Patterns of vegetation in Washington and southern British Columbia are strongly affected by precipitation. Coastal regions in Washington and British Columbia (Fig. 5) were historically forested with a Sitka spruce-dominated floral community, which includes western hemlock, western red cedar, red alder, and Douglas fir as major species. This vegetation type is restricted to coastal regions and river valleys, extending a few kilometers inland only over coastal plains, and to elevations above 150 m only in areas immediately adjacent to the ocean (Franklin and Dyrness 1973). Sitka spruce forests are replaced by western hemlock-dominated forests along the Strait of Juan de Fuca to the north and east. This vegetation type includes western hemlock, Douglas fir, red alder, and western red cedar as major species. The transition point between Sitka spruce and western hemlock along the Strait of Juan de Fuca appears to occur at about the Elwha River on the Olympic Peninsula and Sooke Inlet on Vancouver Island (Figs. 1 and 5). Because of Puget Sound's lower precipitation and glacial soils, drought-tolerant pines, such as western white, lodgepole, and occasionally ponderosa, are more common here than elsewhere in the western hemlock zone.
Only a few zoogeographic groupings of fishes, invertebrates, and amphibians have been identified along the east coast of the North Pacific Ocean within the range considered in this status review. There is a distinct faunal boundary for marine fishes off the northern tip of Vancouver Island (approximately 50oN lat.) (Allen and Smith 1988), but there is no apparent pattern of variation in marine fishes associated with chum salmon in southern British Columbia or in Washington and Oregon. There is only one distinct group of estuarine fishes in the Pacific Northwest, the Fjord group from Puget Sound and Hood Canal (Monaco et al. 1992). Other estuary groupings are less evident and seem to depend more on characteristics of individual estuaries rather than on geographic location. Regional differences are characterized not by unique variation within species (distinct subspecies or populations), but by presence or absence of species. Within the range of chum salmon in the Pacific Northwest, only one major freshwater ichthyogeographic region, the Columbia, has been described (McPhail and Lindsey 1986, Minkley et al. 1986).
The distributions of marine invertebrates show transitions between major faunal communities similar to those of marine fishes (Hall 1964, Valentine 1966, Hayden and Dolan 1976, Brusca and Wallerstein 1979). The primary cause of this zonation is attributed to temperature (Hayden and Dolan 1976), but other abiotic (Valentine 1966) and biotic (Brusca and Wallerstein 1979) factors may also influence invertebrate distribution patterns.
The distributions of many amphibian species appear to begin and end at several common geographical areas within the range of chum salmon in the Pacific Northwest. The Strait of Georgia and Vancouver Island are the northern extents of many amphibian distributions, such as tailed and red-legged frogs, as well as Pacific giant, western long-toed, western red-backed, Oregon, and brown salamanders (Cook 1984). In addition, several amphibians (Olympic torrent and Van Dyke's salamanders) are restricted to the Olympic Peninsula, while other species (Pacific giant and Dunn's salamanders) occur in most areas in western Washington and Oregon, except in the Olympic Peninsula (Leonard et al. 1993).
The U.S. Environmental Protection Agency has developed a system of ecoregions based on patterns of factors such as land use, climate, topography, potential natural vegetation, and soils (Omernik and Gallant 1986, Omernik 1987). Under this system, the range of chum salmon in Washington covers two ecoregions that border on saltwater: the "Coast Range" ecoregion, which extends from the Strait of Juan de Fuca to Monterey Bay and from the ocean to approximately the crest of the coastal mountains; and the "Puget Lowland" ecoregion, which begins in Washington at approximately the Dungeness River near the eastern end of the Strait of Juan de Fuca and extends through Puget Sound to the British Columbia border.
Based on the preceding physiographic, geological, climatic, and biological information, there appears to be evidence for an environmental distinction between the following three areas: the Columbia River and the coast, Hood Canal and the rest of the Olympic Peninsula, and the Olympic Peninsula and the rest of the Pacific Northwest. However, because the gradients in temperature and precipitation within the range of Northwest chum salmon are not sharp, and based on the physical and biological factors examined here, the precise location of an environmental "border" separating these areas is unclear.
Chum salmon have a broader spawning distribution than any other species of Pacific salmon (Bakkala 1970, Salo 1991) (Fig. 4). In Asia, they commonly spawn as far south as the Naktong River in the Republic of Korea (35°N, 129°E) and the Tone River (36°N, 141°E) on Honshu Island, Japan (Walters 1955, Atkinson et al. 1967). In some years, small numbers of chum salmon occur in streams as far south as Nagasaki and Fukuoka Prefectures of Kyushu (about 33°N) (Sano 1966). Historically, the range of the fish extended as far as the Komandorskiye Islands (Smirnov 1975) and the area of present-day northeastern China, although these populations were also reportedly intermittent and sporadic (K. Chew and L. Donaldson8, as cited in Salo 1991). The northern range of chum salmon in Asia extends along the edge of the Arctic Ocean, almost halfway across the Russian Federation to the Lena River (Laptev Sea) (73°N, 125°E) (Soldatov and Lindberg 1930). However, these populations are small and may occur only intermittently. Juvenile outmigration of these northern stocks often occurs beneath the ice of rivers draining the shores of the Arctic Ocean (Sano 1966, 1967).
Historically, the largest and most commercially important run of chum salmon in Asia was located in the Amur River, which flows more than 2,800 km along the border of China and the Russian Federation. The Amur River is also the only site in Asia that has large runs of both early (summer) and late (autumn) returning chum salmon within the same river basin (see "Spawning Migrations"). Other commercially important runs originate in Japan, on Sakhalin Island (the northernmost part of which is opposite the mouth of the Amur River), in the southern Kuril Islands, on the Kamchatka Peninsula north to the Anadyr River, and in continental streams emptying into the Sea of Okhotsk. Sano (1966) divided chum salmon in Asia into five geographic groups: West Kamchatka, East Kamchatka, Okhotsk, Sakhalin-Kurils, and the Amur River.
In North America, chum salmon have been observed as far south as the San Lorenzo River in Monterey, California (about 37°N, 122°W) (Scofield 1916), with spawning populations as far as 322 km upstream in the Sacramento River (37°50'N 122°W) (Hallock and Fry 1967) (Figs. 4 & 5). In recent years, chum salmon have been observed intermittently in northern California (Moyle et al. in press) and southern Oregon (Kostow 1995) (see "Assessment of Extinction Risk" for information on the occurrence of chum salmon in Oregon and California). In North America, the northern range of the species extends to the shores of the Arctic Ocean (Wynne-Edwards 1952), east to the Mackenzie River (69°N, 135°W) (Dymond 1940), and west across the north Pacific Ocean to Attu Island in the Aleutian Archipelago (Holmes 1982).
The geographic range of commercially important runs of North American chum salmon has shrunk in recent years. The northernmost runs that support large commercial fisheries return to Kotzebue Sound (Noatak and Kobuk Rivers) and the Yukon River (Buklis and Barton 1984) (Fig. 4). Before the 1960s, the southernmost commercially important run of chum salmon in North America occurred in Tillamook Bay, Oregon (45° 50' N) (Henry 1953, 1954) (Fig. 5). This was a highly lucrative annual fishery, with chum salmon landings exceeding that of any other salmonid in the Bay. However, Oregon closed this fishery in 1962 because of declining abundances of these fish (see "Assessment of Extinction Risk" for further information). In the 1940s, an important fishery of chum salmon developed in the Columbia River, with commercial landings varying from 1 to 8 million pounds annually.
Since 1959, with a sharp decline in chum salmon abundance, commercial landings were gradually reduced, until presently there are neither recreational nor directed commercial fisheries for chum salmon in the Columbia River (ODFW and WDFW 1995). In the 1990s, the southernmost commercial run of chum salmon in North America was in Willapa Bay and Grays Harbor (Fig. 5). The harvest of chum salmon from these coastal fisheries averaged 96,000 fish from 1988 to 1992 (WDFW 1995).
In both Asia and North America, chum salmon spawn most commonly in the lower reaches of rivers, with redds usually dug in the mainstem or in side channels of rivers from just above tidal influence to nearly 100 km from the sea. In some areas (particularly in Alaska and northern Asia), they typically spawn where upwelled groundwater percolates through the redds (Bakkala 1970, Salo 1991). Some chum salmon even spawn in intertidal zones of streams at low tide, especially in Alaska, where tidal fluctuation is extensive and upwelling of groundwater in intertidal areas may provide preferred spawning sites (Helle9). Bailey (1964) reported that chum salmon eggs in Olsen Creek, Alaska, could survive exposure to tidewater for up to 55% of the time during embryonic development.
Chum salmon are believed to spawn primarily in the lower reaches of rivers because they usually show little persistence in surmounting river blockages and falls. However, in some systems, such as the Skagit River, Washington, chum salmon routinely migrate over long distances upstream (at least 170 km in the Skagit River) (Hendrick10). In two other rivers, the species swims a much greater distance. In the Yukon River, Alaska and the Amur River, the Russian Federation, chum salmon migrate more than 2,500 km inland. Although these distances are impressive, both rivers have low gradients and are without extensive falls or other blockages to migration. In the Columbia River Basin, there are reports that chum salmon may historically have spawned in the Umatilla and Walla Walla Rivers, more than 500 km from the sea (Nehlsen et al. 1991). However, these fish would have had to pass Celilo Falls, a web of rapids and cascades, which presumably was passable by chum salmon only at high water flows.
Spawning migration refers to the movement of adults from seawater into their natal rivers to spawn. Chum salmon may enter natal river systems from June to March, depending on characteristics of the population or geographic location. Groups of fish entering a river system at particular times or seasons are called "runs" (footnotes 1 and 2), and run timing has long been used by the fishing community to distinguish anadromous populations of salmon, steelhead, and sea-run cutthroat trout. Run-timing designations (e.g., summer versus fall or early-fall versus late-fall) are important in this status review, because two of the ESA petitions for chum salmon (PRO-Salmon 1994, Trout Unlimited 1994) used run timing as evidence supporting population distinction.
In ESA status reviews for anadromous Pacific salmonids, run timing, along with information for other biological and ecological characteristics, is evaluated from the perspective of two criteria for defining ESUs--reproductive isolation and contribution to ecological and genetic diversity (Waples 1991). Previous status reviews have shown that simply relying on traditional run-time designations is not sufficient for identifying ESUs; each case must be evaluated individually. For example, NMFS received petitions in 1990 to list three races of chinook salmon (spring-, summer-, and fall-run fish) in the Snake River as threatened or endangered species. The different runs were defined by fixed dates at which adults pass Lower Granite Dam on the lower Snake River. The BRT looked for other evidence to evaluate the biological significance of the nominal run-timing differences, because actual run timing in any given year is variable and may span the cutoff dates. Snake River fall-run chinook salmon spawn later and at a much lower altitude than do the other two forms, and also show substantial differences from spring- and summer-run fish in genetic and juvenile and adult life-history patterns. NMFS concluded on the basis of these differences that Snake River fall-run chinook salmon were in a separate ESU from spring- and summer-run chinook salmon (Waples et al. 1991). Conversely, spring- and summer-run chinook salmon in the Snake River spawn in similar habitats and cannot be reliably separated on the basis of genetic or life-history variation. Therefore, the two forms were considered to be a single ESU (or species under the ESA), i.e., Snake River spring-summer chinook salmon (Matthews and Waples 1991).
Biologists in both Asia and North America have used run-timing differences to separate the species into early- or summer-run stocks and late- or autumn-run stocks11 (Berg 1934, Sano 1966, Bakkala 1970, Salo 1991). Chum salmon return to natal spawning streams on both continents progressively later in southern areas. Chum salmon in northern Asia return only during summer months (June, July, and August); whereas farther south, both summer (July-August) and early autumn-run chum (September-October) salmon occur along the Amur River and streams on Sakhalin Island (Fig. 4). On Honshu Island in Japan in the far south, near the southern limit of their range, only autumn-run (October-November) chum salmon occur (Fig. 4).
The Amur River, which flows along the Russian-Chinese border, is almost 2,900 km in length and is the largest Asian river entering the North Pacific Ocean (Fig. 4). In addition to being the only major river in Asia with both seasonal runs of chum salmon, it has historically supported both the largest runs and runs that migrate farther inland than any other on the continent. Summer-run chum salmon enter the Amur River as early as July and spawn within 100 km of the sea. A second run of fish enters the system later in the fall, and some of these fish are known to migrate 1,000 to 2,500 km inland. The number of autumn-run chum salmon in the Amur River Basin has become so depressed in recent years that the run may be near extinction (Semenchenko and Augerot12).
In Asia, summer-run fish are considered distinct from autumn-run fish. Berg (1934), who first formally described Asian chum salmon, separated Asian chum salmon into seasonal races, and considered the "autumn" chum salmon an "infraspecies," O. k. autumnalis. However, while other Asian investigators have supported Berg's classification (Lovetskaya 1948; Grigo 1953; Birman 1951, 1956; Hirano 1958; Sano 1966), this taxonomy has not found wide support in North America, perhaps because North American chum salmon with different run timings are not as geographically separated as in Asia. Characteristics of autumn-run fish used by Berg (1934) and Sano (1966) to separate the seasonal runs include the following: 1) later entrance into spawning streams, 2) less developed gametes at time of entry into natal streams, 3) later spawning period, 4) larger size (> or = 3.5 kg versus < or = 2.5 kg for summer-run fish), and 5) greater fecundity than summer-run chum salmon.
In North America, return timing of chum salmon also follows a general north-to-south cline of earlier to later returns, similar to that seen in Asia. Only summer-run chum salmon generally enter streams along the northern Bering Sea and Arctic coast; fall-run fish occur in rivers bordering the southern Bering Sea and Gulf of Alaska; and both summer and fall runs commonly occur in southeast Alaska and northern British Columbia (Fig. 4). Fall-run fish occur primarily south of Vancouver Island (Bakkala 1970). Latitudinal changes in run timing occur in both Asia and North America. However, in Asia, geographic separation between fish with fall- and summer-run times appears to be distinct (Bakkala 1970), whereas in North America, many streams have both fall- and summer-run chum salmon populations. Also, in contrast to Asian summer-run chum salmon, many North American summer-run fish mature at older ages, are larger, and have higher fecundities than do fall-run chum salmon in the same region. Overall, the differences between summer and fall runs in North America are not nearly as distinct or well developed as those reported in Asia. This led Salo (1991) to suggest that most summer runs in North America should more accurately be called early fall runs, and that the only North American populations with the same distinguishing characteristics as the Asian summer chum salmon are in the Yukon River.
Alaska run timing--Chum salmon in northern Alaska tend to return earlier than do chum salmon farther south. Although this is similar to the pattern seen in Asia, Alaskan chum salmon in general show greater diversity in run timing than chum salmon across a similar range of latitude in Asia, with many Alaskan basins having runs that return from July through November. For example, tagging studies by Yanagawa13, Bigler (1983, 1985), and Bigler and Burwen (1984) have shown a bimodal return timing of chum salmon in some rivers in the far north in Kotzebue Sound (around 67°N). The spawning run into Kotzebue Sound is primarily composed of fish that spawn in the Noatak and Kobuk Rivers, with the Noatak River supporting about three-quarters of the Sound's total return (Bigler 1983) (Fig. 4). Kobuk River chum salmon enter Kotzebue Sound earlier than Noatak River chum salmon (Yanagawa footnote 12), and chum salmon in Kobuk River spawn in the lower tributaries from late July to late August and in the upper tributaries from late August well into October (Bigler and Burwen 1984). Spawning in the Noatak River showed a single peak in the lower 160 km of the river from mid-August to mid-October.
The Yukon River drainage (river mouth about 63°N) supports populations of chum salmon that not only have one of the most distinctive bimodal summer/fall-run patterns of return for chum salmon in North America, but also make the longest freshwater migration in the world (Buklis and Barton 1984) (Fig. 4). Chum salmon in the Yukon River share many similarities with those of the Amur River of Asia. Both of these large rivers have long (>2,500 km) stretches accessible to anadromous fish, and both have low gradients with no barrier falls (Buklis and Barton 1984, Brannian and Gnath 1988). Both Amur and Yukon River chum salmon consist of summer-run and fall-run fish, but unlike chinook salmon or steelhead trout in the contiguous United States, these run-time designations refer to time of spawning and not to time of river entry. In both rivers, summer-run chum salmon were historically more numerous (and in the Yukon still are more numerous), return in fuller spawning coloration, are less fecund, and are smaller than fall-run chum salmon (Berg 1934, Sano 1966, Buklis and Barton 1984). However, unlike the runs in the Amur River, the two runs in the Yukon River have an overlap in return times (Berg 1934, Buklis and Barton 1984). Yukon River summer-run chum salmon enter the river mouth from early May through mid-July, whereas some Yukon River fall-run chum salmon (e.g., Fishing Branch Creek in the upper river) enter as early as the beginning of June (Buklis and Barton 1984). Chum salmon returning to the Yukon River are considered predominantly fall-run fish that return from mid-July through early September (Brady 1983).
Summer-run chum salmon in both the Yukon and Amur Rivers spawn primarily in tributaries that empty into the lower part of the main stem, whereas fall-run fish spawn farther upstream in spring-fed streams and sloughs (Buklis 1981, Buklis and Barton 1984, Salo 1991). Some late-returning chum salmon in the Yukon River travel almost 3,200 km from the mouth of the river to its headwaters, near Teslin Lake in British Columbia. This freshwater migration is more than 700 km farther than the longest chum salmon migration on the Amur River.
Farther south, in rivers flowing into Bristol Bay (about 58°N), chum salmon return from mid-June through July, with a peak the first week of July (Salo 1991) (Fig. 4). On the Alaska Peninsula, return times are in July and August. In the Kuskokwim River, spawning is in late August and September. In central Alaska (e.g., eastern shore of Prince William Sound), some chum salmon first enter streams in mid-June, but the peak of spawning is primarily in late July and August. Throughout this area, run timing varies greatly between river basins, but is relatively consistent from year to year within basins (Helle14).
In Southeast Alaska, catch statistics from 1984 indicate that nearshore abundance peaked in the first 2 weeks in August, and that the median timing of the escapements occurred from late August to mid-September (Clark and Weller 1986). Spawning chum salmon are difficult to count in the many rivers and streams of this region, because of the large numbers of pink salmon also present in the streams (Helle footnote 14). However, it is clear that the run timings of these chum salmon are diverse, ranging from mid-June through December. As an example, Helle observed that chum salmon first entered some mainland streams in Behm Canal and Portland Canal in southern Southeast Alaska in mid-June, with peak numbers occurring in mid-August. Fish began to enter natal streams on islands in September, and the run did not peak until mid-October (Helle footnote 13). Spawning has also been observed in October near the mouth of the Alsek River in the Yakutat area (Salo 1991). An exceptionally late run of large, fall-run chum salmon return to the Chilkat River near Haines (Cline 1982), and the timing of this run appears to be influenced by upwelling groundwater in sloughs exposed after the glacial waters of the Klehini and Chilkat Rivers recede in the fall (Helle footnote 13). Large numbers of bald eagles congregate in these sloughs in late fall and early winter to feed on the chum salmon carcasses (Helle footnote 13).
British Columbia run timing--Tables 2 and 3 list rivers in British Columbia that support escapements greater than 5,000 chum salmon annually and show run timings (start, peak, end) for each population (Aro and Shepard 1967). Escapements were estimated from catches in inside passages and inlets close to spawning areas. Spawning times were estimated from a variety of stream surveys (e.g., foot, plane, photographs), although Aro and Shepard (1967) cautioned that the dates of runs were approximate because information from different surveys was not always comparable. Since chum salmon spawning areas are spread more evenly over a large number of moderately sized streams in British Columbia, the species is more difficult to survey in spawning areas than are sockeye and pink salmon. Nonetheless, surveys have shown that chum salmon spawn in the lower reaches of more than 880 streams in British Columbia and that this population is fairly well distributed among the streams. Aro and Shepard (1967) calculated that the most productive 58 streams produced only 50% of the estimated total chum salmon escapement, while 75% of the total pink salmon run in British Columbia occurred in the most productive 58 streams, and 75% of the sockeye salmon escapement occurred in only 29 streams.
In general, there is a distinct north-to-south cline of early-to-late return and spawning times in British Columbia, similar to the clines for these variables in Asia and Alaska. Chum salmon pass through inshore fishing areas near the northern mainland of British Columbia in July and August (summer runs), and catches peak in August along the coast of north-central British Columbia, near Bella Bella and Bella Coola. Catches are made primarily in August and September around the Queen Charlotte Islands. Return times of chum salmon into northern rivers around Johnstone Strait and the Strait of Georgia are usually earlier (early October) than they are in southern rivers, such as the Cowichan River (late December). Peak catches north of Vancouver Island occur from mid-July to mid-September, and southern British Columbia peak catches are in October (Beacham 1984).
Aro and Shepard (1967) also found a marked cline for spawning time. They found 64 rivers with "major stocks" of chum salmon north of Vancouver Island (Statistical Area 13), and all but two of these runs appear to spawn in August or September. Farther south (Statistical Area 13 and higher), 47 major runs were listed, but only one spawned as early as August (Toba River, Section 15), and only eight spawned in September. All other runs began spawning in October or November. Most southern runs completed spawning by the end of December, but six apparently continued to spawn into January. In the north, all but six runs completed spawning by the end of October, and these six finished spawning by November.
Canadian biologists do not classify chum salmon runs into "summer" or "fall" categories. However, spawning in southern British Columbia occurs principally from October to January (Tables 2 and 3). Early spawning and summer runs of chum salmon are rare in southern British Columbia, and no July spawners have been reported. Rivers on northern Vancouver Island and those opposite on the mainland with late-summer or August spawning times include the Ahnuhati, Keogh, Kingcome, Quatse, and Waterfall Rivers, and Ahta Valley Creek. Spawning peaks in October in most of these rivers, but continues into November. Farther south, midway down Vancouver Island and on the opposite mainland (Toba River), chum salmon spawning begins in August and continues through December. Spawning peaks in mid-November, not in October as it does to the north. On southern Vancouver Island, chum salmon spawn in the Nanaimo River (average escapement 20,000-50,000 per year) from September to January, with a peak in October. On the outer southern coast of Vancouver Island, chum salmon run from September to November into the Toquart River (20,000-50,000 escapement) and peak in October. Similarly, on the middle outer coast of Vancouver Island, fish in the Zeballo River (5,000-10,000 escapement) spawn from September to November (Aro and Shepard 1967). No data are available to indicate whether there is a break between the early and late portions of the run, but the runs appear to be continuous (Beacham15).
Washington run timing--Most chum salmon in Washington are classified as fall-run fish and generally return to their natal streams from October to November, although distinct summer and winter runs of chum salmon are recognized. WDFW and WWTIT (SASSI: WDF et al 1993) listed return times for 72 runs of chum salmon: 62 are classified as fall runs, 2 as winter runs (both in southern Puget Sound), and 8 as summer runs (4 in southern Puget Sound and 4 in Hood Canal/Strait of Juan de Fuca) (Table 4; and Figs. 7a, 7b, 7c, 7d).
Puget Sound run timing--Only fall (October to late November) runs have been observed in the rivers of northern Puget Sound (Nooksack, Skagit, Stillaguamish, Skykomish, and Snohomish Rivers) (Atkinson et al. 1967; Hiss et al. 1982a; WDF et al. 1993). However, there is a greater variability in river entry and spawning time within these fall runs than runs in British Columbia. Salo (1991) suggested this regional difference was due to the presence of a larger number of moderate-to-large rivers in Washington than in British Columbia, and to the fact that chum salmon spawn farther upstream in Washington rivers. For example, chum salmon were tagged at the mouths of the Stillaguamish and Nooksack Rivers, and spawn timing was calculated from tagged carcasses counted on spawning grounds (Hiss et al. 1982a, 1982b). In the Stillaguamish River, the average time between tagging and spawning-area recovery was 31 days, and some fish spent more than 6 weeks in freshwater before spawning. Also, different parts of the river basins had different spawning times; in general, return times to spawning areas higher in the river basin were progressively later than return times to spawning areas lower in the basin (Tables 5-7) (Hiss et al. 1982a, 1982b).
All three seasonal runs (summer, fall, and winter) occur in Puget Sound (Table 4; Figs. 7a & 7b) (WDF et al. 1993). Summer-run chum salmon spawn from September to mid-October, and WDFW has identified three summer chum salmon runs: Case Inlet, Hammersley Inlet, and Blackjack Creek. A fourth run into Chambers Creek, north of the Nisqually Delta, in summer was listed as extinct (SASSI: WDF et al. 1993). Existing runs exhibit a wide range of spawning times that extend from late August to mid-October (Hammersley Inlet), through mid-September to early November (Case Inlet), to mid-October and mid-November (Blackjack Creek near Port Orchard, Washington). With the exception of Blackjack Creek, these runs have been supplemented with fry from local spawners, which were reared at Johns Creek Hatchery (Hammersley Inlet) and at Coulter Creek Hatchery (Case Inlet).
Chambers Creek, a southern Puget Sound stream that empties into the Tacoma Narrows, also once contained a run of summer fish (WDF et al. 1993). The existence of the run was inferred from weir counts in the 1970s and 1980s at the Garrison Creek Hatchery, which is located on a tributary of Chambers Creek. Hood Canal chum salmon with November return timing were introduced into Chambers Creek in the 1970s. Until the SASSI review, chum salmon returning to Chambers Creek before 10 December were assumed to be a mix of the "early native stock" and introduced fish (Crawford 1997); fish returning after 10 December were considered natural winter-run chum salmon. However, analysis of weir counts at the hatchery revealed an earlier run beginning about the third week in September and peaking about the third week in October. The early run was considered to be extinct by about 1983 when only three fish were observed in the creek. November-returning fish of Hood Canal origin have not been observed in Chambers Creek since 1986, but winter-run chum salmon still return to the creek (see below).
The Chambers Creek winter-run fish enter the river from early December to late January (Table 4) and are reported in SASSI to be geographically and temporally isolated from other Puget Sound runs (WDF et al. 1993). As reported by Crawford (1997:5), "The issue of spawner overlap with winter chum is moot, since summer and 'fall' chum are no longer present."
Winter-run chum salmon (defined in the SASSI report [WDF et al. 1993] as fish with an average peak of spawning after 10 January) also occur in the Nisqually River in southern Puget Sound (Table 4) (WDF et al. 1993). This winter run and the winter run in Chambers Creek may be the latest returning chum salmon in the world. Tagged chum salmon caught by purse seines in the Nisqually River estuary between late November and mid-January from 1974 to 1980 were recaptured from mid-December to late February/early March at weirs on two spawning ground tributaries, Muck and Yelm Creeks (Cole et al. 1986). Small numbers of fall-timed chum salmon are caught in a tribal fishery in the Nisqually River basin, but WDFW does not believe these fish represent a self-sustainable fall run (Turner 1995). Stream surveys or other data for these fall fish are not available. These fall fish may be strays or they may be from hatchery programs in the Nisqually River Basin. These hatchery programs have primarily reared winter-run fish, but summer-run fish from Johns Creek and fall-run fish from Kennedy Creek, Hood Canal, and Bonneville Hatchery Complex on the Columbia River have also been propagated (see Appendix).
Nine groups of natural fall-run chum salmon in southern Puget Sound were identified in the SASSI report (WDF et al. 1993) (Table 4). One of these, the Eld Inlet fall-run chum salmon, contained fish that return to Allison Springs, which was the focus of the petition submitted by the "Save Allison Springs" Citizens Committee (1994). These southern Puget Sound chum salmon spawn from mid-October to late January, with the long duration of spawning apparently caused by differences among the populations. These fish are reported in the SASSI report (WDF et al. 1993) to be isolated by distance, and to some degree by run timing differences, from other groups of chum salmon in Puget Sound.
Eld Inlet is located between Totten and Henderson Inlets, near the southern end of Puget Sound (Fig. 2) and is fed by freshwater streams near its terminus. The primary tributaries emptying into Eld Inlet where fish spawn are McLane, Swift, and Perry Creeks. Chum salmon returning to Eld Inlet enter southern Puget Sound (WDF et al. 1993) from late October to mid-December, with river entry and spawning from late November to early January. Although this is a broad spawning time compared to other fall chum salmon in Puget Sound, WDFW points out that it is "much later than the Puget Sound summer stocks" (WDF et al. 1993, Appendix 1--Hood Canal and Strait of Juan de Fuca, p. 259), the only other chum salmon spawning in the region. Spawn timing of other fall-run chum salmon in southern Puget Sound are shown in Table 4.
Hood Canal summer chum salmon run timing--Co-managers in Washington have classified chum salmon in Hood Canal as summer- and fall-returning stocks (Figs. 7a, 7b, 7c, 7d). Run-timing and other life-history characteristics for the SE populations have been extensively reviewed by Tynan (1997). Information recorded from 1913 to 1914 (Fig. 8) at the Quilcene National Fish Hatchery on the Big Quilcene River in northern Hood Canal (Fig. 3) indicated a separation of almost a month between the end of the summer chum salmon run and the beginning of the fall run (Cook-Tabor and Zajac16). The SASSI report (WDF et al. 1993) states that, at present, the "summer" spawning of the early run occurs from early-September to mid-October, while the "fall" run spawns from early November to late December (Table 4, Fig. 7c). Graphs of spawning survey data collected by WDFW (Figs. 9a, 9b, & 9c) and catch survey patterns developed by the Point No Point Treaty tribes (Figs. 10 and 11; Lampsakis17) also show nearly a month's separation between peak catches of summer-run chum salmon and natural, fall-run chum salmon in Hood Canal (Fishery Management Area 12).
Summer-, but not fall-run, chum salmon also occur in two bays on the eastern end of the Strait of Juan de Fuca (Fig. 1). These fish have run times similar to those of summer-run fish in Hood Canal (Table 4, Fig. 7c). A run of early-returning chum salmon was recently observed in the Dungeness River, but no life-history, abundance, or other data for the run are available (Johnson18).
Hood Canal fall chum salmon run timing--Co-managers in Hood Canal (WDF et al. 1993) identified 10 naturally spawning Hood Canal fall chum salmon stocks: Northeast Hood Canal, Dewatto, Southeast Hood Canal, Lower Skokomish, Upper Skokomish Late, West Hood Canal, Hamma Hamma Late, Duckabush Late, Dosewallips Late, and Quilcene Late. Fall-run chum salmon in the SASSI report are defined as fish that enter the terminal area in Hood Canal (terminal areas are Fishery Management Areas 12A, B, C, and D) (Fig. 3) between the first week in October and the first week in January. These stocks are separated by the co-managers based on two criteria: geographic separation of spawning grounds and differences in run timing (WDF et al. 1993, Appendix 1--Hood Canal and Strait of Juan de Fuca, p. 43):
Spawning [of fall-run chum salmon] begins about the third week in October and may continue through January. This long duration is a result of timing differences among the nine fall chum stocks in the area which are classified for management purposes as "early fall" and "late fall" stocks.Fall chum spawning in this region is isolated from other Puget Sound chum stocks primarily through geographic separation. Hood Canal summer chum are separated by a clear difference in the spawning seasons from the fall stocks. Hood Canal fall chum were separated into 10 stocks based primarily on geographic separation of the spawning grounds. However, in some cases run-timing differences were also large enough to consider separation of stocks.
However, this does not mean fall chum salmon in Hood Canal were classified in SASSI by run timing "primarily for management purposes." Crawford (1997:6) argues that
Differences in run timing of Hood Canal fall chum stocks was considered but in all cases there was significant timing overlap, and none of the fall timed stocks would have been identified as a distinct stock based solely on timing differences. The timing difference criterion was only used in support of spawning distribution, and was always characterized in the individual stock reports as "to some degree by differences in run timing." The reference to SASSI characterizing stocks based on "management differences in run timing" is not correct. Fishery management elements were not a part of the SASSI stock identification or stock rating process. The confusion here probably arises from the fact that WDFW staff feel that because there is so much overlap in timing among the Hood Canal fall chum stocks, the distinction between fall timing and "late" fall timing is only useful in a management sense; to fine tune management dates for fisheries.
Hatchery production of fall chum salmon in Hood Canal is large compared to other locations in Washington or British Columbia. This hatchery production also had an effect on the identification of some fall chum salmon stocks in SASSI (WDF et al. 1993, Appendix 1--Hood Canal and Strait of Juan de Fuca, p. 42):
The West Hood Canal fall chum stock is comprised of chum spawning in small, independent tributaries located on the west side of the canal. These fish were identified as a single stock for two reasons: 1) The proximity of many of these streams allows for the possibility of commingling of the spawners on the spawning grounds, and 2) large hatchery plants and/or major egg box projects using Hood Canal (Hoodsport) Hatchery stock are ongoing in these streams and have established runs with a similar genetic makeup. The Hood Canal Hatchery stock originated from Finch Creek, one of the west-side independent tributaries. Accordingly, these fish were identified as a mixture of native and non-native fish.
Less run-timing information is available on early fall-run chum salmon in Hood Canal (Table 4, Fig. 7c) (WDF et al. 1993) than for most other chum salmon populations. The SASSI report provided "terminal run," "river entry," and "spawning" timings when available for each stock listed in the survey; however, for early fall-run fish in Hood Canal, this information was reported in SASSI for only one of the five early-returning groups reported. For late fall-run fish, all groups have timing reported for all three factors. The early-run group with the most complete record of run timing is the West Hood Canal fall-run chum salmon. These fish are reported to enter the terminal areas between mid-October and late November, begin up-river migrations between early November and mid-December, and spawn between mid-November and mid-December. Only spawn timing is reported for all other early fall-run fish, which is thought to occur from mid-October to about mid-December (WDF et al. 1993). Based upon this reported time of spawning, there is often no clear distinction between "spawn time" in early- and late-run fall chum salmon in Hood Canal. All late fall-run chum salmon begin to enter terminal areas in mid-October and spawn from mid-December to late January (WDF et al. 1993). For example, "Dosewallips Late Fall Chum" stock (WDF et al. 1993, Appendix 1--Hood Canal and Strait of Juan de Fuca, p. 80) reach the entrance to Hood Canal (Terminal Area 12) from mid-October to mid-December, but do not enter the river to spawn until mid-November to mid-January. An early fall run of chum salmon ("Northeast Hood Canal Fall Chum") (WDF et al. 1993), which spawns in a broad swath approximately opposite the mouth of the Dosewallips River, is reported to spawn between mid-November and mid-December (WDF et al. 1993). Clearly, the spawning times of these early and late fall-run fish can overlap.
Washington and Oregon coast run timing--Chum salmon appear to return to tributaries on the Washington and Oregon coasts in October and November and to spawn as late as December, but data are limited (WDF et al. 1993, Kostow 1995). Long-term run times for coastal Oregon chum salmon are available only for Tillamook Bay, which fish enter in October and November. However, spawning may continue into late December, as it does on the Washington coast (Henry 1954, Cooney and Jacobs 1994). Smaller runs of chum salmon are reportedly found farther south in Netarts Bay, Nestucca River, Yaquina River, Siuslaw River, and Coos Bay, although run times for these fish are not available.
Columbia River run timing--Chum salmon are limited to tributaries below Bonneville Dam, with the majority of fish spawning on the Washington side of the Columbia River. Chum salmon have been reported in October in the Washougal, Lewis, Kalama, and Cowlitz Rivers in Washington and to the Sandy River in Oregon (Salo 1991). Only three Washington runs (Grays River, Hamilton Creek, and Hardy Creek) were listed in the SASSI report, and all return in about October (the peak is mid-November), a run time similar to that of chum salmon in rivers along the Washington coast (Table 4) (WDF et al. 1993). Grays River chum salmon enter the Columbia River from mid-October to mid-November, but apparently do not reach the Grays River until late October to early December. These fish spawn from early November to late December. Fish returning to Hamilton and Hardy Creeks begin to appear in the Columbia River earlier than Grays River fish (late September to late October) and have a more protracted spawn timing (mid-November to mid-January). The Oregon Department of Fish and Wildlife (ODFW) cited 25 locations in that state where chum salmon spawn in the lower Columbia River, but run times for these fish are unavailable (Kostow 1995).