A number of salmon stocks in the Columbia River and its tributaries are threatened by regional economic activities. Researchers at Resources for the Future have been assisting regional planners to evaluate proposed combinations of actions to increase the numbers of fish in these stocks. The legislation enacting the region's fish recovery program assigns essentially infinite value to the recovery of fish stocks, so the researchers developed a cost-effectiveness analysis that illustrates the trade-offs between the economic costs and biological effectiveness of recovery alternatives. Through this analysis, they are attempting to identify unique combinations of actions, taken throughout the Columbia River basin, that can efficiently meet various biological objectives.
Encompassing nearly a quarter-million square miles and spread over parts of seven states, the U.S. portion of the Columbia River basin is now the site of one of the country's most complex environmental challenges—namely, how the states in the basin can continue to make economic use of the Columbia River system and still protect the health of the basin's ecosystem. Protection of one of the integral components of the ecosystem—fish—has been of particular concern in recent years. While human exploitation of the resources of the Columbia River and its tributaries for hydroelectric-power production, recreation, irrigation, and navigation has mushroomed in the last fifty years, fish populations have declined precipitously. In the last year, one salmon stock has been declared endangered and two other stocks have been declared threatened under the federal Endangered Species Act. (Salmon stocks are identified as salmon that spawn in a particular subbasin or portion of the river system at a particular season and that generally do not interbreed with salmon in other locations or with salmon in the same location in other seasons.) The American Fisheries Society has identified many additional fish stocks in the Columbia basin that it believes are at risk of extinction.
Scientists estimate that the annual production of adult salmon in the basin has dropped by about 80 percent over the last 150 years. However, some actions to mitigate declines in the basin's fish populations and to restore these populations threaten to curtail or increase the cost of the hydroelectric-power production, navigation, irrigation, and recreation provided by the Columbia River. By the fall of 1993, for example, utilities that purchase hydro-electric power generated by the dams on the river may face an increase of 5 to 10 percent in wholesale electricity rates as a result of the mitigation and recovery actions that the region has decided to implement.
Problems for fish
A number of activities have brought about the decline in fish populations in the Columbia River basin. Logging, mining, and grazing have hindered the reproduction of anadromous fish (fish that ascend rivers from the sea for breeding) by promoting the erosion of soil, which then settles on the gravel streambeds, and by altering the topography and degrading the water quality of many spawning and rearing areas. Irrigation for agricultural production has also impeded reproduction by necessitating diversion dams that block access to these areas, intake pipes that draw fish into irrigation canals and onto fields where they are left stranded, and water withdrawals that have dried up some streams. Commercial and sport fishing kill more than 80 percent of the adult fish of some stocks.
Many smolts are killed by the turbines of downstream dams, and the slower water flow in the pools above the dams increases the time it takes smolts to migrate downstream, increasing their susceptibility to predation.
Production of hydroelectric power has also played a large role in the decline of fish populations. The extensive system of dams on the Columbia and Snake rivers and their major tributaries provide inexpensive electric power and flood protection to the region, but the dams also interfere with the migration and the reproduction of anadromous fish such as Pacific salmon and steelhead. These fish must make two migrations during their lifetime: one migration downstream from tributary streams as juveniles and one migration upstream from the ocean as adults. Unfortunately, hydroelectric-power dams are obstacles to both downstream and upstream migration.
Chief Joseph Dam on the Columbia River and Hells Canyon Dam on the Snake River completely block upstream migration, since adult salmon traveling back to tributary streams to spawn cannot get over or around them. Absent construction of immensely costly fish ladders, areas upstream of these dams will remain inaccessible to the salmon for the foreseeable future. Other, smaller dams on these rivers have fish ladders that allow upstream migration; although the majority of the adult salmon successfully traverse the ladders, the mortality rate at each of the dams is 5 to 15 percent.
Salmon migrating downstream to the ocean as juveniles (smolts) face other dangers. Mortality rates can be high as smolts pass through the turbines of as many as nine downstream dams. Under poor conditions at a single dam, up to 30 percent of the smolts that pass through the dam can die. The fish that successfully avoid the turbines are slowed in their migration to the ocean by the slack water in the pools above the dams. For example, a trip from Idaho or northeastern Washington that took smolts 7 to 14 days before the construction of the dams now averages 20 to 30 days. The increase in travel time makes the smolts more susceptible to predation and may prevent them from reaching the ocean during the time frame in which they can make the physiological transition from fresh water to salt water.
Ironically, hatchery programs, which were intended as part of the solution to dwindling fish runs, may be part of the problem. Many researchers believe that hatchery programs have several pernicious effects on natural stocks. For instance, the programs may foster overly high harvest rates in mixed-stock fisheries, where natural and hatchery-bred fish are indistinguishable. They may also promote genetic mixing of hatchery-bred fish and wild fish, thereby threatening the long-term viability of wild fish populations by reducing the chances that the desirable genetic traits of wild fish will continue to be passed on from generation to generation. In addition, hatchery programs may force wild fish and hatchery-bred fish to compete with each other for food and other resources and increase the possibility that diseases of hatchery-bred fish will be transmitted to naturally spawning fish.
Efforts to increase fish populations in the Northwest
Since the passage of the Pacific Northwest Electric Power Planning and Conservation Act (often referred to as the Northwest Power Act) in 1980, efforts to increase the number of salmon have intensified. The Northwest Power Planning Council, which was established by the act, has designed and adopted a fish and wildlife program that contains a variety of actions to accomplish this goal. These include passage actions, which facilitate migration through the mainstem Columbia and Snake rivers; harvest actions, which reduce the number of fish that can be caught in the ocean, mainstem rivers, and tributaries; and propagation actions, which mitigate degradation of fish habitats in subbasins or increase the number of fish through hatchery programs.
Passage actions facilitate migration; harvest actions reduce the number of fish that can be caught; and propagation actions improve fish habitats and increase numbers of fish through hatchery programs.
Unlike harvest and propagation actions, passage actions can affect all fish stocks. Individual passage actions are designed to accomplish one of four objectives: (1) to guide smolts around powerhouse turbines at major mainstem dams, (2) to move smolts downstream more rapidly, (3) to reduce their susceptibility to predation while migrating downstream, or (4) to facilitate the upstream migration of adults. In the Columbia River basin, bypass facilities are being installed to accomplish the first objective. To accomplish the second objective, smolts are being transported downstream in barges, and water velocities are being increased by raising the volume of water flows or lowering the elevation of reservoirs. To accomplish the third objective, programs to reduce the population of squawfish and other fish that prey on salmon have been implemented. To accomplish the fourth objective, fish ladders have been built at many dams.
Harvest actions can be an effective but politically charged method for increasing fish runs. Fisheries managers can regulate harvests to some degree by adjusting the timing and location of harvests, changing fishery quotas, and controlling which stocks may be caught by commercial and sport fishers. There has been some discussion about creating an incentive to reduce fish harvests in the Columbia River through a program to purchase the fishing licenses of harvesters. Unfortunately, harvest actions are not well integrated with passage or propagation actions, in part because the administration of harvest management across the Columbia River basin is highly fragmented.
The majority of actions already implemented and proposed to increase the numbers of fish are propagation actions. Widely accepted practices for mitigating habitat problems in the Columbia River subbasins include the removal of barriers to migration, the improvement of stream habitat, and the screening of irrigation-canal intakes. A more controversial propagation action is the breeding of fish in hatcheries, because of the problems that potentially arise when hatchery-bred fish are mixed with naturally spawning fish. Recently proposed hatchery programs take much greater account of these problems.
In general, passage actions are the most expensive actions to implement. For example, due to the decrease in hydropower production that it would entail, the proposal by the Northwest Power Planning Council to increase water velocities by increasing water flows would cost in excess of $70 million per year. Extensive predator control also can be quite costly, although it is estimated to be nearly one order of magnitude less expensive than some water-flow options. The costs of commercial fish-harvest reductions in the ocean and rivers are unknown since large-scale reductions that include compensation for commercial fishers have not taken place. The overwhelming majority of propagation actions require expenditures of $5,000 to $250,000 per year.
Assessing the cost-effectiveness of possible strategies
For the past seven years, researchers at Resources for the Future (RFF) have been assisting the Bonneville Power Administration and regional planners to evaluate the trade-offs among proposed combinations of actions to increase fish populations in the Columbia River basin. (These combinations of actions are referred to as strategies. There are three types of strategies: passage strategies, each of which consists of a different combination of passage actions; harvest strategies, consisting of a different combination of harvest actions; and propagation strategies, consisting of a different combination of propagation actions.) Their approach has been to assess the cost-effectiveness of each recovery alternative, which is a combination of a passage strategy, a propagation strategy, and a particular harvest rate. In contrast to cost-benefit analysis, which attempts to compare the economic value of benefits with the economic cost of attaining them, cost-effectiveness analysis attempts to reveal the least-cost way of achieving prescribed objectives, thereby avoiding the economic evaluation of benefits. Such analysis is consistent with the Northwest Power Act—the legislative backbone of the region's fish recovery program—which explicitly states that if there are equally effective means of achieving the same biologically sound objective, the alternative that costs the least should be implemented. This statement suggests that infinite value should be assigned to the recovery of fish stocks and that, once biological objectives have been articulated, the major question is: which recovery alternative will meet objectives at least cost?
Cost-effectiveness frontier of hypothetical fish recovery alternatives
Other concerns besides cost and effectiveness can be allowed to constrain the types of actions to be considered in a cost-effectiveness analysis.
Obviously, other concerns besides cost and effectiveness play important roles in choosing among alternatives. For example, the genetic mixing of hatchery-bred stocks and wild stocks allowed by some hatchery programs may make these programs unacceptable even though they might be superior to other fish-recovery programs in terms of cost and short-term effectiveness. However, genetic mixing and other concerns can be allowed to constrain the types of actions to be considered in a cost-effectiveness analysis or to shape the objectives that are sought prior to this analysis. Thus it is possible to explore trade-offs between the costs and the effectiveness of actions to achieve less generic objectives (such as minimizing the risk of genetic mixing, promoting natural flow conditions in rivers, and rebuilding critically low stocks) as well as more generic objectives (such as providing specific numbers of fish to be harvested and to remain unharvested).
Examining the basinwide effects of recovery alternatives
To date, RFF researchers have used the cost-effectiveness framework to analyze millions of alternatives for restoring more than 100 naturally spawning and hatchery-bred fish stocks in the Columbia River basin above Bonneville Dam, which is located on the Columbia River 40 miles upsiream from Portland, Oregon. The enormous number of alternatives is the result of the number of stocks involved, the number of harvest rates in tributaries that might be desired, and the fact that each passage and propagation action can be implemented in conjunction with other actions. Thus there are many combinations of actions, target stocks, and harvest rates to be taken into account.
Taking a broad perspective in analyzing the propagation, harvest, and passage actions that could be implemented across the Columbia River basin is the only way to promote the maximum effectiveness of these actions at the lowest cost. Absent analysis of how a passage action would affect the entire basin, there is no way to know whether a decision to maximize the survival of one stock would jeopardize the survival of other stocks. Moreover, it is only by examining each action in the context of other actions that analysts can determine which combinations of strategies will meet objectives and will do so in the most cost-effective manner. A propagation strategy that appears cost-effective in conjunction with one passage strategy may not be cost-effective in conjunction with another passage strategy. For example, a propagation strategy that involves a large number of relatively expensive propagation actions might be undertaken if it is assumed that the passage strategy will not greatly enhance the survival of stocks. If it is assumed that another passage strategy will be more effective in enhancing survival, a different, less expensive propagation strategy might be implemented.
The results of analysis
The RFF analysis of passage, propagation, and harvest-rate alternatives in the Columbia River basin is based on the integration of information derived from two sources: computer modeling of the life cycle of Pacific salmon and steelhead populations in the basin and a database developed primarily from fish restoration plans produced by the Northwest Power Planning Council and the Columbia Basin Fish and Wildlife Authority for each major subbasin of the basin. The goal of the analysis is to inform decision makers about which combinations of passage and propagation strategies and harvest rates would be most cost-effective and to give them an appreciation of which assumptions are critical to the results and conclusions of the analysis.
Although measures to increase water flows do not appear on the cost-effectiveness frontier, they are popular because their costs are evenly distributed among people who use the resources of the Columbia River.
In their efforts to find least-cost basinwide alternatives, RFF researchers examined eight passage strategies and more than 2,000 propagation strategies. They attempted to identify combinations of these strategies that would meet the objectives set by regional planners for both terminal harvest (the number of fish in each stock that would be available for harvesting in tributary rivers) and spawning escapement (the number of naturally spawning fish in each stock that would be left after commercial, sport, and ceremonial tribal harvests). The researchers also attempted to address the concern that some alternatives might impose unacceptable risks by promoting the genetic mixing of hatchery-bred stocks and naturally spawning stocks. To do this, they eliminated from consideration any hatchery action that would introduce hatchery-bred fish into a wild population. The elimination of these actions did not change the basinwide alternative that the researchers identified as meeting the planners' objectives at least cost as long as three salmon stocks were not considered in the analysis. They found that no combination of genetically acceptable actions would achieve the terminal harvest and spawning escapement objectives for these stocks.
The combination of strategies recommended by the RFF analysis to meet the planners' objectives is, by itself, not very intriguing. Because these objectives are to some extent arbitrary and could be subject to change, a more interesting and useful output of the analysis is the cost-effectiveness frontier that emerged when the original terminal-harvest and spawning-escapement objectives for each stock were lowered by as much as 50 percent. This frontier shows the way in which total costs increase as the number of fish specified for terminal harvest and spawning escapement for each stock rises from 50 to 100 percent of the number originally specified by planners for each stock. These costs rise sharply at two points (see figure, p. 15). The sharpest increase—$8 million per year—occurs when the number of fish for terminal harvest and spawning escapement for each stock increases from 60 percent to 65 percent of the number originally specified. This increase is due to the introduction of an expensive passage action aimed at predator control. Another sharp increase—$3 million per year—occurs when the number of fish for terminal harvest and spawning escapement for each stock increases from 85 percent to 90 percent of the number originally specified. In this case, the increase is due to a large increase in propagation costs.
The shape of the cost-effectiveness frontier is somewhat contrary to expectations. The marginal costs incurred in achieving each 5-percent increase in the number of fish above the 65-percent level are not always higher than the marginal costs incurred in achieving each 5-percent increase in the number of fish from the 50-percent to the 65-percent level. This oddity results in part from the modeling methods used in the RFF analysis, but it also reflects the fact that an expensive predator-control action is needed to ensure the survival of at least 65 percent of the number of fish originally specified for terminal harvest and spawning escapement for each stock. Once this action is implemented, no additional passage actions (beyond those already planned before the recent changes in the Northwest Power Planning Council's fish and wildlife program) are required. Instead, propagation actions, which are less costly than passage actions, can drive further increases in the numbers of fish.
Cost-effectiveness frontier of alternatives for recovering salmon and steelhead stocks in the Columbia River basin
One revelation of the RFF cost-effectiveness analysis is of particular interest: the Northwest Power Planning Council's proposed water-flow measure, which calls for water flows to be increased, never appears on the cost-effectiveness frontier. Instead, all the alternatives on the frontier include actions that maintain current (1989–1991) water flows. In facilitating the downstream migration of smolts, these actions are considerably less costly than and nearly as effective as the council's water-flow measure. Actions to maintain current water flows appear cost-effective even when the beneficial effects of non-flow passage actions are assumed to be much lower. For example, when estimates of the effectiveness of both proposed predator-control and existing smolt-transport programs are reduced by 50 percent, current water flows are still the most cost-effective way to achieve the planners' original terminal-harvest and spawning-escapement objectives. However, to compensate for the decrease in the assumed effectiveness of these programs, additional propagation actions have to be implemented, resulting in a 40-percent increase in the total costs of these actions.
Because the data on the effectiveness of all water-flow measures are both limited and open to a wide range of interpretations, there is a need to investigate further the sensitivity of the results of the RFF analysis to changes in assumptions about the relationship between water flows and the survival rate of smolts. However, even with more sensitivity testing, the debate over the merits of each alternative flow measure is likely to continue. Flow measures, such as the one proposed by the Northwest Power Planning Council, enjoy a high level of support among key interest groups in the region, in part because the costs of these measures are relatively evenly distributed among all the peo-ple who use the resources of the Columbia River.
The importance of objectives
A cost-effectiveness analysis provides information about how to achieve a set of objectives at least cost. It is important to remember that the objectives are almost never set by the scientists conducting the analysis but emerge from some administrative or social process. Thus the Pacific Northwest region must clearly articulate specific objectives concerning the fish populations of the Columbia River basin.
As clearly shown in the cost-effectiveness analysis conducted by the RFF researchers, the least-cost alternative can change if the number of fish specified in an objective changes. It can also change if the type of objective changes. For instance, doubling the size of the run of each stock and doubling the aggregate run size of all stocks are two distinct objectives, and the accomplishment of each may require the implementation of different fish-recovery alternatives. Certain objectives may even eliminate some alternatives without regard to their cost or effectiveness. For example, if the objective is to increase the number of fish in wild stocks but to do so without genetic mixing of hatchery-bred fish and wild fish, all propagation strategies that rely on hatchery operations would be rejected. The point is that without clear articulation of possible objectives by the policymaking community, analysts are often hard-pressed to provide relevant information to policymakers interested in exploring the trade-offs between the cost and the effectiveness of actions to achieve those objectives.
However, cost-effectiveness analysis is valuable even when the policymaking community is unable to articulate clear objectives. The RFF researchers' cost-effectiveness analysis of alternatives for restoring stocks of anadromous fish in the Columbia River basin demonstrates that such analysis can provide a framework for articulating objectives as well as for developing information about the costs and effectiveness of alternatives to meet these objectives. Cost-effectiveness analysis can also illustrate the trade-offs among possible objectives and alternatives for achieving them. In addition, it can allow decision makers to explore how economic and biological uncertainties can influence the choice of alternatives. Thus, in the context of regionwide decision making, cost-effectiveness analysis is wasted if it is used only to identify a single least-cost alternative to meet an objective.
Kris Wernstedt, Jeffrey B. Hyman, and Charles M. Paulsen are fellows in the Quality of the Environment Division at RFF. This article is based on ongoing research conducted by the authors and senior fellows Allen V. Kneese and Walter 0. Spofford, Jr.
A version of this article appeared in print in the October 1992 issue of Resources magazine.
A version of this article appeared in print in the October 1992 issue of Resources magazine.
