Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
4 Status of Salmon The status of many specific salmon populations in the Pacific Northwest is uncertain, and exceptions exist to generalizations with regard to overall status. Nevertheless, a general examination of the evidence of population declines over broad areas is helpful for understanding the current status of species with differ- ent life cycle characteristics and geographical distributions. Using a geographic information system based on state agencies' status re- ports from the American Fisheries Society and other sources, the Wilderness Society (1993) estimated current extinction risks of Pacific Northwest salmon throughout their historical ranges (Table 4-1J. Classification of stocks followed Nehlsen et al (19911; "endangered" was equivalent to "high risk of extinction," "threatened" was equivalent to "moderate risk of extinction," and "of special concern" referred to populations not currently at risk, but not as secure as "healthy," for various known reasons, or for which there was incomplete infor- mation but the suggestion of depletion. Populations not known to be recently declining, often classified as "healthy" in state assessments (Nickelson et al. 1992, WDF et al. 1993), were also included. Percentages in Table 4-1 should be viewed as provisional, in that they were based on population assessments that were in many instances uncertain; however, they provide a picture of the relative status of different salmon species and runs. The following generalizations can be made from this information: . Pacific salmon have disappeared from about 40% of their historical breeding ranges in Washington, Oregon, Idaho, and California over the last century, and many remaining populations are severely depressed in areas where they were formerly abundant. If the areas in which salmon are threatened or 75
76 UPSTREAM: SALMON AND SOCIETY IN THE PACIFIC NORTHWEST TABLE 4-1 Current Status of Pacific Salmon over Their Known Historical Geographic Range in Washington, Oregon, Idaho, and California Status, To of Historical Range Species (Run) Extinct Endangered Threatened Special Not Known To Concern Be Declining Fall chinook 19 18 7 36 20 Spring/summer chinook 63 8 16 7 6 Coho 55 13 20 5 Chum 37 16 14 11 Sockeye 59 7 3 16 Pink 21 5 < 1 0 Sea-run cutthroat 6 4 61 29 0 Winter steelhead 29 22 7 18 24 Summer steelhead 45 5 5 97 18 Overall 40 13 14 17 16 7 15 73 Source: Data from The Wilderness Society 1993. endangered are added to the areas where they are now extinct, the total area with losses is two-thirds of their previous range in the four states. Although the overall situation is not as serious in southwestern British Columbia, some popu- lations there also are in a state of decline, and all populations have been com- pletely cut off from access to the upper Columbia River in eastern British Colum- bia. Even if the estimate of population losses of about 40% is only a rough approximation, the status of naturally spawning salmon populations gives cause ~ . . for pessimism. · Coastal populations tend to be somewhat better ofi~than populations in habiting interior drainages. Species with populations that occurred in inland subbasins of very large river systems (such as the Sacramento, Klamath, and Columbia rivers) spring/summer chinook, summer steelhead, and sockeye- are extinct over a greater percentage of their range than species limited primarily to coastal rivers. Salmon whose populations are stable over the greatest percent- age of their range (fall chinook, chum, pink, and winter steelhead) chiefly inhabit rivers and streams in coastal zones. Populations near the southern boundary of species' ranges tend to be at greater risk than MortherM populations. In general, proportionately fewer healthy populations exist in California and Oregon than in Washington and British Co- lumbia. The reasons for this trend are complex and appear to be related to both ocean conditions and human activities. · Species with extended freshwater rearing (up to a year)-such as spring/ summer chinook, coho, sockeye, sea-run cutthroat, and steelhead are generally extinct, endangered, or threatened over a greater percentage of their ranges than .
STATUS OF SALMON 77 species with abbreviated freshwater residence, such as fall chinook, chum, and oink salmon. . In many cases, populations that have not declined are composed largely or entirely of hatchery fish. An overall estimate of the proportion of hatchery fish is not available, but several regional estimates make clear that many runs depend mainly or entirely on hatcheries. This committee assessed the current status of Pacific salmon from the Fraser River Basin in southern British Columbia to Monterey Bay in central California. This portion of the range includes nearly all salmon populations caught by fishers in Washington, Oregon, Idaho, and California. Historical records of population abundance vary greatly across this region; some river basins have extensive catch and escapement records extending back nearly a century; run sizes for other river systems are poorly known. Historical data on commercially caught species tend to be much more complete than those on species not heavily fished or species caught only by recreational angling (e.g., sea-run cutthroat trout). The Endangered Species Committee of the American Fisheries Society (Nehlsen et al. 1991) identified 214 salmon stocks as being at risk of extinction and over 100 populations as being recently extinct in Washington, Oregon, Idaho, and California. The report was based on admittedly incomplete data, but it pointed out the widespread nature of salmon declines and the seriousness of the current problem. Although several populations already had been petitioned for protection under the Endangered Species Act (ESA) when the report was pub- lished, public interest in the "salmon problem" was heightened greatly by its appearance, and various state and federal agencies rapidly began to develop management plans to address the conservation needs of potentially listed popula- tions. Nehlsen et al. (1991) also catalyzed efforts to assess further the current condition and causes of declines of salmon. The remainder of this chapter discusses some of the difficulties in evaluating the status of wild populations and how these problems have been addressed in recently published status reports. It then summarizes regional trends in salmon populations in the Fraser River Basin, Puget Sound, the Columbia River Basin, and coastal river basins of Washington, Oregon, and California. Finally, general patterns in the overall condition of the species are presented on the basis of their geographical distribution and life-cycle requirements. INTERPRETING HISTORICAL RECORDS It is tempting to conclude that all salmon populations are depressed, but the picture is complicated. Salmon can be difficult to count; even if accurately counted, their populations are inherently variable owing to continual changes in freshwater and marine factors or to the cyclic nature of some species (e.g., pink and sockeye salmon). Long time-series records of catch or escapement, often
78 UPSTREAM: SALMON AND SOCIETY IN THE PACIFIC NORTHWEST spanning decades, are required for the statistical detection of large changes (50% or greater) in population abundance (Lichatowich and Cramer 19801. In addition to inherent variability, long-term changes in abundance might not be continuous or linear and so might not be clearly revealed with simple regression methods (Bledsoe et al. 1989~. Short periods of record might suggest population increases or decreases when, in fact, long-term trends are in a different direction. Population status can also be difficult to determine if catch statistics are the only source of information, in that these data can be misleading. One example is the often-displayed graph of salmon landings in the lower Columbia River (e.g., Kaczynski and Palmisano 1993:150~. The graph ranges from nearly 50 million pounds before 1920 to less than 5 million in 1990. The authors noted that in-river fisheries were supplemented by offshore trolling in the middle to late part of this century, but casual readers might assume that the depiction of reduced catch reflects the magnitude of the decline of salmon in the Columbia River without recognizing that in-river catch has been largely supplanted by large ocean fisher- ies. Catch statistics tend to be biased in favor of large populations, either because fisheries tend to target them and ignore small populations with unusual run tim- ing or because trends in catches of large populations mask simultaneous but sometimes opposite trends in smaller ones. An example is the trend in Puget Sound sockeye catches early in the twentieth century (Bledsoe et al. 1989:59~. Dramatic declines in sockeye landings by Puget Sound fishers were apparent after 1913, but they were caused by rockslides on the Fraser River, in which most of the Puget Sound sockeye catch originated (Quinn 1994~. Puget Sound fisher- ies on Fraser River sockeye masked trends of Puget Sound sockeye populations unless more refined catch data were examined. THE STOCK CONCEPT In the jargon of Pacific salmon fisheries, managers refer to stocks of salmon, i.e., populations or groups of salmon populations that are recognizable for man- agement. The term stock has been used in various ways, but in some form the stock concept has been fundamental in understanding the population structure and management of Pacific salmon (Moulton 19399. A stock is considered the basic unit of salmon management (Moulton 1939) and has been defined as a recognizable or manageable group of animals (Larkin 1972, Waples 1991J. The basic concept is that these local populations are largely reproductively isolated and over time become adapted to local conditions. In management practice, however, each of these local populations is not identifiable and stock usually refers to larger, recognizable groups of these basic population units. Salmon biologists have recognized this "fine-grained" structure of salmon populations. However, few management policies explicitly recognize the need to protect salmon at the level of local spawning populations, and most stock assessments are conducted at a much coarser geographic scale. The status reports in this
STATUS OF SALMON 79 chapter reflect this limitation and, in most cases, we do not know the status of the smaller local population units. In general, we use the term stock only when the report being cited specifically used that term. Stocks have been identified in connection with geographical areas as small as individual tributaries (e.g., the winter steelhead of Oregon's Illinois River) and as large as entire ecoregions (e.g. sea-run cutthroat trout from Oregon coastal streams). For example, Nehlsen et al. (1991) recognized two stocks of sea-run cutthroat trout from all of Washington, excluding the Columbia River: one stock consisted of Hood Canal and Grays Harbor fish, and the other included all other populations. One reason for the inconsistency is that the available information differs greatly among species: commercial species tend to be better known than species fished only for sport, populations in rivers with dams and counting facili- ties tend to be better known than populations in streams without such facilities, populations close to urban centers tend to be better known than those in remote locations, and so on. Where detailed information is available on run sizes, salmon are often split into many stocks; but where information is lacking populations are often lumped together into only a few stocks. As a result, practical application of the stock concept has included several levels of genetic organization (Chapter 6 discusses this in more detail). An additional difficulty in applying the stock concept to status assessment has arisen from the fact that many Pacific salmon demes are small. Although some demes number in the hundreds of thousands or even millions of fish in exceptional cases such as the Adams River, B.C., sockeye, others may consist of only a few adults that spawn in geographically limited areas, such as small headwater streams and portions of lake shorelines. These very small demes might be substantially isolated from other such demes and even possess local adaptations that distinguish them from other members of the species, but they are rarely considered stocks. Some small populations (e.g., Redfish Lake sockeye and Sacramento River winter chinook) have been treated as separate stocks for ESA purposes, but they are the remnants of much larger runs. In theory, the stock concept makes no allowance for deme size or metapopulation structure, in which populations consist of locally reproducing groups connected by some gene flow within a larger area. The management of stocks by state fisheries agencies has generally not recognized the geographical structure of salmon populations at such fine scales. Few management policies explicitly recognize the need to protect salmon at the level of individual demes, so most stock assessments are carried out on a much coarser geographic scale. In most cases, we do not know the status of salmon on the scale of local domes. RISK ASSESSMENT Beginning with Nehlsen et al. (1991), recent status reports have assigned salmon stocks to various risk categories on the basis of population trends and
80 UPSTREAM: SALMON AND SOCIETY IN THE PACIFIC NORTHWEST potential for loss of genetic integrity from introduction of maladapted genes from nonnative or artificially propagated populations. Graduated categories of risk have been used in which stocks in the highest-risk categories were considered at greatest peril of extinction. Different reports sometimes assessed the condition of the same stocks according to common data sets. However, different criteria of risk categories were often used, and that makes direct comparisons among reports difficult. The risk of extinction of stocks classified as of "special concern" in one report, for example, did not always correspond to the extinction risk of stocks with the same classification in another report. A comparison of the criteria used by Nehlsen et al. (1991), Higgins et al. (1992), Nickelson et al. (1992), WDF et al. (1993), and The Wilderness Society (1993) is given in Table 4-2. Compilations of stock status must be read and compared carefully for several reasons (Quinn 19941. Compilations differ in subtle but important aspects of their definitions, and conclusions of the assessments can be distinctly different in areas of geographical overlap. For example, in the extreme case of stock extinc- tions, WDF et al. (1993) listed only one stock as recently extinct in Washington state whereas Nehlsen et al. ~ 1991) listed 42 recent stock extinctions in Washing- ton. Interestingly, the early chum salmon run to Chambers Creek in Puget Sound was said by Nehlsen et al. (1991) to be "at low levels . . . but appears to be rebuilding" but was classified as extinct by WDF et al. (1993~. There are also major differences in how regional groups of stocks are judged. For example, WDF et al. (1993) listed 17 stocks of chum salmon in southern Puget Sound and classified 15 as healthy, one as extinct, and one of unknown status. In contrast, The Wilderness Society (1993) stated that "chum salmon are depleted or extinct in the rivers of southern Puget Sound . . .", and a petition to invoke ESA protec- tion for chum salmon in some portions of southern Puget Sound was recently filed with the National Marine Fisheries Service. Differences in judgment about the status of stocks are sometimes confounded by disagreement regarding the definition of stock or the number of populations being considered as a single unit. For example, Nehlsen et al. (1991) concluded that "Lake Washington" sockeye did not meet the criteria of risk of extinction, but WDF et al. (1993) recognized three stocks in the watershed (Cedar River spawners, Lake Washington and Lake Samammish tributary spawners, and Lake Washington beach spawners) and classified all of them as depressed. Putting aside the differences and discrepancies among the reports, it is clear that a substantial number of wild salmon populations are in some jeopardy and that the status of many others is poorly known. We must first ask whether the loss of populations should concern us. In addition to the ESA's provision for distinct population segments to be protected, Alkire (1993) described the sub- stantial economic value of salmon resources for both commercial (native and nonnative) and recreational fishers. They are also of great symbolic importance to native peoples and more recent settlers as well, representing clean water, forests, and the wonders of animal migration. And small populations might
STATUS OF SALMON TABLE 4-2 Comparison of Criteria Used to Assign Pacific Salmon Stocks to Different Status Categories 81 Nehlsen et al. (1991), Higgins et al. (1992), The Wilderness Society (1993): High Risk of Extinction . Populations' spawning escapements are declining; less than one adult fish returns from each parent spawner. . Populations' recent escapements (within last 1-5 years) are under 200 in the absence of evidence that they were historically small. · Population is likely candidate for listing as endangered under ESA. Moderate Risk of Extinction · Populations' spawning escapements appear to be stable after previously declining more than natural variation would account for and are generally in range of 200-1,000 spawning adults. · Population is likely candidate for listing as threatened under ESA. Special Concern · Populations are believed to be vulnerable to minor disturbances, especially if a specific threat is known. · Insufficient information on population trend exists, but available evidence suggests depletion. · Continuing releases of nonnative fish are relatively large and potential for interbreeding with native population exists. . _ Population is not at risk but requires attention because of unique characteristic. Nickelson et al. (1992): Special Concern . Population is probably composed of less than 300 spawners, or · Substantial risk exists for interbreeding between the population and stray hatchery fish in excess of standards established by Wild Fish Management Policy (Oregon). Depresseda . Available spawning habitat has generally not been fully seeded, or · Abundance trends have declined over the last 20 years, or · Abundance trends in recent years have been generally below 20-year averages. Healthy · Available spawning habitat has generally been fully seeded, and · Abundance trends have remained stable or increased over last 20 years. Unknown WDF et al. (1993): Insufficient data are available to judge population status. Critical · Stock has declined to point where it is in jeopardy of significant loss of within-stock diversity or, in worst case, extinction. Depressed · Stock's production is below levels expected on basis of available habitat and natural variations in survival rates but above where permanent damage to stock is likely.
82 Healthy . UPSTREAM: SALMON AND SOCIETY IN THE PACIFIC NORTHWEST Stock is experiencing stable escapement, survival, and production trends and not displaying pattern of chronically low abundance. · Stoclc is experiencing production levels consistent with its available habitat and within natural variations in survival for stock. · Stocks have a wide ran; ,e o~t conditions, from "robust to those without surplus production for harvest." Unknown Information is insufficient information to determine status. Category includes "historically small populations [that] could be especially vulnerable to any negative impacts.?' Extinct · Stocks are known to have become extinct "during recent times." · Only one stock (Chambers Creek summer chum) classified as recently extinct, but a number of stocks were not called extinct because of lack of agreement on whether they existed. aThis category supersedes the "special concern classification"; i.e., a population classified as depressed might also fit one of the criteria applied to the special concern category. contain valuable genetic traits that could not be restored if the populations were lost and that could have great value to the aquaculture industry (Scudder 1989, Riddell 1993b). Status reports have been published by agencies (Konkel and McIntyre 1987, Nickelson et al. 1992, WDF et al. 1993), scientific societies (Nehlsen et al. 1991, Higgins et al. 1992), industry organizations (Kaczynski and Palmisano 1993, Palmisano et al. 1993), and an environmental group (The Wilderness Society 19931. Although all the reports were written by fishery scientists, the committee notes that political pressures to classify salmon stocks as at risk or healthy have raised doubts among scientists as to the accuracy of the risk assessments. Stocks classified as healthy might not put fish and wildlife agencies in an unfavorable light or limit future management options; stocks classified as at risk could justify criticism of present policies or support ESA petitions. There are no simple means of verifying the accuracy of status reports. We conclude that the most prudent approach to risk assessment is to examine broad regional trends in populations and manage accordingly (and conservatively), rather than to rely on incomplete information where interpretation is open to question. Taken in total, the status reports are useful for identifying broad trends but should be viewed with caution at the level of individual populations. FRASER RIVER BASIN The Fraser River produces more salmon than any other river in the world (Northcote and Larkin 1989:172-204), including the seven species of anadro- mous salmon on which this report focuses. The Fraser River is smaller than the
STATUS OF SALMON 83 Columbia but shares many biogeoclimatic features with it. Like the Columbia, the Fraser River Basin consists of wet coastal lowlands, canyon regions through the coastal mountains, and a dry, high interior plateau. The rivers have differed substantially in their development, most notably in the absence of dams on the mainstem Fraser River. Historical catches of Pacific salmon in or near the Fraser River tend, however, to show an abundance pattern similar to that observed historically in the Columbia. Because of these shared features between the Fraser and the Columbia and because so many Fraser River salmon are caught in U.S. waters, we describe it here. Historical Fraser River catches by species are summarized in Figure 4-1 (Argue et al. 19861. In general, the early development of the fishery is evident: historically large catches during the early l900s were followed by substantial declines in the 1920s and 1930s. That trend must be interpreted cautiously, however, in that it represents only the catch in the terminal area. For example, catches of chinook salmon in ocean troll fisheries would certainly reduce the rate of decline evident in the figure. However, it is generally true that salmon produc- tion in the Fraser River was declining throughout the basin until the middle 1970s. The Fraser River has escaped the development of major dams on the mainstem, but it has not escaped impacts of human development. Some of the major point-source impacts occurred early in the development of British Colum- bia. Roos (1991' identified major impacts early in this century from gold-mining in the Quesnel drainage, extinction of the upper Adams River sockeye (Williams 1987) due to logging dams, siltation from early logging, and railway develop- ment through the Fraser canyon (Hell's Gate slide, Feb. 1914~. The Hell's Gate rockslide accounts for the major loss of sockeye catch that started in 1916, and the loss of pink salmon populations in the Thompson River (Ricker 1989~. Habi- tat impacts have, of course, accumulated with later development. For example, about 80% of the Fraser River delta wetlands have been lost to agriculture, urbanization, and flood control (Environment Canada 19861; and although the Fraser has avoided mainstem dams, more than 800 dams are licensed for agricul- tural water use in the upper Fraser drainage. It is difficult to assess the impact of those habitat changes on salmon production, particularly because they were si- multaneous with the overfishing of the salmon populations. During the 1980s, however, salmon production has been rebuilding for most species. Information on steelhead production is sparse, but steelhead are consid- ered to be depressed. Recent returns of five species are summarized in Figures 4- 2a to 4-2e. The sockeye, pink, and chum salmon figures are estimates of the total return of these Fraser resources (catch plus spawning escapements), but the chinook and coho figures account only for the terminal catch plus an index of spawning escapements. The indices are based on visual counts of spawners and undoubtedly underestimate the actual number of spawners. The figures for each species also include production from enhancement programs, but these are rela
84 o g (D o o U) Cat I C) to to o to to to to - o to o U) UPSTREAM: SALMON AND SOCIETY IN THE PACIFIC NORTHWEST SOCKEYE 8 to o . COHO 1880 1900 1920 1940 1960 1980 PINK .....~ I no. 1880 1900 1920 1940 1960 1 980 CHUM U. N U) 1880 1 900 1920 1940 1960 1980 YEAR 1880 1900 1920 1940 1960 1980 CHINOOK 1880 1900 1920 1940 1960 1 980 STEELHEAD I 1880 1900 1920 1940 1960 1 980 FIGURE 4-1 Historical commercial catch of Pacific salmon (numbers landed) in the Fraser River area (Statistical areas 28 and 29) between 1876 and 1985. Before 1951, the numbers were estimated from records of canned pack and product statistics; after 1951, the numbers are from records of fish sales (Canadian Department of Fisheries and Oceans annual catch statistics). Source: Argue et al. 1986.
STATUS OF SALMON 4-2a 25 v) .o 20 . ~ 15 I g tn 10 o r, z 4-2c Me ~ 2000 o E tic c o - z 1 500 1 000 160 140 1 20 _ 1 00 -Go 80 O 60 ~ 1 E do 20 85 ·Spawners A O Total catch 25 B O Total Catch ~2b i i 1946 1953 1960 1967 1974 1981 1988 Return Year C ·Spawners -OTotai Catch O i 1 i 1 1 ~ 1951 1956 1961 1966 1971 1976 1981 1986 Return Year spawning Index OTerminal Catch O _ 1952 1962 1972 1982 1992 Return Year 20 10 ·Spawners O- , 1959 1965 1971 t 977 t 983 1989 Return Year 42d 350 -Total Index (+Falls) 250 -Spring/summer esc ClTerminal Catch g 200 ldlOnNl 111 nil a_ nn ° 1 00 z 50 o 1 952 1 962 1 972 1 982 1 992 Return Year FIGURE 4-2a to 4-2e Long-term changes in catch and escapement from Fraser River, B.C. Source: (a), modified from Bisson et al. (1992), (b-e), committee generated from dam provided by Canadian Department of Fisheries and Oceans.
86 UPSTREAM: SALMON AND SOCIETY IN THE PACIFIC NORTHWEST lively small portions of the species production except for recent chum returns and a few years of coho.1 Increases in production of sockeye salmon reflect reduced catch rates in fisheries to increase spawning-population sizes, good marine sur- vival over almost 20 years, and the protection of the water courses. Similarly, the increased pink and chum production reflects management actions to increase spawning-population sizes. Terminal runs of chinook increased substantially after the Pacific Salmon Treaty (1985), but conservation programs had been implemented in the terminal area since 1980. The best representation of the terminal run of chinook is the solid line in Figure 4-2d; this line represents the terminal catch plus spawning index for spring and summer chinook and the estimated spawning population for the Harrison River fall chinook population. The escapement of chinook to the Harrison River has been estimated quantita- tively since 1984 and constitutes one of the largest chinook populations in North America. The remaining concern for Fraser salmon is probably the returns of coho. As evident in Figure 4-2e, the return is highly variable and associated with the strict 2-year cycle of Fraser pink salmon production. Fraser pinks return in the odd years, and fisheries on Fraser pinks incidentally capture returns of Fraser coho. The Department of Fisheries and Oceans in Canada is developing a coho- conservation program for the Strait of Georgia and the Fraser River. The production of salmon from the Fraser River is still less than occurred historically, and economic development and overfishing have resulted in some extinctions. However, the resource base (fish populations and habitat' is appar- ently adequate to support an increase in salmon production. PUGET SOUND The Puget Sound Basin includes river systems in Puget Sound, Hood Canal, and the Strait of Juan de Fuca. Summaries of the status of salmon in Puget Sound (WDE et al. 1993, Quinn 1994J suggest a wide range of population conditions, ranging from healthy to critical. Overall, 93 of 209 salmon stocks (44.5~o) identified by WDF et al. (1993) were classified as healthy, 44 (21.2%) were depressed, 5.3% were in critical condition, and the status of 60 (28.8%) was unknown (Table 4-31. Only one stock, early chum salmon in Chambers Creek, was considered by WDF et al. (1993) to be recently extinct, although there have been many stock extinctions over the last 150 years (Nehlsen et al. 19911. Chum salmon populations were most often classified as healthy, followed by pink salmon. However, other species had about half or more of their stocks classified as depressed or critical (Table 4-31. Puget Sound was considered to have fewer depressed stocks than the Columbia River basin but more than the Washington 1Data on the total runs of Fraser sockeye and pink salmon were provided by the Pacific Salmon Commission.
STATUS OF SALMON TABLE 4-3 Status of Puget Sound Salmon, as Summarized by WDF et al. (1993)a 87 Status Chinook Chum Coho Pink Sockeye Steelhead Total Healthy 10 38 20 9 0 16 93 Depressed 8 1 16 2 3 14 44 Critical 4 2 1 2 1 1 11 Unknown 7 13 9 2 0 29 60 aStocks listed in this table do not include 40 salmon stocks from Washington believed by Nehlsen et al. (1991) to have been extirpated in last 150 years. Also, one stock (of chum salmon) classified as extinct and five stocks of chinook in disputed status were not included. TABLE 4-4 Overall Status of Washington Anadromous Salmon, as Summarized by WDF et al. (1993)a Puget Sound Coast Columbia River All Washington StatusNo. % No. % No. % No. % Healthy93 44.7 65 56.5 29 26.1 187 43.1 Depressed44 21.2 8 7.0 70 63.1 122 28.1 Critical11 5.3 0 0 1 0.9 12 2.8 Unknown60 28.8 42 36.5 11 9.9 113 26.0 Total 208 100.0 115 100.0 111 100.0 434 100.0 aStocks listed in this table do not include 40 salmon stocks from Washington believed by Nehlsen et al. (1991) to have been extirpated in last 150 years. Also, one stock (of chum) classified as extinct and five stocks of chinook in disputed status not included. coastal region (Table 4-4~. However, almost all the state's critical stocks were believed by WDF et al. (1993) to be from Puget Sound. The Puget Sound region was also intermediate between coastal and Columbia River basins in the propor- tion of stocks whose status was unknown. Because there is such a great diversity of conditions in Puget Sound, it is difficult to draw generalizations about the overall status of salmon populations. Some river systems have productive habitat and large, relatively stable salmon populations; other rivers have been heavily altered and have few or no healthy populations. For example, the river basins of northern Puget Sound had the fewest depressed salmon runs of any region in Washington, Oregon, Idaho, and California, according to The Wilderness Society (1993~. However, the Elwha River on the north coast of the Olympic Peninsula has more extirpated salmon
88 UPSTREAM: SALMON AND SOCIETY IN THE PACIFIC NORTHWEST populations than any other river in Washington, except the Columbia River (Nehlsen et al. 1991), partly as a result of two mainstem dams without fish- passage facilities constructed a short distance above the river mouth early in the 1900s. Slightly more than half the Puget Sound chinook stocks whose status was known were believed to be in a depressed or critical condition (Table 4-3~. In addition, five stocks of spring chinook (Stillaguamish, Snohomish, Green, Skokomish, and Elwha rivers) were in "disputed" status because state agency biologists felt that they were extinct but tribal biologists felt that they still existed. The spring chinook from the Elwha River were particularly notable because this population is reputed to have included exceptionally large fish-up to 45 kg- and the Elwha River has been the focus of considerable attention regarding the possible removal of two mainstem dams. In general, spring chinook were more likely to be at risk than fall chinook. Spring chinook stocks classified as critical occurred in the North Fork Nooksack River, the South Fork Nooksack River, the White River (a Puyallup River tributary), and the Dungeness River. Those stocks were included in a recent ESA petition to list nine of the 12 critical stocks identified by WDF et al. (1993~. Habitat loss and overfishing were cited in the petition as primary contributors to stock declines. About half the chinook popu- lations in Puget Sound, including many of the larger populations, were supported (at least in part) by hatchery propagation. Sixteen of the 37 coho salmon stocks from Puget Sound whose status was known were listed as depressed and one stock (Discovery Bay coho) was classi- fied as critical (Table 4-3~. According to WDF et al. (1993), only two U.S. coho stocks in the Puget Sound region Skagit River and Deer Creek, a tributary of the Stillaguamish River were purely wild; all other stocks have been supple- mented with hatchery fish of nonnative origin at some time. About half the total number of coho stocks were being supported by hatchery production, although most of the coho in southern Puget Sound were of hatchery origin and most of the coho from Hood Canal and the Strait of Juan de Fuca were wild. Virtually all the status reports that discuss causes of declines (Nehlsen et al. 1991, Palmisano et al. 1993, The Wilderness Society 1993) cite habitat loss, overfishing, and negative influences from hatchery fish as contributors to the dwindling of wild coho populations. Protection of depleted populations of wild coho in Hood Canal tributaries has been an important reason for extensive recent cutbacks in salmon fisheries along the Strait of Juan de Fuca, through which Hood Canal coho must pass when returning to spawn. Almost all the chum salmon stocks in Puget Sound whose status was known were healthy, according to WDF et al. (19931. The only three stocks classified as critical or depressed (Hood Canal, Discovery Bay, and Sequim Bay) were early- returning populations termed "summer" chum. Depression of early-returning chum populations is most likely related to overfishing (including incidental catch in fisheries that target other species) and to poor spawning conditions. About
STATUS OF SALMON 89 one-third of the chum populations in the Puget Sound region are supported by hatchery production. There are fewer than 15 pink salmon populations in Puget Sound because this area is at the southern edge of the distribution of the species in North America, except for highly disjunct (and now extinct) populations in northern California. They have 2-year life cycles and most return in odd-numbered years to spawn (the single exception is a small, even-year spawning population from the Snohomish River). Most of the pink salmon stocks in Puget Sound have been classified as healthy (Table 4-3), but two stocks in Hood Canal and the Strait of Juan de Fuca were listed as depressed (Dosewallips River and upper Dungeness River), and two stocks were considered critical (lower Dungeness River and Elwha River). Virtually all the pink salmon populations are of wild origin, and none is currently supported by hatchery production. The largest salmon popula- tions in Washington are pinks from northern and central Puget Sound rivers, where in odd-numbered years spawners can run in the hundreds of thousands (WDF et al. 1993J. Like other salmon, pinks are vulnerable to disturbance of spawning areas and overfishing of depleted populations. Four sockeye salmon stocks were identified by WDF et al. (1993) in Puget Sound-three from the Lake Washington system. Nehlsen et al. (1991) referred to an extinct sockeye population from Mason Lake in southern Puget Sound, but this population is not mentioned by WDF et al. (19933. All three Lake Washing- ton stocks, which were partially of nonnative origin, were classified as depressed (Table 4-3), and the sockeye stock from the Baker River, a Skagit River tributary, was listed as critical. The Baker River sockeye is a native population now blocked from historical spawning areas by two dams. Adults are now captured at the lower dam and trucked to artificial spawning beaches in Baker Lake. Accord- ing to WDF et al. (1993', as few as 92 sockeye have returned to the adult collection facility in recent years. More stocks of steelhead (60) than of any other anadromous salmon are recognized in Puget Sound, but the status of half of them is unknown (Table 4-33. Half the stocks whose status is known were listed as depressed or critical. Sur- prisingly, most of the steelhead stocks from heavily urbanized portions of Puget Sound were classified as healthy or unknown by WDF et al. (1993~; there were a few notable exceptions, including Lake Washington winter steelhead, a depressed population that is subject to heavy predation by sea lions at the Ballard Locks (Box 2-1), and Deer Creek summer steelhead, a critical stock that was recently petitioned for ESA protection and was made famous by the writings of Zane Grey. Few hatcheries propagate steelhead in Puget Sound, but large numbers of hatchery juveniles have been released throughout the basin. Most depressed steelhead populations occur in rivers of northern Hood Canal and along the Strait of Juan de Fuca; the origin of these stocks was unresolved between state and tribal biologists (WDF et al. 1993), but most are maintained solely by natural production.
9o UPSTREAM: SALMON AND SOCIETY IN THE PACIFIC NORTHWEST Relatively little is known about the current status of sea-run cutthroat trout in Puget Sound. The species was not included in the WDF et al. (1993) stock inventory, and population trends are monitored at few locations. At least three separate sea-run cutthroat stocks in Puget Sound have been identified on the basis of electrophoretic comparisons of specimens from different areas: a northern Puget Sound stock, a northern Hood Canal stock, and a southern Hood Canal stock (Compton and Utter 1987~. The species was considered at moderate risk of extinction by Nehlsen et al. (19911. In addition to the negative influences of habitat loss and overfishing, sea-run cutthroat appear to be highly vulnerable to competitive displacement by other salmon and might be affected by widespread coho and steelhead supplementation programs (Trotter et al. 19931. COLUMBIA RIVER BASIN Historical Trends The Columbia River provides a case history for some of the most severe declines in salmon populations along the Pacific Coast. There the late spring and summer runs of chinook populations were historically the most abundant and heavily fished salmon in the basin (Thompson 1951, Van Hyning 1968, Chapman 1986~. Chapman (1986) estimated peak runs in 1881-1885 at about 2.7 million spring-summer chinook. For all Columbia River salmon species, the Pacific Fishery Management Council (PFMC 1979) estimated Redevelopment runs at 6.2 million (including all species except steelhead and sea-run cutthroat) on the basis of habitat availability. That number approximates Redevelopment salmon (chinook, coho, sockeye, chum, and steelhead) estimates of 7.5 million by Chapman (1986) on the basis of peak catches and probable exploitation rates. The Northwest Power Planning Council (NPPC 1986) adopted a predevelopment- run estimate of 10-16 million fish about double the numbers estimated by PFMC (1979) and Chapman (1986~. Overfishing in the lower Columbia River rapidly depressed sockeye and chinook populations and then steelhead, coho, and chum. Habitat degradation and loss accelerated the decline. The number of naturally produced salmon in the Columbia River declined by the late 1900s to about one- eighth their Redevelopment abundance, while total runs along the Pacific North- west coast decreased by about two-thirds. Before 1933, only catch data were available to indicate run sizes. Beginning with the completion of Rock Island Dam on the Columbia River in 1933, dam counts provided an additional data source. Completion of Bonneville Dam in 1938 provided a central location for estimating returns to the middle and upper Columbia River (WDEW and ODFW 1994~. Coupled with in-river catch records, dam counts permitted calculations of approximate total run size. However, those estimates were better for some populations than for others. They permitted reasonable assessments of total adult populations of sockeye and steelhead, nei
STATUS OF SALMON 91 ther of which were caught in great numbers by ocean fisheries. For coho and chinook salmon, counts at dams and records of in-river catches yielded only partial adult estimates because information on ocean fisheries was lacking. Much later, in the 1970s, coded-wire tag information permitted managers to estimate ocean catch for some populations, so that total run size could be reconstructed. But those reconstructions were accurate, or even possible, only for particular hatchery populations. Ocean fishing of wild salmon could not be calculated, and distribution of hatchery fish could not properly be assumed to apply to wild fish. Snake River Sockeye salmon of the Snake River are listed as endangered under the provi- sions of the ESA. Sockeye passage at Ice Harbor Dam has varied widely but declined sharply after 1976 (Figure 4-31. The only production area now used by Snake River sockeye is Redfish Lake in the headwaters of the Stanley basin. Only four adults returned to the outlet trap there in 1991, one in 1992, and five in 1993; the Idaho Department of Fish and Game (IDFGJ incorporated these adults in captive broodstock programs rather than permitting them to spawn naturally. Sockeye survivors of downstream migrants from Redfish Lake that were cap- tured at the lake outlet and cultured in hatchery facilities to adulthood were released to spawn in Redfish Lake in 1993. Those downstream migrants might have consisted in part or wholly of the progeny of a group of residual sockeye or kokanee found in lake-shoal spawning areas. Those fish have an October matu- ration similar to that reported by Bjornn et al. (1968) for sockeye in the 1960s. 20 10 o lo lo -100 ~ -200 7 O ~00 . ~............ .................. _, _M _. ~ ~ ~ _ , fit -I ~- Escapement--- ~ - ~ Escapement Goal ~ ......... . ~ ~ Cum Escape ~ ---~;;;;;;;,;;;;;1;;;;;,;;;;;,;.,;;,;,,;,,W,l,;;;;41 Moo 1940 1950 1960 1970 1980 1990 Year FIGURE 4-3 Counts of adult sockeye at Ice Harbor Dam (_ symbols), escapement goal (horizontal line), and the cumulative deviation of actual escapement from the escapement goal after 1970 (+ symbols).
92 1.2 cn o o o C!5 0.8 i cn ~ 0.6 3 C o.4 I ~ 0.2 UPSTREAM: SALMON AND SOCIETY IN THE PACIFIC NORTHWEST A:::::: ~ -~-----~ A i':::::::::::::: / \/ \ ~ \ / . ................................... . ~...... ~,,,, 1,,,, 1,,, 1 1,,,, 1, ...... .~ to 92 62 67 72 77 YEAR 82 87 - SNAKE RIVER L. GRAN. SNAKE RIVER ICE HAR. - 25 20 C) 1 5 CO 10 m FIGURE 4-4 Fishway counts of fall chinook at Lower Granite and Ice Harbor dams. The number of fall chinook returning to the Snake River, as counted at Ice Harbor Dam, declined from 1968 to 1976, remained low but stable through 1985, and then increased slightly (Figure 4-4), probably in response to production from Lyons Ferry Hatchery and straying from the Umatilla and Columbia rivers (Chapman et al. 19911. Adult counts at Lower Granite Dam, available since 1975, have ranged from about 350 to 1,000 and for the most recent 5 years averaged about 600 (Figure 4-49. High catch rates of fall chinook reduce the numbers of adults passing Lower Granite Dam. In the late 1980s, ocean fisheries took about 35% of fish, and river catch took 44-63~o of the in-river run (i.e., the survivors of the ocean fishery), so fisheries took nearly 75% of adult recruits (Chapman et al. 1991~. Catch rates have declined in recent years: perhaps 60% of fall chinook are caught. Fall chinook is the only ESA-listed Snake River salmon with high fishing rates. Redd (spawning-bed or nest) counts for wild spring and summer chinook indicate declining numbers of spawners over the last 3 decades (Figure 4-Sa and b). Redd counting in reasonably consistent index areas extends to 1957, although methods have not always remained the same (Chapman et al. 19913. The few counts before that date (see Welsh et al. 1965) cannot be compared directly with post-1957 counts (P. Hassemer, IDFG, personal communication). The spring chinook run of 1994 appears severely depressed and is predicted to total about 20,000 fish at Bonneville Dam, whereas the escapement goal is about 80,000 fish. Counts of spring chinook jacks in 1994 were very low,
STATUS OF SALMON 2500 2000 1 600 1000 600 6 2 1 lo NUMBER OF REDDS L...................... Use WILD IDAHO SPH CHIN O _1 1' ' 1' ' ' ' - 57 62 67 72 77 ' ' ' 1 ' ' ' ' 1 ' '-Y ' 1- ' -' ' ' 1 ' ' ' ' ~ 82 87 92 YEAR NUMBER OF REDDS (Thousande) WILD IDAHO SUMMERS . Y ~............................................ ' ' ' ' 1 ' ' ' ' 1 ' ' ' ' 1- WI ,~1 '~j1 57 62 67 72 93 77 82 87 92 YEAR FIGURE 4-5 i a) Index area read counts for wild spring chinook in Idaho. (by Index area read counts for wild summer chinook salmon in Idaho. foreshadowing a greatly diminished adult run in 19959. Most of the adults returning in 1995 will have migrated downstream as smelts in 1993, when fewer smelts were artificially transported downstream, because so much water was spilled over the dams in that year. Bathe adult return for 1995 appears to be about half the 1994 run at press time.
94 UPSTREAM: SALMON AND SOCIETY IN THE PACIFIC NORTHWEST For summer chinook, the highest read counts in several Snake River index spawning areas occurred in 1957, the year in which completion of The Dalles Dam flooded Celilo Falls. Tribal fisheries at Celilo tended to take more summer and fall chinook than the earlier-migrating spring chinook because fishing access and efficiency increased as flows dropped. In other index areas, high counts occurred during 1957-1961, before other fisheries intensified (see WDF and ODFW 1992, Table 29~. Summer chinook read counts in the Snake River system declined in the 1970s and then rebounded from 1980 to 1988 until the effects of the extended drought that began in the winter of 1987-1988 again depressed the runs. The IDFG counts of Parr in late summer in index areas of streams that produce spring-summer chinook reflect escapement declines (Figure 4-5a and b). They also indicate habitat degradation in some streams, e.g., Bear Valley Creek (Figure 4-5a; Rich et al. 1992~. Chamberlain Creek, a wilderness stream with no commodity-production uses, held more chinook Parr than other Salmon River tributaries (Figure 4-5b). Declines in Parr abundance reflect the combined effects of reduced spawner densities, drought in rearing areas, and conditions in the migration corridor during smolt emigration. Richards and Olsen (1993) compared trends in Snake River spring chinook with trends in ocean production, emphasizing the 1980s. They concluded that trends were similar and that ocean ecological or interrelated inland climatic con- ditions affected both groups. Their analysis covered, at most, the years 1971- 1991. That tends to mask the effects of reductions from the late 1950s to 1971 caused by intensive dam construction on the main Columbia and Snake rivers. Only Lower Granite and Little Goose dams remained to be constructed on the Snake River after 1971. Two groups of steelhead use the Snake River. The first, called the A group, includes fish that pass Bonneville Dam before the end of August. They tend to spend 1-2 years at sea and be shorter than 81 cm (32 in). B-group steelhead pass Bonneville Dam mostly in September and are generally longer than 81 cm, hav- ing spent 2-3 years at sea. Some of the largest steelhead ever caught have belonged to the B group. Total numbers of wild and hatchery steelhead passing Ice Harbor Dam reached a low point in the middle 1970s but since have trended upward (Figure 4-61. Wild A- and B-group steelhead have been distinguished from hatchery fish at Bonneville Dam and in Zone 6 (upstream from Bonneville Dam) fishery only since 1984, when adipose fin removal permitted identification of hatchery fish. A-group steelhead use both the mid-Columbia and Snake rivers. Most B-group fish enter the Snake River. Numbers of B-group steelhead escap- ing the Zone 6 fishery have trended upward very slightly since 1986 (Figure 4-61. Beginning in 1991, efforts to protect wild fall chinook that originate in the Snake River have increased the B-group steelhead.
STATUS OF SALMON 160 140 120 100 80 60 40 20 95 FISH COUNT 1000S "''"\~\'/~'''''''''''''''''''''"'''''''''A"''""'''''"''"'"''''''''/'''''''''''''''''''""''''''~''V''''''' ...4 A: ~ ~ . , ~ , 3 ~- 1 1 1 1 1 1 1 1 1 ~ _ _ 62 67 72 77 YEAR - ICE HARBOR DAM ~ ZONE ~ ESC. WILD 82 87 92 I ZONE ~ ESC. WILD A FIGURE 4-6 Steelhead escapements, upriver summer steelhead. Dam count June 1 October 1. Source: WDF/ODFW 1992. Middle and Upper Columbia River Sockeye salmon in the Yakima River system were extirpated by construction of irrigation storage reservoirs without provision for fish passage. More recently, the National Marine Fisheries Service and the Bonneville Power Authority have investigated re-establishment of sockeye in some Yakima River reservoirs. In general, however, although Snake River sockeye declined sharply over the last 3 decades, Columbia River sockeye have sustained themselves better (Figure 4-7~. Snake River sockeye pass eight-dams, Wenatc~hee River sockeye pass seven, and Okanogan sockeye pass nine to reach spawning areas in the Okanogan River upstream from Wells Dam (Figure 4-8~. It is not completely clear why abun- dance trends for Snake River and Columbia River sockeye differ, but several hypotheses exist. The Columbia River carries about one-fourth to one-third as many yearling or older hatchery salmon and steelhead smelts as does the Snake River. Fishery agencies release ten times more steelhead from Snake River hatcheries than from middle Columbia River hatcheries upstream from Priest Rapids Dam. Total releases of chinook salmon and steelhead from Snake River hatcheries average near 20 million fish. In the early 1960s, Raymond (1979) estimated that fewer than 6 million wild fish were produced from the Snake River drainage. It is possible that large populations of hatchery-produced fish may interact in negative ways with Snake River sockeye.
96 350 300 - U) 250 CE I By UPSTREAM: SALMON AND SOCIETY IN THE PACIFIC NORTHWEST 200 150 100 50 J ~ ~ v/ O 38 1 11,',~ 1,; ,1,1',\1 all Il`V ~11 ~\\ A ~ _ ~ , ~ 1 x, 1 I_ /~'',: \1~ ,~ ., _~' `1 1 A_ 78 83 88 93 43 48 53 58 63 68 73 YEAR Sockeye Run ---- Sockeye Harvest FIGURE 4-7 Estimated numbers of sockeye entering Columbia River and total sockeye catch. The flow of the Columbia River during spring substantially exceeds that of the Snake River. Fish-deflection screens, vertical barrier screens, transportation, and conventional bypasses are absent from Columbia River dams upstream of the mouth of the Snake River but were present in the two uppermost Snake River dams (Little Goose and Lower Granite) during the rapid decline of the Snake River sockeye. Sockeye appear to suffer more damage in bypass systems than do other spring migrants (Johnsen et al. 1990~. Sockeye of the upper Snake River might have originated as residual sockeye in Stanley basin lakes after the partial removal of Sunbeam Dam in the 1920s, whereas Columbia River sockeye have had continuous access to the sea and to their natal areas. Residual sockeye might be less fit for the rigors of anadromy. All or some of the factors just discussed or variants of them might explain the demise of Snake River sockeye. Although summer and fall chinook were treated separately by WDF et al. (1993), Utter et al. (1993) considered summer and fall chinook of the main Columbia River to consist of one population unit from the Hanford Reach through upriver areas. Thus, the summer-fall unit would include fish that managers have termed "upriver bright fall chinook" (URB) and "summer chinook" that spawn as far upstream as the middle reaches of the Wenatchee River, Methow River, and lower Similkameen River. A few summer and fall chinook also spawn in the
STATUS OF SALMON 97 ~ _ ~1 Y~SHINGTON 1 l V ,4~ COLUMBIA RIVER ~ ~ In +~O a OREGON odd+ -TV ~: 1 .~N . ~ CLEAR - TER RIVER ~] - ' -~: l I SALOON RIVER , SNAKE RIVER 1 ~ 'a, +<o+°~ IDAHO FIGURE 4-8 Location of mainstem hydroelectric dams on Columbia and Snake rivers in areas still accessible to salmon. main Okanogan River. Summer-fall chinook populations in the middle Colum- bia River are more abundant now than they were in the 1930s, when counts became available at Rock Island Dam. However, hatcheries now contribute substantially to some of the runs. Five of the 14 summer and fall chinook stocks identified by WDF et al. (1993) from the middle and upper Columbia River were listed as depressed; the other nine were considered healthy. There has not been an in-river fishery specifically for summer chinook in the Columbia River since 1964. Before then, the total catch of these populations probably amounted to over 70%, although the fish had to pass four to six dams that also killed some juveniles and adults. Ocean catch of these fish continues at about 35%.- The runs are now not as resilient as they were in the 1930s, in that losses at hydroelectric dams cause the death of fish that formerly would have been available for fisheries or escapement. The strongest population is the natu- rally spawning fall run that uses the Hanford Reach between McNary Pool and Priest Rapids Dam, the last free-flowing reach in the middle Columbia River. From 1983 to 1991, numbers of spawners in the Hanford Reach ranged from
98 UPSTREAM: SALMON AND SOCIETY IN THE PACIFIC NORTHWEST about 50,000 to 165,000 (WDF et al. 19931. The reach is the largest natural production area for chinook salmon in the entire Columbia River Basin. Summer-chinook runs at Rock Island Dam have been variable, averaging about 16,000 adults and jacks from 1933 to 1992 (Grant, Douglas, and Chelan Public Utility distncts, personal communication; Figure 4-9~. This population has been proposed for ESA listing. Over the same period, combined summer and fall chinook averaged 22,722 adults and jacks. The 1993 summer-chinook count at Rock Island was 13,401 adults and jacks (Fish Passage Center, September 17, 1993~. Mullan (1990) stated that the average summer-fall chinook adult return in 1967-1987 to middle Columbia River tributaries was 15,497, consisting of the Wenatchee (12,012), Entiat (100), and Methow (3,385J rivers. In the same pe- nod, the total number of summer-fall chinook passing Rock Island Dam was about 16,500 adults. Thus, the hatchery supplement to the summer and fall run at Rock Island Dam was about 1,000 adults, or about 6% of the total run. Numbers of middle Columbia River summer-fall chinook between McNary Dam and Pnest Rapids Dam were fairly stable from 1962 to 1982 but then climbed rapidly and peaked in 1987 (Figure 4-10), when escapement reached 96,400 fish (PFMC 19934. The high returns in the 1980s apparently benefited from reduction in ocean fishenes, a result of the U.S.-Canada treaty, and perhaps from relatively high ocean survival. In-river fishing was also cut back to protect the summer component of the summer-fall chinook run and to reduce the catch of steelhead and fall chinook destined for the upper Snake River. Spring chinook populations in the middle Columbia River were low in the 1930s and 1940s (Figure 4-11), when commercial fisheries took up to 86% of the 40 of C 30 o. c z o 3 An 20 10 hi Aid ~ \ \ \ - / ROCK ISL SUMS ER/FALL ,1~ ~ \ \ ol I 171 I I I I I I I I I I I,,, I I I I I I I I I I I I I I I I I I I I I I I I I I I I i I I I I I,, I 1933 1938 1943 1948 1953 1958 1963 1968 1973 1978 1983 1988 1993 YEAR FIGURE 4-9 Fishway counts of adult summer and fall chinook at Rock Island Dam, 1933-1992.
STATUS OF SALMON 180 u' 160 o o 140 5 120 *c 100 z 80 ~ 60 3 do E 20 99 -MCN'RY DAM _ ~ PRIEST RAPIDS DAM .. -HANFORD REACH _ ..... ~4 If\ ............. ,~ Em-i' A o ~ , ,,, _ 62 67 72 77 YEAR . ~\ . ~_ ;;,, ~\ ' ~ 82 87 g2 FIGURE 4-10 Estimated escapement of adult summer and fall chinook at McNary and Priest Rapids dams. Numbers of fish spawning in Hanford Reach was difference between counts at two dams. runs. Populations grew with the Grand Coulee Fish Maintenance Project and with reduction in catches during the 1940s. Some of the increase resulted from absence of directed fisheries in the Columbia River and cessation of sport fishing in tributaries. Dam-related mortality of smalls and losses of adults between dams substituted for fishing mortality. Despite the upward trend in numbers of spring chinook passing Rock Island Dam during the 1980s, the overall number of upriver spring chinook entering the Columbia River has declined (Figure 4-11~. Of the 16 native stocks of spring chinook from the middle and upper Colum- bia River identified by WDF et al. (1993), 15 were classified as depressed and one (Asotin Creek) as critical. Many of the depressed stocks contained fewer than 500 spawning adults. Some of the stocks remained stable through the middle 1980s but have experienced severe short-term declines. No naturally spawning coho populations remain in the Columbia River basin upstream of The Dalles Dam, although they once inhabited many subbasins in the middle Columbia and Snake River systems (The Wilderness Society 1993~. They were extirpated by a combination of habitat loss, overfishing, and agency policy that relied on hatcheries. Likewise, chum populations have disappeared from the middle Columbia and Snake rivers. Nehlsen et al. (1991) listed chum extinctions in the Umatilla and Walla Walla rivers but did not note when the extinctions occurred or their causes. Pink salmon are not known from the Columbia River in historical times. In the middle Columbia region, summer steelhead are now managed for both
100 UPSTREAM: SALMON AND SOCIETY IN THE PACIFIC NORTHWEST 400 3so 300 250 cn 200 - cn 3 o S cn ~_ lL 30 O 25 a) Q 20 7 15 10 O Minimum 150 so ,: ~Upriver~ Run I - _ ~ 1 /~ W ~ O- 1 1 1 1 ~ I I I I I I I I I I I I I I I ~ I ~ I ~ I I I I ~ I I I I I I I I I I I I I I I I I I I I I 1935 1 945 1955 1965 1975 1985 Rock Island Dam Counts I i I I I I I ~ I r I r I T r r rr r T rT ~ ~ I I I I I rErtl T rr~ Tr~ rr rTrT I T r rttrn 1935 1945 1955 1965; 1975 1985 FIGURE 4-1 1 Estimated minimum number of upriver spring chinook entering Columbia River and ascend~ng Rock Island Dam (dam counts not available for 1968-1972). From Oregon Department of Fish and Wildlife and Washington Department of Fisheries 199 lc and unpublished data, Chelan County Public Utility District. natural and hatchery production, a change from 10-15 years ago, when hatchery production dominated the management strategy of the Washington Department of Game (now the Washington Department of Fish and Wildlife). Adult summer steelhead for hatchery broodstock are trapped at Wells Dam, and progeny are distributed to the Wenatchee, Methow, and Okanogan rivers. Natural spawning
STATUS OF SALMON 35 30 a) to O 25 - 20 o O ]5 Q 10 ~ V A /~\ 0- , I l I l I l I l l I l l I l l l l I l I l I l I l l l l l rl I I l I l l l l l l l--r--r~l I l l l I l l l l I l l 1 933 1941 1949 1957 1965 1973 1981 1989 101 FIGURE 4-12 Number of steelhead ascending Rock Island Dam, 1933-1991. Unpub- lished data from Chelan County Public Utility District. Occurs, but only marked hatchery fish can be retained by sport anglers. Figure 4- 12 shows trends in counts of steelhead passing Rock Island Dam. Steelhead in the middle Columbia are more abundant now than they were in the 1930s as a consequence of extensive hatchery production. However, Nehlsen et al. (1991) listed 11 extinct stocks of summer steelhead from the middle and upper Columbia basin and from the Snake River system, and 13 of 15 wild stocks identified by WDF et al. (1993) were considered depressed. The mid-Columbia steelhead are currently under consideration for ESA listing as threatened or endangered. Win- ter steelhead now occur only as far inland as Fifteenmile Creek near The Dalles, Oregon. Sea-run cutthroat trout apparently are not found above Bonneville Dam and are now extinct in the Wind and Klickitat rivers (Nehlsen et al. 1991~. Lower Columbia River Many of the salmon populations inhabiting the lower Columbia River (all tributaries below Bonneville Dam) have been altered by hatchery and dam prac- tices, and the total number of salmon and steelhead hatcheries operating in the
102 UPSTREAM: SALMON AND SOCIETY IN THE PACIFIC NORTHWEST lower Columbia River system is probably the greatest of any area throughout the range of Pacific salmon. The National Marine Fisheries Service (Johnson et al. 1991) believed that native coho salmon in lower Columbia River tributaries had become extinct, partly because of widespread population transfers and broodstock selection prac- tices (e.g., selection of early returning adults) at hatcheries. Although of mixed hatchery and wild origin, 17 naturally spawning stocks of coho in Washington tributaries of the lower Columbia River were recognized by WDF et al. (19931. All were considered depressed, but spawning counts were available for only three of these stocks. In contrast, only two of the 17 stocks of chinook salmon from Washington tributaries of the lower Columbia were considered depressed by WDF et al. (19933; all others were classified as healthy on the basis of stable or increasing population figures. The two depressed stocks (South Fork Toutle and Green River fall stocks) were both part of the Toutle River system that drains the northern and western slopes of Mount St. Helens and were both affected by the 1980 volcanic eruption. Most of the other stocks were heavily supported by hatchery production. Chum salmon populations are seriously depressed in the lower Columbia. Nehlsen et al. (1991) stated that only about 2,000 chum were spawning in lower Columbia River tributaries, a figure that represents 0.5% of their historical abun- dance. WDF et al. (1993) listed three Washington chum stocks from the lower Columbia and considered two of them depressed. Forestry and agricultural prac- tices, urbanization, pollution, and overfishing in mainstem fisheries directed at coho and fall chinook were believed by Nehlsen et al. (1991) to have contributed to chum declines. Five summer and 18 winter steelhead stocks were identified from Washing- ton tributaries of the lower Columbia River (WDF et al. 1993~. Of the summer steelhead stocks, two were considered depressed, and the status of the other three was unknown. Of the winter steelhead stocks, 12 were listed as depressed, two as healthy (the only two for which escapement goals had been established), and four as of unknown status. Nehlsen et al. (1991) observed that hatchery programs dominated steelhead in the lower Columbia. Although they did not list the native stocks in the Washougal, North Fork Lewis, Cowlitz, and White Salmon rivers as extinct, Nehlsen et al. ~ 1991) felt that few native steelhead existed in these drain- ages. The overall depressed condition of steelhead in the lower Columbia River differs somewhat from the condition of coastal populations in Washington and Oregon, where greater percentages of populations are not declining. Little is known about the status of sea-run cutthroat populations in the lower Columbia River. Sea-run cutthroat are known to spawn in a number of Washing- ton tributaries, but long-term data are lacking. The species spends an extended period in freshwater and estuarine environments before migrating seaward and is sensitive to habitat degradation and estuary loss (Trotter 19893. It might also be
STATUS OF SALMON 103 vulnerable to negative interactions with hatchery fish and to overfishing (Trotter et al. 19933. Nehlsen et al. (1991) considered sea-run cutthroat to be depressed throughout the region. COASTAL WASHINGTON, OREGON, AND CALIFORNIA WDF et al. (1993) inventoried 32 chinook salmon stocks from the Washing- ton coast: 22 fall and 10 spring-summer populations. Overall, about two-thirds of the stocks were classified as healthy according to the criteria in Table 4-2. How- ever, a greater percentage of fall chinook (72%) than spring-summer chinook (40%) were considered healthy. Only a single fall chinook stock was listed as depressed, whereas 40% of the spring-summer chinook stocks were depressed. According to WDF et al. (1993), one-fourth of the coastal chinook stocks in Washington are supported by mixed wild and hatchery production. Nickelson et al. (1992) recognized 55 wild populations of coastal chinook in Oregon, ranging from the Nehalem River on the north coast south to the Winchuck River near the California border. Life-history characteristics of coastal Oregon chinook vary greatly. Timing of spawning runs ranged from April to December, and age at maturity varied from 2 to 6 years. Juvenile chinook in Oregon coastal rivers usually spend less than a year in freshwater, and extended estuary resi- dence is well documented. Populations from the Elk River northward usually rear in nearshore waters off British Columbia and southeastern Alaska; those from the Rogue River southward tend to rear off southern Oregon and northern California (Nickelson et al. 19921. Of the 55 known naturally spawning chinook populations, 30 were consid- ered healthy, eight were of special concern, eight were depressed, and the status of nine populations was unknown. Although the overall status of northward- migrating chinook seemed to be favorable along the Oregon coast (Lichatowich 1989), southward-migrating populations were in decline. According to Nickelson et al. (1992), 70% of the southward-migrating chinook populations i.e., most populations from the Rogue River southward were judged to be of special concern or depressed. In addition, runs of southward-migrating chinook have experienced wide variation in abundance over the last decade. For example, spring and fall chinook salmon entering the Rogue River increased from 30,000 in 1983 to about 200,000 in 1988 and then declined to about 30,000 in 1992. Although many northward-migrating chinook populations from coastal Or- egon were classified as healthy by Nickelson et al. (1992), there is evidence that populations are less abundant now than at the time of colonization. Lichatowich (1989) suggested that apparent increases in northward-migrating chinook could be an artifact of the period of record; if production at the time of the first stock inventories was already severely depressed by habitat destruction, comparisons between historical and present-day populations would imply that runs are stable or increasing. The Wilderness Society report (1993: Table A-2, A-4) compared
04 UPSTREAM: SALMON AND SOCIETY IN THE PACIFIC NORTHWEST estimates of natural runs of coastal chinook and coho in about 1900 (virtually all of which were wild fish) with estimates of current runs (which include both wild and hatchery fish). Of the 10 Oregon river systems examined, chinook in eight are less abundant now than in 1900, and seven of the eight populations were estimated to have declined to about half of their former abundance or less. The Wilderness Society ~ 1993) comparison of current and historical chinook abundances in selected Oregon rivers, as well as a similar comparison for coho, must be viewed somewhat cautiously because estimates of abundances now and in 1900 usually rely on comparisons of periods for which continuous records of abundance are lacking. Turn-of-the-century population estimates were based on reported catch weights, not numbers of fish caught; this introduced two potential sources of error: inaccuracies in reported total catch weights and assumptions concerning conversion of weight to numbers of fish. Nevertheless, decreases of 50% or more from estimated turn-of-the-century abundances to current abun- dances should probably be considered indicative of substantial overall declines in naturally spawning populations, especially because current populations include many hatchery fish. The committee recognizes, however, that 1900 may have been a time of relatively high ocean productivity for salmon in the Pacific North- west (Smith and Moser 1988, CalCOFI Rep. 29, Baumgartner et al. 1992, CalCOFI Rep. 33), and that some of the differences in chinook and coho abun- dance along the Oregon coast between 1900 and the present might reflect differ- ences in ocean conditions. Moyle (1976) noted that chinook salmon once spawned as far south as the Ventura River in southern California but now occur only in the Sacramento-San Joaquin and some other river systems of northern California. The number of extant wild chinook populations in California has apparently not been estab- lished, but virtually all known populations have been reported to be at some risk of extinction (Higgins et al. 1992, Frissell 1993, The Wilderness Society 19931. Five chinook stocks from the Sacramento-San Joaquin River system and four stocks from the upper Klamath River system (actually spawning in southern Oregon) were declared by Nehlsen et al. ~ 1991 ~ to have become extinct within the last 150 years. According to The Wilderness Society (1993), all remaining spring and summer chinook populations in northern California are at high risk of extinc- tion. Fall chinook, although not as imperiled in most areas, nevertheless are seriously depressed. Coho salmon in Washington's coastal streams were considered by WDF et al. (1993) to be in generally good condition. Of the 26 known stocks, 17 were classified as healthy and nine as of unknown status. However, about 70% of the coho populations contained both wild and hatchery-spawned fish; hence there was a heavy reliance on hatchery production of coho on the Washington coast. Coastal Oregon coho populations from the Necanicum River in the north to the Winchuck River in the south have been termed Oregon Coastal Natural (OCN) coho. Although 94 spawning populations are known (Nickelson et al. 1992), the
STATUS OF SALMON 105 most abundant OCN coho occur between the Nehalem and Coquille Rivers. Overall, OCN coho constitute the largest aggregate of coho populations in the United States outside Alaska. About half the 94 recognized populations occur in small streams that drain directly to the ocean. Some of these small streams were grouped by the Oregon Department of Fish and Wildlife according to geographic region and yielded a total of 55 coastal "populations." On the basis of spawner surveys over the last 20 years (Nickelson et al. 1992), six were classified as healthy, two were of special concern, 41 were depressed, and six were of un- known status. Since the middle 1970s, escapement of OCN coho to Oregon's coastal rivers and lake systems has usually been below the 200,000-adult target escapement established by the Oregon Department of Fish and Wildlife. The commercial and sport catch of these fish has also dropped markedly over the last 2 decades (Figure 4-131. Estimated declines of OCN coho since the turn of the century have been even larger. The Wilderness Society (1993) stated that runs in 10 coastal rivers declined from about 50,000-440,000 in 1900 to 5,000-55,000 fish currently, an average reduction of more than 80% in spite of extensive recent hatchery support for OCN coho. Although Nehlsen et al. (1991) noted only one coastal Oregon coho extinction, The Wilderness Society report (1993:25) stated that "in the past few years, detailed stream-specific surveys in Oregon and Cali- fornia have documented widespread extinctions of coho salmon in streams known to have supported spawning populations as recently as the 1960s and 1970s." Specific locations of the extinctions were not given, and it is not clear whether the statement referred to population units in small tributaries or to entire spawning runs. Coho salmon in California apparently do not undertake extensive oceanic migrations but remain within a few hundred kilometers of their natal streams while at sea. They occur primarily in small to mid-size coastal rivers and creeks as far south as Monterey Bay. Attempts to increase runs of coho salmon in the Sacramento River by planting hatchery-produced fry in the 1950s were largely unsuccessful (Moyle 1976~. Nehlsen et al. (1991) treated coastal coho popula- tions from San Francisco to Oregon as a single stock with a moderate risk of extinction. Higgins et al. (1992) identified 20 separate coastal populations, about one-third of which were felt to be at high risk of extinction. The Wilderness Society (1993) considered native Sacramento River coho to be extinct and most of the coastal populations to be either threatened or endangered. Brown et al. (1994) reviewed the status of coho populations in California and concluded that coho had undergone dramatic declines from historic levels statewide. Coho salmon no longer pass Benbow Dam on the Eel River, although that river once supported runs of 5,000-25,000 fish. Brown et al. (1994) noted that the decline occurred despite substantial hatchery programs to raise coho at five hatcheries in California. Overall, although historical coho runs in California once totaled in the hundreds of thousands, estimated numbers of naturally spawning adults are now fewer than 5,000 (The Wilderness Society 19933.
106 in . _ 11 o in o UPSTREAM: SALMON AND SOCIETY IN THE PACIFIC NORTHWEST 1 600 1 400 1200 1 000 800 600 400 200 0q total run size Escapement ;\ \ \ l ''art __] _ 1955 1960 ~ 965 1970 1975 1980 1985 1990 Adult Year FIGURE 4-13 Estimated numbers of Oregon Coastal Natural (OCN) coho, 1953-1993. Area under upper line represents total number of adult or maturing adult fish; shaded area under lower line is number of naturally spawning salmon in Oregon's coastal streams. Difference between upper and lower lines represents commercial and sport catch, includ- ing in-ever sport catch. From 1953 to 1959, escapement results were unavailable for Oregon coastal lake systems; therefore, abundance of lake-denved coho for each year was assumed to be 7.5% of total and 7.1% of escapement averages of lake contributions to total run size and escapement during 1960-1970. Data courtesy of Oregon Department of Fish and Wildlife. Chum salmon have been recorded along the West Coast south to the San Lorenzo River in Monterey Bay (Moyle 1976), but the current southernmost population of chum salmon is in the lower Smith River of northern California (Higgins et al. 19923. According to The Wilderness Society (1993), the Smith River contains the sole remaining self-reproducing California population, but adults occasionally stray into other rivers. There are 26 known wild chum popu- lat~ons on the Oregon coast, most of which vary greatly in abundance from year to year (Nickelson et al. 19921. Overall abundance is considerably less than in the early twentieth century. The chum run to Tillamook Bay in 1928 was esti- mated to be about 650,000 adults, but since the 1960s the maximum estimated
STATUS OF SALMON 107 run has been only 26,000 fish. Over the last 20 years, some runs have been relatively stable, although at reduced levels. Partly because there were no clear declines over this period in some Oregon rivers, Nickelson et al. (1992) declared 10 of the 26 chum salmon populations to be healthy, but 12 were listed as vulnerable because of their small run size. Of the 14 chum salmon populations known from the Washington coast, about two-thirds were classified as healthy by WDF et al. (1993), and the others were unknown. Pink salmon are absent from the Washington and Oregon coast, and it is not clear whether they were found in abundance in any coastal streams within the last 2 centuries. Pink salmon do not now exist in California, although Nehlsen et al. (1991) reported that extinctions of Klamath and Sacramento pink salmon runs were recent. As late as the 1970s, a few stray pink salmon were observed in some northern California rivers, but the only reproducing population occurred in the Sacramento River system (Moyle 1976~. The Sacramento River population was surely a zoogeographic enigma, being over 1,000 km south of the southern limit of pink salmon distribution in Puget Sound. Sockeye salmon depend on lakes for much of their freshwater rearing. A few lakes are present in some coastal Oregon and California river systems, but no sockeye populations occur there now. Historically, small sockeye runs might have existed in the Sacramento River and in the upper Klamath River (Moyle 1976), but Nehlsen et al. (1991) did not list them among recent sockeye extinc- tions. Three sockeye stocks from the Washington coast are known; one is consid- ered healthy, one is listed as depressed, and the status of the third is unknown (WDF et al. 19931. The status of fully half the steelhead stocks in Washington coastal streams was unknown, according to WDF et al. (19931. Summer steelhead were the least understood; eight of the nine stocks were listed as of unknown status. About half the winter steelhead stocks were classified as healthy. Only two of the 31 coastal stocks were considered depressed. Only about 10% of the coastal steelhead populations are partially supported by hatchery production. Winter steelhead occur in nearly all coastal Oregon streams. Summer steel- head are native only to the Siletz, Umpqua, and Rogue rivers. Some Rogue River steelhead exhibit an unusual life-history variation termed "half-pounders"; these spend only 3-4 months at sea before returning to the river immature. The only other populations of steelhead in which this life-history variation occurs are in the Klamath and Eel rivers of California. Otherwise, summer and winter steelhead exhibit a typically variable life cycle of 1-4 years of freshwater residence and 1- 4 years at sea, although 2 years in freshwater and 1-3 years at sea is the norm. Steelhead can reproduce more than once in a lifetime (a phenomenon called iteroparity); however, the incidence of repeat spawning is low, ranging from 3- 20% in coastal streams (Nickelson et al. 1992~. Juvenile steelhead entering the ocean do not stay in coastal waters but move offshore, where they migrate to the Gulf of Alaska and the western North Pacific Ocean.
08 UPSTREAM: SALMON AND SOCIETY IN THE PACIFIC NORTHWEST The Oregon Department of Fish and Wildlife identified a number of wild Oregon coastal steelhead populations, some of which were very small and poorly known. However, its 20-year analysis of current status generally included only streams in which the annual sport catch exceeded about 200-300 fish (Nickelson et al. 1992~. Of the 24 wild winter steelhead populations selected for analysis, 19 were considered depressed and five healthy. Two-thirds of the summer steelhead populations were also considered depressed. A strong correlation was noted among winter steelhead population trends from 1980 to 1990 in most streams. After higher than average runs in the mid-1980s, in the 3-year period from 1988- 1990, 20 of the 24 winter steelhead populations were below the 20-year average in abundance for all 3 years, and four of five summer steelhead populations (two populations were established from nonnative populations) were below average for all 3 years. Historically, steelhead occurred in coastal rivers throughout the entire length of California, ranging as far south as the Tijuana River in northern Mexico. Summer steelhead are now limited to the Eel River and streams northward; winter steelhead are now found as far south as the Ventura River in southern California. Nehlsen et al. (1991) recognized five summer steelhead stocks at moderate to high risk of extinction in northern California. Higgins et al. (1992) identified 11 summer steelhead stocks at risk, eight of which were classified as endangered. Winter steelhead populations in northern California are generally in much better condition than summer steelhead. Most are thought to have been fairly stable in recent decades, but nearly all populations have declined somewhat over the last 5 years (Wilderness Society 1993~. In central and southern Califor- nia, winter steelhead are at low levels, although the populations might always have been relatively small in this warm, dry region. Light (1987) assessed coastwide abundance of steelhead and estimated an average annual abundance of about 1.6 million steelhead adults, of which half were wild and half were of hatchery origin ~ Figure 4-143. However, the propor- tion of wild steelhead was greatest in areas with lower human populations (Alaska and British Columbia) and in California, where huge areas of steelhead habitat have been lost in the Sacramento and San Joaquin rivers. Essentially, California now relies on coastal rivers and streams for most steelhead production. Light (1987) noted that abundance in the 1980s was about the same as it had been in the 1970s but that the fraction of hatchery fish was greater in the later period. Richards and Olsen (1993) found similar production trends in steelhead, citing the Washington Department of Wildlife (1992~. They concluded that ocean conditions were primarily responsible for recent declines in steelhead abundance. Sea-run cutthroat trout are distributed along Pacific Northwest coastal streams south to the Eel River. Like steelhead, sea-run cutthroat rear in fresh water for up to several years, but unlike steelhead, they sometimes spend ex- tended periods in estuaries and most do not travel far from natal streams while at sea (Trotter 1989~. Repeat spawning occurs in sea-run cutthroat trout, perhaps to
STATUS OF SALMON 100 80 60 40 20 o PERCENT WILD ALASKA CANADA COAST - SH.COL. RIVER COAST ORE. CALIFORNIA 109 FIGURE 4-14 Percentages of wild steelhead in Alaska, British Columbia, coastal Wash- ington and Puget Sound, Columbia River Basin, coastal Oregon, and California. Source: Light 1987. a greater extent than in steelhead. Anadromous adults sometimes mate with nonmigratory stream-dwelling cutthroat. Of all species of Pacific salmon, sea- run cutthroat trout are perhaps most poorly known, because they are not commer- cially fished and recreational anglers are not required to maintain punchcard records. No recent extinctions of sea-run cutthroat trout in Washington, Oregon, or California were noted by Nehlsen et al. (1991), but these authors believed the species to be in decline. Ninety two sea-run cutthroat populations from the Oregon coast have been identified. Nickelson et al. (1992) believed that many of these populations had declined before 1980, in that population surveys of central Oregon streams be- tween 1980 and 1990 showed no substantial trends of decline. Sea-run cutthroat from the North Umpqua River have recently been petitioned for status review under the ESA. Returning adults at Winchester Dam declined from about 1,000 in 1946-1956 to fewer than 100 wild fish by 1960 and have remained at ex- tremely low levels. Higgins et al. (1992) identified four stocks from northern California that were felt to be declining, and The Wilderness Society (1993) considered all coastal sea-run cutthroat to be threatened. The status of sea-run cutthroat was not assessed by WDF et al. ~ 1993) for Washington's coastal streams, but Trotter et al. (1993) believed the species to be in general decline in Washing
0 UPSTREAM: SALMON AND SOCIETY IN THE PACIFIC NORTHWEST ton, in part because of supplementation of streams with hatchery-produced coho salmon. WILLAPA BAY A CASE STUDY Although no single river basin is representative of the Pacific Northwest coast as a whole, the Willapa Bay Basin is an instructive example of the interac- tions among habitat alteration, fisheries, and hatchery management in a major drainage system without dams. The Willapa Bay Basin is a large, highly produc- tive estuary and watershed on the southwestern Washington coast. At mean high water, the bay covers about 120 mi2 (311 knit), and the surrounding watershed encompasses 1,060 mi2 (2,745 km21. There are 745 rivers and streams with over 1,470 linear miles (2,366 km) in the basin (Phinney and Bucknell 1975, Suzumoto 19921. There are at least 15 known stocks of salmon in the basin (WDF et al. 1993), mostly fall chinook, chum, coho, and winter steelhead. Land- and water-management activities in the basin have been typical of coastal areas with neither dense human populations nor large hydroelectric projects. Much of the watershed is commercially forested and has been logged since before the turn of the century. Lowlands have been farmed and diked for pasture. Habitat in many streams in the basin has been altered because of those management actions (Pyle 1986), but to date there are no comprehensive surveys of stream conditions in Willapa Bay tributaries. A fishing fleet has operated in the basin since early in the century, catching salmon originally by trapping and netting and later primarily by gill netting. There are about 19,000 inhabitants of the basin, most of whom depend on the natural resources of this productive area. In spite of the typical panoply of human impacts, the Willapa Bay ecosystem remains in remarkably good condition. The bay is one of the cleanest estuaries in the contiguous United States and commercial shellfish production on the exten- sive mudflats is among the highest in the world. Of the 15 salmon stocks identified by WDF et al. (1993), nine were consid- ered healthy, five of unknown status, and only one (North River chinook) de- pressed. By most measures, then, Willapa Bay is believed by many to be a healthy, productive system for salmon and other natural resources. However, wild salmon inhabiting the Willapa Bay Basin are in serious trouble. Nehlsen et al. (1991) listed native Willapa Bay coho at high risk of extinction from high catch rates in mixed-stock fisheries and negative interac- tions with hatchery fish. The primary fisheries-management objective in the Willapa Bay Basin is the production and capture of hatchery fish, and hatcheries have been present there for almost a century. Of the six salmon hatcheries constructed since 1895, three are still operational. Early in the l900s, combined catches of chinook, coho, and chum salmon in Willapa Bay occasionally ex- ceeded 250,000 returning adults; by the 1960s, the terminal fishery captured only 30,000-40,000 fish (Suzumoto 19921. An aggressive program to increase hatch
STATUS OF SALMON 111 cry production was undertaken, and by the 1980s, total catches were 100,000- 240,000 salmon, with the greatest numerical increases occurring in coho and chum. Total numbers of eggs taken at the hatcheries from returning adults increased from fewer than 5 million in the l950s to more than 40 million eggs in the 1980s, and the number of hatchery smelts went from fewer than 5 million to more than 20 million in the same period. Willapa Bay Basin hatcheries recently produced 11% of the fall chinook and 7% of the coho released by all state- operated hatcheries in Washington (Suzumoto 1992~. Many of the hatchery fish are released directly from rearing facilities, but chinook, coho, and chum fry have all been planted in the basin's streams to increase returns to hatchery trapping locations and to bolster declining wild runs. Weirs at the hatcheries prevent many fish from spawning naturally. In some years, fish are allowed to pass upstream, but only after the annual egg-take quota is met. Some of the hatchery releases were offspring of fish from outside the basin. At least six nonnative coho populations and eight nonnative chinool: populations are known to have been cultured and released from the Willapa, Nemah, and Naselle hatcheries from 1952 to 1990 (Suzumoto 1992J. Chum salmon at the hatcheries have been native Willapa Bay Basin populations, although hatchery chum propagation was discontinued in the 1980s. Streamside egg boxes were used in the 1970s and 1980s to augment chum and coho in a number of basin streams; more recent enhancement projects have used in situ egg-incubation chambers (plastic buckets) that hold up to 10,000 eggs. Catch rates of Willapa Bay Basin chinook and coho have been estimated to be in excess of 70% (Table 4-5 J. a high fishing rate that has occurred among salmon populations elsewhere along the Pacific Northwest coast in rivers with large hatchery outputs (Nickelson et al. 1992~. Although most of the catch has TABLE 4-5 Average Catch and Catch Rates of Coho and Chinook in the Willapa Bay Basin for the Years 971_1991a Coho Chinook Escapement31,60016~500 Catch in Willapa Bay44,50014,200 Washington mixed-stock catch42.00017~000 Alaska and B.C. catch24,00019,000 Total run size142,10066,700 Total catch1 1 1,00050,000 Catch rate78%75~o aCatches in Willapa Bay and escapement from Suzumoto (1992). Interception by fisheries in British Columbia and Alaska based on assumed interception rates of 40% for fall chinook and 20% for coho.
2 UPSTREAM: SALMON AND SOCIETY IN THE PACIFIC NORTHWEST TABLE 4-6 Comparison of Historical Numbers (Based on Early 20th-Century Estimates) of Spawning Salmon with Recent Escapement Goals and Hatchery Use of Returning Adults for the Willapa Bay Basina Recent Number Number Recent Goal (% of Hatchery Naturally Spawning Historical Escapement Historical Take Spawning per Escapement Goal Run Size) (ho of Run)Adults Kilometer Chinook 72 000-122,000 15,000 12-21% 80% 3,000 2.4 Coho 64,000- 108,000 8.000 7- 12% 80% 1,600 0.6 Chum 215,000-366,000 35,400 10-16% 0% 35,400 42.0 aAbout 2,450 km of streams was assumed to be available to coho, 1,170 hen to chinook, and 830 km to chum salmon. Source: Suzumoto 1992. been of hatchery origin, exploitation rates of wild fish have also been high. Compared with historical estimates of salmon returns, which occasionally ex- ceeded 500,000 fish, recent escapements are only a small fraction of former runs (Table 4-61. Coho appear to be most severely depressed, but both chinook and chum salmon are also far below estimated escapements of early the twentieth century. Large numbers of coho and chinook are removed from streams for artificial propagation, so the actual number of naturally spawning salmon in the basin is less than 10% of the historical runs. Naturally spawning coho have not been counted annually routinely, because the basin has been managed for hatchery production of this species; but spawning counts of chinook and chum are available. From 1968 to 1991, chum spawning counts for the Willapa Bay Basin averaged about 28,000 which agrees reasonably well with the escapement goal of 35,000 (Table 4-61. However, average annual chinook spawning counts from 1987 to 1991 averaged about 15,000 (Suzumoto 1992), a figure considerably greater than the estimate of 3,000 naturally spawn- ing fish in Table 4-6. Suzumoto (1992) pointed out that most of the chinook and coho observed in recent spawning surveys were strays from hatcheries. Never- theless, the estimate of naturally spawning chinook in Table 4-6 might not accu- rately reflect the most recent survey information. Even allowing for error in the estimates of naturally spawning chinook and coho, the number of salmon using the basin's streams is still far below what the ecosystem is capable of supporting. For example, the Oregon Department of Fish and Wildlife recommends spawning densities of 24 adult coho per kilometer (about 40 adults per mile) for adequate seeding of Oregon's coastal streams.
STATUS OF SALMON 113 Recent coho escapement to the Willapa Bay tributaries might be only about one- tenth that target (Table 4-6~. Management of Willapa Bay salmon for hatchery production has resulted in an underuse of the natural rearing capacity of the drainage system. If most of the naturally spawning coho and chinook in streams with artificial production facilities are hatchery strays, the number of wild fish and the genetic integrity of native populations must indeed be low. In addition to causing substantial declines of wild populations of salmon in the Willapa basin, the combination of habitat alteration, high catch rates, and removal of fish at hatcheries might be depriving the aquatic ecosystem of an important seasonal source of nutrients. To judge by reductions in naturally spawning fish over the last century, the basin has lost more than several thousand metric tons of salmon tissue each year (Table 4-7~. Present loadings of salmon carcasses and their nutrients are now generally less than 10% of historical levels (Tables 4-7 and 4-8J. It is likely that this absence of nutrient capital has further reduced the capacity of the Willapa basin to produce fish, shellfish, and other important aquatic resources and has led to a long-term decline in ecosystem productivity. The current condition of wild salmon in the Willapa Bay Basin illustrates a systemic problem along the Pacific Northwest coast: habitat, hatchery, and fish- ery management decisions have failed to protect the natural capacity of these areas to produce salmon. Large annual investments in artificial-production facili- ties coupled with various degrees of habitat losses and high exploitation rates have driven wild populations down while increasing the importance of hatchery runs. So dependent is the basin on hatchery salmon production that if the flow of hatchery smelts were stopped, the runs would probably experience major de- clines. TABLE 4-7 Comparison of Historical and Recent Return of Salmon-Carcass Biomass to Willapa Bay Basin Streamsa Historical Recent Mean Biomass Biomass Body Historical Returned to Recent Returned to Weight Escapement Streams Escapement Streams (kg) to Streams (metric tons) to Streams (metric tons) Chinook 8.9 72,000- 122,000 641 - 1,086 3,000 27 Coho 3.7 64,000- 108,000 237-400 1,600 6 Chum 5.0 215,000-366,000 1,075-1,830 34,500 172 Total 351,000-596,000 1,953-3,316 39,100 205 aHistoric run-size information from Suzumoto (1992). Current salmon run sizes based on fishery escapement goals and assumation that 80~o of chinook and coho are taken for hatchery use and removed from ecosystem.
4 UPSTREAM: SALMON AND SOCIETY IN THE PACIFIC NORTHWEST TABLE 4-8 Histoncal and Recent Annual Loading of Salmon-Carcass Phosphorus, Nitrogen, and Total Biomass to Willapa Bay and Its Tnbutanesa Delivery to Streams (kg/km of stream length) Delivery to Willapa Bay (kg/ha of surface area) Historical Recent Historical Recent Phosphorus 3.0-5.0 0.30 0.23-0.38 0.02 Nitrogen 82-140 9.0 6.3-10.70 0.69 Total biomass 823-1,400 86.0 63-107 6.90 aCarcass biomass assumed to be 0.364% phosphorus and 10.0% nitrogen by wet weight.