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

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

Thirteen spawning populations of pink salmon have been identified in Washington (WDF et al. 1993). For 12 of these populations, spawning occurs only in odd years; the sole even-year population exists in the Snohomish River in Puget Sound (Fig. 4)(Footnote 12). Four odd-year populations occur in the Nooksack, Skagit, Stillaguamish, and Snohomish Rivers in northern Puget Sound, where most pink salmon are produced in Washington. Two odd-year populations occur in southern Puget Sound in the Puyallup and Nisqually Rivers. Three odd-year populations occur in Hood Canal in the Hamma Hamma, Duckabush, and Dosewallips Rivers. In addition, three odd-year populations have been identified by WDF et al. (1993) on the Strait of Juan de Fuca: upper Dungeness ( summer population), lower Dungeness (fall population), and Elwha Rivers (Fig. 4).

Pink salmon have been observed periodically in other Washington rivers, including the Skokomish River on Hood Canal and the Bogachiel River on the western Olympic Peninsula (Williams et al. 1975). Elwha River pink salmon now appear to be extinct, as no adult fish have been observed there since 1989 despite extensive annual surveys for chinook salmon. Because of the temporal overlap in spawning distributions of chinook and pink salmon in the Elwha River, pink salmon should have been observed there if they had been present. Pink salmon apparently occurred historically in the Green/Duwamish River system in Puget Sound; the Washington Department of Fisheries (WDF) (1916-64) reported these fish as very scarce in the Green River, and absent from Burns, Newaukum, Spaight, and Soos Creeks. The highest annual number of adult pink salmon observed in the Green River over the last several decades is 13. Up to 16 adults have been observed in the Cedar River in the Lake Washington watershed, but these are believed to be strays (J. Ames - Footnote 13).


Figure 4
Figure 4. Principal pink salmon spawning streams in northwestern Washington. Stream names are in boldface. Figure modified from IPSFC (1958).

Several aspects of the life history and ecology of pink salmon in Washington and southern British Columbia are poorly understood in detail. Topics for which data were insufficient to assist the BRT in evaluating possible ESUs for pink salmon include variation in fecundity, development rate, growth and mortality, length of fry residence and migration in the nearshore marine environment, behavior, and physiology, and several aspects of spawning habitat. The remainder of this section concentrates on topics that were important in determining proposed ESUs for pink salmon in Washington and southern British Columbia.

Spawning Habitat

In northwestern Washington rivers, pink salmon generally spawn in relatively fast- flowing shallow water in small, clear water drainages and in both clear water and turbid large drainages (Bonar et al. 1989, WDF et al. 1993). Some intertidal spawning may also occur, but it is probably not extensive in this region. Several populations spawn a considerable distance upstream or under unusual river conditions. For example, summer Dungeness River (Strait of Juan de Fuca) pink salmon return earlier than fall Dungeness River pink salmon and spawn above RKm 16, about 11 km upstream from the fall fish spawning in the lower river (Johnson 1973, WDF et al. 1993). Some Dosewallips River pink salmon (Hood Canal) migrate more than 15 km upstream to spawn. In southern Puget Sound, pink salmon spawn in the Puyallup River primarily above RKm 20, in South Prairie Creek and in other tributaries as well as the main stem. Spawning by pink salmon in the Nisqually River occurs primarily in the main stem between RKm 35 and RKm 65.

In northern Puget Sound, pink salmon spawning occurs in the following locations: Nooksack River (all three forks), up to RKm 40 in both main stem and tributaries; Skagit River, between RKm 37-149 (especially above RKm 124) and also in the Sauk River up to RKm 64; Stillaguamish River, between RKm 9-54 in the South Fork (primarily RKm 30-48) and throughout the North Fork; and Snohomish River, between RKm 21-34 for both odd-year and even-year populations, especially above RKm 29 (both populations also spawn in the Skykomish River, but use of the river for spawning is more extensive for odd-year fish; only the odd-year fish are known to spawn in the Snoqualmie River) (WDF et al. 1993).

Spawning can occur in water carrying substantial amounts of glacial silt in the Nisqually and Nooksack Rivers. Although the Puyallup River also contains substantial glacial runoff, nearly all known pink salmon spawning occurs in two clear water tributaries, South Prairie and Kapowsin Creeks (WDF et al. 1993).

Migration, Spawning, and Emergence Timing

Migration--Little detailed information is available on the migration routes and timing of different pink salmon populations from Washington and southern British Columbia. Neave et al. (1967) reported that the return migration of maturing fish in the eastern North Pacific occurs mostly between June and September. The marine migration of pink salmon in this region has been inferred from tagging data to be less extensive than that of fish from more northern populations (Takagi et al. 1981, Ogura 1994). Pink salmon migrating from Washington and British Columbia are believed to remain east of about 150°W longitude and south of about 59°N latitude in the Gulf of Alaska. Thus, they probably overlap at sea with pink salmon originating from southeastern and central Alaska (Takagi et al. 1981). Ogura (1994) summarized recoveries of pink salmon tagged in Washington and southward and found that these recoveries were restricted to the east of 137°W longitude and between 48o and 55°N latitude.

Several early tagging studies were conducted to determine the final migratory pathways of pink salmon returning to British Columbia and Washington (Pritchard and DeLacy 1944, DeLacy and Neave 1948, IPSFC 1958, WDF 1959). A graphical summary of the information gathered in these studies is given in Figure 5. Most pink salmon returning to Washington are believed to travel east through the Strait of Juan de Fuca, but some fish probably also use the Johnstone Strait/Strait of Georgia corridor during their return migration (IPSFC 1958). In addition, returning Fraser River and other southern British Columbia pink salmon may migrate through either corridor.

Although available information is insufficient to discriminate fish from different populations as they migrate into nearshore marine waters in Washington and southern British Columbia to reach spawning grounds, WDF et al. (1993) have outlined some prominent patterns. Generally, differences in migration timing of maturing pink salmon in this region are small; differences among populations within areas can be larger than differences among populations from different areas. Even-year pink salmon returning to the Snohomish River in Puget Sound begin appearing in catches in early June and drop off sharply after mid- to late September. Peak abundance off the river mouth generally occurs in late August (WDF et al. 1993).

For odd-year pink salmon returning to Washington rivers, timing of arrival to terminal areas (i.e., near estuaries and river mouths) shows no clear geographic pattern. Maturing fish returning to streams in northern Puget Sound generally arrive at these areas between mid- August and late September. The earliest run appears to occur to the Nooksack River, the latest to the Skagit River. In southern Puget Sound, fish arrive at terminal areas primarily between mid-July and mid-August. In Hood Canal, fish arrive between mid-July and early September (WDF et al. 1993). Pink salmon appear off the Dungeness River in the Strait of Juan de Fuca earlier than pink salmon returning to other Washington rivers. Fish that migrate to the Dungeness River begin arriving at the river s mouth as early as mid-June and continue to show up until early August, with abundance off the river mouth declining after river entry between late July and early September. These fish constitute the two distinct runs that return to the upper and lower river, respectively, and show considerable temporal differentiation in average spawn timing (WDF et al. 1993).


Figure 5
Figure 5. Principal migration routes for pink salmon returning to southern British Columbia and Washington. Figure adapted from WDF (1959).
Spawn timing--Pink salmon populations can vary considerably in their arrival timing at spawning grounds (Sheridan 1962, WDF et al. 1993), and some evidence exists for substantial differences in spawn timing within a single river system. For example, Taylor (1980) and Gharrett and Smoker (1993a) identified early and late spawning populations in Auke Creek, Alaska that were also differentiated by their patterns of fry outmigration timing. These populations have not yet been distinguished by allozymic variation (Lane et al. 1990). Because this sort of life history variability can have consequences for fitness (Taylor 1980, Mortensen et al. 1991), it may be instrumental in maintaining adaptive diversity within and among populations. The fact that such life history variation may not be detected with routine allozyme surveys indicates the importance of considering this variation when attempting to identify distinct population units (Gharrett and Smoker 1993b, Hard 1995a). However, the interpretation of life history variation is complicated because this variation may also be quite sensitive to environmental variation (Waples 1991a, Hard 1995b).

In Washington and southern British Columbia, river entry occurs from July to October, and spawning is generally observed from August to October (Neave 1963, WDF et al. 1993). In their SASSI document, WDF et al. (1993) reported that average timing of river entry and spawning are generally earlier in northern Puget Sound and in the upper Dungeness River on the Olympic Peninsula than elsewhere in northwestern Washington. As noted in an earlier section, the Dungeness River has two populations of pink salmon: summer run fish enter the river in ocean-bright condition in July and August, hold in the river for up to a few weeks before final maturation, and spawn above RKm 16 in August and September; fall run fish enter the river already exhibiting much of their spawning morphology and coloration in August and September, spawning soon afterward in the lower river. These dark fall-run fish spawn below RKm 5 an average of 2 weeks later than summer-run fish (WDF et al. 1993, J. Uehara - Footnote 14).

The Washington Department of Fish and Wildlife has surveyed pink salmon escapement (counts of live and dead adults) in many northwestern Washington rivers since the late 1950s. The Department uses an area under the curve method (Ames 1984) to estimate escapement from many of these surveys and from an estimate of spawner life in the stream (generally, 10 or 15 days). Escapement curves generated with this method can also be used to derive median spawn timing for different rivers in various years. One means of deriving an estimate of median spawn timing for a given reach or drainage in a particular year is to determine the date at which half the area under the curve is reached. Since these curves are actually a series of connected trapezoids, it is a straightforward matter to calculate this date. This method may be preferable to using peak counts because the median spawn date should be less sensitive than peak count to variation in the shape of the spawning distribution.

Data from WDF spawning surveys conducted between 1983 and 1993 were used to calculate median run timings and examine their patterns. Table 2 summarizes these data and compares them to the range of spawning dates observed in these systems and summarized by WDF et al. (1993). It should be noted that 1) no data were available from Snohomish River even-year pink salmon for this analysis, 2) the Elwha River statistics were based on a single survey in 1985, and 3) few surveys have been made in the Nisqually River. It should also be noted that spawning curves for some river reaches that may contain significant spawning populations are not yet available. In some cases, several surveys within a river and year were combined to determine the median date for that river. For these reasons, data generated from the spawning curves should be examined with caution.

± Nevertheless, a complex pattern for odd-year pink salmon run timing is evident from this limited analysis. In general, the run timing estimates derived from spawning curves correspond well to the midranges of spawning reported by WDF et al. (1993); the exceptions are estimated median run timings for Puyallup River and Elwha River pink salmon, which are both earlier than the midranges interpolated from the WDF et al. (1993) report. However, it should be noted that the Puyallup River median timing is within the range reported by WDF et al. (1993), and the Elwha River median timing is based on only a single year.

In general, it appears that pink salmon spawn earliest in the upper Dungeness River (average of the median calendar dates ± SE = 245 ± 0.8, where day 245 = 2 September), followed by the Nooksack River (North Fork, 254 ± 0.8), lower Dungeness River (265 ± 2.7), Hood Canal (major rivers only, 269 ± 6.5), Puyallup River (270 ± 4.1), Skagit River (275 ± 0.8), Stillaguamish River (279 ± 1.3), Snohomish River (280 ± 0.7), and Nisqually River (283 ± 1.3). The general pattern appears to be that peak spawning is earliest for Strait of Juan de Fuca populations and the Nooksack River, with peak spawning in Hood Canal occurring 2-3 weeks later on average, and peak spawning in Puget Sound (excluding the Nooksack and Nisqually Rivers) occurring another week after that.

Run timing for Nisqually River pink salmon is the latest for all Washington populations examined, based on 2 years of surveys. Timing of peak spawning of even-year pink salmon in the Snohomish River is about 3-4 weeks earlier than that of odd-year fish, even though these two groups of fish use some of the same habitat (D. Hendrick - Footnote 15). Even-year Snohomish River fish spawn during about the same time that even-year fish spawn farther north in central British Columbia (Aro and Shepard 1967; Gould et al. 1988; Stefanson et al. 1989, 1991).


Table 2. Comparison of estimates of median run timing for Washington pink salmon (generated from 1983-93 spawning curves; WDFW, unpubl. data) to spawn timing ranges and approximate midranges (WDF et al. 1993). Estimates are in calendar days, where 1 September = day 244 and 15 October = day 288. N refers to the number of spawner surveys in the sample.

1983-93 Median run timing

Spawn timinga
Drainage Avg. SD
N Start Midrange End
Northern Puget Sound
Nooksack (NF) 254 3.7 24 233 251 269
Thompson Cr. 254 2.1 6 -- -- --
Cornell Cr. 254 6.4 6 -- -- --
Gallop Cr. 253 2.6 6 -- -- --
Other (pooled) 254 3.2 6 -- -- --
Nooksack (SF) -- -- -- 269 275 281

Skagit 275b 3.8 23 238 270 303
Upper Skagit 275 2.5 6 -- -- --
Mid Skagit 270 5.1 6 -- -- --
Lower Skagit 278 4.8 5 -- -- --
Sauk 260 -- 1c -- -- --
Other (pooled) 276 3.4 6 -- -- --

Stillaguamish 279 6.1 23 -- -- --
North Fork 273 3.6 6 251 273 296
South Fork 281 9.0 6 256 279 303
Pilchuck Cr. & 287 3.0 6 -- -- --
MS Stillag
Other (pooled) 276 8.2 6 -- -- --

Snohomish 280 2.5 12 254 278 303
(odd year)
Wallace 276 2.5 5 -- -- --
MS Snohomish 283 4.0 3 -- -- --
Elwell Cr. & 281 2.1 5 -- -- --
other
Snohomish -- -- -- 254 266 278
(even year)
Southern Puget Sound
Puyallup & 265 3.8 5 251 275 299
S. Prairie Cr.
Kapowsin 276 5.0 4 -- -- --
Nisqually 283 1.9 2 254 278 303
Hood Canal
Dosewallips 272 7.1 5 244 270 298
Duckabush 269 7.5 5 244 270 298
Hamma Hamma 267 6.6 5 244 270 298
Lilliwaup Cr. 271 7.8 5 -- -- --
Skokomish 272 11.1 5 -- -- --
Strait of Juan de Fuca
Lower Dungeness 265 6.0 5 244 265 286
Upper Dungeness 245c 2.4 9 218 242 265
East Fork 245 2.4 4 -- -- --
Grey Wolf 245 2.7 6 -- -- --
Forks 240 -- 1c -- -- --
Elwha 235 -- 1e 254 268 283

a - Estimates for the lower and upper Dungeness River are revised from those in the Washington State Salmon and Steelhead Stock Inventory (J. Uehara, Washington Department of Fish and Wildlife, P.O. Box 43151, Olympia, WA 98501, pers. commun., April 1995).
b - Does not include Sauk River data.
c - 1993 data.
d - Does not include Forks data.
e - Based on 87 adults returning in 1985.
WDFW = Washington Department of Fish and Wildlife.
WDF = Washington Department of Fisheries.
NF = North Fork.
SF = South Fork.
MS = main stem.


Emergence timing--WDFW biologists have also collected unpublished observations on development of pink salmon embryos during redd surveys in many of the Washington rivers that support these fish. However, the collection of this information has become infrequent in recent years. Examination of the available information indicates that hatching generally occurs between late January and early March, but the patterns are highly variable both among river systems and within these systems among years. No clear interdrainage patterns are evident from these data. Emergence timing is not directly assessed with these data but would appear to occur primarily in March and April. Because of large gaps in the data and the sensitivity of embryonic development to environmental conditions in the river, the BRT concluded that this information is not likely to be useful in aiding ESU determinations for Washington pink salmon.

Outmigration and Estuarine Use

Juvenile pink salmon generally begin migrating downstream immediately upon emerging from the gravel. Based on work in the Fraser River by Vernon (1966) and in Hooknose Creek, British Columbia by Hunter (1959), pink salmon fry migrate downstream primarily in March and April in this region, although migration can extend into May. Pink salmon outmigrating from rivers draining into Puget Sound and Hood Canal appear to use nearshore areas in these embayments extensively for early rearing (Jewell 1966). The degree of habitat overlap between odd-year populations is unknown.

Little is known about the migratory and feeding habits of juvenile pink salmon outmigrating into the Strait of Juan de Fuca. Hiss (1994) found that juvenile pink salmon began migrating into Dungeness Bay in 1994 by April, peaked in their migration in late April, and had largely left the estuary by the third week in May. Fry ranged in length from 35 to 75 mm over the sampling period; some fry captured in early April were more than 45 mm long, suggesting that they had entered the estuary in March.

In addition, preemergent fry samples collected in the Dungeness River in late February and early March showed that 80-90% of fry were near yolk-sac absorption, suggesting an early outmigration (WDFW, unpubl. data). It was unclear from Hiss (1994) study whether there was any spatial or temporal separation between juvenile pink salmon from the lower and upper Dungeness River. The study did not address where these pink salmon go after they leave Dungeness Bay; it is unclear whether these fish migrate west through the Strait of Juan de Fuca or (possibly) east and north through the Strait of Georgia.

Resident Fish

Like pink salmon in many populations from more northern regions, pink salmon in British Columbia and Washington often spend an extended period in the nearshore marine environment feeding and growing rapidly before they move offshore (Manzer and Shepard 1962, Phillips and Barraclough 1978, Healey 1980). This period may be as long as 2 to 3 months (reviewed by Heard 1991). However, most pink salmon migrate to the open ocean by late summer or early fall.

An exception to this pattern appears in some pink salmon from Puget Sound (and possibly Hood Canal) that spend their entire marine phase in the nearshore environment. This behavior has been inferred from tagging studies described by Jensen (1956) and Hartt and Dell (1986). Jensen (1956) tagged 873 small (1-1.5 kg) pink salmon in the Tacoma Narrows in June 1955 and sampled these fish in various parts of Puget Sound for the next 4 months to ascertain their origin. These fish were presumed to be resident to Puget Sound on the basis of their small size. Tagged fish remained in the general area for about 1 month, then moved generally north along the eastern shore of Puget Sound. Most of the tags were recovered by sport fishers. The results were largely inconclusive, but three of the four recoveries were made on spawning grounds in the Stillaguamish River, and Jensen suggested the Stillaguamish River population as a possible source of resident pink salmon. Small pink salmon have been repeatedly observed spawning in Pilchuck Creek, a Stillaguamish River tributary, as long as 20 years ago (W. Waknitz - Footnote 16).

The Suquamish Tribe sampled immature pink salmon in central Puget Sound in October and November 1976, catching hundreds of them to both the north and south of Bainbridge Island (Hartt and Dell 1986). These fish were about 23 cm long at capture, a size described as larger than most pinks that had migrated hundreds of miles in the open sea by September (Hartt and Dell 1986). The captured fish would probably not migrate offshore that late in the year. It is unlikely that this reticence to move to the open ocean is related to their size; Jensen (1956) believed these fish adopted residence in Puget Sound for the marine phase of the life cycle. So far as is known, this unusual practice of forgoing migration on the high seas is not documented for any populations of pink salmon outside Washington (W. Heard - Footnote 17). Resident chinook and coho salmon also occur in Puget Sound (Haw et al. 1967).

Presumably resident pink salmon (based on their small size of 35-45 cm) supported a sport fishery in Puget Sound in odd-years from the late 1940s until the early 1960s (Haw et al. 1967). These fish were harvested primarily in southern Puget Sound, especially in the Tacoma Narrows and Commencement Bay areas, with gear that incorporated an attractant device composed of a large spinner blade preceded by a rudder to control line twist (known as shovel and rudder gear; F. Haw - Footnote 18).

Body Morphometry, Fecundity, and Egg Size

Body morphometry--Beacham (1985) and Beacham et al. (1988) observed significant variation in body size and shape both within and among broodlines of British Columbia pink salmon. Variation in morphometric characters (particularly head size, caudal peduncle thickness, and fin size) tended to show greater variation among fish in different broodlines returning to the same river drainage than among fish of different populations within broodlines. Because larger fish with larger heads, thicker caudal peduncles, and larger fins were usually associated with large rivers, the authors interpreted these results as evidence for morphometric adaptation to natal river conditions.

To assist in stock separation analyses, WDFW and its predecessor WDF have been collecting tissue samples from adult pink salmon for protein electrophoresis in several Washington rivers since 1985. Staff biologists have also measured the body lengths of sampled adults (WDFW, unpubl. data). Although these samples may not be representative of the entire spawning populations, sampling in each river was generally conducted in more than 1 year and often in several different locations (S. Young - Footnote 19). Early measurements were of snout-to-fork (SF) length (a measure commonly used by U.S. fishery biologists) to the nearest centimeter, but in more recent sampling the measurements have been of postorbital-to-hypural plate (POH) length (a measure more commonly used by Canadian fishery biologists). The POH measurement has the advantage that it is unaffected by snout and tail erosion in spawning adults.

All even-year fish sampled from the Snohomish River were collected in 1990, and all fish measurements were POH lengths. For odd-year populations, fish measurements were either SF or POH lengths, depending on the year of sampling. Separate linear regression equations were necessary to convert SF to POH for both males and females. Regression equations were not available to convert SF to POH; instead, regression equations to convert SF to mideye-to-hypural plate (MEH) length were generated from relationships between SF and mideye-to-fork (MEF) length and between MEF and MEH given for Alaskan pink salmon by Nickerson (1979). Although MEH is smaller than POH, the use of MEH can be justified because 1) MEH and POH are highly correlated, 2) the difference between them is small (< 1 cm), and 3) all measurements analyzed here are rounded to the nearest centimeter. Thus, the practice of using both MEH and POH in this analysis should have little potential to bias the results. The equations (including their correlation coefficients and sample sizes) relating MEH and SF for odd-year pink salmon are

males: MEH = 2.4301 + 0.8179 SF; r2 = 0.974 (n = 27);

females: MEH = 0.6777 + 0.8652 SF; r2 = 0.972 (n = 38).

The results of these manipulations of WDFW s unpublished size data are summarized in Table 3. Analysis of variance of the data indicated highly significant (P < 0.0001) effects of sex and sampling year. The sex effect reflects the strong sexual dimorphism that is characteristic of pink salmon. Multiple comparisons (Newman-Keuls test, Sokal and Rohlf 1981) among sampling years, after data among drainages were pooled, indicated a substantial general decline in MEH for odd-year pink salmon over time (1993 < 1991 < 1989 < 1987 < 1985; P < 0.0001). The data in Table 3 indicate that the size decline over time is common to all the populations sampled in more than one year. Data are available for Snohomish River even-year pink salmon from only 1 year, but these were the smallest fish observed in all Washington samples (Table 3).

Significant variation among drainages was also observed in odd-year pink salmon when sampling years and sexes were pooled. The relationship among drainages, in order of increasing adult MEH length, was

Nooksack < lower Dungeness < Nisqually < upper Dungeness < [Skagit, Snohomish,
Stillaguamish, Dosewallips] < [Puyallup, Duckabush] < Hamma Hamma.

where group differences were significant (P < 0.05). Hood Canal Hatchery fish were similar in length to Skagit, Snohomish, Stillaguamish, and Dosewallips River fish. In general, the smallest fish appear to exist in cold, turbid rivers in Puget Sound (Nooksack and Nisqually Rivers) and along the Strait of Juan de Fuca (lower and upper Dungeness River), and the largest fish tend to exist in Hood Canal. Although the Puyallup River is also heavily glacially influenced, all fish were sampled from a clear water tributary, South Prairie Creek, where the majority of spawning in that river is observed (WDF et al. 1993, J. Uehara - Footnote 20). No size data are available for Elwha River pink salmon.

Beacham and Murray (1985) and Beacham et al. (1988) calculated POH length data for several even- and odd-year pink salmon populations in British Columbia, and these data provide some basis for comparison with the Washington data. Beacham and Murray (1985) found that even-year fish were shorter (P < 0.01) than odd-year fish among the central and southern British Columbia populations they examined, a general result previously reported for body weight by Godfrey (1959). Snohomish River even-year adult pink salmon sampled by WDF in 1990 are similar in size to pink salmon in the even-year populations from the central mainland and northern Vancouver Island described by Beacham and Murray, although the Snohomish River fish tend to be smaller than the even-year British Columbia average. The length estimates for odd-year populations in Washington taken in recent years (Table 3), with the exception of the Snohomish and Skagit River populations, tend to be smaller than the average for odd-year British Columbia populations reported by Beacham and Murray (1985). However, many of the lengths taken in 1985 and 1987 in Washington are larger than the British Columbia average, which is based on 1981 and 1983 collections.


Table 3. Means and standard errors (SE) of mideye-to-hypural plate lengths (odd-year fish) and postorbital-to-hypural plate lengths (even-year fish) for adult pink salmon sampled in Washington (WDFW, unpubl. data). For each drainage, multiple samples collected in the same year have been pooled to increase sample sizes; mean sample size was 105.5 fish (range, 18-274). Measurements are in centimeters.

Drainage

Sex
1985
Mean ± SE
1987
Mean ± SE
1989
Mean ± SE
1990
Mean ± SE
1991
Mean ± SE
1993
Mean ± SE
Odd-year
Nooksack m -- 42.4 ± 0.4 -- -- 36.7 ± 0.3 36.6 ± 0.6
f -- 42.9 ± 0.3 -- -- 38.7 ± 0.3 37.8 ± 0.6

Skagit m 48.9 ± 0.8 46.3 ± 0.4 43.9 ± 0.3 -- 42.4 ± 0.7 --
f 46.4 ± 0.5 44.3 ± 0.3 42.4 ± 0.2 -- 41.1 ± 0.3 --

Stillaguamish m 48.8 ± 2.3 46.9 ± 0.5 -- -- -- 40.5 ± 0.4
f 48.2 ± 0.4 44.7 ± 0.3 -- -- -- 39.6 ± 0.3

Snohomish m 52.4 ± 0.5 47.3 ± 0.4 44.4 ± 0.5 -- 43.0 ± 0.5 --
f 46.9 ± 0.4 43.8 ± 0.5 41.9 ± 0.5 -- 41.4 ± 0.3 --

Puyallup m 49.8 ± 0.7 44.2 ± 0.6 -- -- -- --
f 48.5 ± 0.6 43.9 ± 0.4 -- -- -- --

Nisqually m -- -- -- -- 41.5 ± 0.5 --
f -- -- -- -- 40.7 ± 0.2 --

Dosewallips m 53.2 ± 0.7 47.6 ± 0.7 42.5 ± 0.4 -- 43.8 ± 1.5 41.1 ± 0.3
f 50.6 ± 1.2 43.4 ± 0.3 40.7 ± 0.6 -- 42.8 ± 0.7 39.7 ± 0.3

Duckabush m 49.6 ± 0.8 48.7 ± 0.7 -- -- -- --
f 45.9 ± 0.8 43.0 ± 0.2 -- -- -- --

Hamma m 50.7 ± 0.7 47.8 ± 0.8 -- -- -- --
Hamma f 48.3 ± 0.9 43.9 ± 0.4 -- -- -- --

Upper m -- 46.3 ± 0.4 43.6 ± 0.5 -- -- 40.2 ± 0.6
Dungeness f -- 43.4 ± 0.4 40.8 ± 0.4 -- -- 37.7 ± 0.3

Lower m -- -- 42.0 ± 0.7 -- 40.7 ± 0.4 41.3 ± 1.3
Dungeness f -- -- 40.2 ± 0.3 -- 38.6 ± 0.2 38.4 ± 0.6
Even-year
Snohomish m -- -- -- 37.2 ± 0.4 -- --
f -- -- -- 37.1 ± 0.7 -- --
WDFW = Washington Department of Fish and Wildlife.

Snohomish River even-year pink salmon, based on the 1990 sample, are most similar in length to even-year males sampled from large rivers on the central British Columbia coast and to even-year females sampled from small rivers on Vancouver Island (Beacham et al. 1988). For odd-year males, the greatest length similarities between Washington and British Columbia pink salmon samples are Nooksack River and small Fraser River tributaries; all other Puget Sound rivers and large Fraser River tributaries, except for Nisqually River, which is most similar to large south coast and Vancouver Island rivers; all Hood Canal rivers and large Fraser River tributaries; upper Dungeness River and large Fraser River tributaries; and lower Dungeness River and small Fraser River tributaries.

For odd-year females, the greatest similarities exist between the following: Nisqually River and both large and small north coast rivers as well as small Fraser River tributaries; lower Dungeness River and large Skeena River and Vancouver Island tributaries; and all other Puget Sound rivers and large Fraser River tributaries. These odd-year comparisons are based on Washington collections made in 1985 (the last year sampling was conducted by Beacham et al. 1988), except for the Nooksack (1987), Nisqually (1991), upper Dungeness (1987), and lower Dungeness Rivers (1989). In Washington, there is no apparent relationship between the length of adult pink salmon and river size, as reported for the more comprehensive survey of British Columbia populations by Beacham et al. (1988). No data are available for other morphometric characters in Washington pink salmon.

Two caveats are important to emphasize in interpreting these comparisons. First, the possibility that sampling was not random within drainages cannot be excluded and could bias these results. Second, because much of the Washington and British Columbia length sampling was done in different years, these comparisons should be regarded with caution, as adult pink salmon size can vary significantly between years (Ricker et al. 1978).

Fecundity and egg size--Beacham and Murray (1993) summarized geographic patterns of fecundity and egg size in pink salmon. When fecundity was standardized to a body length of 435 mm, fecundity in both even- and odd-year fish tended to increase with latitude over the range of about 49°N to 61°N in North America. However, available information was meager for Alaskan populations, and considerable variability existed among the British Columbia populations sampled. For odd-year pink salmon in British Columbia, Beacham and Murray (1993) found that fecundity was generally lower in Vancouver Island and upper Fraser River populations than in other British Columbia populations. For even-year British Columbia pink salmon, fish from the Queen Charlotte Islands had the highest fecundities.

Pink salmon egg size (weight and diameter) showed considerable geographic variation within both broodlines but showed no clear relationship with latitude for either broodline. Egg size (when standardized by body length) tended to be larger in even- than in odd-year pink salmon in British Columbia (Beacham et al. 1988, Beacham and Murray 1993). In British Columbia, egg weight and diameter tended to be largest in Skeena River populations for odd- year fish and in the Queen Charlotte Islands for even-year fish. Southern British Columbia populations showed a considerable degree of scatter in egg size (Beacham and Murray 1993). Among several odd-year populations of British Columbia pink salmon sampled for egg size, populations with larger eggs produced heavier alevins with greater amounts of yolk, but the relationship between egg and alevin size depended on temperature under controlled incubation conditions (Beacham and Murray 1986). No clear relationship between egg or alevin size and embryonic survival was observed by these workers.

Homing and Straying

Homing and straying are prominent features of Pacific salmon biology that can have significant effects on population structure. Consequently, these related behaviors are relevant to ESU determinations for these species. Pink salmon have a widespread reputation for straying at higher rates than other species of Pacific salmon (e.g., Horrall 1981). If true, then if straying pink salmon also reproduce successfully, the elevated rate of gene flow would be expected to result in a less conspicuous population structure and, potentially, reduced opportunity for local adaptations to be maintained in this species. (The analysis of genetic variation provided in the next section sheds some light on the level of population structure in pink salmon.)

Few well-designed studies have been carried out to estimate straying rates in Pacific salmon, and the empirical evidence supporting the hypothesis that pink salmon stray at relatively higher rates than other species of Oncorhynchus is mixed (Quinn 1993, Altukhov and Salmenkova 1994). Early work by Davidson (1934) in Washington (Hood Canal) and southeastern Alaska and by Pritchard (1939) in British Columbia produced estimates of straying of about 10%, based on recovery of tagged adults, but Ricker (1962) concluded that this rate was probably rarely exceeded by natural pink salmon.

Helle (1966) moved adult pink salmon from Olsen Creek in Prince William Sound, Alaska to a location 5 km away and recovered 91% of them in Olsen Creek. Gharrett (1985) detected no straying of genetically marked odd-year pink salmon from Auke Creek, Alaska to two other pink salmon streams within 10 km of Auke Creek; however, he did provide some evidence for higher straying of even-year fish to both another run within Auke Creek and another drainage. Altukhov and Salmenkova (1994) reviewed pink salmon straying data from marked fry released from North American and Asian hatcheries and found that estimates ranged from 0.1 to 11.5%. Although most estimates of straying are based on recoveries of adults from streams in the vicinity of the natal stream, Heard (1991) discussed some data that indicate pink salmon can travel several hundred kilometers from the natal stream during their adult migration. Quinn (1993) summarized these studies in this way: the conclusion that [pink salmon] stray more commonly than other salmon species seems premature.

Nevertheless, the rapid colonization of systems newly available to pink salmon indicates that this species has an unusual ability to expand into suitable habitat when conditions are favorable. The rapid expansion of pink salmon into the Great Lakes after their accidental introduction (Kwain and Laurie 1981); the recolonization of Sashin Creek, southeastern Alaska, by even-year pink salmon after nearly complete experimental removal of the even-year run (Merrell 1962, Heard 1991); and the immediate recolonization of the upper Fraser River after almost 35 years of blocked access (Vernon 1962) are testaments to this ability. In addition, the reports of pink salmon periodically spawning south of northwestern Washington, in the absence of any evidence for permanent spawning populations in these southern areas, suggest that pink salmon homing behavior is highly plastic. Finally, maturing pink salmon show a strong propensity to move around between neighboring streams, at least in certain areas, prior to spawning (Jones and Thomason 1983).

Straying in pink salmon may depend strongly on spawning location and on conditions at time of spawning. In a recent study in Prince William Sound, Sharp et al. (1994) estimated straying rates between 9 and 53% in coded-wire tagged odd-year wild and hatchery pink salmon. Several factors may contribute to the relatively high straying observed: 1) Prince William Sound is a highly dynamic geological zone, having experienced two major earthquakes in the last century that destroyed many streams and created others; 2) a large fraction of Prince William Sound pink salmon spawn intertidally--approximately 72-77% in even years and 35-57% in odd years prior to the 1964 Good Friday earthquake (Noerenberg 1963); and 3) the southwestern part of Prince William Sound was heavily affected by the 11- million gallon Exxon Valdez oil spill in 1989 (S. Rice - Footnote 21). In addition, because coded-wire tagging of juvenile pink salmon must occur at the emergent fry stage, some question has been raised as to whether the tagging itself may impair the homing of these fish (Morrison and Zajac 1987; J. Seeb, Alaska Department of Fish and Game, pers. commun., cited in Mathisen 1994).

Levels of natural straying in pink salmon are therefore unclear and may vary widely among populations and within populations under different conditions. More importantly, the genetic consequences of straying in pink salmon are not well understood.


Table of Contents