Chum salmon juveniles, like other anadromous salmonids, use estuaries to feed before beginning long-distance oceanic migrations. However, chum and ocean-type chinook salmon usually have longer residence times in estuaries than do other anadromous salmonids (Dorcey et al. 1978, Healey 1982).21 The period of estuarine residence appears to be the most critical phase in the life history of chum salmon and appears to play a major role in determining the size of the subsequent adult run back to freshwater (Mazer and Shepard 1962, Bakkala 1970, Mathews and Senn 1975, Fraser et al. 1978, Peterman 1978, Sakuramoto and Yamada 1980, Martin et al. 1986, Healey 1982, Bax 1983a, Salo 1991). Bax (1983b) determined that the extent of juvenile mortality within 4 days after a hatchery release into the Hood Canal estuary was 31-46%. The most important determinant of estuarine survival may be the timing of entry into saltwater because of the strong seasonality of plankton in estuaries (Gunsolus 1978, Helle 1979, Gallagher 1979, Simenstad and Salo 1982).
For these reasons, the estuarine life-history phase has been the most intensely studied period of chum salmon early life history. In all countries with large chum salmon hatchery programs, research has focused on determining the optimal size and time for release of juvenile chum salmon to enhance adult returns. How changes in estuaries affect the use of estuaries by chum salmon, and how such changes influence the time of seawater entry, run and spawn timing, and growth and maturation, may be important in delimiting ESUs.
This term denotes the life-history stage in which juvenile salmonids lose their parr marks, turn silvery, and migrate from freshwater into seawater. In summary, smoltification is perhaps the most intensively studied aspect of salmonid life history. Groot et al. (1995) contains an extensive review of both historical and recent studies on this process. Coho, stream-type chinook, lake-type sockeye salmon, and steelhead have a distinct "smolt" stage that can be identified visually among fry, parr, and fingerlings (Hoar 1958). Chum and pink salmon, however, do not have clearly defined smolt stages, but are nonetheless capable of adapting to seawater soon after emerging from gravel. Chum salmon also usually retain parr marks when they first enter seawater. In Japan, chum salmon fry weighing less than 2 g maintained normal levels of plasma sodium (Na+) when they moved from freshwater into seawater (Iwata 1982). This ability, however, declines slightly with continued residence in freshwater. The capability of chum salmon fry for early osmoregulation in seawater may be important for adult homing back to natal streams. For example, hatchery coho salmon were 10 times less likely to stray within a river system if they were released into the river as fingerlings rather than as smolts (McHenry 1981, cited in Lister et al. 1981).
Chum salmon fry from various spawning populations and adult runs in a river system tend to enter seawater at a similar time, one that maximizes their chance of survival. The most critical factor for survival of fry within an estuary appears to be fish size (Healey 1982). Similar entry timing into an estuary by fish from different rivers may be an adaptation to temporally variable food resources because plankton abundance in estuaries is highly seasonal (Gunsolus 1978, Helle 1979, Gallagher 1979, Simenstad and Salo 1982). Walters et al. (1978) developed a model of optimal timing for downstream migration and entry into estuaries to maximize early marine survival. The parameters for this model included 1) zooplankton production, 2) diet and growth of young salmon, 3) size-dependent survival, and 4) timing of fry outmigration into saltwater. For juvenile chum salmon in the Fraser River estuary, this model demonstrated a close correlation between timing of seawater entry and early chum salmon survival.
Chum salmon juveniles of early-returning adults tend to enter estuaries before juveniles of late-returning fish (Koski 1975). Unlike some other species--sockeye salmon, for example, which move immediately into deep water after entering an estuary--chum salmon tend to remain in shallow eelgrass beds or other productive areas within the estuary from January to July (Healy 1982). Residence times are known for only a few estuaries, even though residence timing has been studied since the 1940s (reviewed in Congleton 1979, Healey 1982, Simenstad et al. 1982, Bax 1983a). Observed residence times range from 4 to 32 days, with a period of about 24 days being the most common (Table 10).
Migration patterns of juvenile chum salmon have been studied intensively in areas such as Hood Canal by following marked juveniles from hatchery populations of fall-run chum salmon and by monitoring outmigration (Bax 1982, 1983a,b; Bax et al. 1979, Bax et al. 1980; Bax and Whitmus 1981; Schreiner 1977; Whitmus and Olsen 1979; Whitmus 1985; Salo et al. 1980). Some fry remain near the mouth of their natal river when they enter an estuary, but most disperse within a few hours into tidal creeks and sloughs up to several kilometers from the mouth of their natal river. In the Nanaimo and Fraser River estuaries, juveniles spend up to 3 weeks feeding in the inner estuary, with little local movement (Healey 1979, Levy et al. 1979). Chum salmon juveniles in the Nanaimo, Yaquina, Cowichan, and Courtenay estuaries are most abundant in nearshore areas during April and May, but are most abundant in the outer estuary during May and June (Myers 1980, Healey 1982).
Chum salmon fry show daily tidal migrations in the Fraser and Nanaimo Rivers, which have large deltas and marshlands (Healey 1982). However, fry in Hood Canal have not been observed to display daily tidal migrations (Bax 1983a), most likely because rivers entering Hood Canal do not have extensive delta or tidal marsh systems (with the exceptions of the Quilcene and Skokomish Rivers).
Although in general, movements of chum salmon fry in Hood Canal appear to follow a pattern that depends on the time of release from hatcheries, release time is not the only causative factor influencing migratory patterns (Bax 1982, 1983a). Chum salmon fry released into Hood Canal in early February and March have spread out over a large area, but fish released in April and early May tended to remain inshore initially, moving offshore in summer. These movements were apparently associated with prey availability. Fish initially fed inshore on epibenthic organisms, then offshore on plankton later in the season. Foraging success (growth), as well as age, appeared to be the major factors in offshore movements of chum salmon juveniles in the Strait of Georgia in early June and July (Healey 1980).
A seasonal change in the swimming speed of juvenile chum salmon was also observed in Hood Canal (Bax 1982). In February and March, chum salmon fry (<40 mm fork length) usually moved 8-14 km/day, but in May and June they moved 3-7 km/day. Fish released from the Hood Canal Hatchery at Hoodsport in June took 3 weeks to arrive at the mouth of Hood Canal (Bangor Annex), but arrived after only 1 week when released in April. However, this pattern may not be consistent from year to year: Bax (1982) found that in 2 of the 3 years of a study (1977-1979), average migration speed was lower in February than in March. Larger fish tended to move faster than small fish, but the overall rate of movement decreased as the season progressed (Bax 1983a).
Larger fish also tended to move out of an area first when both large and small chum salmon fry were simultaneously released from hatcheries on the Skokomish River in southern Hood Canal (Bax and Whitmus 1981, Whitmus 1985). In southern British Columbia, larger fish also tended to migrate offshore first, but by mid-July all chum salmon juveniles had left the estuaries, regardless of size (Healey 1980). A similar timing of outmigration was observed in the Yaquina River estuary in Oregon, except that most chum juveniles had moved out of the estuary by mid-May (Myers 1980).
Reasons for the differences in movement patterns among areas are unclear. One reason may be that genetically based physiological differences between runs produce different behavior patterns (Whitmus 1985). Migration may be facilitated by both active swimming and passive movement in currents, and seasonal changes in river discharge and surface flow in Hood Canal may contribute to different migration patterns among populations (Bax 1982). The rate of movement, especially early in the season, appears to be correlated with surface outflow in Hood Canal (Bax 1982, 1983a). Movement of chum salmon juveniles apparently is also influenced by the abundance of pink and chum salmon juveniles of the same size class in Hood Canal (Salo 1991).
Hood Canal fall-run chum salmon--Establishing riverine or estuarine residence times of hatchery fall-run chum salmon is particularly important because of possible overlap with summer-run chum salmon juveniles in Hood Canal. Chum salmon juveniles released from the Hood Canal and Quilcene Hatcheries in Hood Canal showed little delay in their migration out of Hood Canal, traveling in distinct groups past the Bangor Annex sampling site in the northern part of Hood Canal (Whitmus 1985). However, some portion of large and small hatchery chum salmon fry released at the same time from hatcheries in the southern part of Hood Canal resided in the large Skokomish River delta region for up to 4 weeks (Whitmus 1985). Although the hatchery fish released into southern Hood Canal were derived from the Hood Canal and Quilcene hatcheries, they not only had longer residence times near their release site but failed to pass the Bangor Annex as a distinct group (Whitmus 1985).
Hood Canal summer-run chum salmon--No experimental mark-and-release studies have been conducted on natural or fall chum in Hood Canal, including summer-run chum salmon. However, the outmigration of chum was monitored before the release of hatchery fish into Hood Canal, and small peaks of outmigrants have been observed in February and March at sites on both the east and west sides of Hood Canal (Bax et al. 1979, Bax et al. 1980, Bax 1982, 1983a).
Juveniles from early-spawning adults at Big Beef Creek were observed passing Bangor in Hood Canal (Fig. 1) one week after peak outmigration from Big Beef Creek, but juveniles from late-run fish, which had emerged in April, took two weeks to cover the same distance (Bax 1982). These differences were interpreted to result from differences in surface outflow in Hood Canal rather than from any intrinsic behavior in the chum salmon juveniles (Bax 1982).
While these results indicate that summer-run chum salmon quickly migrate up the Canal and into the main body of Puget Sound, preliminary data from ongoing snorkel and beach-seine surveys by the U.S. Fish and Wildlife Service (USFWS) have revealed the presence of natural chum salmon juveniles in Quilcene Bay, Hood Canal, from mid-January to mid-April (Tabor-Cook et al.22). These observations would suggest that either these fish emerge from streams over an extended period or that juveniles remain in Quilcene Bay for several weeks. Washington fisheries co-managers are currently conducting studies to clarify residence times and the timing of juvenile migration for summer-run juveniles in Hood Canal (Tynan23).
Based on the time when summer-run chum salmon spawn, emergence of fry would be expected at least a month prior to fall-run chum salmon emergence, but Koski (1975) found that summer-run chum salmon in Big Beef Creek delayed embryonic development by an average of 12 days compared to the length of time it took fall-run chum salmon to emerge from the gravel. However, in so doing, the fry apparently sacrificed their robustness (mass) and lipid reserves. Therefore, if size at ocean entry is similar for all chum salmon (Peterman 1978), summer-run chum salmon would be expected to spend more time in a nearshore estuarine environment, either Hood Canal or the main body of Puget Sound, until they reached this optimal size (Salo 1991).
Japan--In Japan, prey availability may also influence the estuarine migration patterns of juvenile chum salmon. The warm summer Oyashio Current off northern Japan moves inshore in May and June, forcing cold-water oceanic plankton populations to move far offshore: Many fish are then forced to move offshore also. Kaeriyama (1986, 1989) divided juvenile chum salmon from this region into three groups: 1) "river" type, which remain in the river until they are large enough to migrate offshore as the warm current approaches; 2) "foraging" type, which move into offshore feeding areas in February and March before the warm currents arrive; and 3) "escape foragers," which migrate to low-salinity inshore and estuarine waters, where they feed until the Oyashio Current retreats in June and July.
Growth greatly influences the survival and migration timing of juvenile chum salmon. Chum salmon grow rapidly in estuaries, and even though growth rates vary substantially between areas, there are some consistent patterns. Salo et al. (1980) and Bax and Whitmus (1981) measured growth of chum salmon fry in Hood Canal and found a daily gain of 5.7-8.6% body mass, but found gains as high as 10.1% body mass in the first 4 days of estuarine residence. Chum salmon in the Skagit River salt marsh grew 6% of their body mass per day (Congleton 1979, Congleton et al. 1982). Marked chum captured in the Nanaimo River estuary on Vancouver Island also grew at about 5.7% of their body mass per day (Healey 1979), whereas unmarked chum salmon in the Nitinat Lake (British Columbia) estuary grew only about 2.7% body mass per day. However, Healey (1982) suggested this latter growth rate may have been underestimated because larger fish migrated seaward from the lake. He suggested that the true growth rate was closer to 3.5% body mass per day, but was still significantly less than the growth rates in the Nanaimo estuary.
In some areas, such as the Fraser River estuary, chum salmon fry captured high in the estuarine marshes were smaller than those captured in other estuaries (Levy and Northcote 1981), but the fry were larger farther seaward in the Fraser River (Goodman 1975). Fry also increased in size more slowly in the Squamish (British Columbia) and Yaquina (Oregon) River estuaries than in the Nanaimo estuary (Levy and Levings 1978, Myers 1980). However, in all of the above studies, chum salmon juveniles captured in estuaries were heavier than those captured in rivers, and the difference increased with time. Also, while some fry remained for considerable time in the inner estuary, all fry moved seaward by mid-May, with heavier fish migrating first.
Juvenile salmon, particularly chum and chinook, depend heavily on benthic organisms for food in estuaries, but in outer areas they depend more on planktonic organisms. Detritus-based food webs and juvenile chum salmon production in estuaries are closely linked (Sibert et al. 1978). In the 1970s and 1980s, chum salmon feeding and food-chain relationships were examined in the above studies along with migrational timings of juveniles in fresh and estuarine waters.
Simenstad et al. (1982) summarized the diets of juvenile salmonids in 16 estuaries and concluded that small (< or = 50-60 mm FL) juveniles of chum salmon fed primarily on such epibenthic crustaceans as harpacticoid copepods, gammarid amphipods, and isopods, whereas larger juveniles (>50-60 mm FL) in neritic habitats fed on drift insects and on such plankton as calanoid copepods, larvaceans, and hyperiid amphipods. This diet is broader than that of similarly sized pink salmon juveniles, which feed only on neritic zooplankton similar to those consumed by large chum salmon juveniles, even in shallow sublittoral habitats (Healey 1982, Simenstad et al. 1982). However, the early diet of juvenile chum salmon at some localities also consists exclusively of neritic zooplankton.
Juvenile chum salmon in the Nanaimo River, British Columbia (Healey 1980) and in Auke Bay, Alaska (Landingham 1982) fed only on harpacticoid copepods; but at other localities, such as at some localities in the Skagit River (Congleton 1979) and in some estuaries of Vancouver Island (Mason 1974), fry fed only on dipterans, primarily chironomids. Comparisons between juvenile chum and chinook salmon in marsh habitats in the Fraser River estuary indicated that the diet of chum salmon varied less from place to place than did the diets of chinook salmon (Levy and Northcote 1981). Feeding preferences among juvenile salmon appear to correlate with the degree to which a species depends on estuarine habitats (Healey 1982). Chum salmon fry may exploit a greater variety of prey, because they can withstand greater changes in salinity than other salmonids. For example, chum salmon fry in the Skagit River fed on freshwater, estuarine, and marine organisms during a single tidal cycle (Congleton 1979).
Migration of chum salmon juveniles out of estuaries appears to be closely correlated with prey availability. Chum salmon juveniles move offshore as they reach a size that allows them to feed on the larger neritic plankton, and this movement normally occurs as inshore prey resources decline (Salo 1991). This transition has taken place at 45-60 mm FL in Puget Sound and Hood Canal, Washington (reviewed in Simenstad and Salo 1982, Salo 1991), but at 60 mm FL in Prince William Sound, Alaska (Cooney et al. 1978).
Do different groups of chum salmon, or chum salmon compared to other salmon species, use estuarine habitats differently? In the few studies published on this subject, juvenile chum and pink salmon apparently occupy shallow sublittoral habitats before moving into neritic habitats. They appear to remain in the shallow areas until they reach 45-60 mm FL, after which they move into neritic habitats. Chum salmon apparently prefer exposed cobble or gravel beaches in nearshore areas (Miller et al. 1977, 1980), especially within embayments. Chum salmon also school in shallow habitats during daylight, but disperse into smaller groups at night (Salo et al. 1980). Juveniles from most runs of chum salmon migrate in schools through northern Puget Sound and into the Strait of Juan de Fuca (Fresh 1979). In contrast, juvenile coho salmon move directly into neritic waters after entering an estuary and school much less than do pink and chum salmon.
Chum salmon interact with other salmonids in several ways (reviewed in Gallagher 1979, Salo 1991). Most notable is the observation that returns from odd-year broods of chum salmon in the Pacific Northwest and from even-year broods in Alaska tend to be lower when pink salmon juveniles coexist with chum salmon juveniles in estuaries (Rounsefell and Kelez 1938, Smirnov 1947, Lovetskaya 1948, Noble 1955). Ivankov and Andreyev (1971) modeled chum salmon populations in southeastern Russia and found that when pink salmon juveniles were abundant, predicted feeding and growth rates of juvenile chum salmon were lower. Ames (1983) also hypothesized that competition for food and predation between pink and chum salmon juveniles in estuary and nearshore marine habitats may cause distinct odd- and even-year cycles in natural chum salmon populations in Puget Sound. The diets of pink and chum salmon in Hood Canal overlapped up to 84% (Simenstad et al. 1980). Chum salmon diet also shifted in years when pink salmon were abundant (Gallagher 1979), but this may have been due in part to a more diverse diet in chum salmon juveniles than in pink salmon juveniles. Invertebrates not eaten by pink salmon juveniles may be more available to chum salmon in years when pink salmon are also abundant.
Culture techniques may also influence interactions between chum salmon juveniles and other salmonids. Estuarine predation on natural and hatchery pink and chum salmon by larger, piscivorous salmon--such as coho and chinook salmon smolts--may have caused declines in some Puget Sound pink and chum salmon populations (Johnson 1973, Simenstad et al. 1982).
Estimates of mortality of chum salmon in freshwater after emergence from the gravel have been made by several authors, including Neave (1953) and Hunter (1959) in British Columbia, Semko (1954) in the Russian Federation, and Beall (1972) and Fresh and Schroder (1987) in Big Beef Creek, Washington. Although estimates of mortality in a study can vary greatly (e.g., mortality varied from 5 to 60% in the study by Fresh and Schroder (1987)), the average for all of the above studies was about 45%. Mortalities in natural habitats can be influenced by such short-term physical factors as extreme cold, water diversions, and flooding. Mortality is also affected by long-term factors such as the cumulative effects of habitat degradation, climatic changes, and urbanization, and by biotic factors such as disease, interspecific competition, bird and other predation, and the introduction of exotic predators and competitors. Fish in most experimental studies are usually monitored in relatively stable environments over a short period of time. Most of the above natural sources of mortality are absent, and the only natural cause of mortality reported is usually predation. Nonetheless, estimates of mortality under such experimental conditions can still be helpful for identifying vulnerable life-history stages. Predation by juvenile coho salmon was the primary cause of mortality to chum salmon in all the freshwater studies reviewed here. In Big Beef Creek on Hood Canal, size selection of chum salmon juveniles by coho salmon was identified by Beall (1972), but in a later study (Fresh and Schroder 1987) size selection by coho salmon and rainbow trout was not observed.
Mortality of chum salmon juveniles, especially those from natural populations, is difficult to estimate in estuaries. In studies on fluorescently marked juvenile chum salmon released from the Enetai Hatchery in Hood Canal, Bax (1983a, b) estimated average daily mortalities between 31 and 46% over a 2- and 4-day period. In a study on releases of equal numbers of fish of two different sizes, Whitmus (1985) estimated that small fish suffered higher mortalities than did large fish. About 58% of the small fish died over 2 days, and of the fish remaining after 10 days only 26% were small fish. This mortality was apparently due to predation by cutthroat trout and marine birds. However, predator selectivity on fish size may have been due to the distribution of the differently sized fish rather than to selective behavior: Large fish avoided predation in the study area by emigrating out of the area sooner than small fish.
It is unclear how long chum salmon juveniles remain in estuarine areas, but chum salmon in the Washington and southern British Columbia generally entered the ocean earlier than did more northern and western populations (Hartt 1980, Hartt and Dell 1986). In studies of juvenile chum salmon (300-400 mm FL) captured and tagged in June in central Puget Sound, Jensen (1956) found that juveniles moved northward to the Strait of Georgia and the west coast of Vancouver Island shortly after release. They appeared to migrate northward along the coast in a narrow band about 32 km in width. Hartt (1980) and Hartt and Dell (1986) summarized available data on the distribution, migration, and growth of chum salmon in their first year at sea and found that chum, pink and sockeye salmon juveniles tended to group together and remained nearer shore (within 36 km) than juvenile coho and chinook salmon and steelhead. Later in the season, pink salmon were extensively caught offshore by Canadian longline gear in November and December, but no juvenile sockeye or chum salmon were caught in these offshore waters. As groups of chum salmon reached Alaska, they moved offshore in a generally southwestern direction, although movement was variable and appeared to be strongly influenced by currents (Hartt 1980, Hartt and Dell 1986). A difficulty in these studies is that few numbers of tagged fish were recovered. In the tag recoveries summaries by Hartt and Dell, over 110,000 juvenile salmon and steelhead were caught and 35,259 tagged, of which 4,412 were chum salmon, although only 6 chum salmon, or 0.1%, were recovered.
A second factor that obscures patterns of oceanic distribution and migration is the extent of delayed ocean migrations and residualism by chum salmon. In the tagging studies by Jensen (1956) juvenile chum salmon remained in nearshore waters beyond the usual time of ocean migration, although the extent of this residualism was unclear (Jensen 1956, Hartt 1980, Fresh et al. 1980, Hartt and Dell 1986). Not all of the chum salmon juveniles tagged in Hood Canal and Puget Sound moved northward toward British Columbia; some remained in Puget Sound throughout the summer, perhaps not leaving until the next spring (Jensen 1956). In November, Hartt and Dell (1986) found juvenile chum salmon in central Puget Sound and in Hecate Strait that averaged 230 mm in length, an indication of good growth. It has been hypothesized that these fish may not make an extended northwest migration along the British Columbia/Alaska coast, but may instead proceed directly offshore into the north Pacific Ocean (Hartt and Dell 1986).
Oceanic distributions of salmonids have been used to differentiate salmonid populations for ESA considerations (Waples et al. 1991, Weitkamp et al. 1995). The International North Pacific Fisheries Commission (INPFC) has collected a large amount of information since 1952 on the distributions and origins of high-seas chum salmon. Studies by the INPFC have focused on tagging experiments and scale pattern analysis (see Davis et al. 1990), and more recently on mixed-stock identification (MSI) (see Winans et al. 1994). There have also been several regional studies on the identification of high-seas chum salmon. These studies have used a variety of techniques including gonad development, scale characteristics, age, morphology, allozyme patterns, mtDNA variation, and DNA fingerprinting (e.g., Altukhov et al. 1980, Okazaki 1986, Nikolaeva 1987, Ishida et al. 1989, Park et al. 1993, Taylor et al. 1994). However, until recently, these studies focused on estimating the continent of origin for chum salmon, and little information has been developed on the distributions of specific regional populations.
Tagging and scale studies by the INPFC showed that although chum salmon from both Asia and North America are distributed throughout the North Pacific Ocean and Bering Sea, Asian chum salmon apparently migrate farther across the Pacific Ocean than do North American fish. Neave et al. (1976) reported that North American chum salmon were rarely found west of the mid-Pacific Ocean (beyond long. 175°E), whereas Asian chum salmon were routinely encountered far east of this line. Asian chum salmon have extended their distribution in recent years into the central and eastern North Pacific Ocean, perhaps because of the large increase in releases of hatchery fish in Japan (Kaeriyama 1989, Salo 1991), and because of the change from high-sea to inshore fisheries by Japan's fishing industry (Kaeriyama 1989, Ogura and Ito 1994). Bigler and Helle (1994) and Helle and Hoffman (1995) suggested that the overlap of continental groups may be detrimental to North American chum salmon because maturing chum salmon in the North Pacific Ocean may be at or above carrying capacity.
Limited information exists on stock- or population-specific migrational patterns, and distributions of chum salmon during their oceanic phase are limited. Maturing chum salmon in the North Pacific begin to move coastward in May and June and enter coastal waters from June to November (Neave et al. 1976, Fredin et al. 1977, Hartt 1980). No region-specific information on chum salmon migrations to Washington and Oregon has been reported. Whether the large populations of chum salmon that once inhabited the Columbia River (Rich 1942) and Tillamook basins (Henry 1953, 1954) had oceanic distributions similar to Puget Sound chum salmon is unknown. As landings in coastal Oregon historically excluded landings on the Oregon side of the Columbia River (Henry 1953), one speculation is that these fish had a more southern distribution, like the present distribution of Columbia River coho salmon (Sandercock 1991), and may have returned northward along the Oregon coast as do some Columbia River coho salmon.
Age and growth in adult chum salmon have been extensively studied (reviewed in Bakkala 1970 and Salo 1991). Although clear trends are difficult to detect because most studies have relied on commercial catches of mixed stocks (Helle 1984), regional differences in growth rate, age at maturity, and size at maturity are evident.
Growth rate--Asian chum salmon grow faster in the marine environment (higher instantaneous growth rate) than do North American chum salmon. North American fish, however, are usually larger at each stage of marine life. Salo (1991) suggested that the faster growth rate in Asian chum salmon may be a genetic (bioenergetic) adaptation to poor habitat conditions in the western Pacific Ocean. LaLanne (1971) and Ricker (1964) suggested that the growing season may be longer for North American fish because scale analysis showed that growth began earlier and ended later in the Gulf of Alaska than it did in Asian waters.
In general, average size for chum salmon of all ages increases from north to south, and age at maturity decreases from north to south (reviewed in Salo 1991). This is attributed to the longer growing season and earlier maturation of southern populations. However, Helle (1984) demonstrated that in individual populations from the Noatak River in northern Alaska to Whiskey Creek in central Oregon, 4-year males during the same year were largest in southern southeast Alaskan mainland (Fish Creek at the head of Portland Canal) and decreased in size both to the north and to the south. The average mass of chum salmon in Alaska was less than in British Columbia, but the average sizes of chum salmon in northern British Columbia did not differ significantly from average sizes in southern British Columbia (Ricker 1980). The greater number of 3-year-old fish in southern catches and greater variability in fish age in northern British Columbia may explain these results (Ricker 1980). Four-year-old chum salmon were longer in northern Puget Sound (average 78.3 cm FL in 1964, and 75.7 cm FL in 1970) than in the southern Puget Sound (from Discovery Bay to Tacoma) (average 74.3 cm FL in 1963-1966 and 72.4 cm FL in 1970) (Pratt 1974).
Age at maturity--Age at maturity also appears to follow a latitudinal trend in which a greater number of older fish occur in the northern portion of the species' range. Age at maturity has been investigated in many studies, and in both Asia and North America, it appears that most chum salmon (95%) mature between 3 and 5 years of age, with 60-90% of the fish maturing at 4 years of age. However, there is a higher proportion of 5-year-old fish in the north, and a higher proportion of 3-year-old fish in the south (southern British Columbia, Washington, Oregon) (Gilbert 1922, Marr 1943, Pritchard 1943, Kobayashi 1961, Oakley 1966, Sano 1966). Helle (1979) has shown that the average age at maturity in Alaska is negatively correlated with growth during the second year of marine life, but not with growth in the first year, and that age at maturity is negatively correlated with year-class strength. A few populations of chum salmon also show an alternation of dominance between 3- to 4-year-old fish, usually in the presence of dominant year classes of pink salmon (Gallagher 1979).
Differences in age at maturity might be expected between summer and fall chum salmon in such rivers as the Yukon and Amur, where distinct differences in the life history of the runs occur. However, Buklis and Barton (1984) found little difference in the age at return for Yukon River summer-run and fall-run chum salmon in commercial catches between 1973 and 1983. Four-year-old fish made up 72.7% of the fall-run fish and 70.5% of the summer-run fish; age-5 fish were next most abundant, followed by age-3 and age-6 fish. An important consideration in all studies of age composition is the large year-to-year variation common to all large data sets on chum salmon.
Several authors have shown that the fluctuations observed in age composition are explained by differences in abundances between brood years (Helle 1979, Buklis and Barton 1984). For example, in Olsen Creek, Prince William Sound, Alaska, 62% of the male chum salmon from the 1956 brood came back as 3-year olds in 1959 (Helle 1979). The 1971 brood produced the next largest percentage (32%) of 3-year-old males during the 20-year study. The percentage of production of 3-year-old males ranged from 0 to 62%. Production of 4-year-old males from the 1956-1972 brood years ranged from 33 to 94%. Production of 5-year-old males from 1956-72 brood years ranged from 1 to 64% (Helle 1979). On the Amur River in the Russian Federation, Lovetskaya (1948) reported data collected in the late 1920s and early 1930s on age at maturity for summer-run and fall-run fish, and found that 4-year-old fish made up 73.0% of the summer-run and 68.6% of the fall-run chum salmon. However, age structure varied widely between years. Only 26.6% of the summer-run fish in 1929 were 4-year-old fish, but 98.1% in 1930 were 4-year-old fish. No data are available for fall-run fish in 1930, although 73.7% in 1929 were 4-year-old fish.
Adult chum salmon have been decreasing in size since the early 1980s (Fig. 12). Helle and Hoffman (1995) found that the average mass of 4-year-old fall chum salmon returning to the Quilcene National Fish Hatchery on the Quilcene River in Hood Canal, Washington and to Fish Creek at the head of Portland Canal in Southeast Alaska declined about 46% between the early 1970s and early 1990s (Helle footnote 14). The average age at maturity for fish at both localities has also increased as size has decreased. Helle and Hoffman (In press) added four more years (1993-96) of data on size changes in chum salmon to their previously published data set from Fish Creek (1972-92) in southern southeast Alaska and Quilcene National Fish Hatchery (1973-92) on the Quilcene River in Hood Canal in Washington (Helle and Hoffman 1995) (Fig. 12). These new data on size changes in chum salmon show an abrupt significant increase in mean length in 1995 and 1996 at Fish Creek and a significant increase in mean length for chum salmon at Quilcene National Fish Hatchery in 1994, 1995 and 1996. Because population abundance of chum salmon in both Asia and North America increased during these years, the authors suggest that the climate regime of the North Pacific Ocean may be undergoing another change (Helle and Hoffman In press).
Earlier, Bigler et al. (1996) examined mean length and mass of chum salmon captured in commercial fisheries throughout North America from 1970 to the early 1990s, and most of the catches showed declines in size. In addition, Japanese and Russian chum salmon have declined in average size and have increased in average age at maturity during the past 15 years (Ishida et al. 1993). The similarity of changes in size and age of chum salmon in North America, as well as in Asia, suggests that these changes are the result of one or more factors in the ocean experience of chum salmon.
A major change in ocean climate in the North Pacific Ocean occurred during 1976-77 (McLain 1984, Miller et al. 1994). Sea surface temperatures cooled in the central North Pacific Ocean and warmed along the coast of North America. These conditions continued through the mid-1990s. Although these oceanic changes enhanced survival and increased abundance, particularly in Alaska and Asia, the relationship of these changes to the decline in size is uncertain. The declining size and older age at maturity may be the result of density-dependent factors (see Kaeriyama 1989, Ishida et al. 1993, Helle and Hoffman 1995, Bigler et al. 1996). Density-dependent growth of salmonids in the ocean is complex and poorly understood (see Peterman 1978, Rogers 1980, Peterman and Wong 1984).
Life-history characteristics are complex attributes of biological organisms, derived from the interaction of genetic and environmental influences. They may reveal population-specific differences among local populations, useful in the identification of ESUs, although life-history differences are often obscured by regional trends, environmental perturbations, and selective sampling. In the BRT's review of chum salmon life-history traits, only run timing appeared to be important in defining ESUs, especially for populations of summer-run and winter-run chum salmon, which show unusual run timings.
Chum salmon show several distinct run-timing patterns: 1) In both Asia and North America, there is a bimodal distribution in return timing, with an early or summer peak and a late or fall peak; 2) these seasonal peaks of returning fish appear to follow a north-to-south cline, from earlier to later spawning times; 3) populations with early run times are common in Alaska and in northern British Columbia, but are rare farther south; and 4) south of southeast Alaska there is a distinct break between natural summer-run and fall-run chum salmon.