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

NOAA-NMFS-NWFSC TM-33: Sockeye Salmon Status Review (cont)
Back to Previous Section

Information Specific to Sockeye Salmon Populations Under Review

Oregon

The only river systems in Oregon, besides the mainstem Columbia, where anadromous O. nerka are consistently seen each year are the Deschutes and Willamette Rivers.

Deschutes River, Oregon

Small numbers of sockeye salmon are consistently seen each year and trapped at the base of the re-regulating dam below Pelton Dam (which forms Lake Simtustus) on the Deschutes River. These fish are subsequently released at the same location and it is unknown whether they spawn below the Pelton/Round Butte Dam complex (ODFW 1995a, Kostow 1996b). Historically, sockeye salmon occurred in the Deschutes River sub-basin, migrating up the Columbia River to the Deschutes River and then up the Metolius River to Suttle Lake.

Fulton (1970) reported that a 1.2-m-tall power dam and upright screen were installed at the outlet of Suttle Lake in 1930 (Mullan 1986). Nielson (1950) reported that "Blueback salmon formerly ascended to Suttle Lake, but none have been seen for a number of years." Nielson (1950) also reported that a fish passage survey of the Deschutes River in 1942 revealed that

There is a concrete power dam, 4 feet high, at the outlet of the stream from Suttle Lake. This dam . . . may have been responsible for the disappearance of the blueback salmon run. The spillway has a 15 inch flashboard at the upper end of a sloping concrete apron 11 feet long that would be impassable except under very favorable circumstances. The 3-step fishway is too small for large fish and is blocked at the upper end by a stationary screen. Two rotary screens prevent the escapement of fish from the lake to the creek. The diversion to the small power plant is screened.

Nehlsen (1995) also reported on this dam at Suttle Lake, and added that a swimming pool dam (built between 1925 and 1938) and power dam (built between 1925 and 1942) were installed at Lake Creek Lodge on Lake Creek, the outlet stream of Suttle Lake, and both likely hindered or blocked upstream and downstream fish passage.

Several subsequent authors (CBFWA 1990, Olsen et al. 1994, ODFW 1995a) indicated that sockeye salmon continued to return to the Metolius River and spawned below Suttle Lake after fish passage to Suttle Lake was blocked. ODFW (1995a) suggested that sockeye salmon persisted in the Metolius River after construction of the Suttle Lake barrier, until construction of Pelton Re-regulating Dam and Pelton Dam in 1958 and Round Butte Dam (which formed Lake Billy Chinook) in 1964. Sockeye salmon may have persisted by continued spawning in the Metolius River, with juvenile rearing occurring in the Deschutes River or Columbia River, or by return of outmigrants of residual sockeye salmon or kokanee that had escaped over the Suttle Lake barrier. However, Gunsolus and Eicher (1962) stated that, "The spawning of blueback salmon is confined to the Suttle Lake area of the Metolius River and the run is composed, for all practical purposes, of hatchery fish which the Oregon Fish Commission has planted in an attempt to generate a run."

Presently, two kokanee populations occur above the dams: one population resides in Suttle Lake and spawns in the lake inlet stream (Link Creek), and a second population resides in Lake Billy Chinook and spawns in the upper Metolius River (ODFW 1995a). Both kokanee populations have a distinctive blue-black body coloration and are distinguished from hatchery kokanee reared in Lake Simtustus and Deschutes River basin hatcheries by their color pattern (ODFW 1995a). The Lake Billy Chinook/Metolius kokanee reportedly spawn about the same time that Deschutes River sockeye salmon arrive at the Pelton Dam hatchery trap, whereas the Suttle Lake/Link Creek kokanee spawn 2 to 3 weeks later (ODFW 1995a). Sockeye salmon enter the Deschutes River from July to September.

ODFW (1995a) and Kostow (1996b) suggested that sockeye salmon that are consistently trapped in the Deschutes River may derive from 1) a self-sustaining sockeye salmon population that spawns below the Pelton/Round Butte Dam complex and rears in mainstem Columbia River reservoirs, 2) strays from elsewhere on the Columbia River, or outmigrating smolts of "kokanee-sized" fish that escape over the Pelton/Round Butte Dam complex and return as sockeye salmon.

Artificial propagation data (see Appendix Table D-2) indicate that over 740,000 sockeye salmon fry and fingerlings from the Leavenworth National Fish Hatchery and 15,000 smolt from the Bonneville Hatchery were released into Suttle Lake between 1937 and 1958. Additionally over 478,000 sockeye salmon fry, fingerlings, and smolts of mixed Metolius, Leavenworth, and unknown parentage were released in the Metolius River or its tributaries between 1948 and 1961 (see Appendix Table D-2). Many of the reported returns of sockeye salmon to the Deschutes River prior to the 1960s may have been derived from these juvenile sockeye salmon releases.

Willamette and Santiam Rivers

Foy et al. (1995a) and Chapman et al. (1995) reported that small numbers of adult sockeye salmon currently return to the Willamette, Middle Santiam, and South Santiam Rivers. Juvenile sockeye salmon were introduced into several reservoirs in the upper reaches of the Willamette and Santiam Rivers in the 1950s (see Appendix Table D-2), and presumably the downstream migration of some individuals derived from these transplants led to returns of anadromous sockeye salmon (Foy et al. 1995a, Chapman et al. 1995, p. 21).

Washington

The following nine spawning populations of sockeye salmon have been identified in Washington by WDF et al. (1993): 1) Baker River, 2) Ozette Lake, 3) Lake Pleasant, 4) Quinault Lake, and 5) Okanogan River, classified as native stocks; 6) Cedar River classified as a non-native stock; 7) Lake Wenatchee classified as having mixed stock origin; and 8) Lake Washington/Lake Sammamish tributaries, and 9) Lake Washington beach spawners, classified as having unknown stock origin. Chapman et al. (1995) listed four additional spawning aggregations of sockeye salmon that appear consistently in Columbia River tributaries: the Methow, Entiat, and Similkameen Rivers and Icicle Creek in the Wenatchee River drainage.

Sockeye salmon have been periodically observed in other Washington rivers that lack accessible lake habitat, including the Nooksack, Samish, mainstem Skagit, Sauk, Stillaguamish, Green, Skokomish, Dungeness, Calawah, Hoh, Queets and North Fork Lewis Rivers. Reportedly, several sockeye salmon are observed yearly during spawner surveys in almost every river in Puget Sound; this phenomenon is more common, and numbers of sockeye salmon are higher, in north Puget Sound rivers than in south Puget Sound rivers (J. Ames17).

Okanogan River

Okanogan River sockeye salmon rear in Lake Osoyoos, which is composed of three connected basins: north, middle, and south. WDF et al. (1993) reported that Okanogan River sockeye salmon bound for Lake Osoyoos begin migrating up the Columbia River in mid- to late-June and peak in early July. In contrast, Chapman et al. (1995) reported that sockeye salmon bound for the mid-Columbia River begin entering the Columbia River in April and May, peaking at Bonneville Dam in the third week of June and at Rock Island Dam in the third week of July. Chapman et al. (1995) compared sockeye salmon run-timing data from dam counts in 1933-1947 to similar counts in 1988-1992 and found that current run-timing is about a week earlier than it used to be. This change in timing was believed to be due to reduced water velocities in mainstem reservoir reaches of the Columbia River, with later velocities allowing for more rapid upriver fish migration (Chapman et al. 1995). Quinn and Adams (1996) also reported that sockeye salmon upriver migration timing is about 6 days earlier now than it was in 1949. Based on scale pattern analysis, Fryer and Schwartzberg (1994) suggested that Okanogan River sockeye salmon migrate past Bonneville Dam later than the population bound for Lake Wenatchee. Major and Mighell (1966) reported that most adult sockeye salmon begin migrating up the Okanogan River in mid- to late-July and enter Lake Osoyoos in August, although in some years sockeye salmon may reach Lake Osoyoos as early as mid- to late-July. WDFW (1996) stated that Okanogan sockeye begin migrating slightly later than the Wenatchee stock, based on scale analysis at Bonneville Dam which shows Wenatchee fish dominating the early portion and shifting to Okanogan stock later. Okanogan sockeye probably begin their entry in early to mid-June and peak at Bonneville Dam in early July.

Migration may be impeded by as much as 3 weeks in some years by high water temperatures during mid-summer in the Okanogan River (Major and Mighell 1966, Allen and Meekin 1980, Mullan 1986, Swan et al. 1994, Chapman et al. 1995). Sockeye salmon congregate at the confluence of the Okanogan and Columbia Rivers when water temperatures exceed 21.1oC and only migrate up the Okanogan River when temperatures fall below this level (Major and Mighell 1966, Allen and Meekin 1980, Chapman et al. 1995).

Swan et al. (1994) reported that upon reaching Lake Osoyoos, sockeye salmon stay in the lake from less than 1 day to 46 days, with a median of 28 days, prior to moving upstream to the spawning grounds. WDF et al. (1993) indicated that this population spawns upstream from Lake Osoyoos in the Okanogan River but below the Southern Okanagan Lands Project Dam (=Oliver Diversion Dam = Vaseux Dam) during late September through October. According to Chapman et al. (1995), spawning occurs primarily from about 1 to 23 October, with a peak about the third week in October.

Burner (1951) observed a few sockeye salmon redds on the shoreline of Lake Osoyoos. Allen and Meekin (1980) observed about 1,200 sockeye salmon spawning on the shoreline of Lake Osoyoos in October of 1971, whereas only a "few" to none were observed in 1972-1974.

Lake Osoyoos has been variously characterized as eutrophic (Mullan 1986) and as displaying the range of conditions known as mesotrophic (see Appendix Table B-2) (Rensel 1995, cited in Chapman et al. 1995). Lake Osoyoos is atypical of sockeye salmon rearing lakes, which are typically oligo- or ultra-oligotrophic (Mullan 1986, Chapman et al. 1995). From data provided in Mullan (1986), the morphoedaphic indices for the northern, middle, and southern basins of Lake Osoyoos were estimated as 10.91, 4.69, and 14.74 respectively; these values are at the high end of the scale for sockeye salmon nursery lakes and indicate the potential for high primary production (see Appendix Table B-2). Lake Osoyoos has been ranked as one of the most productive of all sockeye salmon rearing lakes, based on phytoplankton and zooplankton abundance (Foerster 1968, Allen and Meekin 1980, Chapman et al. 1995). A strong thermocline develops in Lake Osoyoos during the summer, when surface temperatures can reach 25oC and the hypolimnion becomes anoxic, leaving only a narrow 1- to 2-m sub-surface layer of water in the south basin with conditions suitable to sockeye salmon survival. These conditions indicate that sockeye salmon juveniles may be limited to the north and middle basins of Lake Osoyoos during summer months (Rensel 1995, cited in Chapman et al. 1995).

High plankton productivity has led to the production in Lake Osoyoos of "some of the largest sockeye salmon smolts reported in the literature" (Mullan 1986). The average length of known age-1+ sockeye salmon smolts from Lake Osoyoos has ranged over a number of years from 94 to 114 mm with a median of about 110 mm (Allen and Meekin 1980, Chapman et al. 1995), a length exceeded only by sockeye salmon smolts from Lake Washington, Baker Lake, and Ozette Lake (see Appendix Table C-4). Age composition data presented in Allen and Meekin (1980) and Chapman et al. (1995) show that in some years an unusually large percentage of adult spawners in the Okanogan River sockeye salmon population are 3-year-old fish, whereas 3-year-olds are extremely rare in the Lake Wenatchee population (see Appendix Tables C-1 and C-2). Okanogan River sockeye salmon are thought to have the youngest average age at maturity for sockeye salmon throughout their range (Chapman et al. 1995).

Fry emergence and migration downstream to Lake Osoyoos has been reported to occur mostly at night, beginning in early March (prior to the Lake Wenatchee migration), peaking in mid-April, and concluding by the third week in May (Allen and Meekin 1980, Shepherd and Inkster 1995 as cited in Chapman et al. 1995). Data presented in Chapman et al. (1995) indicate that currently sockeye salmon smolts leave Lake Osoyoos in mid- to late May and migrate past Rock Island Dam in May (Peven 1987). In contrast, Wenatchee-origin sockeye salmon smolts typically arrive at Rock Island Dam in April (Peven 1987). Chapman et al. (1995) pointed out that currently, sockeye salmon smolts appear to arrive at downstream dams on the Columbia River earlier than they did in the 1940s through 1960s, although the reasons for this earlier run-timing are not clear.

Between 1939 and 1943, all adult sockeye salmon returning to the Columbia River above the confluence with the Snake River were trapped at Rock Island Dam on the Columbia River as part of the Grand Coulee Fish Maintenance Project. A total of 19,795 of these trapped adult sockeye salmon of mixed Okanogan River, Lake Wenatchee, and Upper Columbia River heritage were transported to and released in Lake Osoyoos. Appendix Table D-2 shows that between 1940 and 1968, about 395,000 fry resulting from a mixed-stock spawning of Rock Island Dam and Quinault Lake stock, and over 4.2 million fish descended from original spawners collected at Rock Island and Bonneville Dams, were released into Lake Osoyoos (Mullan 1986). In the brood years 1992 and 1993, 73,000 and 110,500 pen-reared juvenile sockeye salmon (adults captured at Wells Dam) were released in Lake Osoyoos (Chapman et al. 1995). No adult returns from the releases in 1992 and 1993 have been noted (Chapman et al. 1995).

Sockeye salmon and kokanee-sized O. nerka are reported to spawn at the same time and place in the Okanogan River, often with overlapping redds, although it is unknown whether peak spawn timing of these two groups of fish are the same. Kokanee-sized fish reportedly acquire a drab-olive spawning coloration, whereas sockeye salmon in this population have the typical spawning color pattern (L. LaVoy18). Dark colored residual O. nerka presumably occur to some degree in all years on the sockeye salmon spawning grounds of the Okanogan River (Chapman et al. 1995, p. 21).

Kokanee stocking history in Lake Osoyoos includes the release of 195,550 kokanee fry from an unnamed source into Lake Osoyoos between 1919 and 1920 (WDFG 1921a) (see Appendix Table D-5). Further kokanee stocking information was not obtained from either U.S. or Canadian sources.

Lake Wenatchee

WDF et al. (1993) reported that sockeye salmon bound for Lake Wenatchee begin migrating up the Columbia River in mid-June, peaking in early July, and enter Lake Wenatchee in late July to early August. It was stated in WDFW (1996) that "Wenatchee sockeye enter the Columbia in May and peak at Bonneville Dam in late June or early July." Based on scale pattern analysis, Lake Wenatchee sockeye salmon appear to migrate past Bonneville Dam earlier than the population bound for the Okanogan River (Fryer and Schwartzberg 1994). As mentioned above for Okanogan River, Chapman et al. (1995) reported that sockeye salmon bound for the mid-Columbia River begin entering the Columbia River in April and May, peaking at Bonneville Dam in the third week of June and at Rock Island Dam in the third week of July. Chapman et al. (1995) reported that comparison of run-timing data from dam counts in 1933-1947 and 1988-1992 indicate that current sockeye salmon run-timing is about a week earlier than it used to be. Quinn and Adams (1996) also reported that sockeye salmon upriver migration timing is about 6 days earlier now than it was in 1949. Run-timing of sockeye salmon in the Wenatchee River, as measured at Tumwater Dam, appears to be as much as a month earlier at the present time than it was in the 1930s (Chapman et al. 1995). Factors contributing to this run-timing change may include improvements to fish ladders at Tumwater and Dryden Dams on the Wenatchee River, lower river flows in recent years, and faster within-reservoir migration in the Columbia River since modern dam construction (Allen and Meekin 1980, Mullan 1986, Chapman et al. 1995).

The Wenatchee population spawns from mid-September through October in the Little Wenatchee, White, and Napeequa Rivers above Lake Wenatchee (WDF et al. 1993). According to Chapman et al. (1995), main spawning activity currently occurs from mid-September to about the beginning of October, with a peak in the third week of September. Gangmark and Fulton (1952) reported two lakeshore seepage areas in Lake Wenatchee that were used by spawning sockeye salmon. Mullan (1986) indicated that only limited shore spawning occurs in Lake Wenatchee. Although no active surveys targeting beach-spawning sockeye salmon have been undertaken, shoreline spawning has not been observed in recent years in Lake Wenatchee (L. LaVoy19).

Lake Wenatchee has been characterized as a typical oligotrophic or ultra-oligotrophic sockeye salmon nursery lake: clear, cold, well-oxygenated, and with low productivity (Allen and Meekin 1980, Mullan 1986, Chapman et al. 1995). Lake Wenatchee has an estimated metric morphoedaphic index of 0.51, which is within the range of MEI typical for sockeye salmon nursery lakes (see Appendix Table B-2) and is considerably lower than the MEI for Lake Osoyoos (Mullan 1986). Water residence time in Lake Wenatchee was estimated at the relatively rapid rate of 2.2 exchanges per year (Mullan 1986). A strong thermocline does not apparently develop in Lake Wenatchee in the summer, and dissolved oxygen and temperature conditions allow sockeye salmon to use all depths of the lake (Chapman et al. 1995, p. 83). Thompson and Tufts (1967) identified Dolly Varden and northern squawfish as predators of sockeye salmon juveniles in Lake Wenatchee, although only 12% of Dolly Varden and 1% of northern squawfish collected had consumed wild sockeye salmon fingerlings.

The average size of known age-1+ sockeye salmon smolts from Lake Wenatchee have ranged from 65 to 124 mm fork length, with a median of about 88 mm (Allen and Meekin 1980, Chapman et al. 1995) (see Appendix Table C-4). Peven (1987) indicated that sockeye salmon smolts from Lake Wenatchee are generally smaller than 100 mm, whereas Okanogan River smolts are generally larger than 100 mm. Age composition data show (see Appendix Tables C-1 and C-2) that although an unusually large percentage of adult spawners in the Okanogan River sockeye salmon population are 3-year-old fish, very few Lake Wenatchee sockeye salmon exhibit this age pattern. Chapman et al. (1995) pointed out that sockeye salmon from Wenatchee show a stronger tendency to spend 2 years in freshwater prior to smoltification than do members of the Okanogan River population.

Dawson et al. (1973) found that sockeye salmon fry were entering Lake Wenatchee between March and May, while Chapman et al. (1995) deduced from data in Gangmark and Fulton (1952) that fry emerge from redds in the Wenatchee River by mid-March.

Peven (1987) showed that Wenatchee-origin sockeye salmon smolts typically arrive at Rock Island Dam in April. As mentioned above for Okanogan River sockeye salmon, Chapman et al. (1995) pointed out that currently, sockeye salmon smolts appear to arrive at downstream dams on the Columbia River earlier than they did from the 1940s through 1960s, although the reasons for this earlier run-timing are not clear.

Between 1939 and 1943, all sockeye salmon entering the mid-Columbia River were trapped at Rock Island Dam, and over 32,000 mixed Lake Wenatchee, Okanogan River, and Arrow Lakes adult sockeye salmon were released into Lake Wenatchee as part of the Grand Coulee Fish Maintenance Project. Between 1940 and 1968, over 2.4 million fry derived from original Quinault Lake stock, and over 52.8 million fry descended from original spawners collected at Rock Island and Bonneville Dams, were released into Lake Wenatchee (see Appendix Table D-2). Starting with the 1989 brood year, between 167,500 and 372,100 pen-reared Lake Wenatchee-origin juvenile sockeye salmon have been released yearly into Lake Wenatchee. From the 1990 release, an estimated 4,133 sockeye salmon returned in 1994, for a fry-to-adult survival rate of 1.6% (survival estimate based on scale pattern analysis) (Chapman et al. 1995).

Kokanee are reportedly native to Lake Wenatchee (Crawford 1979). Sockeye salmon and kokanee have been seen to spawn at the same time and place in tributaries of Lake Wenatchee (the forms may have overlapping redds in the White, Napeequa, and lower end of the Little Wenatchee Rivers), and the kokanee reportedly acquire a drab olive spawning coloration, whereas Wenatchee sockeye salmon have the typical spawning color pattern (L. LaVoy20). Residual O. nerka reportedly occur on the spawning grounds with Lake Wenatchee sockeye salmon (Chapman et al. 1995). Between 1934 and 1966, 22.5 million Lake Whatcom kokanee were released in Lake Wenatchee (Mullan 1986) and approximately 0.5 million kokanee of the same broodstock origin were released in 1983 (Knutzen 1995) (see Appendix Table D-5).

Quinault Lake

This sockeye salmon population is the most southerly coastal population of this species in North America. WDF et al. (1993) indicated that sockeye salmon, or blueback salmon as they are known locally, begin entering the lower Quinault River in small numbers in January and continue to the end of July, peaking in late May to early July. Sockeye salmon have been known to enter the Quinault River as early as December and as late as August (QIN 1981). The duration of this run is unusually long for sockeye salmon, lasting over 7-9 months (Burgner 1991, p. 9). Johnson (1977) stated that it takes sockeye salmon approximately 3 days to migrate between the mouth of the Quinault River and Quinault Lake. Sockeye salmon adults may remain in Quinault Lake for 3-10 months without feeding (QIN 1981) prior to moving upstream to spawn from November through February, primarily in the upper Quinault River and its tributaries (WDF et al. 1993). Sockeye salmon spawn timing for the Quinault stock is unusually protracted; observed duration has been 7 months, from August through March, although peak spawning occurs from November through January (QIN 1981).

The majority of sockeye salmon in the Quinault system take on a drab gray-green (D. Boyer, Jr.21) or olive (Storm et al. 1990) spawning coloration, in contrast to the typical red body coloration of sockeye salmon, but are very red-fleshed with high oil content when they enter the river (D. Boyer Jr.22). Storm et al. (1990) stated that a small segment of early spawners take on the more typical coloration of spawning sockeye salmon, and these may represent a unique strain. Due to carotenoid metabolism, spawning sockeye salmon may contain up to 65% less carotenoid than pre-spawning sockeye salmon taken at sea (Crozier 1969). The loss of tissue carotenoids in Quinault Lake sockeye salmon may result from the prolonged non-feeding adult lake-residence period prior to maturation and spawning. Although a green coloration at spawning is not common for sockeye salmon, spawning individuals of two sockeye salmon stocks in British Columbia (Weaver Creek (lower Fraser River) and Alastair Lake (Skeena River)) also reportedly appear more green than red (C. C. Wood23).

Smolt outmigration occurs in May and June (Davidson and Barnaby 1936) or April and May (Tyler and Wright 1974), and takes place during the hours of darkness (Tyler and Wright 1974) (see Appendix Table C-5). The percentage in each age group, and length and weight of sockeye salmon captured in the fishery for various years, are presented in Appendix Tables C-1 and C-8. Figure 4 illustrates temporal changes in freshwater and saltwater age composition of Quinault Lake sockeye salmon by return year.

The Quinault sockeye salmon or blueback has always been culturally and economically important to the Quinault Indians, and its flavor has often been remarked upon.

Lestelle and Workman (1990) stated that

Culturally, this salmon run links Quinault people to their rich heritage as nothing else does. The salmon was always the very lifeblood of Quinault society, and the blueback was the most sacred of the various fish runs.

Brown (1982, p. 32) related that

The Chinook tribe . . . esteemed the Quinault sockeye so highly that they used it as an all-purpose term of excellence. Whites . . . picked this up and mistakenly applied the name to the most prized of the Columbia's runs, the salmon known as the Chinook. For half a century Chinook salmon were known as "Quinnat" . . .

Numerous early references to the superior quality of Quinault Lake sockeye salmon exist (Willoughby 1889, Curtright 1979), and the unusual quality of the flesh has often been attributed to the stored energy reserves necessary to maintain these fish through the long lake residence period prior to maturation (QIN 1981, Lestelle and Workman 1990). This population of sockeye salmon has long supported a commercial set-net fishery operated by the Quinault Indian Nation near Taholah on the lower Quinault River.

The U.S. Bureau of Fisheries operated a fish hatchery from 1914 to 1947 at Falls Creek on Quinault Lake (this hatchery was referred to as the "Quinault, Washington Station," and should not be confused with the present-day Quinault National Fish Hatchery). This hatchery program utilized native broodstock for the most part; however, out-of-basin transplant history includes the transfer from Alaska of about 20 million sockeye salmon eggs from 1916 to 1921 and 260,000 kokanee eggs from Lake Whatcom in 1925 to the Quinault, Washington Station on Falls Creek (see Appendix Tables D-1 and D-5). Kokanee do not currently inhabit Quinault Lake, although over 300,000 kokanee fry from unnamed sources were released in Quinault Lake between 1917 and 1922 (WDFG 1919, 1921a, 1923) (see Appendix Table D-5).

A portion of the lower watershed above Quinault Lake was logged in association with early homesteading and Sitka spruce harvest for war plane construction during the First World War. Around this period of time, severe erosion of the banks of the upper Quinault River occurred, although it is unknown whether logging caused this accelerated erosion, or whether the erosion was part of a natural process (D. Boyer, Jr24). Today, much of the upper Quinault River below the North Fork, where most sockeye salmon spawning occurs, is a braided stream subject to severe meander (Davidson and Barnaby 1936, Brown 1982, WDF et al. 1993). Davidson and Barnaby (1936) reported that

The early settlers and inhabitants of this region describe the upper Quinault River as a large stream that flowed between two rather narrow heavily wooded banks. . . . the logging off of the watersheds of the river has caused excessive washing to the extent that there is no definite river bed but a wide river valley through which the stream frequently changes its course with the winter and spring freshets.

Severe storm runoff problems in the upper Quinault River in the fall of 1990 and winter of 1990-1991 led to a prolonged period of lake turbidity (S. A. Chitwood and D. Boyer, Jr.25).

Ozette Lake

Migration of adult sockeye salmon up the Ozette River and into Ozette Lake occurs between dusk to dawn from April to early August (WDF et al. 1993) (see Appendix Table C-6) or May to August (Dlugokenski et al. 1981). Kemmerich (1945) counted sockeye salmon past a weir constructed in the Ozette River in 1924, 1925, and 1926 between 27 May and 8 August, 8 June and 15 September, and 28 May and 8 September, respectively. Jacobs et al. (1996) noted that the tribal sockeye salmon fishery in the lower Ozette River that operated between 1948 and 1957, began in mid-April and peaked from 2 to 15 June. Fifty sockeye salmon were seen moving up the Ozette River on 20 October 1989 following a rise in the lake level (LaRiviere 1991).

High water temperatures in Ozette Lake and River and low water flows in the summer may create a thermal block to migration and influence timing of the sockeye salmon migration (LaRiviere 1991). Recorded water temperatures in late-July and August in the Ozette River near the lake outlet have exceeded the temperature range over which sockeye salmon are known to migrate (J. Meyer26).

Currently, spawning is restricted to submerged beaches where upwelling occurs along the lakeshore or to tributary outwash fans (Dlugokenski et al. 1981, WDF et al. 1993). Spawning has been variously reported to occur from mid- to late November through early February (WDF et al. 1993) and from late November to early April (Dlugokenski et al. 1981) (see Appendix Table C-6). Dlugokenski et al. (1981) suggested that discreet sub-populations may be present in the lake, as evidenced by disjunct spawning times between beach spawners in different parts of the lake.

The two principle shoreline spawning beaches for sockeye salmon in Ozette Lake are Olsen's Beach (or Olsen's Landing) (north of Siwash Creek on the lake's eastern shore) and the beach area north of Allen's Bay on the lake's western shore (WDF et al. 1993, Jacobs et al. 1996). Reportedly, some spawning has also been seen recently on the south shore of Baby Island at the southern end of Lake Ozette (Jacobs et al. 1996). Historically, sockeye salmon reportedly spawned in tributary creeks of Ozette Lake, on the shoreline north of Umbrella Point, and in Ericson's Bay (Dlugokenski et al. 1981, WDF et al. 1993, Jacobs et al. 1996). A small degree of outlet spawning may occur in the Ozette River or in Coal Creek, a tributary of Ozette River below Ozette Lake (WDF et al. 1993, Jacobs et al. 1996, E. Currence and D. Dailey27 ). A number of sockeye salmon fry were inadvertently released in Umbrella Creek near the tribal hatchery in 1987, and 13 adult sockeye salmon were noted spawning in this creek 4 years later, in 1991. Over 8,000 sockeye salmon fry of the 1991 brood year were released in Umbrella Creek in 1992 and approximately 30-50 sockeye salmon redds were counted in Umbrella Creek in the fall of 1995 (Jacobs et al. 1996, E. Currence28).

Kemmerich (1945) reported that during his work with sockeye salmon at Ozette Lake in the years 1923-1926, "there was no evidence that they ascended any of the tributaries of the lake to spawn." In reference to Ozette Lake sockeye salmon, Kemmerich (1939) in a letter to R. E. Foerster stated that

We made no special investigations of spawning beds during the years covered . . . but merely observed from time to time that most of the spawning seemed to be along the lake shore in suitable places and especially at the mouths of the several creeks. I do not recall that any sockeyes ascended any of the creeks to spawn but it seems to me that spawning took place during the latter part of September and October.

Abundance of sockeye salmon outmigrant smolts from Ozette Lake was estimated in 1977 at 9,600 (Dlugokenski et al. 1981), in 1990 at 7,942, and in 1992 at 2,752 (Jacobs et al. 1996). Based on these numbers and adult returns 2 years later (see Jacobs et al. 1996, their table 3), ocean survival of broodyears 1975, 1990, and 1991 were 5.6%, 18%, and 27%, respectively (Jacobs et al. 1996).

A total of 13 species of fish occur in Ozette Lake (see Appendix Table B-4). Dlugokenski et al. (1981) and Blum (1984) listed potential competitors with sockeye salmon juveniles in Ozette Lake, including kokanee, red sided shiner (Richardsonius balteatus), northern squawfish (Ptychocheilus oregonensis), yellow perch (Perca flavescens), and peamouth (Mylocheilus caurinus). Potential predators listed by these same authors included cutthroat trout (Salmo clarki), northern squawfish (Ptychocheilus oregonensis), and prickly sculpin (Cottus asper). Beauchamp et al. (1995) showed that competition is unlikely to limit the sockeye salmon population in Ozette Lake; however, predation on juvenile sockeye salmon, which was 25 times greater by individual cutthroat trout than by individual squawfish, may be limiting, although total predator abundance has yet to be assessed.

Harbor seals (Phoca vitulina) migrate up the Ozette River into Ozette Lake and have been seen feeding on adult sockeye salmon off the spawning beaches in Ozette Lake. The numbers of seals and the number of salmon taken by each seal is unknown. Seal predation on sockeye salmon at the river mouth and during the salmon's migration up the Ozette River may also be occurring. The upriver migration of harbor seals to feed on adult sockeye salmon is common in British Columbia, occurring 100 miles upriver on the Fraser River at Harrison Lake and up to 200 miles inland on the Skeena River (Foerster 1968). Sockeye salmon migrate up to Ozette Lake in less than 48 hours and the majority of adults travel at night (Jacobs et al. 1996).

Chamberlain (1907, p. 40) reported that "dwarf sockeye" were present in Ozette Lake around the turn of the century, and it is likely that kokanee were present prehistorically in Ozette Lake. Between 5,000 and 10,000 kokanee spawn in small tributaries to Ozette Lake, and Dlugokenski et al. (1981) and Beauchamp et al. (1995) thought that these numbers of kokanee were insufficient to deplete food resources for sockeye salmon. Dlugokenski et al. (1981, p. 34) reported that kokanee spawn not only in tributaries, but also spawn interspersed with sockeye salmon on the lakeshore in mid-November to early December. Over 108,000 kokanee fry from the Lake Crescent Trout hatchery were planted in Ozette Lake in 1940 (Kloempken 1996, see Appendix Table D-5). An unknown number of kokanee from an unknown stock were reportedly planted in Ozette Lake in 1958 (Dlugokenski et al. 1981).

Kemmerich's (1945) escapement counts for sockeye salmon to Ozette Lake were 3,241 in 1924 (a portion of the run was missed), 6,343 in 1925, and 2,210 in 1926. No information relative to the down-river catch of sockeye salmon in the Makah Tribal fishery was available for this time period. Dlugokenski et al. (1981) reported that smolt outmigration occurs during the hours of darkness and peaks around 6 May, and that Ozette Lake sockeye salmon have the third largest yearling smolt size of any population reported in the literature. Data on smolt size and age are presented in Appendix Table C-4 and smolt outmigration period in Appendix Table C-5.

In 1937, almost 450,000 sockeye salmon fingerlings cultured at the U.S. Bureau of Fisheries Quilcene Hatchery derived from eggs received from the Birdsview Hatchery on Grandy Creek (Skagit River Basin) were released into Ozette Lake (Kemmerich 1945, Boomer 1995) (see Appendix Table D-2). Sockeye salmon of the 1936 brood-year at the Birdsview Hatchery were composed primarily of Baker Lake broodstock and a probable Fraser River and Quinault Lake component (Kemmerich 1945). In 1983, 120,000 Quinault Lake sockeye salmon fry were released into Ozette Lake (MFMD n.d., Hill 1984). Between 1984 and 1995, almost 0.5 million Ozette Lake-origin sockeye salmon fry were reared at the Makah Tribal Hatchery on Umbrella Creek, a tributary of Ozette Lake, and released into the Ozette Lake drainage (MFMD n.d.) (see Appendix Table D-2). Spawning stock for this hatchery effort have been captured on the lakeshore spawning grounds (WDF et al. 1993).

Outside of that portion in Olympic National Park, virtually the entire watershed of Ozette Lake has been logged (Blum 1988). A combination of past overfishing and spawning habitat degradation, due to stream and tributary outwash fan siltation, associated with timber harvest and road building, have been cited as major causes of this stock's decline (Bortleson and Dion 1979, Dlugokenski et al. 1981, Blum 1988, WDF et al. 1993). McHenry et al. (1994) found that percent fine sediments (<0.85 mm) averaged 18.7% in Ozette Lake tributaries (although these levels may be partly attributable to the occurrence of sandstones, siltstones, and mudstones in this basin) and fine sediment levels were consistently higher in logged watersheds than in unlogged watersheds on the Olympic Peninsula, as a whole.

During low water levels in summer, much of the beach habitat may become exposed (Bortleson and Dion 1979). The exotic plant, reed canary grass (Phalaria arundinacea), has been encroaching on sockeye salmon spawning beaches in Ozette Lake, particularly on the shoreline north of Umbrella Creek, where sockeye spawning has not occurred for several years. This plant survives overwinter submergence in up to 3 feet of water and may possibly provide cover for predators of sockeye salmon fry (J. H. Meyer29). Suitable lakeshore spawning habitat for sockeye salmon is reported to be extremely limited in Ozette Lake (Blum 1984, Pauley et al. 1989).

Lake Pleasant

Sockeye salmon that spawn and rear in Lake Pleasant enter the Quillayute River and migrate up the Sol Duc River in May to September. Normally, this stock remains, throughout the summer, in the Sol Duc River at the confluence with Lake Creek (the Lake Pleasant outlet stream) until the creek receives sufficient water input to allow the fish to migrate up to Lake Pleasant. Sufficient stream discharge to allow upstream migration does not usually occur until late October to early November (WDF et al. 1993, J. Haymes30). Sockeye salmon spawn predominantly on lakeshore beaches from late November to early January. Little spawning has been observed in streams tributary to Lake Pleasant (WDF et al. 1993). In describing a survey of potential habitat for sockeye salmon introductions on the Olympic Peninsula undertaken prior to 1932, Kemmerich (1945) stated that

It was found that a small run of sockeye or blueback salmon already enters Lake Pleasant by way of the Sol Duc River and Lake Creek and these natural run fish were found to be in individual size comparable with the size of the fish of the Lake Quinault and Columbia River runs.

Currently, sockeye salmon in the Lake Pleasant stock are said to weigh no more than about 2 to 3 pounds (0.9 to 1.4 kg) (J. Haymes31), which is considerably less than sockeye salmon from Quinault Lake, with the exception of the very few jacks and jills recorded from Quinault Lake (see Appendix Table C-8). Limited data on smolt size are presented in Appendix Table C-4. Average length of spawners collected in 1995 and 1996 for genetic analysis were 451 mm (n=10) and 460 (n=72) for males, and 459 mm (n=5) and 456 (n=28) for females, respectively (see Appendix Table C-8). This stock has the smallest average adult body size of any sockeye salmon stock in Washington.

The following out-of-basin introductions of sockeye salmon into Lake Pleasant occurred in the 1930s: 1) in 1933, 210,000, and in 1937, 75,000, fingerlings derived from the Birdsview Hatchery were released into Lake Pleasant (assuming a 4-year return cycle, the 1932 and 1936 Birdsview Hatchery broodstock were descended from mixed releases of the progeny of Fraser River sockeye salmon in 1908 and 1912; Quinault Lake sockeye salmon in 1916; and Baker Lake sockeye salmon in 1920 and 1928), and 2) in 1934, 175,000 fingerlings from the Birdsview Hatchery were released into Lake Pleasant (the 1933 broodyear was composed of Baker Lake broodstock) (Kemmerich 1945, Boomer 1995) (see Appendix Table D-2).

Fisheries biologists of the Washington Department of Fisheries undertook a survey of Lake Pleasant in July and September of 1952, in part to determine its suitability for sockeye salmon rearing (Smoker et al. 1952, Heg 1953). These two reports were written with the assumption that sockeye salmon were absent from, or very rare in, Lake Pleasant at this time. Smoker et al. (1952) stated that

The suitability of Lake Pleasant for the rearing of sockeye can be better determined by further examination. Its carrying capacity can only be learned by making a plant and watching the results. . . . Local residents speak rather vaguely of occasional "bluebacks" being taken in the lower creek. These could be either sockeye or sea-run cutthroats. . . . Upper Lake Creek would provide good spawning for early-run sockeye. Spawning in the lake itself would be negligible.

Heg (1953) stated that

Local residents report that Lake Pleasant used to support a small run of sockeye salmon. However, in view of the large scrap fish population, the unfavorable temperature conditions, and the occurrence of dry years of extreme low flows in the outlet stream, it does not appear likely that this lake can be developed into an important sockeye producer.

In 1956, Lake Pleasant was treated with rotenone by the Washington Department of Game in an attempt to eliminate all the resident fish in the lake and its tributaries in anticipation of using Lake Pleasant as rearing habitat for winter steelhead. Since the rotenone treatment only occurred in a single year, anadromous fish with multiple broodyear life histories were probably less affected by this program than resident fish. A box-lattice type fish trap was operated in the outlet creek (Lower Lake Creek) between 1958 and 1962 or 1963 during the winter steelhead run; however, this trap reportedly did not impede the adult sockeye salmon migration (J. Ayerst32).

Crutchfield et al. (1965) reported that an adult trap was operated in Lake Creek in the fall of 1960, 1961, and 1962. During these years, total counts of sockeye salmon at this trap were 1,223, 1,485, and 763, respectively. In the spring of 1958, 64,946 juvenile sockeye salmon smolts were counted in a downstream migrant trap placed in Lake Creek (Crutchfield et al. 1965). Migration of sockeye salmon up Lake Creek may have been interrupted for a few years in the early to mid-1970s due to operation of a weir to trap chinook salmon for artificial propagation (S. A. Chitwood33).

Kokanee-size O. nerka currently occur in Lake Pleasant, although their origin is uncertain. Smoker et al. (1952) stated that

The State Game Department made apparently unsuccessful plants of silver trout [kokanee] in 1936, 1937 and 1938.

No further information on historical or recent kokanee introductions was available. Both kokanee-sized fish and sockeye salmon have been observed spawning at the same place and time on lakeshore beaches of Lake Pleasant (R. Gustafson, NMFS, Pers. observ., November 1995). Most Lake Pleasant sockeye salmon display a dirty red coloration on the spawning grounds (J. Haymes34).

Baker River

Rathbun (1900) indicated that, historically, sockeye salmon began arriving at the mouth of the Baker River in the middle of June and reached Baker Lake chiefly during July. Spawning occurred both in the lake and in Noisy Creek and "Sutter River" (upper Baker River?) beginning near the end of August or early September (Rathbun 1900, p. 269). The State of Washington established a hatchery, principally for sockeye salmon propagation, at Baker Lake in 1896. Baker Lake Hatchery was sold to the U.S. Fish Commission in 1899 and continued propagating the majority of returning sockeye salmon to this system until the end of 1933 (Kemmerich 1945).

Early reports of the Baker Lake Station included in Reports of the U.S. Commissioner of Fish and Fisheries (Ravenel 1901, 1902; Titcomb 1904), indicated that spawning sockeye salmon occurred both along the shoreline and in the Upper Baker River. Around the turn of the century, gill nets were used to capture adult sockeye salmon for hatchery broodstock along lake shore spawning beds, and racks were placed in the upper Baker River in an attempt to prevent sockeye salmon from ascending the river (Ravenel 1901, 1902; Titcomb 1904). Ravenel (1902) estimated that in 1900 over 25% of adult sockeye salmon in Baker Lake ascended the upper Baker River to spawn. Surveys conducted by WDF personnel in 1954 and 1955 showed that 95% of the sockeye salmon in the Baker River system at that time, spawned in Baker Lake on shoreline beaches (Quistorff 1954a,b,c; Quistorff 1959; PRO-Salmon 1994). Quistorff (1955) stated that

Spawning sockeye salmon were observed in heaviest concentrations along the mid- south shore of Baker Lake where a condition of underground water movement was found.

Some spawning sockeye salmon were also observed by Quistorff (1955) in Channel Creek and in the main upper Baker River one quarter mile downstream of Channel Creek.

Hamilton and Andrew (1954) stated that sockeye salmon spawned in the upper Baker River without mention of shoreline spawning. However, Wayne (1961) stated that "the natural spawning areas for sockeye salmon . . . had been located along the north shore of the lake." It is unknown whether sockeye salmon that originally spawned on the shoreline of Baker Lake, and to some degree in tributaries of Baker Lake, consisted of a single genetic stock or multiple stocks. However, as pointed out in Hendry and Quinn (1997), between 1899 and 1933 hatchery operations "thoroughly mixed the descendants of any subpopulations that might initially have been present."

Construction of Lower Baker Dam just above the town of Concrete on the Baker River in 1924 to 1927 created Lake Shannon Reservoir (Wayne 1961). During dam construction in 1925, approximately 8,000-10,000 adult sockeye salmon were blocked from reaching Baker Lake, and only 40 sockeye salmon were successively lifted over the dam and eventually reached Baker Lake (Kemmerich 1945). Between 1926 and 1957, sockeye salmon were trapped at the base of Lower Baker Dam and transported over the dam in small steel tanks on a 244-m-long highline cableway and then chuted into the reservoir (Wayne 1961).

Pre-spawner mortalities occurred below the dam (Wayne 1961), and escapement records for this period are for sockeye salmon that actually passed over the dam; Kemmerich (1945) estimated that 20-25% of sockeye salmon counted over the dam between 1926 and 1933 never reached Baker Lake due to mortalities resulting from this handling.

Following cessation of propagation efforts in 1933, and until construction of Upper Baker Dam in 1956, sockeye salmon that reached Baker Lake were allowed to spawn naturally. At that time, outmigrating sockeye salmon either passed over the surface spillway of Lower Baker Dam (where mortality was estimated at 64% when one spillway was open) or through the turbines (where mortality was estimated at 34%) (Hamilton and Andrew 1954). Hamilton and Andrew (1954) estimated that the sockeye salmon population had declined by 55% since dam construction. Use of a ski-jump spillway, first installed in 1955, considerably decreased spillway mortalities (Regenthal 1955) but resulted in loss of potential hydroelectric power (Wayne 1961).

Construction of Upper Baker Dam in 1959 inundated the original Baker Lake and created New Baker Lake (Upper Baker Reservoir), submerging the natural lakeshore spawning beaches and most of the potential tributary spawning areas beneath more than 18 m of water (Wayne 1961). Today this reservoir is commonly referred to as Baker Lake.

Fish handling facilities were updated in the late 1950s with construction of a new barrier dam and fish trap 0.8 km downstream of Lower Baker Dam. These events induced the use of tanker trucks for transporting adult sockeye salmon to Baker Lake, construction of artificial spawning beaches adjacent to Channel Creek above Baker Lake, the installation of turbine-pump-operated smolt collecting barges ("gulpers") at the head of each dam, and 20 to 25-cm-diameter fish transportation pipes that guided smolts from the gulpers through the face of each dam to be deposited into the tailrace channel below each dam (Wayne 1961, Quistorff 1966). Use of the transportation pipe through Upper Baker Dam was discontinued in 1987, and outmigrating juvenile sockeye salmon trapped since then at the Upper Baker Dam gulper have been trucked from Upper Baker Dam to the Baker River below Lower Baker Dam. Guide nets have subsequently been installed at both dams to discourage outmigrating fish from going over the spillways or through the turbines.

Artificial sockeye salmon spawning beaches 1, 2, and 3 were constructed in 1957, 1959, and 1966, respectively above Baker Lake off Channel Creek. Beach 1 ceased operation in 1965, but the spawner capacity of beaches 2 and 3 continues at 1,500 adult sockeye salmon each. Sockeye salmon fry from beaches 2 and 3 currently leave the beaches on their own volition through outlets into Channel Creek and from there into Baker Lake. The future of spawning beaches 2 and 3 is uncertain. According to WDFW (1996), "the state and tribes favor the continued use of beaches 2 and 3, but Puget Power and the Forest Service would like to close them." Spawning beach 4, on Sulphur Creek, below Upper Baker Dam, began operating in 1990 with a spawner capacity of 3,000. Fry leaving beach 4 are captured and hauled by tanker truck to Baker Lake.

Since the construction in 1958 of a barrier dam and fish trap below Lower Baker Dam, adult sockeye salmon have been trapped and hauled by tanker truck to the spawning beaches or to Baker Lake. Since 1986, a portion of the fry leaving the spawning beaches have been collected and reared in net-pens in Lake Shannon Reservoir prior to being released as smolts through the Lower Baker Dam gulper. Total smolt releases between 1987 and 1992 were over 400,000 (WDF et al. 1993) (see Appendix Table D-2 and "Artificial Propagation" section). It was stated in WDFW (1996) that

The future of the net pen program is uncertain. There have been three consecutive years of major IHN outbreaks in the net pens. Although the program did increase egg-to-smolt survival, it did not increase smolt-to-adult survival. In 1996, sockeye fry (brood year 1995) are not expected to be taken to the net pens.

Currently, adult sockeye salmon return to the Baker River trap from mid-June to mid-August and spawn in the artificial beaches from late September through December, peaking from late October to late November (WDF et al. 1993). In addition to releases into the artificial spawning beaches, significant numbers of adult sockeye salmon were released into Baker Lake in 1967, 1972, and 1994, and most likely in 1962, 1963, and 1964, as well (WDFW 1996). For instance, over 1,000 sockeye salmon (25% of the fish trap count) were released in Baker Lake in 1967 (Orrell 1969). The return of almost 16,000 sockeye salmon to Baker River in 1994 far exceeded the 4,000-fish capacity of the spawning beaches. Following consultations between WDFW and the Skagit System Cooperative (representing Skagit River area tribes) regarding a fishery on this stock, the remaining 12,000 adults were liberated into Baker Lake in an experiment to determine survival rate and production potential for natural spawning of sockeye salmon in Baker Lake (J. Ames35).

The lower portion of the upper Baker River is now a braided stream subject to severe meander and, like shoreline spawning habitat, is subject to effects of reservoir drawdown in winter and spring (G. Sprague36). Baker Lake and beaches 2 and 3 near Channel Creek commonly freeze over in winter, but Lake Shannon and beach 4 never freeze over (G. Sprague37).

Kokanee are present in the system, although Ward (1929, 1930, 1932) and Kemmerich (1945) thought that "kokanee-sized" O. nerka in Baker Lake were derived from sockeye salmon residuals. Following construction of Lower Baker Dam, resident O. nerka were observed spawning in Baker Lake for the first time and were presumed to have originated from sockeye salmon residualizing in Lake Shannon Reservoir (Ward 1929, 1930, 1932, Kemmerich 1945).

Between 1991 and 1994, 1,158,200 hatchery reared Lake Whatcom kokanee were released in Lake Shannon (Knutzen 1995). Lake Whatcom kokanee were reportedly released into Baker Lake in the past,38 although we were unable to locate stocking records. In average years, 40-100 "kokanee-sized" O. nerka spawn in the outlet channel that drains the two upper sockeye salmon spawning beaches and flows into Channel Creek (W. Steuer39). It is possible that a portion of the kokanee that have been recently planted in Lake Shannon Reservoir from Lake Whatcom stock may have outmigrated through the Lower Baker Dam gulper. It is unknown whether these potential outmigrating kokanee have returned as sockeye salmon; however, if they have, they would presumably have been placed together with native sockeye salmon in the spawning beaches (G. Sprague40).

Lake Washington

The following historical overview of changes to the Lake Washington Basin has been compiled from Evermann and Meek (1898), Ajwani (1956), Woodey (1966), Larson (1975), Stickney and McDonald (1977), Corsaletti (1981), Chrzastowski (1983), and Buerge (1985, 1989).

Between 1911 and 1916, construction of the Lake Washington Ship Canal, and associated engineering projects, profoundly altered both the natural drainage patterns of Lake Washington and potential migratory routes of anadromous fish native to the basin. During this period, the Cedar River was diverted to discharge into Lake Washington, the level of Lake Washington was lowered approximately 3 m, the outlet into the Black River ceased to exist, the Sammamish River channel was widened and deepened, and the newly constructed Lake Washington Ship Canal became the new lake outlet.

Historically, Lake Washington drained to the south through the Black River, which flowed for 5.3 km to its confluence with the White River (now the Green River) to form the Duwamish River and then flowed into Puget Sound. The Cedar River entered the Black River less than 1 km below the Lake Washington outlet, and Lake Washington's principal tributary was the Sammamish River (historically called Squak Slough). The Black River had an average depth of 1.2 m and ranged in width from 15 to 46 m. At flood stage, the Cedar River commonly reversed the flow of the upper segment of the Black River, causing Cedar River water to flow into Lake Washington. At these times, the Black River had water flowing in opposite directions at its two ends, north into Lake Washington and west into the Duwamish River. This is why the Black River was called "Mox La Push," meaning "two mouths," in the Chinook jargon. Prehistorically, the Cedar River may have been a major tributary to Lake Washington (Chrzastowski 1983).

In 1911, construction on the locks (now the Hiram M. Chittenden Locks), dam, and Fremont and Montlake Cuts began. According to Chrzastowski (1983) and Buerge (1985, 1989), the Cedar River was permanently diverted into Lake Washington in the summer of 1912 by excavation of a channelway 24 m wide and 610 m long. This diversion was precipitated by severe flooding on the Cedar River in the winter of 1911 that required the evacuation of the city of Renton. The locks were completed in the spring of 1916, and by 25 July 1916 the level of Salmon Bay had been raised to equal that of Lake Union. The lowering of the level of Lake Washington to that of Lake Union was gradual, occurring over a 4-month period from July to October 1916 (Stickney and McDonald 1977, Chrzastowski 1983). The lowering of Lake Washington also increased the gradient of the Sammamish River, making it too shallow and narrow for navigation, and leading to the widening, deepening, and channelization of the Sammamish River in 1916 by the U. S. Army Corps of Engineers (Stickney and McDonald 1977, Chrzastowski 1983). The opening of the Lake Washington Ship Canal was celebrated on 4 July 1917.

The lowering of Lake Washington in 1916 left the channel of the Black River high and dry, while the diversion of the Cedar River into Lake Washington in 1912 caused the Cedar River to become Lake Washington's principal tributary and approximately doubled the flow of freshwater into Lake Washington. Ajwani (1956) stated that

Whether the runs of fish occurring at the time of these diversions were eliminated or what the degree of their reduction was, cannot be determined because of lack of data.

Reports in the literature are equivocal as to whether sockeye salmon were historically present in the Lake Washington/Lake Sammamish Basin prior to 1916, although kokanee were numerous. Prior to construction of the Lake Washington Ship Canal, fishing for "silvers" and "trout" was reported to be at its best near the confluence of the Black River and Lake Washington (Larson 1975, Slauson 1976). Silver salmon is a common local name for coho salmon, while in the early part of this century kokanee were called "silver trout" (and are still so designated in Washington State fishery regulations). Hammond (1886), in reference to Lakes Washington and Sammamish, stated that

The only fish in them is a species of trout, very few in number, the largest of which are about a foot in length.

Seale (1895) reported the collection of "six large specimens" of "dark" O. nerka taken in Lake Washington on 7 November 1892 and two others, "more silvery in color," taken on 30 June 1895. No dimensions for these fish were recorded, although the species was reportedly "very abundant" (Seale 1895). Woodey (1966) surmised from the dates of collection and coloration of the specimens that the O. nerka reported by Seale (1895) were most likely kokanee. Four of Seale's (1895) specimens of O. nerka collected in Lake Washington on 7 November 1892 are currently deposited in the California Academy of Sciences fish collection (D. Catania41). The relatively short fork lengths of these 4 specimens (241-249 mm) indicate that Seale's (1895) "six large specimens" of O. nerka were kokanee, not sockeye salmon.

Jordan and Evermann (1896) reported that Prof. O. B. Johnson observed large "redfish," presumably sockeye salmon, at Lake Washington. Evermann (1896) reported that the "small redfish" was found at Lake Washington, and goes on to say that
Prof. O. B. Johnson found the small form spawning in Lake Washington near the last of November, 1888, and on October 8, 1889.

Evermann and Meek (1898), in reference to Lake Washington, stated that

Salmon are said to enter the lake through the Black River early in the fall, but none was seen. They are probably the large form of the redfish or sockeye (Oncorhynchus nerka). Redfish are said to run up into shallow places during the latter part of October and a part of November . . .

All stocks of sockeye salmon presently found in Lake Washington complete their migration into Lake Washington before the end of August (WDF et al. 1993), suggesting that the salmon Evermann and Meek (1898) reported as entering the Black River in the fall were not sockeye salmon, or that sockeye salmon had radically different run-timing in Lake Washington in the 1890s compared to the present day, and that this stock is now extinct. While seining in Lake Washington, Evermann and Meek (1898) collected 17 "small redfish" ranging in length from 24 to 27 cm, but did not collect anadromous-sized O. nerka.

In regard to Lake Sammamish, Evermann and Meek (1898) stated "no information could be obtained as to what kind of salmon enter the lake," but reported that local residents said that redfish were plentiful in "Squak Slough" (Sammamish River) and that "salmon run with the redfish." Evermann and Meek (1898) presumed that redfish, or "grayling" as they were called locally, spawned from the latter part of October to early or mid-November.

Subsequent authors either stated that a small population of sockeye salmon occurred in Lake Washington (Rathbun 1900; Evermann and Goldsborough 1907; Cobb 1911, 1914, 1930) or that Baker River had the only population of sockeye salmon in Puget Sound (Cobb 1927, Rounsefell and Kelez 1938, Royal and Seymour 1940, Kemmerich 1945). Pratt and Jewell (1972) reported that no record has been located of "sea-run" sockeye salmon in Lake Washington prior to their introduction in 1935. Surveys conducted on the Cedar River, Big Bear Creek, Cottage Lake Creek, and Evans Creek on 2 and 3 September 1930 did not report the occurrence of sockeye salmon (WDFG 1932). Currently, early September is near the beginning of sockeye salmon spawn timing for these streams (WDF et al. 1993). In reviewing the history of O. nerka in the Lake Washington/Lake Sammamish drainage, Hendry (1995) concluded that "limited runs of sockeye salmon . . . were probably present at the turn of the century," and that

The status of Lake Washington sockeye salmon during this period (1917-1937) will probably never be fully determined but it is certainly unlikely that large populations were present.

Sockeye salmon vertebral remains were identified in prehistoric fish remains from the Duwamish No. 1 archeological site (45-KI-23), located 3.8 km upstream from Elliot Bay on the Duwamish River, utilized by aboriginal humans between A.D. 15 and A. D. 1654 (Butler 1987). Fish remains from two archeological sites on the former Black River, Tualdad Altu (45-KI-59, Earlington site) and Sbabadid-D (45-KI-51-D), revealed numerous Oncorhynchus sp. remains, but identification to the species level was not undertaken in this study (Chatters 1988, Butler 1990).

Smith (1940), reporting on cultural interviews with local tribal elders, stated that "when asked about the red salmon (O. nerka) informants said the silver side might be called that as it turned red in freshwater, but they knew of no separate species by this name." Smith (1940) goes on to say

A small salmon was said to live permanently in Lake Washington spawning in the creeks which emptied into the lake. The Duwamish of that section and even those at the intersection of the White and Green Rivers were said to prefer this salmon to that which entered the rivers from the Sound.

This reference to land-locked salmon most likely refers to the numerous kokanee then present in Lake Washington.

The undisputed historic presence of kokanee in Lake Washington indicates that sockeye salmon existed in Lake Washington, at least in prehistoric times. Several factors may have favored subsequent evolution of kokanee (and non-anadromy) at the expense of anadromous O. nerka in Lake Washington. Chrzastowski (1983) stated

For most of the year, Lake Washington in its natural state was a poorly flushed lake, and water quality reportedly worsened noticeably during the dry season (July-Sept.) when the lake was relatively stagnant. Average residence time for the lake water in the natural state probably was about twice the present-day value, or nearly 5 years.

In addition, spring floods on the Cedar River that commonly backed up the Black River into Lake Washington (preventing the lake from draining for a time) would probably have occurred during the period of potential smolt outmigration (March to early June). Both these factors (low flushing rate and difficulty of locating the outlet during flood stages) may have inhibited smolt outmigration.

Foerster (1937) found that when surface epilimnion temperatures rose above 10oC in Cultus Lake, a physiological temperature barrier was formed that terminated downstream migration of young sockeye salmon. The observations of Foerster (1937) may be used to support the hypothesis that prior to diversion of the Cedar River into Lake Washington, which more than doubled the lake's water budget, relatively low outflow and seasonal development of a deep epilimnion of warm water may have presented a physical and/or physiological impediment to downstream sockeye salmon smolt migration. Recent historical changes in the drainage pattern of Lake Washington may have created conditions that were more favorable to development of anadromous O. nerka. However, it should be noted that currently sockeye salmon smolts in Lake Washington are known to continue to outmigrate into June through 17oC temperature water (Warner 1997).

WDF et al. (1993) recognized three separate stocks of sockeye salmon currently in the Lake Washington/Lake Sammamish drainage: Cedar River, Lake Washington/Lake Sammamish Tributaries, and Lake Washington beach spawning. WDF et al. (1993) indicated that sockeye salmon stocks that spawned in the Cedar River were of non-native origin, and stocks that spawned in other Lake Washington/Lake Sammamish tributaries and on lakeshore beach habitat in Lake Washington were of unknown origin and were perhaps native to the drainage.

Available artificial propagation data and transplantation records provide evidence that the current Cedar River and Issaquah Creek (a tributary of Lake Sammamish) sockeye salmon are introduced populations (Royal and Seymour 1940, Kolb 1971, Burgner 1991) (see Appendix Table D-2). The majority of Cedar River sockeye salmon spawn from mid-September into January (a few are still spawning in February), with a peak in mid- to late October (WDF et al. 1993). The Cedar River population was believed by Kemmerich (1945), Royal and Seymour (1940), Kolb (1971), and Pratt and Jewell (1972) to be derived from the direct planting of over 1 million fry and fingerlings between 1935 and 1944 (see Appendix Table D-2 and " Artificial Propagation" section). These introductions originated from a sockeye salmon stock perpetuated at the U.S. Bureau of Fisheries Birdsview Hatchery on Grandy Creek in the Skagit River Basin.

The Birdsview Hatchery stock was started in 1908 from Fraser River sockeye salmon captured in commercial traps at Point Roberts, and egg takes in 1912 and 1916 indicate that substantial numbers of adult fish returned from these initial releases in Grandy Creek and Grandy Lake (Kemmerich 1945). In 1916, fry derived from Quinault Lake were used to supplement the Birdsview Hatchery stock (Kemmerich 1945, see Appendix Table D-2). However, over the years 1914-1945 the parent stock for this hatchery program was overwhelmingly Baker Lake sockeye salmon (Kemmerich 1945).

Out-of-basin releases in the Lake Washington/Lake Sammamish Basin totaled over 3.4 million fry and fingerlings from releases of 1) an unknown stock in 1917, 2) Birdsview Hatchery stock between 1935 and 1945, and 3) Cultus Lake stock released in 1944, 1950, and 1954 (Woodey 1966, Kolb 1971, Hendry 1995) (see Appendix Table D-2). From 1947 to 1970, adult sockeye salmon returning to the Issaquah Hatchery provided broodstock for numerous additional fry and fingerling releases to Issaquah Creek, and limited releases to Lake Union and the Cedar River (Kolb 1971).

Sockeye salmon in the Lake Washington/Lake Sammamish tributaries stock spawn primarily in Big Bear Creek, Cottage Lake Creek, and East Fork Issaquah Creek, with minor numbers in other Lake Sammamish tributaries (WDF et al. 1993), such as Laughing Jacobs and Lewis Creeks (Ostergaard et al. 1994). Spawning in these creeks extends from early September through November, with a peak in mid- to late October, depending on stream flow. Issaquah Creek received sockeye salmon fry and fingerling plants of over 1.6 million Birdsview Hatchery-origin fish between 1935 and 1945, and over 59,000 Cultus Lake-origin fish in 1950 and 1954. North Creek, a Sammamish River tributary, received over 23,000 Cultus Lake sockeye salmon fry planted in 1944 (Kolb 1971).

Kemmerich (1945), reporting on the effectiveness of introductions of sockeye salmon of the 1934 broodyear from the Birdsview Hatchery into the Cedar River and Issaquah Creek, stated that two sockeye salmon were found "in the Bear Creek fish trap of the State Game Department" on 5 October 1938. No sockeye salmon had been planted in Big Bear Creek up to this point, with the exception of fry planted in the spring of 1937 which would not have reached maturity in 1938. However, Kemmerich (1945) pointed out that 76,000 sockeye salmon fry from Baker Lake had been planted in Issaquah Creek in the summer of 1935. Out-of-basin releases of O. nerka fry into Big Bear Creek and its two tributaries, Cottage Lake Creek and Evans Creek, included 576,000 sockeye salmon fry, primarily of Baker Lake origin, stocked in Big Bear Creek in 1937 (Royal and Seymour 1940, Kemmerich 1945, Kolb 1971). In addition, over 34 million Lake Whatcom kokanee were stocked in Big Bear and Evans Creeks between 1917 and 1969 (Pfeifer 1992), and over 177,000 kokanee from an unknown source population stocked in Big Bear Creek in 1917 (Pfeifer 1992) (see Appendix Tables D-2 and D-5).

Kokanee used for stock transfers in the early part of this century were most commonly derived from either Kootenay Lake, British Columbia or Lake Whatcom, Washington (Pfeifer 1992). Pfeifer (1992) stated that

I cannot rule out the possibility that kokanee from Kootenay Lake were among the many early introductions for which the egg or fry source was not explicitly recorded. In the Lake Washington system, the kokanee found in Big Bear Creek exhibit a spawn timing intermediate to that of the Kootenay and Whatcom strains, and are found spawning alongside the anadromous form.

Currently, kokanee in the Lake Washington/Lake Sammamish Basin can be separated into two groups based on very different spawn timing; 1) a group of early-entry kokanee in Issaquah Creek (a tributary at the southern end of Lake Sammamish) that exhibit late July to early September spawn timing, and 2) kokanee in the Sammamish River and Lake Sammamish tributaries, including a second run of kokanee in Issaquah Creek that spawn in September/October in Big Bear Creek, October/November in Issaquah Creek, and late November/December in Laughing Jacobs and Lewis Creeks (Pfeifer 1992, Ostergaard et al. 1995). Ostergaard et al. (1995) stated that early entering kokanee in Issaquah Creek are known to be native, while kokanee in other tributaries to Lake Sammamish and the Sammamish River are believed to be non-native, based on their later run-timing. Ostergaard (1996) listed 8 creeks, tributary to the east and south shores of Lake Sammamish, that historically supported native early entering kokanee. Ostergaard (1996) estimated the 4-year (1992-1995) total spawning population of these kokanee in Issaquah Creek at 81 fish.

Kokanee once existed in streams tributary to Lake Washington, other than the Sammamish and Cedar Rivers. Shultz and Students (1935) observed kokanee spawning in Swamp Creek, a tributary of the lower Sammamish River, from September to November 1933. The University of Washington Fish Collection has specimens of kokanee collected in Swamp Creek on 30 August 1920, 28 October 1928, and 27 November 1933. Since these observations and collections were made before the first recorded transplants of kokanee to Swamp Creek (see Appendix Table D-5), it is apparent that not all kokanee native to the Lake Washington/Lake Sammamish Basin had exclusively early run-timing.

Spawning sockeye salmon intermingle with spawning Big Bear Creek and late-entry Issaquah Creek kokanee (Pfeifer 1992), as well as with kokanee spawning in Laughing Jacobs and Lewis Creeks (Ostergaard et al. 1995). Kokanee are also reported to spawn together with sockeye salmon in the Cedar River (Pfeifer 1992), although the coloration of Cedar River kokanee-sized fish is typical of the coloration shown by residuals in other lake systems (J. Ames42).

Pfeiffer (1992) stated that fish traps were operated in Big Bear Creek by the King County Game Department and the Washington Department of Game in the 1930s. In reference to Big Bear Creek, Ajwani (1956, p. 67-68) stated that

A wooden weir was constructed across the stream in 1925, when the County Game Commission was in operation. At that time the county would take all the eggs obtained from the silver trout run and either plant or trade these eggs elsewhere. . . . For its size, this stream is . . . one of the largest producers of silver trout in the state.

The run-timing of kokanee in Big Bear Creek is essentially concurrent with that of anadromous sockeye salmon: from early September to late November with a peak in the first to second week of October (Ostergaard et al. 1995). Therefore, kokanee fish traps and weirs operated in Big Bear Creek in the 1920s and 1930s would presumably have impeded migration of sockeye salmon that may have been in the system at that time. Prior to the single recorded transplant of 576,000 sockeye salmon fry into Big Bear Creek in 1937 and the recorded return of 2 adults in October 1938 and another 2 adults in October 1940, no mention of sockeye salmon in Big Bear Creek occurs in the published literature.

Surveys by King County Surface Water Management Division in 1992, 1993, and 1994 recorded only 242, 23, and 9 kokanee, respectively in the Big Bear Creek drainage (Ostergaard et al. 1995). In addition, Ostergaard et al (1995) stated that the 9 fish seen in 1994 may have been residual sockeye salmon or sockeye salmon x kokanee hybrids. Past results of electroshocking in Big Bear Creek by WDFW have indicated that the number of kokanee visually observed is a small fraction of the actual number of fish present (WDFW 1996), therefore it is probable that the kokanee population in Big Bear Creek during 1992-1994 was larger than the numbers in Ostergaard et al. (1995) suggest. Ostergaard (1996) estimated an escapement of 317 kokanee to Big Bear Creek in the fall of 1995, based on WDFW survey numbers.

Beach spawning sockeye salmon are found in both Lake Washington and Lake Sammamish; WDF et al. (1993) considered the beach spawners in Lake Washington a separate stock, but the status of the Lake Sammamish beach spawners was undetermined due to lack of information. Berggren (1974) reported that the numbers of beach-spawning sockeye salmon in Lake Sammamish between 1969 and 1972 ranged from a low of 125-200 in 1969 to a high of 1,400-1,900 in 1971. Recent estimates of Lake Sammamish beach-spawning sockeye salmon were unavailable. Lake Washington beach spawning occurs primarily between October and January. Spawning has been observed in many locations around the perimeter of Lake Washington, but primarily at Pleasure Point Beach on the southeast shoreline, in the Bellevue area, near Juanita Point, along Enatai Beach (Buckley 1965), and around the shoreline of Mercer Island (Woodey 1966, WDF et al. 1993).

Riverine spawning sockeye salmon in Washington

Methow River-Prior to the hatchery program at the Winthrop National Fish Hatchery (NFH), sockeye salmon were apparently not present in the Methow River (WDF et al. 1938, Mullan 1986, Chapman et al. 1995). Over 1.8 million sockeye salmon of Rock Island and Bonneville Dam origin were released in the Methow River from Winthrop NFH between 1945 and 1957 as part of the GCFMP (Mullan 1986, Chapman et al. 1995) (see Appendix Table D-2). Chapman et al. (1995) indicated that small numbers of sockeye salmon continue to return to the Methow River every year and this population appears to be self-perpetuating. Allen and Meekin (1973) reported that, based on weir counts, about 1% of the sockeye salmon passing Wells Dam in 1965 and 1966 entered the Methow River. French and Wahle (1960) and Fryer and Schwartzberg (1993) reported that sockeye salmon spawning occurred between river kilometers 57 and 64 on the Methow River downstream of Twisp. Langness (1991) reported that sockeye salmon were observed spawning in the Methow River from 1987 to 1990 and that the distribution of spawning was essentially the same as reported in French and Wahle (1960). It has been postulated that sockeye salmon that spawn in the Methow River may rear in mainstem reservoirs on the Columbia River (Chapman et al. 1995).

Allozyme data presented by Chapman et al. (1995) indicate a closer association of Methow River sockeye salmon with Lake Wenatchee sockeye salmon than with Okanogan River sockeye salmon. However, recent analysis of allozyme data based on sockeye salmon collected in 1994 (Okanogan River fish collected at Wells Dam) indicate that sockeye salmon from the Methow River, Wenatchee River, and Okanogan River (Wells Dam) belong to a common gene pool (Utter 1995). The apparent genetic similarity between Wenatchee and Okanogan River sockeye salmon reported in Utter (1995) is inconsistent with findings of significant genetic distance between Wenatchee and Okanogan populations as reported in Utter et al. (1984), Brannon et al. (1994), Thorgaard et al (1995), and Winans et al. (1996).

Entiat River-Prior to the hatchery program at the Entiat National Fish Hatchery, sockeye salmon had not been observed in the Entiat River (WDF et al. 1938, Mullan 1986, Chapman et al. 1995). Approximately 161,787 juvenile sockeye salmon derived from Quinault Lake stock were released into the Entiat River in 1942 and 1943 (Chapman et al. 1995) (see Appendix Table D-2). In addition, 22,341 Lake Chelan kokanee, derived from Lake Whatcom stock, were released into the Entiat River in 1944 (Mullan 1986) (see Appendix Table D-5).

Barnaby (1946) indicated that of 60,010 marked sockeye salmon juveniles of the 1941 brood released in the Entiat River on 8 May 1943, 93 were recovered as adults in 1944 and 658 in 1945. In 1945, 33 marked sockeye salmon were recovered in the Entiat River, 3 in the Wenatchee River, and 622 in the Columbia River commercial fishery (Barnaby 1946). In contrast, Fulton and Pearson (1981) indicated that 670 adults from this experiment were recovered in the lower river fishery, with only one recovered in the Entiat River and 3 in the Wenatchee River. Apparently, these introductions established a small sockeye salmon population that provided enough returning adults to provide broodstock in the 1950s for release into Lake Wenatchee, Lake Osoyoos, and Icicle Creek (Mullan 1986, Chapman et al. 1995) (see Appendix Table D-2).

Although Mullan (1986) believed that transplants of Quinault Lake stock established sockeye salmon in the Entiat River, he postulated these three alternate hypotheses to explain their occurrence: 1) inadvertent inclusion of sockeye salmon with other species of salmon trapped at Rock Island Dam and released in 1939-1940 in the Entiat River, 2) escape of juvenile sockeye salmon from the Entiat Hatchery, and 3) straying from other stocks. Since natural sockeye salmon stocks had not become established in the Entiat River prior to the GCFMP, Mullan (1986) discounted straying as a possible origin for Entiat River sockeye salmon. Currently small numbers of sockeye salmon are observed in the Entiat River almost every year (Chapman et al. 1995). Chapman et al. (1995) considered these fish either as strays from Lake Wenatchee or Okanogan River or as artifacts of the hatchery stocking program carried on during the 1940s and 1950s. It was postulated that sockeye salmon that spawn in the Entiat River rear in mainstem reservoirs on the Columbia River (Chapman et al. 1995).

Similkameen River-Although the Similkameen River, which originates in British Columbia, is considered the main tributary of the Okanogan River downstream from Lake Osoyoos, it is considerably larger than the Okanogan, contributing some 3 to 4 times the water volume of the mainstem Okanogan (WDF et al. 1938, Bryant and Parkhurst 1950, Mullan 1986). Fulton (1970) and Allen and Meekin (1980) listed Palmer Lake and its inlet tributary Sinlahekin Creek as historical sockeye salmon habitat. In contrast, Mitchell (1980) suggested that prior to construction of Enloe Dam in 1920, the original Squantle (Similkameen) Falls was 7.6 to 9.1 m high and would have acted as a block to upstream migration of sockeye salmon to Palmer Lake. Bryant and Parkhurst (1950) reported that 500 dead unspawned sockeye salmon were found in the Similkameen River in 1936 and that these fish may have been part of the population that normally spawned above Lake Osoyoos. French and Wahle (1954) observed sockeye salmon below Enloe Dam on the Similkameen River from early to mid-August, but not in late August or September of 1954. Other authors reporting the occurrence of sockeye salmon in the Similkameen River included Chapman (1941), French and Wahle (1960, 1965), CBFWA (1990), Langness (1991), and Chapman et al. (1995).

Currently, small numbers of sockeye salmon are seen almost every year below Enloe Dam on the Similkameen River (Chapman et al. 1995). The origin of these sockeye salmon is uncertain; hypotheses proposed include straying of sockeye salmon from the Okanogan River and returns of anadromous individuals derived from kokanee in upstream Palmer Lake (Chapman 1941, Rounsefell 1958a, Fulton 1966). WDFG (1921a, 1921b) recorded the release of 132,500 kokanee (silver trout) into Palmer Lake in 1919-1920. In 1966, Fulton (1966) reported that 45,000 kokanee and 87,000 sockeye salmon were released in Sinlahekin Creek and Palmer Lake (see Appendix Tables D-2 and D-5 ); 15 thermally marked outmigrants from this release were captured at Priest Rapids Dam.

Icicle Creek-Icicle Creek is a tributary of the Wenatchee River below Lake Wenatchee and is also the site of Leavenworth National Fish Hatchery. Over 1.5 million juvenile O. nerka were released directly into Icicle Creek between 1942 and 1969: 1.1 million of Rock Island Dam heritage, over 270,000 of Entiat River heritage (progeny of Quinault Lake stock), over 44,000 of Methow River heritage, over 100,000 of Lake Wenatchee heritage, about 3,000 from an unknown British Columbia sockeye salmon stock, and over 29,000 Lake Wenatchee kokanee (Chapman et al. 1995, NRC 1995) (see Appendix Table D-2). Chapman et al. (1995) stated that currently, small numbers of adult sockeye salmon are observed in Icicle Creek almost every year. Since the Leavenworth NFH is located only 4.5 km from the Wenatchee River, Mullan (1986) suggested that observations of sockeye salmon in Icicle Creek could represent some residual attraction of hatchery-reared fish to the water they were reared in before their release into Lake Wenatchee, or it could have represented straying into the wrong tributary. As pointed out by Chapman et al. (1995), no sockeye salmon have been reared at Leavenworth since the mid-1960s; however, "generally less than a few dozen" sockeye salmon are still seen in Icicle Creek each year.

Nooksack River-In reference to the glacially influenced Nooksack River, Rathbun (1900) stated that

The sockeye have been said to enter it, but the evidence to that effect is not conclusive.

Kershaw (1902) stated that "the sockeye occasionally ascend the river in small numbers." In reference to sockeye salmon, Crawford (1907) stated

a few have been known to enter the Nooksack River and spawn in one of its small tributaries . . . those from the Nooksack . . . were noticed during the great run of 1905 when the sockeyes ran closer to the shore on the Sound than has ever been known before. Last season a great many salmon ascended the Nooksack River.

FWTC (1970) stated that

At least one section of the Nooksack system supports a small run of sockeye salmon. It is a half-mile-long side channel of the North Fork, located 3.5 miles upstream from the town of Glacier. Other stream sections, and some tributaries, in both the North and South Fork Nooksack, also receive limited sockeye runs.

Williams et al. (1975) also reported that sockeye salmon spawn along a half-mile side channel of the North Fork Nooksack River about 3.5 miles above the town of Glacier and below Lookout Creek (RKm 100.5). Small numbers of spawners are still seen each year in the North and South Forks of the Nooksack and in Maple Creek (D. Hendrick43). As summarized in Appendix Table C-7, WDFW Salmon Spawning Ground Survey Data (Egan 1977, 1995, 1997) indicated several locations, dates, and peak numbers of spawning sockeye salmon in the Nooksack River.

Sockeye salmon are caught in the Nooksack River as by-catch in the Nooksack Tribal coho harvest during the months of September through October. This freshwater fishery occurs from the confluence of the North and South Forks of the Nooksack River, downstream to the mouth (D. Grieggs44). Recent tribal harvest has ranged from 15 in 1992 to 386 in 1991 (Hoines 1995). Run-timing of sockeye salmon caught in the Nooksack Tribal fishery is significantly later than either Baker River or Lake Washington sockeye salmon stocks, which terminate by mid-August, but Nooksack River sockeye salmon run-timing does overlap the timing of several lower Fraser River stocks (see Appendix Tables C-6 and C-7).

Several anecdotal reports indicated that early hatchery supplementation of sockeye salmon occurred at the Nooksack Hatchery on Kendall Creek (Pacific Fisherman 1905a), and that introductions into the Nooksack River of a small number of out-of-basin sockeye salmon fry also occurred (Pacific Fisherman 1905b, 1906).

Samish River-WDFW Salmon Spawning Ground Survey Data (Egan 1977, 1995, 1997) indicated several locations, dates, and peak numbers of spawning sockeye salmon in the Samish River (see Appendix Table C-7). Anecdotal records indicated that extensive culture of sockeye salmon, taken in fish traps in Puget Sound, occurred at the Samish State Hatchery at least in the years 1915-1917 (Pacific Fisherman 1915a, 1915b, 1916, 1918). WDFG (1916a, 1917, 1920) recorded the release into Lake Samish and Cain Lake of almost 9 million sockeye salmon fry between 1915 and 1918. They reported that the fish released were derived from sockeye salmon captured on the west side of Lummi Island (off Bellingham Bay) (see Appendix Table D-2). In 1920, over 165 sockeye salmon spawned naturally above the Samish Hatchery racks (WDFG 1921b).

Between 1934 and 1937, over 0.5 million sockeye salmon fry from the Birdsview Hatchery on Grandy Creek were released in Lake Samish (Royal and Seymour 1940, Kemmerich 1945) (see Appendix Table D-2). Several sockeye salmon were observed in 1937 and 1938 at the Samish State Hatchery and in the Samish River (Kemmerich 1945). An estimated 300-400 sockeye salmon returned to the Samish Hatchery in the fall of 1940 (Royal and Seymour 1940, Kemmerich 1945).

Skagit River Basin-In reference to riverine-spawning sockeye salmon in the glacially influenced Skagit River Basin, WDF et al. (1993) stated

They are consistently found in very small numbers in the upper Sauk River and the mainstem Skagit near Newhalem. Whether these represent strays from the Baker or other river systems or are small self-sustaining populations of a few individuals is unknown.

WDFW Salmon Spawning Ground Survey Data (Egan 1977, 1995, 1997) indicated several locations, dates, and peak numbers of spawning sockeye salmon in the Skagit River Basin (see Appendix Table C-7). Juvenile sockeye salmon displaying parr marks have been observed in the mainstem Skagit River near the town of Lyman (D. Hendrick45).

In the 1930s, extensive sockeye salmon transplants were made from Birdsview Hatchery into the following Skagit River tributaries: Day Creek, Illabot Creek, Bacon Creek, and Diobsud Creek, as well as Lake McMurray, McMurray Creek, Big Lake, and Clear Lake on Nookachamps Creek (Kemmerich 1945) (see Appendix Table D-2). No returns were noted from these plantings to Nookachamps Creek, Illabot Creek, or Day Creek. About 300 sockeye salmon were seen in Bacon Creek in 1936, and 20 and 6 in Diobsud Creek in 1936 and 1937, respectively.

Stillaguamish River-WDFW Salmon Spawning Ground Survey Data (Egan 1977, 1995, 1997) indicated several locations, dates, and peak numbers of spawning sockeye salmon in the Stillaguamish River (see Appendix Table C-7). Between 1929 and 1937, 322,175 juvenile sockeye salmon were released into Lake Cavanaugh and Pilchuck Creek (see Appendix Table D-2). In the fall of 1935, 1936, 1937, and 1938, returning sockeye salmon adults counted at the base of the falls on Pilchuck Creek amounted to 40, 3,000-4,000, 1,000-2,000, and 200-300, respectively (Kemmerich 1945). Recent tribal freshwater harvest information recorded 186 sockeye salmon taken in 1989 on the Stillaguamish River, but none in other years (Hoines 1995).

Duwamish River/Green River-WDFW Salmon Spawning Ground Survey Data (Egan 1977, 1995, 1997) indicated several locations, dates, and peak numbers of spawning sockeye salmon in the Green River (see Appendix Table C-7). Sockeye salmon have been observed spawning below Howard Hanson Dam on the Green River (E. Warner46). Recent tribal freshwater harvest of sockeye salmon in the Duwamish-Green River Basin has ranged from 0 in 1987 to 278 in 1984 (Hoines 1995). At least 392,050 sockeye salmon fry derived from Green River, Quinault Lake, and unspecified Alaska stocks were released into the Green River from the Green River State Hatchery between 1925 and 1931 (WDFG 1928, 1930, 1932) (see Appendix Table D-2).

Puyallup River-Anadromous fish trapped at the base of Mud Mountain Dam on the White River, a tributary of the Puyallup River in Puget Sound, are trucked around the dam and placed in the White River above Mud Mountain Dam. Small numbers of sockeye salmon have been reported in the yearly Mud Mountain Fish Haul Reports beginning in 1983, when 19 adult sockeye salmon were counted over the dam. Since 1985, when 378 sockeye salmon were counted at Mud Mountain Dam, small numbers ranging from 5 to 114 have been counted each year at this facility (MMDFHR 1996). Mud Mountain Reservoir is a run-of-the-river flood-control reservoir, and as such does not provide lake-rearing conditions for sockeye salmon. Information on possible spawning locations for sockeye salmon released above Mud Mountain Dam was not located.

Nisqually River-WDFW Salmon Spawning Ground Survey Data (Egan 1977, 1995, 1997) indicated that 19 sockeye salmon and 6 sockeye salmon redds were observed in August 1966 at river kilometer 20 on the Nisqually River (see Appendix Table C-7). In addition a few sockeye salmon have been reported in Mashel River and Ohop Creek, although none have been reported in Nisqually River surveys since 1982. A very few sockeye salmon are reported in the tribal freshwater harvest statistics for the Nisqually River (Hoines 1995).

Lewis River-According to WDFW (1996),

Anadromous size sockeye are occasionally observed in the North Fork Lewis River downstream of Merwin Dam.

Hamilton and Rothfus (1963) reported that 890,000 sockeye salmon fry were released in Lake Merwin in 1961 and that over 3,000 sockeye salmon smolt were counted in a downstream migrant trap in spring 1962. Appendix Table D-2 also shows that large numbers of sockeye salmon fry (over 900,000) were released in the Lewis River in 1961, and an additional 38,000 fry were released in a tributary of Lake Merwin in 1965. Returns of sockeye salmon to the Lewis River below Lake Merwin reported in WDFW (1996) may represent a remnant of these transplants.

Dungeness River-Brannon (1996) reported that sockeye salmon have been observed coming back to the same spawning ground on the Dungeness River and that "their timing was earlier than other sockeye in Washington, suggesting they were not strays." WDFW Salmon Spawning Ground Survey Data (Egan 1977, 1995, 1997) indicated that 1-5 sockeye salmon are observed in the Dungeness River during the months of August-September in most survey years (see Appendix Table C-7).

Quillayute River/Calawah River-WDF (1973) reported that the Calawah and Bogachiel Rivers supported "a small run of sockeye salmon that must rear in the stream." Phinney and Bucknell (1975) reported that a small number of "river-race" sockeye salmon spawn in the lower reaches of the North Fork and South Fork Calawah Rivers as well as in several small tributaries. During the 1960s, 3 to 6 sockeye salmon were reportedly seen every year in the same place during July to August on the South Fork Calawah River near Hyas Creek (J. Ayerst47). Houston (1983, 1984) suggested that the Quillayute River may have a "river dwelling" population of sockeye salmon "of fewer than 10 fish per year average."

Hoh River-Wendler and Deschamps (1955) reported that small numbers of sockeye salmon are taken in the Hoh River in June and July. Houston (1983, 1984) suggested that the Hoh River may have a "river dwelling" population of sockeye salmon "of fewer than 10 fish per year average." Up to 50 sockeye salmon were observed schooling around RKm 47.5 on the Hoh River in mid-September 1985, and sockeye salmon were also observed around RKm 45.8 in 1994 and 1995 (J. Haymes48). The Hoh River receives a large glacial melt-water input and is milky in the summer, making fish identification difficult. Other indications are that a self-sustaining spawning population of sockeye salmon does not occur in the Hoh River (J. Jorgensen49). Recent tribal freshwater harvest of sockeye salmon in the Hoh River has ranged from 0 in 1991 to 26 in 1992 (Hoines 1995).

Queets River/Clearwater River-In some years a large tribal fishery catch of sockeye salmon occurred in the Queets River (Wendler and Deschamps 1955, Brix and Kolb 1971). Wendler and Deschamps (1955), citing the fact that there are few, if any, accessible lakes in the Queets River system, suggested that "sockeye salmon caught in the Queets River are probably strays from the nearby Quinault River." Wendler and Deschamps (1955) also stated that

In general, when the Quinault River has a good run of sockeyes, many are caught in the Queets. Also, the converse is true.

Dipping-in of Quinault Lake sockeye salmon into the Queets River most likely explains the bulk of the large sockeye salmon catch in the Queets River (Wendler and Deschamps 1955, S. A. Chitwood and D. Boyer, Jr.50).

Houston (1983, 1984) suggested that the Queets River may have a "river dwelling" population of sockeye salmon "of fewer than 100 fish per year average." Brown (1982, p. 30) observed mature sockeye salmon in Paradise Creek on the Queets River. WDF (1973) stated that "limited numbers of sockeye reportedly spawn in the Clearwater River" and Phinney and Bucknell (1975) stated that "sockeye salmon reportedly spawn in the mainstem of the Clearwater River and several tributary streams." The Clearwater River is a tributary of the Queets River. Analysis of over 300 Queets River sockeye salmon scales collected between 1975 and 1993 has revealed only one sea-type sockeye salmon in the Queets River fishery (QIN 1995a), however, river-type sockeye salmon (that do not outmigrate as underyearlings) cannot be differentiated by scale age from lake-type sockeye salmon (see "Life History of O. nerka" section). The Queets River receives a large glacial melt-water input from the Olympic Mountains. Both Edie (1975) and Cedarholm et al. (1978) stated that the Clearwater River, a tributary of the Queets River, had small populations of sockeye salmon.

British Columbia

A total of 917 anadromous sockeye salmon stocks have been identified in British Columbia (Slaney et al. 1996). Major sockeye salmon stocks on Vancouver Island are as follows: 1) Cheewhat Lake; 2) Hobiton River/Hobiton Lake; 3) Henderson Lake; 4) Sproat Lake and Great Central Lake in the Somass River Basin; 5) Kennedy Lake, Upper Kennedy River, Clayoquot River, Cold Creek, and Muriel Lake in the Kennedy River System; 6) Mahatta River/O'Connell Lake in Quatsino Sound; and 7) Woss Lake, Nimpkish Lake, and Vernon Lake in the Nimpkish River Basin (Aro and Shepard 1967).

Major sockeye salmon stocks in the Queen Charlotte Islands are these: 1)Mathers Lake, 2) Copper Creek/Skidegate Lake, 3) Yakoun Lake, 4) Mercer Creek/Mercer Lake, 5) Awun Lake, 6) Ian Lake, and 7) Naden River/Eden Lake. Major coastal sockeye salmon stocks in central to south mainland British Columbia include: 1) Sakinaw Lake, 2) Heydon Lake, 3) Phillips River, 4) Mackenzie Lake, 5) Klinaklini River, 6) Kakweiken River, 7) Long Lake/Smokehouse Creek, 8) Rivers Inlet/Owikeno Lake, 9) Koeye Lake, 10) Atnarko River/Tenas Lake, 11) Tankeeah Lake, 12) Kimsquit Lake, and 13) Port John Lake. Major coastal sockeye salmon stocks in north mainland British Columbia are as follows: 1) Kitlope Lake, 2) Canoona Lake, 3) Banks Lake, 4) Mikado Lake, 5) Devon Lake, 6) Lowe Lake, 7) Curtis Lake, and 8) Bonilla Lake (Aro and Shepard 1967).

The following are major sockeye salmon stocks in the Fraser River: 1) Cultus Lake, 2) Upper Pitt River/Pitt Lake, 3) Weaver Creek/Harrison Lake, 4) Harrison River Rapids (river/sea-type), 5) Birkenhead River/Lillooet Lake, 6) Seymour Creek/Shuswap Lake, 7) Scotch Creek/Shuswap Lake, 8) Lower Adams River/Shuswap Lake, 9) Lower Shuswap River/Shuswap Lake, 10) Gates Creek, 11) Raft River, 12) Fennel Creek, 13) Chilko Lake, 14) Taseko River, 15) Horsefly River, 16) Mitchell River, 17) Nadina River/FranÁois Lake, 18) Stellako River/Fraser Lake, 19) early Stuart (= Takla Lake/Trembleur Lake/Stuart Lake), 20) late Stuart = Trembleur Lake/Stuart Lake, and 21) Bowron River (Aro and Shepard 1967).

Lake-type sockeye salmon in British Columbia inhabit nursery lakes that can be categorized as either coastal or interior, and as clear, humic-stained, or glacial. Coastal lakes are thermally stratified in summer and become continuously mixed in winter following turnover (monomictic); they experience cool, wet winters and warm, dry summers on the south coast and wetter, colder summers on the north coast. Interior lakes have episodes of mixing, before and after ice formation, and become thermally stratified in both summer and winter (dimictic). However, these lakes experience a more typically continental climate on the leeward side of the Coastal Mountains.

Coastal lakes experience peak flow and nutrient input in winter, when sunlight and temperatures are low, a pattern leading to low nutrient concentrations and low productivity. These lakes are generally classified as oligotrophic. Interior lakes, such as those upstream from Hell's Gate on the Fraser River, experience maximum water and nutrient input in spring when light intensity and water temperatures are increasing. These lakes consequently have higher nutrient and productivity levels, and are classified as oligo-mesotrophic (Stockner 1987).

Many coastal lakes in central British Columbia are humic-stained and smaller than clear, larger lakes on the south mainland coast and Vancouver Island. Humic substances reduce light penetration and diminish the depth of the euphotic zone (Stockner 1987). Glacial lakes include both coastal and interior types, and their high turbidity is imparted by suspended silts and clays (glacial flour) carried down by tributaries during summer glacial-melt.

Glacial lakes produce some of the smallest 1-year-old sockeye salmon smolts recorded (Goodlad et al. 1974, Hyatt and Stockner 1985). Generally sockeye salmon smolts leaving interior lakes of British Columbia as yearlings are larger than similar age smolts from coastal systems. This may be due to the greater rearing area of interior lakes or to the higher productivity of interior lakes (Foerster 1968, Hyatt and Stockner 1985).



On to Next Section

Table of Contents