NMFS policy (Hard et al. 1992; NMFS 1993) stipulates that in determining 1) whether a population is distinct for purposes of the ESA, and 2) whether an ESA species is threatened or endangered, attention should focus on "natural" fish, which are defined as the progeny of naturally spawning fish (Waples 1991a). This approach directs attention to fish that spend their entire life cycle in natural habitat and is consistent with the mandate of the ESA to conserve threatened and endangered species in their native ecosystems. Implicit in this approach is the recognition that fish hatcheries are not a substitute for natural ecosystems.
Nevertheless, artificial propagation is important to consider in ESA evaluations of anadromous Pacific salmonids for several reasons. First, although natural fish are the focus of ESU determinations, possible effects of artificial propagation on natural populations must also be evaluated. For example, stock transfers might change the genetic bases or phenotypic expression of life-history characteristics in a natural population in such a way that the population might seem either less or more distinctive than it was historically. Artificial propagation can also alter life-history characteristics such as smolt age and migration and spawn timing (e.g., Crawford 1979, NRC 1996). Second, artificial propagation poses a number of risks to natural populations that may affect their risk of extinction or endangerment. These risks are discussed below in the "Assessment of Extinction Risk" section. Finally, if any natural populations are listed under the ESA, then it will be necessary to determine the ESA status of all associated hatchery populations. This latter determination would be made following a proposed listing and is not considered further in this document. The remainder of this section is intended to provide a summary of the nature and scope of artificial propagation activities for west coast chinook salmon and to identify influences of artificial propagation on natural populations.
The focus of the Artificial Propagation section concerns the culture of chinook salmon in individual ESUs. To provide some perspective with respect to the magnitude of propagation efforts along the West Coast, a brief review of chinook salmon culture in areas outside the continental United States will be given here. In addition, we will provide a short review of important events in the history of artificial propagation of chinook salmon in the Columbia River Basin will be presented, as 7 of the 15 chinook salmon ESUs are located in this large river system.
Japan--Although spawning chinook salmon have been observed in Japanese streams (Healey 1991), there appear to have been few, if any, large-scale chinook salmon programs in Japanese hatcheries, although experimental releases of Washington State chinook salmon have occurred (McNeil 1977).
Russia--Spawning populations of chinook salmon are found in large rivers of eastern Russia; however, the overwhelming majority of effort regarding artificial propagation has been devoted to sockeye and chum salmon (Atkinson 1960, Konovalov 1980). Experiments to investigate the effects of hatchery culture on chinook salmon biology have been conducted (Pisarevsky 1978, Smirnov et al. 1994) with the goal of developing hatchery chinook salmon for harvest (Smirnov et al. 1994).
New Zealand--Attempts to introduce chinook salmon to New Zealand waters in the 1870s were not successful; however, transplants of Sacramento River chinook salmon in 1901 successfully established self-sustaining anadromous and landlocked populations, as well as providing broodstock for subsequent artificial propagation programs (McDowall 1994). By 1925, the naturalized chinook salmon had produced 1.5 million eggs for distribution in New Zealand streams (Lever 1996). Artificial propagation of chinook salmon in New Zealand remains an important component of management of the species (Unwin 1997).
Alaska--Hatcheries in Alaska have been used to mitigate overharvest and to provide harvest opportunities, whereas hatcheries in the lower 48 States have usually been operated to mitigate for destruction and blockage of habitat. In the early days of the Alaskan salmon fishery, hatcheries were used as a means of assurance against the adverse effects of commercial fishing (Roppel 1982). The first federal hatchery in Alaska was built on a lake at Yes Bay in Southeast Alaska in 1905, and a second federal facility was built on Afognak Island in 1908 (Roppel 1982). During this period, legislation in Alaska required canneries to operate hatcheries, although few companies complied. Nonetheless, by 1920 there were at least four private hatcheries in the state, as well as several federal facilities inovlved in the propagation of Pacific salmon (Heard 1985, Heard et al. 1995). Hatchery efforts were directed primarily at the premier commercial species in Alaska, sockeye salmon; other salmon species, including chinook salmon, were reared on an experimental basis.
Occasional attempts to establish runs of non-native chinook salmon were made in Alaska. Between 1923 and 1926, chinook salmon originating from the Columbia River and unspecified locations in Washington State were released into lakes and rivers near Cordova, (571,000 "Washington" chinook salmon), Seward (1,387,000 "Washington" chinook salmon) and near Ketchican (1,952,000 Kalama River, 972,500 "lower Columbia River," and 1,819,000 "Washington" chinook salmon) (Roppel 1982). Not long afterward, Alaska abandoned the concept of using hatcheries to augment natural production, as hatchery releases had not resulted in increases in fish abundance. This may have been related to the poor hatchery practices of that era and general large-scale increases in harvest (Roppel 1982). After a hiatus of two decades, chinook salmon production was resumed at several hatcheries in 1955 in Southeast Alaska and near Anchorage (Wahle and Smith 1979), although production numbers for the state have been relatively low until recently. For example, between 1975 and 1982, a total of 4.7 million fish, or about 597,000 chinook salmon juveniles annually, were released in Alaskan waters. Since 1983, total hatchery production has increased to 73 million fish, or about 7.3 million fish per year (Fig. 27). Much of the increased production has resulted from legislation permitting the operation of private, non-profit hatcheries (McNair 1996). As of 1992, seven private, three state, and one federal hatchery accounted for almost all chinook salmon hatchery production in Alaska (NRC 1996). In Alaska, the majority of chinook salmon stocks exhibit a stream-type life-history, therefore the majority of hatchery fish are released as yearling smolts (NRC 1996).
British Columbia--The first British Columbia salmon hatchery was constructed in 1884 near Westminster, on the Fraser River. Although sockeye salmon were the principal focus of this and other early hatcheries in this province, a few chinook salmon were released as well (Wahle and Smith 1979). Between 1903 and 1927, 72 million chinook salmon were released into British Columbian waters, three-quarters of these into the Fraser River Basin (Cobb 1930). Production during this period peaked in 1908 with the release of 7.5 million chinook salmon (Cobb 1930). However, as in Alaska, there was no apparent increase in the abundance of sockeye salmon, and it became apparent that the artificial propagation of sockeye salmon in British Columbia did not result in a significant increase in efficiency over natural production in areas where there was a reasonable expectation of successful natural propagation (Foerster 1968). By 1930, salmon hatcheries were no longer operating in British Columbia (Foerster 1968, Wahle and Smith 1979). Economic restrictions resulting from the Great Depression and World War II further constrained the ability of the provincial government to initiate hatchery programs. Hatchery production of salmonids was not reestablished in British Columbia until 1967 with the construction of the Big Qualicum Hatchery on Vancouver Island (Wahle and Smith 1979). Artificial propagation efforts accelerated after the launching of the Salmonid Enhancement Program (SEP) in 1977, which was designed to double harvest levels and preserve, rehabilitate, and enhance natural salmonid stocks (Winton and Hilborn 1994). Since that time, the total chinook salmon hatchery effort in British Columbia has expanded to include 50 major (>40,000 juvenile fish released annually) and about 20 minor (<40,000 juvenile fish released annually) fish rearing facilities (NRC 1996). Total chinook salmon production for the period 1975 to 1982 was about 94.7 million juveniles for an average of just under 12 million fish per year. However, to meet expanding harvest demands, hatchery production between 1983 and 1992 increased to 562 million fish, about 56 million fish annually. New propagation/release strategies are being employed to rebuild or enhance British Columbia chinook salmon stocks, especially in lower Georgia Strait streams. These new methods include rearing juveniles to smolt in net-pens in lakes, extended rearing of smolts in sea pens, and maintaining captive broodstocks in sea pens to increase egg availability (Cross et al. 1991). Unlike many chinook salmon hatcheries in the United States (see below), British Columbia hatchery broodstocks have been established using local stocks, although, in some cases, centralized hatcheries are used for the enhancement of many different river-specific stocks within a region (Cross et al. 1991). The contribution from SEP hatcheries varied between 5.3% and 18.6% of the total British Columbia chinook salmon catch from 1978 through 1989 (Winton and Hilborn 1994).
Columbia River Basin--Artificial propagation in the Columbia River basin initially developed following the expansion of the commercial fishery, with the first Columbia River hatchery built in 1876 on the Clackamas River and operated by a cannery interest (CBFWA 1990b). State and federal hatchery operations to enhance commercial fisheries began soon afterward, and by the 1890s, many hatcheries and egg-taking stations were in operation between the Chinook River at the mouth of the Columbia River and the Little Spokane River in the upper basin (CBFWA 1990b). By 1905, about 62 million fry were released annually; however, due to poor returns to these hatcheries, support for Columbia River hatcheries waned shortly thereafter (CBFWA 1996). After the late 1930s, the negative effects of agricultural development, timber activities and other land use practices, and the initial development of the Columbia River dam complex, resulted in an increased need to mitigate for reduced natural production (CBFWA 1990b). Between 1957 and 1975, eleven new mainstream dams were constructed on the Columbia and Snake Rivers, resulting in further loss of habitat and increased migrational mortality. Although fish passage facilities were generally successful at low dams, their efficacy was not great at high dams, which constituted most of the dams built during this later period (CBFWA 1990b). Therefore, artificial production appeared to be the only means available to fish managers to compensate for fish losses and the resulting decline in fish available for harvest. Several of these mitigation programs will be briefly discussed here.
Grand Coulee Fish Maintenance Project--After the construction of the Grand Coulee Dam (RKm 959) in 1939, which completely eliminated passage of anadromous salmon above that point, the federal government initiated the Grand Coulee Fish Maintenance Project (GCFMP), which lasted from 1939 to 1943. The GCFMP sought to maintain fish runs in the Columbia River above Rock Island Dam (RKm 730) by two means: 1) improving salmonid habitat, and 2) establishing hatcheries (Fish and Hanavan 1948).
Adult chinook salmon passing Rock Island Dam from 1939 to 1943 were taken either to USFWS hatcheries on the Wenatchee or Methow Rivers for artificial spawning or to fenced reaches of the Wenatchee or Entiat Rivers for natural spawning. Juveniles derived from adults passing over Rock Island Dam were reared at USFWS hatcheries and transplanted into the Wenatchee, Methow, and Entiat Rivers.
Fish trapping operations began in May 1939, and continued through late fall each year until 1943. A total of five brood years were affected. Early-run fish (stream type) were treated separately from late-run fish (ocean type), but few distinctions were made regarding either the so-called "summer" or "fall" components of the late run, as all late-run fish were captured. The GCFMP continued for five years and intercepted all chinook salmon passing Rock Island Dam, including those destined for now inaccessible spawning areas in British Columbia. As a result, all present day chinook salmon above Rock Island Dam are the progeny of the mixture of chinook salmon collected at Rock Island Dam from 1939 to 1943 (Waknitz et al. 1995).
Chinook salmon spawning channels--Artificial spawning channels for ocean-type chinook salmon were operated during the 1960s and 1970s near Priest Rapids (1963-71), Turtle Rock (1961-69), and Wells Dam (1967-77), but were discontinued in favor of more traditional hatchery methods due to high pre-spawning mortality in adult fish and poor egg survival in the artificial spawning beds (CBFWA 1990b, Chapman et al. 1994).
Mitchell Act--In 1938, in response to the construction of Bonneville and Grand Coulee Dams, Congress passed the Mitchell Act, which required the construction of hatcheries to compensate for fish losses caused by these dams and by logging and pollution (Mighetto and Ebel 1994). An amendment to the Mitchell Act in 1946 led to the development of the Lower Columbia River Fishery Development Plan (CRFDP) in 1948, which initiated the major phase of hatchery construction in the Columbia River Basin (CBFWA 1990b). In 1956, the CRFDP was expanded to include the upper Columbia River and Snake River Basins. Although the majority of lost natural salmonid production to be mitigated by the Mitchell Act was located in the upper Columbia River and Snake River basins, only 4 of the 39 facilities eventually authorized by this Act were constructed above Dalles Dam on the lower Columbia River, partly due to concerns regarding the ability of fish to bypass dams in the upper basin, and partly because the primary goal was to provide fish for harvest in the ocean and lower river (CBFWA 1990b, 1996).
Lower Snake River Fish and Wildlife Compensation Plan--The Lower Snake River Fish and Wildlife Compensation Plan (LSRCP) was authorized by Congress in 1976 to replace lost salmonid production caused by fish passage problems at four U.S. Army Corps of Engineer (COE) dams in the lower Snake River (CBFWA 1990b). To date, 22 facilities have been constructed under the LSRCP, including hatcheries and acclimation ponds. In general, LSRCP facilities have had more success in increasing the abundance of steelhead than chinook salmon (Mighetto and Ebel 1994).
U.S. Army Corps of Engineers--The Corps of Engineers (COE) has funded the construction or expansion of 19 hatcheries as mitigation for fish losses caused by COE hydroelectric programs throughout the entire Columbia River basin, including the building of 12 dams in the Willamette River basin between 1941 and 1968 (CBFWA 1990b). Many hatcheries constructed under the Mitchell Act were funded by COE.
Public and private power generators--These non-governmental entities have funded the construction and/or operation of 16 artificial propagation facilities in the Columbia River basin as compensation for lost fish production due to their water-use projects. Utilities and companies participating in Columbia River fish culture operations include Chelan, Douglas and Grant County PUDs in Washington (ESUs 12 and 13), Idaho Power Company (ESUs 14 and 15), Portland General Electric (ESUs 9 and 11), Tacoma City Light (ESU 9), and Pacific Power and Light (ESU 9) (CBFWA 1990b).
West Coast hatchery production of chinook salmon is summarized in Table 6, with data taken from a database developed under contract to NMFS (NRC 1996). Some release information presented here dates back to the turn of the century, but any data prior to 1950--when hatchery records became more reliable--should be considered incomplete.
The ratio of hatchery- to naturally-produced chinook salmon on the West Coast varies from region to region, as well as from watershed to watershed, within a particular ESU, with chinook salmon populations dominated by hatchery production in some areas and maintained by natural production in others (Howell et al. 1985, WDF et al. 1993, Kostow 1995). Large hatchery programs have produced substantial numbers of fish relative to natural production in many West Coast regions, especially in areas where hatcheries have been used to create or enhance harvest opportunities. These areas include many locations in Puget Sound, the majority of watersheds in the Columbia River Basin, several Oregon coastal streams, the Klamath River
Basin, and the Sacramento River Basin (Howell et al. 1985; WDF et al. 1993; PFMC 1994,1997; Kostow 1995). A list of the larger chinook salmon artificial propagation facilities operating on the West Coast is provided in Table 7.
Chinook salmon have often been transferred among watersheds, regions, states, and countries, either to initiate or maintain hatchery populations or naturally spawning population in other watersheds. The transfer of non-native fish into some areas has shifted the genetic profiles of some hatchery and natural populations so that the affected population is genetically more similar to distant hatchery populations than to local populations (Kostow 1995, Howell et al. 1985, Marshall et al. 1995).
It is often difficult to determine the proportion of native and non-native hatchery fish released into a given watershed. Table 6 shows estimates of the proportion of non-native fish introduced into each ESU, but in many cases they will be underestimates for two reasons. First, hatchery or outplanted fish that were designated as "origin unknown" in the database (NRC 1996) were counted as native fish, even though in some cases they were probably not native. Second, transplanted hatchery fish routinely acquire the name of the river system into which they have been transferred. For example, spring-run chinook salmon released from the Leavenworth NFH are primarily the descendants of the Carson NFH stock (Marshall et al. 1995), but are designated as Leavenworth stock when released or transferred (NRC 1996). These fish were counted as native fish in this review. Sol Duc River (Washington Coast ESU) spring chinook salmon were derived from a hybrid of two out-of-ESU stocks (WDF et al. 1993), but were identified as Sol Duc stock when released from the Sol Duc Hatchery or when transferred to other ESUs, such as Hood Canal (Puget Sound ESU) (WDF et al. 1993, NRC 1996). Similarly, the Russian River (So. Oregon and Coastal California ESU) receives fall chinook salmon from a number of different hatcheries in other ESUs, which are correctly identified by hatchery of origin at release, but become "Russian River" stock when they return and are propagated for release in subsequent generations at the Warm Springs Hatchery (NRC 1996).
Until recently, the transfer of hatchery chinook salmon stocks between distant watersheds and facilities was a common management strategy (Matthews and Waples 1991, WDF et al. 1993, Kostow 1995). Agencies have instituted policies to reduce the exchange of non-indigenous genetic material among watersheds. In 1991, chinook salmon co-managers in Washington adopted a statewide plan to reduce the number of out-of-basin hatchery-to-hatchery transfers of salmon. This included genetic guidelines specifying which transfers between areas were acceptable. However, these policies applied only to transfers between hatcheries and did not explicitly prohibit introductions of non-native salmonids into natural populations (WDF 1991). At present, co-managers in Washington State are developing guidelines for transfers of hatchery chinook salmon into natural populations (WDFW 1994). In 1992, the Oregon Coastal Chinook Salmon Management Plan was implemented, which provides guidelines for stock transfers (Kostow 1995).