by New Deal leaders. The Columbia Basin Joint Investigations were launched as part of this planning process. The investigations, which involved over 300 people from over 40 organisations during the period from 1941 to 1943, examined 28 potential issues related to the project, including the types of crops that should be grown, the optimum size of farm units, and the annual rate at which lands should be brought into production. The investigations played a key role in shaping CBP.

Acreage Limitations and Anti-speculation Statutes

The Roosevelt administration was concerned that word of the planned irrigation in the CBP area would cause individuals to buy out land with hopes that prices would rise and windfall profits could be made by selling project land at inflated prices. In response, Congress passed the anti-speculation pact, which limited the size of project farms in CBP to 40 acres (16ha) for an individual, and 80 acres (32ha) for a husband and wife. In 1943, these provisions were modified to allow farmers receiving CBP irrigation water to hold as much as 160 acres (64ha). This limit, while thought to be adequate, was low enough to maintain the vision of CBP as a project populated primarily with family farms, not the large corporate farms that have come to be referred to as "agribusiness".

7. Criteria and Guidelines: Policy Evolution and Compliance

The most significant GCD-related policy changes after the mid-1950s were as follows: (i) the Columbia River Treaty, which formalised basin-wide management of operations for hydropower and flood protection; (ii) NEPA, which institutionalised an environmental impact statement (EIS) process that was used to considerable advantage in the context of FCRPS's System Operation Review, and in Reclamation's decision not to irrigate the CBP's "second half" at the current time; (iii) the Northwest Power Act of 1980, which called for a fish and wildlife management programme for the US portion of the basin and opened up the decision-making process related to FCRPS operations to Native Americans and other previously excluded groups; and (iv) the ESA listings and NMFS Biological Opinions related to Columbia River Basin anadromous fish, because those listings and opinions have profoundly affected operations of FCRPS.

8. Lessons Learned

Open Planning Process

Lesson: An open planning process facilitates identifying and resolving conflicts among stakeholders; a closed process serves the opposite purpose.

Evidence: Inadequacies in opportunities for Native Americans and Canadian First Nations to participate in decision-making has fostered conflicts over GCD for several decades.

Managing Debates on Project Operations

Lesson: In a multipurpose water project, it is common for project purposes (eg, flood control and recreation) to be in conflict. Because conflicts among competing purposes are practically inevitable, a process for managing stakeholder contributions to debates on project operations should be institutionalised for future projects.

Evidence: Operation of GCD has changed in response to shifts in social values and changing political and economic circumstances. Some stakeholders concerned with residential fish and recreation feel that they lack a productive forum for advocating their interests related to project operations.

Incorporating Changing Social Values into Operations

Lesson: For future projects, periodic, planned re-evaluations can provide a mechanism for incorporating temporal changes in social values into project operations. To meet social policy objectives, it might be necessary to reduce uncertainties for stakeholders whose decisions would be influenced by results from re-evaluations.

Evidence: Support for the social goal of having small family farms located in the semi-arid Columbia Plateau has faded, but long-term contracts with subsidised prices for irrigation water persist. Support for maintaining wild salmon and steelhead in the upper Columbia River is much stronger today than it was when the project was planned.

Incorporating Changes in Science and Technology into Operations

Lesson: For future projects, periodic, planned re-evaluations provide a mechanism for incorporating changes in science and technology into project operations. To meet social policy objectives, it might be necessary to reduce uncertainties for stakeholders whose decisions would be influenced by the results of re-evaluations.

Evidence: Biologists' views on native vs. hatchery fish have changed. Changes in farm technology have increased pressures for larger farms than anticipated by CBP planners.

Sensitivity Analysis of Economic Parameters

Lesson: Substantial inflation-corrected cost overruns in the GCD and the CBP reflect the uncertainties that surround large construction projects. These uncertainties underscore the need for wide-ranging sensitivity analyses to ensure that project goals and objectives are robust and can be met with available resources. Implicit or indirect subsidies need to be evaluated under alternative market conditions to be sure that the subsidies are in line with a project's social objectives.

Evidence: Inflation-corrected Third Powerplant costs were approximately 55% above planned costs. CPB costs were nearly three times those projected with the result that repayments by beneficiaries is roughly 15% of construction costs rather than the planned 50%. Indirect energy subsidies of the CPB have increased over time as the value of firm power to BPA non-agricultural customers has increased.

Developing a Shared Conceptual Framework for Project Appraisal

Lesson: Stakeholders and planners involved in an open planning process need to work with a common conceptual framework and vocabulary in making formal project appraisals. Of particular importance is the distinction between private and social (economy-wide) perspectives. Failure to develop a shared conceptual framework and vocabulary can lead to unnecessary acrimony.

Evidence: Interviews with and letters from stakeholders indicate that numerous disagreements and misunderstandings resulted because of the absence of a shared framework and vocabulary for appraising projects. Particular sources of difficulty include the distinctions between financial and economic prices and between direct and indirect (secondary) benefits.

Mechanisms for Ensuring Just Compensation

Lesson: In large water resources projects, those who bear the costs may not receive many benefits. Therefore, mechanisms for ensuring just compensation are important. In a project that has impacts that cross international borders, the usual forums for allowing parties to make compensation claims - typically the judicial system in the US - may not be satisfactory, and alternative forums should be considered. Alternative dispute resolution mechanisms may also be able to speed up the settlements of claims normally brought using the court system.

Evidence: In the case of GCD and CBP, there were not adequate opportunities for Native Americans and Canadian First Nations to obtain appropriate compensation for project-related losses in a timely fashion

Limits to Government Planning in a Market Economy

Lesson: Limits exist on the extent to which government plans can be implemented effectively in a market-driven, capitalistic economy.

Evidence: Changes in the economics and technology of farming provide irrigators with incentives to circumvent Reclamation's acreage limitations.

Centralised vs. Decentralised Basin Management Institutions

Lesson: In designing institutions for river basin management, centralisation and decentralisation each have their advantages and disadvantages.

Evidence: Co-ordinated decision-making process by four agencies - BPA, the Corps, Reclamation, and BC Hydro - has been effective in managing flood control and hydropower, but some stakeholders feel left out of the decision-making processes for operations. Decentralised decision-making can mean responsiveness to particular constituencies, but inter-agency co-ordination difficulties in the Columbia River Basin have led to problems in this regard.

Actions Having Significant Irreversible Effects

Lesson: Decisions that introduce significant irreversible effects should only be taken after very careful study and broad input from those affected at the grassroots level.

Evidence: Building GCD without fish passage facilities was, for all practical purposes, virtually irreversible; the decision was made without significant study and participation by all affected parties.

Cumulative Impact Assessment

Lesson: Tools for cumulative impact assessment need to be applied to avoid resource management problems.

Evidence: Failure in the past to account for cumulative impacts of dams is at the heart of many fisheries-related controversies within the Columbia River Basin today. The lack of a cumulative impact assessment for the series of major dams on the Columbia and Snake rivers constituted a failure to recognise a major fisheries management problem before it occurred.

9. Reflections on the Development Effectiveness of GCD and CBP

Criteria for Gauging Effectiveness

An assessment of development effectiveness cannot be undertaken without first delineating the criteria to be used in judging effectiveness. In the context of GCD and CBP, applicable criteria can be categorised as follows: economic efficiency, income redistribution, regional economic development, and environmental quality. Inevitably, significant differences of opinion arise about the value of GCD and CBP because the relative weight attached to different criteria for judging effectiveness varies across individuals and groups. Finding a consensus about weights is complicated further because weights frequently vary over time as a result of changes in cultural values, social norms, and economic conditions. A mechanical calculation cannot be used to determine weights or to combine weighted objectives and make choices; a political process is employed to accomplish these ends.

At a conceptual level, the above-noted factors used to characterise development effectiveness can be defined as follows. "Economic efficiency" refers to the condition in which the difference between the present value of economic benefits of a project and the present value of economic costs is as large as possible. (Economic benefits are not the same as monetary benefits except in the case where markets are reasonably competitive.) "Regional development" refers to the objective of fostering growth in a particular geographic area. "Equity" refers to the fair distribution of a project's economic, environmental, and social effects (both positive and negative) among stakeholders. The lack of widespread agreement on what constitutes a fair outcome makes it difficult to apply this criterion. Project analysts do not have formal calculation procedures for gauging whether project outcomes are distributed fairly; a helpful analysis is one that makes the distribution of project gains and losses clear to decision-makers. "Environmental quality" is a broad category that includes a project's effects on the biological and physical environment as well as effects on social conditions and cultural resources. It is possible to conceive of additional objectives of water resources development projects, such as the maintenance of national food security, but the four categories of factors above are appropriate for characterising development effectiveness in the context of GCD and CBP.

Temporal Shifts in Weights Ascribed to Different Effectiveness Criteria

Each of the above-noted objectives plays a role in water resources planning, but the roles of economic efficiency, regional development, equity, and environmental quality in contemporary US water resources planning are different from those roles at the time GCD and CBP were being planned. In examining the development effectiveness of GCD and CBP, it is useful to review how criteria for gauging the effectiveness of US water projects have changed over the past several decades.

Both the Butler and Reclamation reports were focused on regional development and neither was concerned with economic efficiency in the sense of maximising net economic benefits. When the authors of those reports talked of the economic feasibility of GCD and CBP, they had in mind whether or not the beneficiaries of the project could, collectively, pay for the project's monetary costs. In the view of the US Army Chief of Engineers writing in 1932, "the irrigation of land as pertains to the Columbia River area under consideration is not an economical proposition at this time and should await the future" (USACE, 1933: 4).

Writing in 1932, the Commissioner of Reclamation, Elwood Mead, took exception to the US Army Chief of Engineers. Mead argued that GCD and CBP "will enable the largest single water supply of the arid region to be used to give cheap power to industries, and make feasible the irrigation of the largest and finest body of unreclaimed land left in the arid region" (USACE, 1933: 5). Both the Butler and Reclamation reports recognised that the goal of bringing irrigated agriculture to the Columbia Plateau could not be satisfied without using hydroelectric power revenues to cover a sizeable portion of the investment costs for irrigation.

Because an economic efficiency objective (ie, the condition that economic benefits exceed costs) for water resources projects developed by Reclamation and the Corps did not come about until the late 1930s and early 1940s, this objective had little formal influence on the planning of GCD and CBP. However, concerns about what would now be termed economic efficiency were raised in the context of GCD and CBP. For example, the US Secretary of Agriculture and the Chief of the US Army Corps of Engineers both used economic efficiency arguments to support their opposition to the project. Both were concerned with what they perceived as the absence of sufficient demand for agricultural outputs. However, these critics of the economics of the project did not carry the day. Political factors - including extensive lobbying by local project supporters and Franklin Delano Roosevelt's strategy for using water projects to increase employment and his desire to honour political commitments to the US Northwest - played the key roles in the decision to proceed with GCD.

The overriding significance of regional development as an objective of GCD and CBP has continued, and it is reflected in contemporary project assessments by some stakeholders. For example, the consensus of the 12 individuals we interviewed representing irrigators, PUDs, and local governments in the CBP area was that the net positive impacts of GCD and CBP for the region far outweighed the costs to Native Americans. Such regional development arguments frequently ignore the subtleties involved in making arguments related to economic efficiency. Indeed, some of those who trumpet the economic significance of the project do not recognise either the failure to pay interest on the capital cost of irrigation or the lost power revenues associated with providing below-market price energy to pump irrigation water as signs of economic inefficiency.

The other components of development effectiveness - what we have termed equity and environmental quality - did not become major elements of US water resources planning until the late1960s and early 1970s. This is reflected in passage in 1969 of NEPA, which intended to force all agencies of the federal government to integrate environmental and social concerns into their planning and decision-making. This Act introduced requirements for environmental impact statements, documents that were to disclose fully the environmental and social impacts of proposed federal projects and alternatives to those projects. Implementation rules issued subsequently by the US Council of Environmental Quality required that EISs highlight any irreversible and irretrievable commitments of resources which would be involved in a proposed action should it be implemented.

The period from the1960s through the 1980s witnessed an upsurge in use of the US judicial system by citizens making claims of unjust treatment in cases involving environmental impacts. US Supreme Court decisions in the 1960s and 1970s made it possible (for the first time) for environmental groups to bring suit in US courts in instances where adverse environmental impacts were significant but plaintiffs had not suffered direct monetary damages. Since the late 1960s, citizens and non-governmental organisations have used the courts to press claims centring on the adverse environmental effects of development projects or alleged instances of "environmental injustice" (ie, the inequitable distribution of environmental costs and gains).

GCD and CBP: Trade-offs Between Regional Development and Economic Efficiency

If judged only in terms of regional development goals (without regard to environmental and social impacts), GCD and CBP would be judged as developmentally effective by many people. Indeed, this is clear from our interviews with irrigators and users of the relatively inexpensive hydroelectric power in the US Northwest who feel that GCD and CBP is a great success. They point to the contributions of low-cost electricity to: 1) the booming economy of the US Northwest, 2) the contributions of agriculture to the regional economy and food supply, 3) the numerous monetary benefits associated with recreation and flood control, and 4) the way storage at GCD creates opportunities to generate additional power at downstream dams.

Circumstances exist under which projects advocated from a regional perspective are perfectly compatible with the efficient use of resources as seen from a national point of view. However, when regional development becomes a principal policy objective, governments frequently intervene (directly or indirectly) in markets for inputs to and outputs from production processes. With the exception of instances where such interventions are a response to incomplete or imperfect markets, the distorting policy interventions needed to implement a regional development objective lead to an inefficient allocation of national resources.

The attitudes of regional stakeholders toward such government interventions depend, of course, upon whether the interventions constitute a tax or a subsidy. The formation of these attitudes is complicated by the fact that many taxes and subsidies are implicit or indirect. In our interviews with people affected by GCD and CBP, for example, interviewees acknowledged that direct payments, such as those used to cover a large part of the cost of CBP's construction, were a subsidy. They often failed to see, however, that market interventions that indirectly enhance an industry's financial competitive position are subsidies.

In the case of GCD and CBP, the criterion for efficient energy use has been met, to some degree, by the energy component of the project. Environmental and social issues aside, energy production by GCD, including its associated Third Powerplant, has contributed greatly to the development of the US Northwest. Although below-market energy pricing has reduced revenues to the government from what they might have been, the project has repaid the US Treasury for the cost of constructing hydroelectric power facilities at GCD, and the portion of the project related to hydropower clearly has a positive economic benefit-cost ratio.

Supporters of agricultural development in CBP, past and present, have argued that development of irrigation facilities should also be seen as positive from a national perspective. Although the original planners acknowledged that agricultural development could not pay for itself, those planners maintained that substantial national benefits would accrue from the regional development in the form of secondary benefits. This claim was controversial when it was made in the 1930s, and secondary benefits are now inadmissible in evaluating the economic efficiency of federal water projects.

When GCD and CBP are analysed separately, it is apparent that the magnitude of the direct government subsidy involved in the construction of CBP has been substantial. Irrigators will pay only about 15% of this cost in dollars uncorrected for inflation. The US Treasury has paid the remaining cost. Eventually, this remainder will be repaid (in dollars uncorrected for inflation) by BPA using revenues power sales.

But direct construction subsidies to the CBP covered by power sales have not been the only drain on funds that would otherwise have remained with the US Treasury. Indirect subsidies have been provided to a variety of regional power users (eg, DSIs and PUDs) by contractual and regulatory arrangements that have delivered energy at below-market costs. Also, energy used for pumping water has been provided to the CBP at roughly the cost of production at the GCD, again well below the market price.

Interventions that distort prices also create second order effects as inefficiencies become embedded in private decisions. For example, BPA has been criticised for setting prices of electricity at below-market levels for some consumers (as BPA must as a matter of federal law) because these below-market costs failed to encourage parsimonious use of electricity. Experiences from other areas (eg, California in the drought years of the mid-1980s) indicate that there is substantial room for better water management when water becomes truly scarce. When subsidies keep the cost of irrigation water low, farmers are less likely to give attention to a more water-efficient agriculture.

The regional development objectives of GCD and CBP have, to a considerable extent, been achieved. But they have come at a substantial cost to the rest of the economy, both in terms of direct construction subsidies and in revenues foregone from indirect subsidies in the form of below-market energy prices.

GCD and CBP: Trade-offs BetweenRegional Development, Equity, and the Environment

Trade-offs also exist between regional development and objectives related to equity and the environment. This is clearly shown by the way GCD affected indigenous peoples in the upper Columbia River Basin. In the view of many Native Americans and members of First Nations in Canada, GCD was nothing short of catastrophic. For them, the project had a disastrous affect on the continuance of their culture. In our interviews, Native Americans and members of First Nations highlighted the importance of Kettle Falls and other fishing areas as places where different tribes came together to enhance mutual understanding, share language and stories, and continue rituals and other traditions. These opportunities were lost after GCD blocked the runs of salmon to the upper Columbia River.

There is no calculation procedure that allows a balancing of these negative social impacts and cultural losses against the substantial regional development benefits that the US Northwest has enjoyed as a result of GCD and CBP. Today, US citizens rely on an open planning process tied to NEPA to help decision-makers become aware of trade-offs: how much of one objective, such as the quality of the environment, must be sacrificed when attempting to augment another, such as regional development. However, nothing equivalent to NEPA existed at the time that President Franklin Delano Roosevelt and his administrators decided to proceed with construction of GCD. Moreover, even the open planning prescribed by NEPA has limitations. For example the NEPA process does not necessarily address the consequences of unequal power among stakeholders, a problem that still plagues the anadromous fish recovery and recreational jurisdiction issues associated with GCD and CBP.

Effects on Native Americans and First Nations are not the only ones that would be questioned in gauging the development effectiveness of GCD and CBP using contemporary criteria. The effects on salmon and steelhead upstream of the dam would also be called into question, primarily because the placement of GCD constituted an irreversible commitment to eliminate anadromous fish in the hundreds of miles of habitat upstream of the dam. The fish mitigation programme implemented in the late 1930s and early 1940s also had an irreversible impact: the hatchery and transplantation elements of GCFMP constituted the irreversible elimination of the wild stocks of salmon and steelhead in the mid-Columbia River tributaries.

The observations in this section about the history of decision-making criteria used by federal resources development agencies during the 20th-century in America help put the discussion by stakeholders of the draft version of our report in perspective. After nearly 60 years of project operations, those who have benefited from GCD and CBP have, quite naturally, become focused on maintaining the advantages they have enjoyed as a result of the project - mainly low-cost irrigation water, low-cost electricity, and benefits from flood control and recreation. At the same time, groups that were disadvantaged by the project (ie, Native Americans and First Nations), are continuing their struggles to obtain compensation for what they perceive as broken promises and grave injustices of the past). It is possible that individuals who gain or lose from future water resources projects will be just as tenacious in defending their gains or seeking compensation for their losses many years after basic project decisions have been made.