Executive summary
From 1933 to 1975, development of the hydroelectric system on the mainstem Columbia and Snake rivers involved construction of a series of major dams. Of these dams, Grand Coulee Dam blocked the Canadian portion of the mainstem, and Hells Canyon Dam blocked the upper Snake River to migrations of adult salmon (Figure 1). In the portions of the mainstem and Snake that are not blocked, there are 13 dams with fish ladders that provide passage for adult salmon. These include four on the lower mainstem Columbia reach, five in the mid-Columbia reach, and four in the Snake River. 70 dams located on tributaries in the basin are also part of the coordinated hydroelectric system. Some of these, such as those on the Cowlitz River, are not passable by salmon, and others, such as those above Hells Canyon Dam, lie above impassable dams. Juvenile salmon migrating downstream past those projects with fish passage facilities may use several routes. They either pass through the turbines, spillways, turbine intake bypass systems, navigation locks or ice and trash sluiceways. A few juvenile salmon may pass by way of fish ladders designed for adult passage, but these are not designed, located or operated in ways that will attract juveniles. Full development of the hydroelectric system included provisions for storage of spring runoff by dams in the upper tributaries. These storage operations distributed flow over the year and reduced the volume of spill available for passage of juvenile salmon in the spring, forcing more juveniles to pass through the turbines.
The first measurement of mortality of juvenile salmon in passing through turbines came in the early 1950's by Harlan Holmes. He recorded recovery of marked adult salmon that had been marked as juveniles and released in two groups, one that passed through the turbines and the other that was released below the turbines. Numerous studies since then have recorded measurements of mortality between 2.3 and 19 percent at various projects, averaging about 11 percent overall. With this rate of loss in passing through turbines, fewer than half of the fish migrating downstream from above the uppermost projects in the Snake River or upper Columbia River would survive to below Bonneville Dam without spill or other passage routes. Mortality in spill averages less than 2 percent at projects where losses due to concentrations of predators below the spillway are not a factor.
Numerous methods have been investigated for their potential in diverting the juvenile salmon away from the turbine intakes and into a safe passage route. We summarize those efforts in the text. They include efforts to improve effectiveness of spill, barrier nets, fish "gulpers", salvage from gatewells, electric fields, sound, lights, louvers, ice and trash sluiceways, turbine intake screens, and surface collectors. Of these mechanical devices, only turbine intake screens and surface collectors have proven effective enough to justify full installation.
Spill is effective as an interim measure or as a supplement to a mechanical bypass. Effectiveness of spill is measured in terms of the percentage of fish approaching the dam that is diverted into spill. Spill effectiveness varies from project to project. Application of spill is limited by water quality standards that limit the amount of spill because of concern about production by spill of high gas saturation levels that can kill fish. Several studies suggest that the effectiveness of spill in passing juvenile salmon can be improved. The standard spill gates at projects in the basin are fitted with tainter gates that open from below, usually at a depth of around 50 feet below the surface. Provision of spill from the surface increases the effectiveness of a given volume of water in passing juvenile salmon by a factor of two or more. It has also been learned that spreading spill of a given total volume of water (in acre feet) over a 24 hour period passes more than twice as many fish as the same number of acre feet of water spilled over a 12 hour period.
Tests of prototype turbine intake diversion devices have led to the development, construction and operation of bypass systems at all of the projects on the mainstem Columbia River and in the Snake River except The Dalles, Rocky Reach, Rock Island, Wanapum and Priest Rapids dams. In Appendix A, we document the requirements of the Federal Energy Regulatory Commission (FERC), the Northwest Power Planning Council (NPPC), and the National Marine Fisheries Service (NMFS/NOAA) for installation of bypass systems at each project. We also provide information on the status of those bypass facilities and future installation schedules at each of the 13 projects. Future installation is scheduled at The Dalles, Wanapum and Priest Rapids Dams, and is under study at Rocky Reach and Rock Island dams. Meanwhile, The Dalles Dam uses the ice and trash sluiceway to pass at least 40 percent of the juvenile salmon approaching the dam and more when spill is added. Turbine intake screens have been the primary choice by the Corps of Engineers (Corps) at their projects. The Corps has a schedule for replacement of standard length screens with extended-length screens at all eight of their projects, including The Dalles. As an example of the cost of these measures, in 1992 the Corps budgeted $32 million for development and installation of intake diversion screens. This program has been ongoing since the 1960's (Corps, 1992).
Effectiveness of turbine intake screens is measured by fish guidance efficiency (FGE), which is the percentage of fish approaching the turbine intakes that is diverted by screens. Measured fish guidance efficiency differs from project to project and with respect to other factors, which include the design of the screen, the species of salmon, degree of smoltification, time of day, and progress of the season. Extended length screens have achieved higher measured fish guidance efficiency than standard length screens. Fish guidance efficiency of extended length intake screens, although at times reaching values as high as 93 percent for steelhead and coho and 88 percent for chinook yearlings, are for the most part below 50 percent for subyearling chinook and sockeye. Fish that are not guided will pass through the turbines. Fish guidance efficiency appears to have reached an upper limit that is less than the surface collector at Wells Dam. Intake screens are unlikely to prove 100 percent effective in diverting juvenile salmon (Office of Technology Assessment, 1995, p.127). The 1987 Fish and Wildlife Program of the NPPC set a standard of 90 percent fish guidance efficiency for intake screens - "if it can be achieved." It appears that this standard can not be achieved.
Fish that are successfully guided by the screens into the bypass conduits are subject to injury at the screen and within the bypass system. The Fish and Wildlife Program of 1994 specified a criterion of 98 percent smolt survival within bypass and collection systems from the screen to the end of the outfall. This standard appears to be attainable. However, losses due to predation at the outfalls and in the tailraces can be substantial in some situations.
While fish guidance efficiency focuses on measurement of effectiveness of mechanical devices in diverting fish away from turbine intakes, fish passage effectiveness (FPE) is a measurement of the total percentage of fish that pass a project by routes other than the turbines. Passage can be through spill, the ice and trash sluiceway, or through the intake bypass system with the aid of screen diversion. In other words fish passage efficiency focuses on fish that are diverted away from the turbine intakes and into a safer passage route. Both NMFS/NOAA and the NPPC have established goals of 80 percent fish passage which includes all salmon species in both spring and summer.
The most successful bypass system in the basin is at Wells Dam where a surface collector passes an estimated 89 percent of the juvenile salmon in both spring and summer. Thus, it is the one project in the basin with a bypass that can achieve the 80 percent fish passage goal. Feasibility of using surface collectors at other projects is being investigated by the Corps, and by Chelan and Grant County Public Utility Districts.
In Appendix B, we compare the goals for fish passage of FERC, NPPC, and NMFS/NOAA with what was achieved in 1995. Experience in that year serves as an example illustrating that, except at Wells Dam, the NPPC and NMFS/NOAA goals of 80 percent fish passage can not be achieved without the addition of spill, due to limitations of performance of turbine intake screens or ice and trash sluiceways that are present. Furthermore, the amounts of spill required to meet the goals can not be provided due to gas saturation limits required by water quality standards.
To compare the performance of the fish passage measures as required in the NMFS/NOAA Proposed Recovery Plan with what was actually achieved, estimates of fish passage achieved at each of the Snake River and lower Columbia River projects in 1995 were produced by the Fish Passage Center (1995). (See Appendix B.) Under the NMFS/NOAA requirements, highest fish passage, 78 percent, was achieved at The Dalles Dam (Fish Passage Center, 1995). With the exception of Bonneville Dam at 55 to 62 percent, all of the lower river projects achieved fish passages in the 70 percent range. Snake River projects, with the exception of Ice Harbor Dam, achieved fish passages between 50 to 60 percent. Ice Harbor Dam achieved an estimated 79 to 84 percent fish passage, but with excessive spill.
Therefore, the 80 percent fish passage goals of NMFS/NOAA or the NPPC can not be achieved in the river reaches where endangered Snake River fish are present. The fish passage goals were achieved at Ice Harbor Dam in 1995 but only because turbine outages led to inadvertent spill in amounts that caused gas saturation to exceed limits.
With respect to FERC requirements, which apply in the mid-Columbia reach, the fish passage requirements at Wells Dam of 70 percent in spring and 50 percent in summer were exceeded by the 89 percent measured. The FERC requirement of 50 percent fish passage at Priest Rapids Dam in summer was met (62 percent) by provision of spill alone. Fish passage requirements by FERC at Wanapum Dam and at Priest Rapids Dam in summer could not be met because of limitations on spill due to gas saturation limits for water quality. FERC requirements were met at Rocky Reach and Rock Island dams, as they are specified in terms of spill amounts, not fish passage. The NPPC fish passage requirement of 80 percent, which applies as well as FERC requirements in the mid-Columbia reach, was met only at Wells Dam.
Calculation of the amount of spill required to achieve the 80 percent fish passage goal in the Snake River and lower Columbia River is complicated, requiring assumptions that go beyond available data. Decisions on the appropriate fish guidance efficiency to use depend upon predictions of the mix of species and other factors. Spill effectiveness curves are lacking for most of the Corps projects, requiring an assumption of a 1:1 relationship between percentage of flow that is spilled and the percentage of fish passed. This assumption is not met where adequate data are available for analysis, such as at John Day and The Dalles dams.
The calculated spill amounts in the Detailed Fishery Operating Plan depend upon an assumption (or conclusion) that there is an advantage to spilling 12 hours at night versus 24 hours a day as a benefit to power production. We question this assumption. We believe a more detailed analysis of costs and benefits to fish and power would be warranted. In any case, the spill amounts calculated for use by NMFS/NOAA can not be provided in practice due to limitations on gas saturation levels. This in spite of the fact that for the duration of the juvenile migration period water quality standards were expanded to permit 120 percent saturation. In addition, by 1995, five of the eight federal mainstem and Snake River projects were at least partially equipped with flip lip spillway deflectors designed to reduce gas saturation levels. (See Appendix A.) The NMFS/NOAA Proposed Recovery Plan specifies an upper limit of 115 percent gas saturation in the forebays of the projects. This standard can not be met at all during normal spring runoff.
Analysis by the Fish Passage Center indicated that survival of PIT tagged juvenile chinook salmon was not adversely affected by gas saturation levels experienced in 1995, which were 130 percent to 138 percent from May 25 to June 8 at Ice Harbor Dam (Fish Passage Center, 1995). Further studies are needed of juvenile salmon survival in natural river situations where gas saturation levels are high.
Future developments of juvenile bypass are expected to be in surface collectors, ice and trash sluiceways, and surface spill, all of which take advantage of a natural surface orientation of juveniles. Measures are needed to reduce gas saturation levels associated with spill levels that are required to meet fish passage goals.
