U.S. Dept Commerce/NOAA/NMFS/NWFSC/Publications

NOAA-NWFSC Tech Memo-25: Status Review of Pink Salmon from Washington, Oregon, and California Coast



As outlined in the Introduction above, NMFS considers a variety of information in evaluating the level of risk facing an ESU. Aspects of several of these risk considerations are common to all pink salmon ESUs. These are discussed in general below; more specific discussion of factors for each of the ESUs under consideration here can be found in the following sections. Because we have not taken future effects of conservation measures into acount (see Introduction), we have drawn scientific conclusions about the risk of extinction faced by identified ESUs under the assumption that present conditions will continue. Future effects of conservation measures will be taken into account by the NMFS Northwest and Southwest Regional Offices in making listing recommendations.

Absolute Numbers

The absolute number of individuals in a population is important in assessing two aspects of extinction risk. For small populations that are stable or increasing, population size can be an indicator of whether the population can sustain itself into the future in the face of environmental fluctuations and small-population stochasticity; this aspect is related to the concept of minimum viable populations (MVP) (Gilpin and Soulé 1986, Thompson 1991). For a declining population, present abundance is an indicator of the expected time until the population reaches critically low numbers; this aspect is related to the concept of driven extinction (Caughley 1994).

In addition to total numbers, the spatial and temporal distribution of adults is important in assessing risk to an ESU. Spatial distribution is important both at the scale of river basins within an ESU and at the scale of spawning areas within basins ( metapopulation structure). Temporal distribution is important both among years, as an indicator of the relative health of different brood-year lineages, and within seasons, as an indicator of the relative abundance of different life history types or runs.

Traditionally, assessment of salmon populations has focused on the number of harvestable or reproductive adults, and these measures comprise most of the data available for Pacific salmon. In assessing the future status of a population, the number of reproductive adults is the most important measure of abundance, and we focus here on measures of the number of adults escaping to spawn in natural habitat. However, total run size (spawning escapement + harvest) is also of interest because it indicates potential spawning in the absence of harvest. Data on other life history stages (e.g., freshwater smolt production) can be used as a supplemental indicator of abundance.

Because the ESA (and NMFS policy) mandates a biological review that focuses on viability of natural populations, we attempted to distinguish natural fish from hatchery produced fish. All statistics are based on data that indicate total numbers or density of adults that spawn in natural habitat ( naturally spawning fish ). The total of all naturally spawning fish ( total escapement ) is divided into two components: Hatchery produced fish are reared as juveniles in a hatchery but return as adults to spawn naturally; natural fish are progeny of naturally spawning fish.

Historical Abundance and Carrying Capacity

The relationship of current abundance and habitat capacity to that which existed historically is an important consideration in evaluating risk for several reasons. Knowledge of historical population conditions provides a perspective of the conditions under which present stocks evolved. Historical abundance also provides the basis for establishing long-term population trends. Comparison of present and past habitat capacity can also indicate long-term population trends and problems of population fragmentation.

Although the relationship of present abundance to present carrying capacity is important for understanding the health of populations, the fact that a population is near its current capacity does not in itself mean that it is healthy. If a population is near capacity, there will be limits to the effectiveness of short-term management actions to increase its abundance, and competition and other interactions between hatchery and natural fish may be important considerations because hatchery supplementation will further increase population density in a limited habitat.

Quantitative assessments of habitat are quite rare, although rough estimates of carrying capacity are frequently made for setting management goals. From the evidence available, overall natural production of pink salmon does not appear to be below historical levels for the odd-year ESU considered here, and production in the even-year ESU has been generally increasing over the last 15 years. In the odd-year ESU, however, abundance in some individual populations appears to be severely depressed, and some of these declines could be due in part to habitat degradation in individual drainages (e.g., Dungeness River pink salmon; Lichatowich 1993).

Trends in Abundance

Short- and long-term trends in abundance are a primary indicator of risk in salmonid populations. Trends may be calculated from a variety of quantitative data, including dam or weir counts, stream surveys, and catch data. These data sources and methods are discussed in more detail below, under Approach. When data series are lacking, general trends may be inferred by comparing historical and recent abundance estimates, or by considering trends in habitat quantity or condition.

The role of artificial propagation (in the form of hatcheries) for Pacific salmon requires careful consideration in ESA evaluations. Artificial propagation has implications both for evaluating production trends and in evaluating genetic integrity of populations. Waples (1991a,b) and Hard et al. (1992) discussed the role of artificial propagation in ESU determination and emphasized the need to focus on natural production in the threatened or endangered status determination. Because of the ESA s emphasis on ecosystem conservation, this analysis focuses on naturally reproducing salmon. A fundamental question in ESA risk assessments is whether natural production is sufficient to maintain the population without the constant infusion of artificially produced fish. A full answer to this question is difficult without extensive studies of relative production and interactions between hatchery and natural fish.

When such information is lacking, the presence of hatchery fish in natural populations leads to substantial uncertainty in evaluating the status of the natural population. For Washington pink salmon, hatchery production is small relative to natural production and is localized in southern Hood Canal. Therefore, the presence of hatchery fish among naturally spawning pink salmon is not likely to have a substantial effect on our attempts to evaluate the sustainability of natural production for individual populations in this review.

Factors Causing Variability

Variations in the freshwater and marine environments are thought to be a primary factor driving fluctuations in salmonid run size and escapement (Pearcy 1992, Beamish and Bouillon 1993, Lawson 1993). Recent changes in ocean condition are discussed below. Habitat degradation and harvest have probably made stocks less resilient to poor climate conditions, but these effects are not easily quantifiable.

Threats to Genetic Integrity

In addition to being a factor in evaluating natural replacement rates, artificial propagation can substantially affect the genetic integrity of natural salmon populations in several ways. First, stock transfers that result in interbreeding of hatchery and natural fish can lead to loss of fitness in local populations and loss of diversity among populations. The latter is important to maintaining long-term viability of an ESU because genetic diversity among salmon populations helps to buffer overall productivity against periodic or unpredictable changes in the environment (Fagen and Smoker 1989, Riggs 1990). Ricker (1972) and Taylor (1991) summarized some of the evidence for local adaptations in Pacific salmon that may be at risk from stock transfers.

Second, because a successful salmon hatchery dramatically changes the mortality profile of a population, some level of genetic change relative to the wild population is inevitable, even in hatcheries that use local broodstock (Waples 1991b). These changes are unlikely to be beneficial to naturally reproducing fish.

Third, even if naturally spawning hatchery fish leave few or no surviving offspring, they still can have ecological and indirect genetic effects on natural populations. On the spawning grounds, hatchery fish may interfere with natural production by competing with natural fish for territory or mates. If they successfully spawn with natural fish, they may divert production from more productive natural-by-natural crosses. The presence of large numbers of hatchery juveniles or adults may also alter the selective regime faced by natural fish.

For smaller stocks (either natural or hatchery), small-population effects (inbreeding, genetic drift) can also be important concerns for genetic integrity. Inbreeding and genetic drift are well understood at the theoretical level, and researchers have found inbreeding depression in various fish species (reviewed by Gall 1987 and Allendorf and Ryman 1987). Other studies (e.g., Simon et al. 1986, Withler 1988, Waples and Teel 1990; see also Campton 1995) have shown that hatchery practices commonly used with anadromous Pacific salmonids have the potential to affect genetic integrity. However, we are not aware of empirical evidence for inbreeding depression or loss of genetic variability in any natural or hatchery populations of Pacific salmon or steelhead.

For Washington pink salmon, genetic concerns, particularly those related to stock transfers and small-population effects, are relevant primarily to Snohomish River even-year pink salmon and Hood Canal odd-year pink salmon.

Recent Events

A variety of factors, both natural and human-induced, affect the degree of risk facing salmon populations. Because of time-lags in these effects and variability in populations, recent changes in any of these factors may affect current risk without any apparent change in available population statistics. Thus, consideration of these effects must go beyond examination of recent abundance and trends. Unfortunately, forecasting future effects is rarely straightforward and usually involves qualitative evaluations based on informed professional judgment. A key question regarding the role of recent events is: Given our uncertainty regarding the future, how do we evaluate the risk that a population may not persist?

For example, climate conditions are known to have changed recently in the Pacific Northwest, and Pacific salmon stocks south of British Columbia have been affected by changes in ocean production that occurred during the 1970s (Pearcy 1992, Lawson 1993). Much of the Pacific coast has also been experiencing drought conditions in recent years, which may depress freshwater salmon production. However, at this time we do not know whether these climate conditions represent a long-term change that will continue to affect stocks in the future or whether these changes are short-term environmental fluctuations that can be expected to be reversed in the near future. Possible future effects of recent or proposed conservation measures have not been taken into account in this analysis.

Other Risk Factors

Other risk factors typically considered for salmonid populations include disease prevalence, predation, and changes in life history characteristics such as spawning age or size. We have not found evidence that any of these factors are widespread throughout any pink salmon ESU, except for the apparent decline in body size of adult pink salmon previously discussed under Pink Salmon Populations in Washington. Factors that may be important for individual populations are noted in the ESU summaries below.


Previous Assessments

In considering the status of ESUs, we evaluated both quantitative and qualitative data. Among the qualitative data considered were previous reviews of the status of pink salmon (Nehlsen et al. 1991, WDF et al. 1993). These reviews used different definitions of population and different criteria to assess the status of the populations. Nehlsen et al. (1991) classified populations as at high risk of extinction, moderate risk of extinction, or of special concern. They considered populations at high risk of extinction to have likely reached the threshold for classification as endangered under the ESA. Populations were placed in this category if they had declined from historical levels, were continuing to decline, or had spawning escapements less than 200. Populations were classified as at moderate risk of extinction if they had declined from historical levels but presently appear to be stable at a level above 200 spawners. Nehlsen et al. (1991) felt that populations in this category had reached the threshold for threatened status under the ESA. Populations were classified as of special concern if a relatively minor disturbance could threaten them, if insufficient data were available for them, if they were influenced by large releases of hatchery fish, or if they possessed some unique character. Nehlsen et al. (1991) also listed some populations that they considered as possibly extinct, but did not discuss populations not considered to be at some risk. They classified pink salmon in the Skokomish (southern Hood Canal) and Elwha Rivers as at high risk of extinction, and pink salmon in the Dungeness River as at moderate risk of extinction (Table 7). They listed California runs in the Klamath and Sacramento Rivers as extinct, and the run in the Russian River as possibly extinct.

WDF et al. (1993) classified populations as to origin ( native, non-native, mixed, or unknown ), production ( wild, composite, or unknown ), and status ( healthy, depressed, critical, or unknown ). Status categories were defined as healthy, experiencing production levels consistent with its available habitat and within the natural variations in survival for the stock ; depressed, production is below expected levels...but above the level where permanent damage to the stock is likely ; and critical, experiencing production levels that are so low that permanent damage to the stock is likely or has already occurred.

Of the 15 populations of pink salmon identified by WDF et al. (1993; see Table 7), 9 were classified as healthy, 2 as critical, 2 as depressed, and 2 as unknown. They classified all runs as wild production and all except those in the North and Middle Forks of the Nooksack River as native origin. Pink salmon spawning in these two forks of the Nooksack River are likely to be designated as native runs in the 1995 revision of SASSI, based on new sample collections and interpretation of genetic data (J. Ames - footnote 27, J. Shaklee - footnote 28). Nine of the Puget Sound and Hood Canal populations were classified as healthy and two in the Nooksack River were of unknown status. Populations in the Dosewallips and upper Dungeness Rivers were classified as depressed, and populations in the lower Dungeness and Elwha Rivers were classified as critical (Table 7).

Table 7. Status of pink salmon populations assessed in previous reviews.

WDF et al. (1993)
Populationa Nehlsen et al. (1991)b Originc Production
Snohomish N W H
   North Fork/
   Middle Fork
   South Fork N W U
Skagit N W H
   North Fork N W H
   South Fork N W H
Snohomish N W H
Puyallup N W H
Nisqually N W H
Skokomish A
Hamma Hamma N W H
Duckabush N W H
Dosewallips N W D
Dungeness B
   Upper N W D
   Lower N W C
Elwha A N W C
Klamath X
Russian A+
Sacramento X
a - Tributaries and minor drainages combined.
b - A+ = possibly extinct, A = high risk of extinction, B = moderate risk of extinction, X = extinct.
c - N = native, M = mixed.
d - W = wild.
e - H = healthy, D = depressed, C = critical, U = unknown.

Various problems arise in applying results of these studies to ESA evaluations. One major problem is that the definition of stock or population varied considerably in scale among studies, and sometimes among regions within a study. Identified units range in size from large river basins to minor coastal streams and tributaries. A second problem is the definition of categories used to classify population status. Only Nehlsen et al. (1991) used categories intended to relate to ESA threatened or endangered status, and they applied their own interpretations of these terms to individual populations, not to ESUs as defined here. WDF et al. (1993) used general terms describing the status of populations that cannot be directly related to the considerations important in ESA evaluations. For example, the WDF et al. (1993) definition of healthy could conceivably include a population that is at substantial extinction risk due to loss of habitat, hatchery fish interactions, and/or environmental variation, although this does not appear to be the case for any pink salmon population. A third problem is the selection of populations or stocks included in the review. Nehlsen et al. (1991) did not evaluate (or even identify) populations not perceived to be at risk, so it is difficult to determine the proportion of populations they considered to be at risk in any given area. WDF et al. (1993) included all natural Washington populations of pink salmon in their assessment, as they considered all of these to be substantially wild.

Data Evaluations

Quantitative evaluations of data included comparisons of current and historical abundance of pink salmon, calculation of recent trends in escapement, and evaluation of the proportion of natural spawning attributable to hatchery fish. Historical abundance information for these ESUs is largely anecdotal. Time-series data are available for many populations, but the amount and quality of the data varied among ESUs. We compiled and analyzed this information to provide several summary statistics of natural spawning abundance, including (where available) recent total spawning run size and escapement, percent annual change in total escapement, recent naturally produced spawning run size and escapement, and average percentage of natural spawners of hatchery origin.

Although our evaluation used the best data available, these data have several limitations, and not all summary statistics were available for all populations. For example, we generally did not measure spawner abundance directly for all populations; rather, we often had to estimate abundance from catch (which itself may not always have been measured accurately) or from limited survey data. In many cases, it was difficult to separate hatchery production from natural production.

Quantitative methods--Information on population abundance was compiled from a variety of state, federal, and tribal agency records. We believe it to be complete in terms of long-term adult abundance records for pink salmon in the region covered here. Principal data sources were fishery statistics and stream surveys. Neither of these provides a complete measure of abundance for any of the streams. Specific problems are discussed below for each data type.

Data types--There are two primary sources of abundance data for pink salmon: fishery statistics and spawning escapement surveys. Fishery data span a longer series, but landings in Washington, Oregon, and California are dominated by Fraser River fish and do not accurately reflect abundance of local pink salmon populations. In addition, fisheries harvest mixed populations and no systematic marking or tagging programs can be used to attribute fishery landings to streams of origin consistently. Landings from areas covered by the Pacific Salmon Treaty (PST) have been partitioned by country of origin with genetic stock identification (GSI) techniques since 1985, and prior to that were allocated by assuming that U.S. populations provided a constant percentage of the landings in treaty catch areas (J. Woodey - footnote 29). These GSI techniques are used to estimate landings of fish originating from 23 stocks, but harvest management is currently based on the analysis of contributions from 3 major groups--Fraser River, non-Fraser River British Columbia, and Puget Sound (B. White - footnote 30). Landings within Puget Sound and the U.S. origin catches from U.S. PST catch areas are allocated among individual populations on the basis of the location of the fisheries and the relative abundance of the populations believed to contribute to each catch area.

Implicit in this methodology is the assumption that populations contribute to fisheries in direct proportion to their run size. However, shifts in the timing and migratory pathways of returning populations can substantially alter their contribution rate to preterminal fisheries. Because the Fraser River typically produces many more pink salmon than Puget Sound, this system of run reconstruction inaccurately estimates fishery exploitation rates on individual populations, especially for smaller populations. For these reasons, spawning escapement estimates are the most reliable estimates of population abundance, although they also have inaccuracies and reflect abundance only after the populations have been subjected to variable domestic exploitation and foreign interception rates.

Fishery landings of pink salmon are summarized in the Northwest Indian Fisheries Commission (NWIFC) run reconstruction database (Big Eagle & Assoc. and LGL Ltd. 1995). Run reconstruction attributes catches to stream of origin on the basis of geographic location and assumed migration pathways, but the process is an approximation at best and can provide only reasonably accurate estimates for individual populations. Nevertheless, estimates of Puget Sound exploitation rates for individual populations can be constructed from these data.

The most comprehensive attempt to attribute historical harvest to area of origin was made by Canadian Department of Fisheries and Oceans (CDFO) biologists who reconstructed estimates of landings from many sources using a variety of methods and synthesized a time series of harvest for the United States that spans the period from 1889 through 1981 (Shepard et al. 1985). They also estimated the harvest of U.S. (excluding Alaska) pink salmon from U.S. landings by adding in Canadian interceptions of U.S. salmon and subtracting U.S. interceptions of Canadian salmon. Since there is practically no domestic production of pink salmon outside of Puget Sound, these landings can be attributed to Puget Sound production. However, the partitioning of U.S. and British Columbia landings by country of origin requires a number of assumptions and is therefore imprecise. It is possible to calculate exploitation rates on Puget Sound pink salmon by dividing catch by catch plus escapement for the entire Puget Sound for the years in which both estimates are available. These estimated exploitation rates show an increasing trend with a low of about 35% in 1969 to a high of about 90% in 1981, but the accuracy of these estimates is questionable.

In Puget Sound, most spawning escapement estimates have been made with a few standard methods (see Vernon et al. 1964, Hourston et al. 1965). In the early 1960s an intensive effort was made to estimate the size of pink salmon runs in Puget Sound. Pink salmon were captured near the mouths of rivers and marked with external tags. Spawning grounds were surveyed and spawners were sampled throughout the river systems; effort was concentrated in areas of heaviest spawning. The spawner distribution was used to establish index areas. Carcasses were examined for tags, and Petersen mark-recapture estimates (Ricker 1958) of the total numbers of spawners were made. The Petersen estimate simply expands the number of tags applied to fish by the ratio of unmarked to marked fish observed in the carcass surveys to estimate the number of unmarked fish in the spawning run (Caughley 1977). This procedure assumes that the marked fish are randomly distributed in the spawning population, that they have the same survival rate as unmarked fish or that spawners have an equal probability of being in the sample, that no tags are lost, and that all marked fish encountered in the carcass surveys are recognized.

Violation of these assumptions reduces the accuracy of estimates and tends to bias the estimates upward. Attempts were made by WDF to reduce biases due to mortality of tagged fish and tag loss by rounding down the Petersen estimates (D. Hendrick - footnote 31; see Vernon et al. 1964, Hourston et al. 1965). These Petersen estimates of total spawning escapement are the best estimates available for Puget Sound, and they became the base-year data for subsequent escapement estimates (Johnson et al. 1968, J. Ames - footnote 32). In years following the Petersen estimates, spawning escapement has typically been estimated by multiplying the base-year spawning escapements by the ratio of counts of total live plus dead fish observed at the peak of the run (peak counts) in index reaches to peak counts in the base years. Historical estimates for northern Puget Sound rivers were usually generated by comparing the total number of carcasses counted in index areas to the total number of carcasses in the same areas during the tagging studies. (Except for the Nooksack River, consistent live counts were not made during the tagging years; D. Hendrick - footnote 33).

Since the mid-1970s, area-under-the-curve (AUC) estimates have been used instead of peak counts to scale the base-year Petersen estimates (Ames 1984). These estimates are currently compared for the Nooksack and Stillaguamish Rivers (D. Hendrick - footnote 34). These AUC estimates are made from repeated surveys of the spawning grounds; the sequential counts are plotted against time and used to generate an escapement curve. The area under the curve is calculated to provide an estimate of the number of fish-days composing the spawning run in an index reach.

Computed statistics--To represent current run size or escapement where recent data were available, we have computed the geometric mean of the most recent 5 years reported (or fewer years if the data series is shorter than 5 years). We tried to use only estimates that reflect the total abundance for an entire river basin or tributary, and avoided index counts or dam counts that represent only a small portion of available habitat.

As an indication of overall trend in pink salmon populations in individual streams, we calculated average (over the available data series) percent annual change in adult spawner indices within each river basin. Trends were calculated as the slope (a) of the regression of loge(abundance) against years corresponding to the biological model N(t) = beat, where N(t) is population size in year t, b is a scalar coefficient, and e is the base of the natural logarithm. Slopes significantly different from zero (P < 0.05) were noted. The regressions provided direct estimates of mean instantaneous rates of population change (a); these values were subsequently converted to percent annual change, calculated as 100(ea - 1). No attempt was made to account for the influence of hatchery produced fish on these estimates, so the estimated trends included any supplementation effect of hatchery fish.

In reviewing the status of individual ESUs of pink salmon in Washington and southern British Columbia, we considered the risks posed by artificial propagation to be less important than other risk factors, such as habitat degradation, indicating declines in abundance. This factor was a consideration in the BRT s conclusions that the ESUs for pink salmon are not at risk of extinction or endangerment. However, the BRT s conclusions on this issue should be regarded as preliminary because information about the degree of interactions that actually occur between hatchery and natural fish is still incomplete.

Analysis of Extinction Risk by ESU

Several data series of pink salmon spawner abundance in northwestern Washington exist, but the ultimate sources for all quantitative estimates of spawning escapements are WDFW for data from Washington and CDFO for data from British Columbia. These data have been compiled in an electronic database and submitted to the ESA administrative record (Big Eagle & Assoc. and LGL Ltd. 1995). These data are summarized for even- and odd-year pink salmon in Figures 18-20. Trends and recent averages for spawning escapements of even- and odd-year pink salmon populations in Washington and British Columbia are given in Tables 8 and 9. Additional information on individual pink salmon streams is summarized below.

Even-year ESU

Snohomish River--There are anecdotal and sporadic accounts of even-year runs for some northern Puget Sound rivers (D. Hendrick - footnote 35, W. Waknitz - footnote 36), but the only river for which an even-year population has been documented is the Snohomish River (WDF et al. 1993). The run sizes were estimated by aerial counts of redds, and these estimates were used to scale Snohomish River base-year estimates from odd-year runs. This run has been documented every even year since 1980, but no run-size estimate was made in 1982. The distribution of spawning is thought to be far more restricted for even-year pink salmon than for odd-year pink salmon, with most spawning activity occurring in the mainstem Snohomish River. Scattered redds have been observed in the Skykomish River as far upstream as Sultan. In general, this run has been increasing since 1980. Estimated escapement for 1994 was approximately 1,600 fish, which is less than 1990 and 1992 estimates but greater than those for 1986 and 1988. A few adults are occasionally reported in even years in the Nooksack, Stillaguamish, and Skagit Rivers, but the numbers are very small, and surveys are not conducted in these systems in even years.

British Columbia--All of the Canadian even-year populations for which we received escapement estimates are located in rivers draining into the northern Strait of Georgia or Johnstone Strait. Additional populations may occur on the west coast of Vancouver Island as far south as the Strait of Juan de Fuca (Aro and Shepard 1967), but if so these populations are probably very small.

Retrieve Figure 18: Retrieve Figure 19: Retrieve Figure 20:

Table 8. Average spawning escapement estimates and trends for even-year pink salmon populations in Washington and British Columbia. Average escapements are geometric means calculated from the five most recent spawning escapement estimates. For each system, annual percentage change is calculated for the periods shown in Figures 18-20. See text for further explanation.
Average Annual
Population Production escapement % change
Snohomish natural 1,604 +25.6
British Columbia
Cluxewe natural 28,163 +3.3
Quatse natural 1,316 -14.8*
Keogh natural 1,961 -11.3*
Ahnuhati natural 129,310 +8.5*
Ahta natural 18,161 +2.2
Embley Creek natural 13,000 +0.6
Kingcome natural 33,066 +3.0
Wakeman natural 37,279 +8.7*
Phillips natural 40,735 +10.0*
Campbell mixed 5,021 +5.3*
Glendale Creek mixed 272,385 +5.1*
Kakweiken mixed 335,882 +13.2*
Quinsam mixed 36,110 +16.4*
Puntledge hatchery 5,847 +1.9
* Denotes trends that differ significantly (P < 0.05) from zero.
Table 9. Average spawning escapement estimates and trends for odd-year pink salmon populations in Washington and British Columbia. Average escapements are geometric means calculated from the five most recent spawning escapement estimates. The annual percentage change, calculated for the periods shown in Figures 18-20, includes the effect of the unusually high odd-year escapements observed in 1963 in Washington. See text for further explanation.
Population Productiona Average escapement Annual % change
Nooksack natural 43,573 +0.2
Skagit natural 500,125 +2.1
Stillaguamish natural 109,650 +1.5
Snohomish natural 144,958 +1.1
Puyallup natural 24,142 +1.2
Nisqually natural 2,141 -0.5
Dosewallips natural 16,024 -6.3
Duckabush natural 23,272 -1.2
Hamma Hamma natural 17,651 +7.3
South Hood Canal natural 181 +0.2
Upper Dungeness natural 3,981 -8.6
Lower Dungeness natural 340 -17.0b
Elwha natural 53 --c
Hood Canal Hatchery hatchery 2713 +0.4
British Columbia
Lower Fraser natural 3,828,815 +6.6b
Fraser Canyon natural 59,390 +4.4b
Thompson natural 352,460 +3.8b
Cluxewe natural 1,223 -6.6b
Quatse natural 16,101 +12.9b
Keogh natural 56,234 +2.6
Ahnuhati natural 7,153 +1.8
Ahta natural 7,391 -0.7
Embley Creek natural 500 +12.7
Kingcome natural 7,145 -1.4
Wakeman natural 36,651 -1.3
Phillips natural 23,571 -1.1
Seton-Anderson mixed 705,877 +10.0b
Harrison mixed 540,953 +3.9
Chilliwack-Vedder mixed 150,291 -1.2
Glendale Creek mixed 33,798 -1.4
Kakweiken mixed 163,260 +6.0b
Campbell mixed 1,616 +10.1b
Quinsam mixed 258,749 +15.2b
Puntledge hatchery 6,069 +0.3
a - According to the corresponding management agency.
b - Denotes trends that differ significantly (P < 0.05) from zero.
c - Exponential trend cannot be calculated when recent spawning escapements are zero.

Odd-year ESU

Nooksack River--Because the Nooksack River is fed by glacial meltwater and the fish cannot be counted reliably in the main stem because of the high turbidity, escapement estimates for the Nooksack River are the least reliable in northern Puget Sound. Data collection began in 1959, and Petersen estimates were made in 1959, 1961, and 1963. These base-year estimates have subsequently been scaled by peak counts and AUC estimates. In the base years, most fish spawned in the main stem. This may still be the case, as tributary spawning alone cannot account for total production from the system, but only clear water tributaries are surveyed due to the difficulty observing spawners in the main stem. Because these tributaries contain a relatively small and variable fraction of the spawning escapement, the escapement estimates have potentially large errors.

Skagit River--Skagit River run sizes were estimated in the main stem by comparing total carcasses in mainstem index areas to the number of carcasses sampled during tagging to construct the Peterson estimates in the base years 1959 through 1963. AUC estimates were used in tributaries. Since the base years, the distribution of spawners has shifted somewhat. There is presently more spawning in upper reaches of the river than occurred in the base years, so recent runs may be underestimated. However, this bias should be minimal because tagging- year data are available for the entire main stem and because WDFW currently surveys nearly 65 km of the river s length (D. Hendrick - footnote 37).

Stillaguamish River--Escapement estimates are again based on AUC estimates of observed live spawners. Base years are 1959 through 1965, with additional Petersen estimates from 1967 and 1987. In the original base years the majority of fish spawned in the North Fork Stillaguamish River. This remained the case until 1987, when the Washington Department of Fisheries (WDF) conducted a mark-recapture study on the Stillaguamish River (D. Hendrick - footnote 38). Drought conditions in that year limited spawning in the North Fork, and substantial numbers of fish spawned in the South Fork. WDF biologists did not realize where the fish were until after the peak of spawning occurred, so their use of the 1987 AUC estimate for the South Fork in scaling subsequent escapements is questionable. Since 1989, the run in the North Fork has been rebuilding, but the South Fork still has the majority of spawning activity.

Snohomish River--Runs were estimated by AUC scaling of base-year Petersen estimates for 1959 through 1963 and 1967. Index reaches were in the Snohomish River and the Skykomish River, with a few observational surveys made in the Snoqualmie River. The portion of the run using the Snoqualmie River is unknown but may be increasing. In some years, considerable numbers of spawners have been reported in the Tolt River, but the run is inconsistent and is therefore believed to be just a component of the Snoqualmie River run. Spawning distribution has expanded since the base years. There is presently more spawning in intertidal areas, the Snoqualmie River, and in upper reaches of the Skykomish River and its tributaries (D. Hendrick - footnote 39). Because of this shift, the index reaches may contain a smaller fraction of the escapement than they did in the base years, and recent estimates may be biased downward.

Lake Washington--There is no evidence of sustained runs of pink salmon in the Lake Washington watershed.

Southern Puget Sound and Hood Canal--Recent spawning escapement estimates for southern Puget Sound and Hood Canal are based on AUC estimates of total live spawners. Data series are of variable length, and early estimates were the sum of live and dead fish observed during the peak of the spawning run. Except for the Nisqually River, which is surveyed by Nisqually tribal biologists, spawning ground surveys are generally conducted by WDFW biologists, and the AUC estimates are made by WDFW from WDFW and tribal data. Data are generally reliable, except for those from surveys in the Nisqually River for 1981-85 and 1993 (see below).

The Nisqually River is glacially influenced with a few clear water tributaries, and consequently has presented some special problems in estimating escapement. Escapements have been estimated with several methods. Before 1981, escapements were estimated from visual surveys in the main stem and tributaries. From 1987 to 1991, estimates were made from a combination of mark-recapture studies, visual surveys in tributaries, and fishery catches. Estimates from the tributary surveys are expanded, somewhat arbitrarily, to account for mainstem spawning. In 1981, 1983, and 1985, conditions were unfavorable in the tributaries and the vast majority of fish spawned in the main stem. The number of spawners in the main stem was unknown, but very few fish were seen in the tributary surveys. Peak counts in those years ranged from three to nine fish. In 1993, no pink salmon data were recorded during stream surveys. Thus, for odd years from 1981 to 1985 and for 1993, an escapement of 500 was assigned as a placeholder for run reconstruction (J. Ames - footnote 40, J. Uehara - footnote 41). For all four of these years, actual spawning escapement is unknown and should be considered missing values.

Dungeness River--The Dungeness River supports two different populations of pink salmon: an early run that enters the river in July and August and ascends to areas above RKm 16, and a late run that enters the river in late August and September and spawns below RKm 5. Until 1981, a hatchery rack at about RKm 16 was used to enumerate the early run. The lower (late) run was not surveyed regularly and was assumed to be a fixed proportion of the early run in some years. When the rack was removed in 1981, WDF had to change their estimation method, a change that coincided with the effects of a major winter flood in 1980 on the incubating 1979 brood. WDFW biologists do not believe that the visual survey method significantly underestimates the run size in this case, and it seems unlikely that underestimation could fully account for the observed drop in run size. Survey conditions are consistently good in the Dungeness River when spawner counts are made (J. Ames - footnote 42).

Since 1981, the spawning ground surveys have attempted to estimate the entire run by AUC estimates. Runs were relatively stable until 1981, when 1979-brood fish affected by the 1980 winter flood returned (footnote 43). Both runs have been depressed since that time, with the lower run exhibiting a continued decline. Both runs on the Dungeness River exhibit statistically significant downward trends in spawner abundance.

Elwha River--Data extend back to 1959 and tend to become less reliable in recent years. Estimates in 1959, 1961, and 1963 were based on visual surveys made by WDF as part of an extensive marking program at that time (J. Ames - footnote 44). In 1975, the escapement estimate of 1,500 was based on a single float survey in which it was estimated that 300 spawners were seen. In 1977 and 1979, the Elwha River was not surveyed, but its escapement was estimated to be 14.25% of the total escapement to the Dungeness River. This was based on the average ratio of escapements to the two rivers in the preceding five spawning runs. As did the Dungeness River, the Elwha River experienced severe flooding in the winter of 1979-80. Since 1981, the Elwha River has been surveyed every run year with 5 to 18 surveys conducted each year. These surveys, which were made primarily for chinook salmon, incorporated multiple breakout counts for different river sections. More intensive surveys in 1991 and 1993 failed to find any pink salmon in the Elwha River. As noted earlier in this report, WDFW s management biologists for this area believe the Elwha River run of pink salmon to be extinct.

British Columbia--Data on run sizes for Fraser River populations and populations in the Strait of Georgia and Johnstone Strait have been received from CDFO (Big Eagle & Assoc. and LGL Ltd. 1995). Spawning escapements were estimated with a variety of methods, and managers have been required to provide escapement numbers regardless of the quality of the estimates (Big Eagle & Assoc. and LGL Ltd. 1995). In addition, escapement data from western Vancouver Island may be collected only incidentally during surveys for other species. For these reasons, several escapement estimates for British Columbia populations are of questionable quality. Nevertheless, pink salmon escapement data collected over the last several years by Canadian Department of Fisheries and Oceans staff suggest that western Vancouver Island escapements have declined significantly since the 1970s. Current escapement levels on the west coast of the island are typically small, that is, a hundred to a few thousand adults. Indirect evidence points to declining trends in some populations along the southwestern coast of Vancouver Island, but this evidence is weak because the escapement data are not robust (W. Luedke - footnote 45). Tables 8 and 9 summarize the escapement data that Canadian biologists felt were the most reliable for runs in central and southern British Columbia.


There is no evidence of strong or sustained recent declines in abundance for most pink salmon populations in Washington and southern British Columbia. However, both odd-year pink salmon populations in the Dungeness River are depressed, and the lower river population shows a strong declining trend. Although this latter population also declined substantially from 1973 to 1979, major flooding during the winter of 1979-80 appears to have been an important factor leading to its current, severely depressed state. This flooding in fact appears to have influenced all the Washington odd-year pink salmon populations along the Strait of Juan de Fuca, including the Elwha River population. The single U.S. population of even-year pink salmon in the Snohomish River is small, and although it has been increasing in size over the last decade, the estimated 1994 escapement was a substantial drop over that in 1992.

In addition to abundance, two ecological factors relating to the status of pink salmon populations were of concern to the BRT: low-water conditions upon river entry and spawning, which can limit spawner distribution to suboptimal center-channel areas; and subsequent high- water conditions, which can erode the quality of spawning and incubation habitat and may adversely affect embryonic development. These factors may have been exacerbated by water withdrawals for irrigation and by structures erected for flood control, particularly on the lower Dungeness River (Lichatowich 1993, K. Lutz - footnote 46). In addition, flooding can reduce substrate stability and permeability and subsequent pink salmon productivity (Wickett 1962). Extensive severe flooding in several Puget Sound rivers caused by heavy rains in winter 1995-96 may have substantially reduced the productivity of the 1995 brood in these systems through the mortality of embryos and hatchlings.

Other possible threats to pink salmon in the Puget Sound region include predation on juvenile pink salmon by wild (Hiss 1994) and hatchery (Hiss 1995) coho salmon, and, perhaps in some locations, predation by marine mammals (Knudsen et al. 1990, J. Ames - footnote 47). In addition, the population dynamics of pink salmon are not well understood, and may be affected by intra- and interspecific interactions involving pink and chum salmon (Gallagher 1979, Smoker 1984, Heard 1991). Consequently, the possibility that hatchery production of pink and chum salmon can limit the viability of some Washington populations of pink salmon cannot be excluded.

Finally, oceanic conditions may also affect pink salmon abundance (Davidson and Vaughan 1941; Mysak et al. 1982; Blackbourn 1985, 1990; Mysak 1986; Heard 1989). The widespread declines observed in pink salmon body size (Ricker et al. 1978, Ricker 1981, Marshall and Quinn 1988, Ricker 1989; Table 3) may have resulted from increased salmon density, reduced ocean productivity, or directional selection on body size in both net and troll fisheries. From analysis of the southeastern Alaska pink salmon fishery between 1895 and 1940, Davidson and Vaughan (1941) identified an inverse relationship between the abundance (measured as total pack) of these fish and adult body size, and they suggested that this relationship resulted from greater competition for marine prey. Small adult body size is a cause for concern because it limits reproductive potential (Skud 1973), and there is some evidence for a strong genetic component to body size (Smoker et al. 1994). Growth at sea or the subsequent body sizes of adults may also affect the timing of their spawning migration (Davidson and Vaughan 1941), which in turn may affect fry recruitment. Skud (1973) suggested that the interaction between these factors tends to yield higher fry survival in years when spawners are larger and spawn earlier than in years when spawners are smaller and spawn later. Conclusions for each of the proposed ESUs are summarized below.

Even-year ESU

Because it is unclear whether populations other than that in the Snohomish River are in the even-year pink salmon ESU, the BRT first considered the status of this ESU under the assumption that it included only the single U.S. population. Based on available information, which shows a relatively small population with a generally increasing trend in abundance in recent years, the BRT concluded that this ESU is not at risk of imminent extinction or endangerment. Because even-year populations in British Columbia are generally stable or increasing and are apparently not at low levels compared to historical abundance, the BRT also concluded that an ESU that included populations in British Columbia would not currently be at risk of extinction or endangerment.

However, three factors prompted considerable concern among BRT members about risks the Snohomish River population might face in the near future: 1) it may be strongly isolated from all other even-year populations; 2) although the population has generally been increasing in recent years, it remains at low abundance; and 3) the invariant age structure in pink salmon, coupled with considerable interannual variability in abundance, increases risks faced by small, isolated populations. Clearly, these concerns would be greatest for an ESU that included only the single U.S. population; however, it is also likely that the Snohomish River population would be considered an important part of a larger ESU that contained populations from British Columbia. Therefore, the BRT concluded that the Snohomish River even-year pink salmon population should be closely monitored so that if any factors arise that substantially increase risk faced by this population, they can be identified at an early stage.

Odd-year ESU

Most populations in the odd-year pink salmon ESU appear to be healthy, and overall abundance appears to be close to historical levels. The two most distinctive Puget Sound populations (from the Nooksack and Nisqually Rivers) both show nonsignificant trends in recent abundance (Nooksack River slightly increasing, Nisqually River slightly decreasing), and no other factors were found that would suggest that either of these populations is at immediate risk. The BRT therefore concluded that the ESU for odd-year pink salmon as a whole is not at significant risk of becoming extinct or endangered.

However, two populations on the Strait of Juan de Fuca are clearly at risk, and an additional population (in the Elwha River) already appears to be extinct. These populations contribute substantially to the ecological and genetic diversity of the ESU, and the BRT expressed concern that further erosion of individual populations might result in future risk to a significant portion of the ESU as a whole.

The BRT also identified risk factors that should be monitored in the future. Some evidence exists for recent declines in body length of odd-year fish in Washington, which raises concern about the ability of natural populations (notably those in the Strait of Juan de Fuca) to recover naturally. The BRT was unable to review data on body size in odd-year British Columbia pink salmon since the studies by Ricker et al. (1978) and Ricker (1989) to ascertain whether body size in these populations has been declining to levels similar to those observed in Washington fish in recent years. However, the decline in body length of odd-year Washington pink salmon indicated by the data in Table 3 is qualitatively similar to a decline in length observed in pink salmon returning to Auke Creek, Alaska over the last 20 years (W. Smoker - footnote 48). Data estimated from catches of southeastern Alaska pink salmon suggest similarly declining trends over an even longer period (Marshall and Quinn 1988). Collectively, these patterns suggest that pink salmon from populations over wide geographic areas have been experiencing conditions at sea that restrict adult body size and, consequently, reproductive potential.

Table of Contents