Lower Columbia River (to the Cascade Crest)--The Columbia River is the third largest river system in the United States. The Columbia River exerts a dominant influence on the biota of the Pacific Northwest, although smaller, regional, distinctions exist within the basin. In the lower Columbia River, the Cowlitz, Kalama, Lewis, White Salmon, and Klickitat Rivers are the major river systems on the Washington State side, while the Willamette and Sandy Rivers are foremost on the Oregon State side. Spring-run chinook salmon, which spawn above the Willamette Falls, will be discussed separately because of their geographic and life-history distinctiveness. The Clackamas River is the major tributary to the Willamette River below the Willamette Falls and is included in the discussion of this region.
The fall run is predominant in this region. Fall-run fish return to the river in mid-August and spawn within a few weeks (WDF et al. 1993, Kostow 1995). These fall-run chinook salmon are often called "tules" and are distinguished by their dark skin coloration and advanced state of maturation at the time of freshwater entry. Tule fall-run chinook salmon populations may have historically spawned from the mouth of the Columbia River to the Klickitat River (RKm 290). Whatever spawning grounds were accessible to fall-run chinook salmon on the Klickitat River (below Lyle Falls at RKm 3) would have been inundated following the construction of Bonneville Dam (RKm 243) in 1938 (Bryant 1949, Hymer et al. 1992a, WDF et al. 1993). There is no record of fall chinook salmon utilizing this lower portion of the Klickitat River (Fulton 1968). A significant fall run once existed on the Hood River (RKm 272) prior to the construction of Powerdale Dam (1929) and other diversion and irrigation dams (Fulton 1968); however, this run has become severely depleted and may have been extirpated (Howell et al. 1985, Nehlsen et al. 1991, Theis and Melcher 1995). The Big White Salmon River (RKm 270) supported runs of chinook salmon prior to the construction of Condit Dam (RKm 4) in 1913 (Fulton 1968). Although some fall-run salmon spawning occurs below Condit Dam, there have been substantial introductions of non-native stocks (WDF et al. 1993), and the persistence of a discrete native stock is unlikely. Fall-run fish from the Big White Salmon River were used to establish the nearby Spring Creek National Fish Hatchery (NFH) in 1901 (Hymer et al. 1992a). Spring Creek NFH is one component of the extensive hatchery system in Washington and Oregon producing fall chinook salmon (Howell et al. 1985). "Tule fall-run" chinook salmon begin the freshwater phase of their return migration in late August and October and the peak spawning interval does not occur until November (WDF et al. 1993).
Among other fall-run populations, a later returning component of the fall chinook salmon run exists in the Lewis and Sandy Rivers (WDF et al. 1993, Kostow 1995, Marshall et al. 1995). Because of the longer time interval between freshwater entry and spawning, Lewis and Sandy River fall chinook salmon are less mature at freshwater entry than tule fall chinook salmon and are commonly termed lower river "brights" (Marshall et al. 1995).
The Cowlitz, Kalama, Lewis, Clackamas, and Sandy Rivers presently contain both spring and fall runs, while the Big White Salmon River historically contained both spring and fall runs but presently only contains fall-run fish (Fulton 1968, WDF et al. 1993). The Klickitat River probably contained only spring-run chinook salmon due to falls that blocked access to fall-run chinook salmon during autumn low flows (Fulton 1968). The spring run on the Big White Salmon River was extirpated following construction of Condit Dam (Fulton 1968), while a variety of factors may have caused the decline and extinction of spring-run chinook salmon on the Hood River (Nehlsen et al. 1991, Kostow 1995).
Spring-run chinook salmon on the lower Columbia River, like those from coastal stocks, enter freshwater in March and April well in advance of spawning in August and September. Historically, fish migrations were synchronized with periods of high rainfall or snowmelt to provide access to upper reaches of most tributaries where fish would hold until spawning (Fulton 1968, Olsen et al. 1992, WDF et al. 1993). Dams have reduced or eliminated access to upriver spawning areas on the Cowlitz, Lewis, Clackamas, Sandy, and Big White Salmon Rivers. A distinct winter-spawning run may have existed on the Sandy River (Mattson 1955) but is believed to have been extirpated (Kostow 1995).
Hatchery programs are widespread throughout the region, and most populations, with the possible exception of fall chinook salmon on the Lewis and Sandy Rivers, are maintained to a significant extent via artificial propagation (Howell et al. 1985, WDF et al. 1993, Kostow 1995). The life-history characteristics of spring- and fall-run populations in many rivers have probably been influenced, to varying degrees, by transfers of non-indigenous stocks. This is especially true of the stream-type chinook salmon spring-run established in the Wind River at the Carson NFH and of upriver bright fall-run chinook salmon transferred into various systems.
The majority of fall-run chinook salmon emigrate to the marine environment as subyearlings (Reimers and Loeffel 1967, Howell et al. 1985, Hymer et al. 1992a, Olsen et al. 1992, WDF et al. 1993). A portion of returning adults whose scales indicate a yearling smolt migration may be the result of extended hatchery-rearing programs rather than of natural, volitional yearling emigration. It is also possible that modifications in the river environment may have altered the duration of freshwater residence. The natural timing of spring-run chinook salmon emigration is similarly obscured by hatchery releases of spring-run chinook salmon juveniles late in their first autumn or early in their second spring. Age analysis based on scales from naturally spawning spring-run adults from the Kalama and Lewis Rivers indicated a significant contribution to escapement by fish that entered saltwater as subyearlings (Hymer et al. 1992a). This subyearling smoltification pattern may also be indicative of life-history patterns for the Cowlitz River spring run, because both the Kalama and Lewis Rivers have received considerable numbers of transplanted fish from the Cowlitz River. Life-history data from the Clackamas and Sandy Rivers is very limited, and transplantation records indicated that these rivers have received overwhelmingly large numbers of upper Willamette River spring-run chinook salmon (Nicholas 1995). In 1898, eggs from returning spring-run chinook salmon were collected from the Clackamas River (near Clear Creek) from 15 September to 24 October, and from the upper Clackamas River from 17 July to 26 August (Ravenel 1899). The upper Clackamas River spring-run chinook salmon spawning peak has apparently shifted from mid-August (1899) to the present day peak interval from late September to early October (Nicholas 1995, Willis et al. 1995). This later spawning peak is more consistent with upper Willamette River stocks (Nicholas 1995, Willis et al. 1995). Smoltification patterns for fish from the upper Willamette River are discussed in a later section.
Comparisons of historical data on the age structure of fish returning to the Columbia River are also informative in analyzing natural smoltification traits without the impact of large hatchery programs. Analysis of scales from returning adult chinook salmon sampled in the lower Columbia River and at Bonneville Dam indicate that the proportion of yearling migrants contributing to escapement was much lower for spring-run fish in the 1920s than at present (Fig. 15) (Rich 1925; Young and Robinson 1974; Fryer and Schwartzberg 1991a, 1991b, 1992, 1993, 1994; Fryer et al. 1992). This decrease over time in the proportion of subyearling smolts may be due to increased hatchery releases of yearling smolts, increased use of stream-type spring-run stocks in hatcheries, decline in Columbia River summer-run populations, or the decreased survival/abundance of naturally-reared subyearling smolts related to changing freshwater habitat or smolt passage problems.
Adults return to tributaries in the lower Columbia River at 3 and 4 years of age for fall-run fish and 4 to 5 years of age for spring-run fish. This may be related to the predominance of yearling smolts among spring-run stocks. Marine CWT recoveries for lower Columbia River stocks tend to occur off the British Columbia and Washington coasts, with a small proportion of tags recovered from Alaska.
Upper Willamette River--Willamette Falls (RKm 42) has historically limited access to the upper river and thus defines the boundary of a distinct geographic region. High flows over the falls provided a window for returning chinook salmon in the spring, while low flows prevented fish from ascending the falls in the autumn (Howell et al. 1985). The predominant tributaries to the Willamette River that historically supported spring-run chinook salmon--the Molalla (Rkm 58), Santiam (RKm 174), McKenzie (RKm 282), and Middle Fork Willamette Rivers (RKm 301)--all of which drain the Cascades to the east (Mattson 1948, Nicholas 1995). Since the Willamette Valley was not glaciated during the last epoch (McPhail and Lindsey 1970), the reproductive isolation provided by the falls probably has been uninterrupted for a considerable time period. This isolation has provided the potential for significant local adaptation relative to other Columbia River populations.
Three major populations of spring-run chinook salmon are presently located above Willamette Falls (McKenzie River, and North and South Forks of the Santiam River) (Kostow 1995). Within-basin transfers associated with increased artificial propagation efforts since the turn of the century have reduced the genetic diversity between upper Willamette River stocks (Kostow 1995, Nicholas 1995). Fall-run chinook salmon are present in the upper Willamette River, but these fish are the result of transplants subsequent to the construction of fish passage facilities in 1971 and 1975 (Bennett 1988). Adult spring-run chinook salmon enter the Columbia River in March and April, but they do not ascend the Willamette Falls until May or June. The migration past the falls generally coincides with a rise in river temperatures above 10C (Mattson 1948, Howell et al. 1985, Nicholas 1995). Spawning generally begins in late August and continues into early October, with spawning peaks in September (Mattson 1948, Nicholas 1995, Willis et al. 1995). Recent analysis of scales from returning adults indicated that the majority of fish had emigrated to saltwater as yearlings, but this is certainly biased by the overwhelming hatchery contribution to escapement (90+%) and the hatchery strategy of releasing fish late in their first autumn or in their second spring (Nicholas 1995, Willis et al. 1995). Scales sampled from returning adults in 1941 indicated that the fish had entered saltwater during the autumn of their first year (Craig and Townsend 1946). Mattson (1963) found that returning adults which had emigrated as "fingerling" (subyearling) smolts made up a significant proportion of the 3-year-old age class, with fingerling emigrants making up a smaller proportion of the older age classes. A recent study indicated that Willamette River spring-run chinook salmon have a physiological smoltification window during their first autumn (Beckman6). Large numbers of fry and fingerlings have been observed migrating downriver from the Willamette River and its tributaries (Craig and Townsend 1946, Mattson 1962, Howell et al. 1988). Based on the examination of scale patterns from returning adults, it would appear that these fry do not immediately enter the estuary or do not survive the emigration. Emigrating fry would have been severely affected by the high water temperatures and industrial waste discharges that were common throughout much of this century in the lower Willamette River, especially during periods of low river flow in the late spring and early summer (Craig and Townsend 1946, Mattson 1962, USGS 1993). More recently, fry migrants constitute a relatively small proportion of the smolt emigration (especially when compared to the artificially propagated fingerling and yearling contribution); thus their potential contribution to returning adults would be expected to be quite low. Alternatively, these fry migrants could be rearing in the Columbia River prior to emigrating to the marine environment (Craig and Townsend 1946, Mattson 1962).
In general, Willamette River spring-run chinook salmon mature in their fourth and fifth year of life, with the majority maturing at age 4. Historically, 5-year-old fish comprised the dominant portion of the run (Nicholas 1995, Willis et al. 1995). Marine recoveries of CWT-marked fish occur off the British Columbia and Alaska coasts, and a much larger component (>30%) of the recoveries is from Alaska relative to other lower Columbia River stocks. Age of release (subyearling vs. yearling) does not appear to influence the general oceanic distribution of fish. Morphologically, Willamette River spring-run fish are similar to other lower Columbia River chinook salmon (Schreck et al. 1986). Vertebral counts for several Willamette River "wild" and hatchery samples average 68.3-69.5, which is similar to other ocean-type chinook salmon from the Columbia River, but it is significantly less than vertebral counts for upper Columbia River stream-type spring- and summer-run chinook salmon, 71.3-72.5 (Schreck et al. 1986). These vertebral counts suggest that past transplants of Carson NFH spring-run chinook salmon (a stream-type stock) did not have a significant genetic impact on Willamette River stocks. Although Willamette River spring-run chinook salmon can generally be categorized as Columbia River ocean-type chinook salmon, they do exhibit some distinct life-history attributes relative to other stocks in this general group.
Water diversions, dam placements, and river channelizations may have altered the abundance, spawning and rearing distribution, and smolt timing of populations of spring-run chinook salmon from historical levels. Although the Willamette River was once highly braided with numerous side channels offering ideal rearing habitat for juvenile salmonids (Kostow 1995), approximately 75% of that river shoreline has been lost (Sedell and Froggatt 1984). Irrigation withdrawals began in the 1800s; additionally, timber harvest activities and the construction of splash dams had a severe impact on spawning and rearing habitat access and quality (Kaczynski and Palmisano 1993). Water diversion and hydroelectric dam construction in the 1950s and 1960s limited access to significant portions of the major spring-run chinook salmon bearing tributaries to the Willamette River. In all, water storage projects eliminated access to 707 stream kilometers (Cramer et al. 1996). In addition to loss of habitat, the dams have altered the natural thermal regime. The premature emergence of spring-run chinook salmon fry due to releases of warmer reservoir water in the autumn may have caused high mortalities among naturally spawning fish (Kostow 1995). Furthermore, cooler than normal waters released in the spring limit the growth of naturally rearing fish. Habitat changes may have created selective pressures that would alter the expression of historical life-history traits, primarily impacting naturally spawning and rearing salmonids.
Despite the homogenization of spring-run chinook salmon stocks through intrabasin transfers and the impact of large scale artificial propagation efforts, the distinctiveness of Willamette River spring-run chinook salmon life-history traits relative to other ocean-type populations appears to have been retained.
Columbia River (east of the Cascade Crest)--East of the Cascade Crest, many river systems support populations of both ocean- and stream-type chinook salmon. Fall-run (ocean-type) fish return to spawn in the mainstem Columbia and Snake Rivers and their tributaries, primarily the Deschutes and Yakima Rivers (Hymer et al. 1992b, Olsen 1992). Historically, numerous other Columbia River tributaries in Washington, Oregon, and Idaho supported fall runs, but for a variety of reasons these are now extinct (Fulton 1968, Nehlsen et al. 1991, Hymer et al. 1992a, Olsen et al. 1992, WDF et al. 1993). Fall-run salmon historically migrated as far as Kettle Falls (RKm 1,090) on the Columbia River prior to the completion of Grand Coulee Dam (RKm 961) in 1941 (Mullan 1987). Chapman (1943) observed chinook salmon spawning in deep water just below Kettle Falls in October 1938. Similarly, fall-run chinook salmon migrated up the Snake River to Shoshone Falls (RKm 976), although Augur Falls (RKm 960) probably blocked the passage of most fish (Evermann 1896, Fulton 1968).
Summer-run chinook salmon populations on the Columbia River exhibit an ocean-type life history, while summer-run fish on the Snake River exhibit a stream-type life history (Taylor 1990a, Chapman et al. 1991, Chapman et al. 1994, Matthews and Waples 1991, Waknitz et al. 1995). Summer-run fish return to freshwater in June through mid-August--slightly earlier than the fall-run fish, which return from mid-August through October (Fulton 1968). Summer-run fish were able to ascend Kettle Falls (Evermann 1896, Bryant and Parkhurst 1950) and probably migrated as far as Lake Windermere in British Columbia (Hymer et al. 1992b, Chapman et al. 1994). With the completion of the Grand Coulee Dam in 1941 (RKm 961) and Chief Joseph Dam in 1955 (RKm 877), the farthest that summer-run chinook salmon can migrate upriver is the Okanogan River (RKm 859). Currently, naturally spawning ocean-type summer-run chinook salmon are also found in the Wenatchee (RKm 753) and Methow Rivers (RKm 843) (Waknitz et al. 1995). Summer-run chinook salmon are also reported to spawn in the lower Entiat and Chelan Rivers, in addition to below mainstem Columbia River dams (Marshall et al. 1995); however, it has not been determined whether or not these are self-staining populations.
There are numerous differences between ocean-type fish east and west of the Cascade Crest. Celilo Falls (RKm 320), which was submerged under Lake Celilo following the building of the Dalles Dam (RKm 309) in 1957, was located where the Cascade Crest line intersects the Columbia River and may have historically been a barrier to returning tule (lower river) fall-run chinook salmon. The Cascade Crest also marks the boundary between the maritime ecoregions to the west and the arid ecoregions to the east. Historically, summer-run and "upriver bright" fall-run fish in the Columbia River were not found below this demarcation (Fulton 1968). "Upriver brights" are so named because they enter freshwater prior to the expression of secondary maturation characteristics (darkening of skin and formation of the kype) and 1 to 3 months prior to actual spawning (WDF et al. 1993, Marshall et al. 1995). Among ocean-type Columbia River populations above Celilo Falls, summer-run chinook salmon spawn in the mid- and lower reaches of tributaries with peak spawning occurring in October, whereas fall-run chinook salmon spawn in the mainstem Columbia and Snake Rivers and the lower reaches of the Deschutes and Yakima Rivers with peak spawning occurring in November (Howell et al. 1985, Marshall et al. 1995, Mullan 1987, Garcia et al. 1996). Additionally, fall-run chinook salmon in the mainstem Columbia and Snake Rivers have been observed spawning in water 10 m deep or more (Chapman 1943, Bruner 1951, Swan et al. 1988, Hymer et al. 1992b, Dauble et al. 1995).
Ocean-type fry west of the Cascade Crest emerge in April and May, and the majority rear from 1 to 4 months in freshwater prior to emigrating to the ocean (Mullan 1987, Olsen et al. 1992, Hymer et al. 1992a, WDF et al. 1993, Chapman et al. 1994, Marshall et al. 1995). A small proportion of summer- and fall-run fish remain in freshwater until their second spring and emigrate as yearlings (Chapman et al. 1994, Waknitz et al. 1995). The proportion of yearling outmigrants varies from year to year due, perhaps, to environmental fluctuations. Among summer-run populations, the lowest incidence of yearling outmigrants is found in the Okanogan River, where the waters are relatively warm and highly productive (Chapman et al. 1994).
The age of maturation for ocean-type chinook salmon varies considerably among rivers in this region. Naturally spawning summer-run fish in the Wenatchee, Methow, and Okanogan Rivers mature primarily in their fourth or fifth year (Chapman et al. 1994, Waknitz et al. 1995, Marshall et al. 1995). The age distribution for fall-run chinook salmon returning to the Hanford Reach section of the Columbia River (RKm 292) and the lower Yakima River (below Prosser Dam RKm 75.8) includes higher proportions of 2-year-old "jacks" and 3-year-old adults relative to summer-run fish (Hymer et al. 1992b, WDFW 1995). However, the Hanford Reach and lower Yakima River populations contain higher proportions of 4- and 5-year-old spawners than other fall-run stocks (the Deschutes River and the Marion Drain) found above the Cascade Crest (Hymer et al. 1992b, WDFW et al. 1995). The Deschutes River and Marion Drain systems support fall-runs with very high incidences of 2-year-old "jack" chinook salmon (Hymer et al. 1992b, ODFW 1995, WDFW 1995). A significant proportion of the Snake River fall run is presently reared at the Lyons Ferry Hatchery and limited information is available on naturally spawning fish. The age distribution for fish returning to Lyons Ferry includes a large proportion (20%) of 2-year-old jacks relative to other stocks, although the majority return as 4- and 5-year olds (Hymer et al. 1992b, Marshall et al. 1995). The high incidence of jacks may be related to the release of yearling smolts, which constitute approximately one-half of all releases (Howell et al. 1985, Chapman et al. 1991); however, size distributions for Snake River fall-run fish intercepted at Little Goose Dam (RKm 113) in 1976 (NMFS 1996a) and at Salmon Falls (RKm 922) in 1894 (Evermann 1896) were very similar (Fig. 16) and included a large number of smaller jacks.
Ocean recoveries of CWTs describe two basic patterns. Fall-run fish from the lower Yakima River and summer- and fall-run fish from the mainstem Columbia River and its tributaries (above the confluence of the Yakima and Columbia Rivers) are recovered primarily in Alaska and British Columbia coastal waters. In contrast, a significant number of tagged fall-run chinook salmon from the Snake and Deschutes Rivers are recovered in southerly waters off the Oregon and California Coast, and recovery of CWT-marked Snake and Deschutes River fall-run chinook salmon off Alaska is not large (Howell et al. 1985, Waples et al. 1991b). Thus, among ocean-type populations east of the Cascade Crest, there appears to be some degree of divergence in maturation rates and migration.
Anthropogenic influences have had a great impact on the life history and distribution of ocean-type chinook salmon in the Columbia River Basin. Access to spawning habitat on the mainstem Snake River was blocked to migrating salmonids beginning in 1910 with Swan Falls Dam (RKm 734) and most recently by the Hells Canyon Dam (RKm 459) in 1967 (Fulton 1968, Waples et al. 1991b). An additional four mainstem dams (Ice Harbor Dam [1961; RKm 16], Lower Monumental Dam [1969; RKm 67], Little Goose Dam [1970; RKm 113], and Lower Granite Dam [1975; RKm 173]) on the Snake River have inundated spawning areas and impeded adult and smolt migrations (Fulton 1968, Chapman et al. 1991, Waples et al. 1991b). Nine dams exist on that portion of the mainstem Columbia River that is still accessible to migrating salmon, and numerous historical spawning sites were probably inundated by reservoirs created by those dams upriver from the present Dalles Dam (Smith 1966, Waknitz et al. 1995).
The construction of Grand Coulee Dam and the concurrent Grand Coulee Fish Maintenance Project (GCFMP) also influenced the present distribution of summer/fall-run chinook salmon. To compensate for the loss of spawning habitat above the dam, spring- and summer-run chinook salmon were intercepted at Rock Island Dam (RKm 730) from 1939-43 and either transported to surrogate spawning sites or held in hatchery facilities for artificial propagation (Fish and Hanavan 1948). Returning summer-run adults were transported to enclosed sections of the Wenatchee or Entiat Rivers to spawn naturally (Fish and Hanavan 1948). Captive spawning began in 1940 at the Leavenworth NFH on Icicle Creek and subsequently at other facilities on the Entiat and Methow Rivers. Artificially propagated fry and fingerlings were planted in the Wenatchee, Entiat, and Methow Rivers during the GCFMP, but neither adults nor juveniles were introduced into the Okanogan River. The reintroduction of summer-run fish into the Okanogan River resulted from later transplantations or recolonization by straying fish after the termination of trapping activities at Rock Island Dam in late 1943 (Waknitz et al. 1995). Prior to the GCFMP, Craig and Suomela (1941) reported that summer-run chinook salmon above Rock Island Dam were found in fairly low numbers in the Wenatchee and Okanogan Rivers. Emigrating young-of-year chinook salmon trapped in the Methow River in 1937 (WDF 1938) may have been ocean-type summer-run juveniles migrating to the ocean or stream-type spring-run juveniles moving to winter feeding ground downstream. Given the small numbers of returning adults reported by WDF (1938) and Craig and Suomela (1941) native fish populations were probably swamped by later releases. Another consequence of the GCFMP was the potential mixing of spring-run (stream-type) and summer/fall-run (ocean-type) fish. Runs were discriminated based on a 9 July cut-off date at the Rock Island Dam trap, and no distinction was made between later returns of summer- and fall-run fish (Fish and Hanavan 1948).
Historically, a substantial population of summer-run chinook salmon once existed on the Yakima River; however, the last summer-run redd was observed in 1970 and this stock appears to be extirpated (BPA et al. 1996). A summer run may also have existed on the Deschutes River. Recoveries of returning adults tagged at Bonneville Dam in June and July (a migration timing that is generally associated with summer runs) were made in the Deschutes and Metolius (a tributary to the upper Deschutes River) Rivers (Galbreath 1966). Jonasson and Lindsay (1988) speculated that a distinct summer run spawned in the upper Deschutes River prior to the construction of Pelton Dam (RKm 166) in 1958 and Round Butte Dam (RKm 177) in 1964, and that subsequently the run was eliminated or assimilated into the fall-run. Presently, fall-run chinook salmon on the Deschutes River return much earlier than any other fall-run stock on the Columbia River (Olsen et al. 1992), suggesting that some assimilation may have taken place.
Fall-run chinook salmon populations have been extirpated in the John Day, Umatilla, and Walla Walla Rivers (Kostow 1995). Information on the historical life-history traits for these rivers is limited. Rich (1920b) remarked that Umatilla River fall chinook salmon were unusually small, with average weights of 4.5-5.5 kg compared to 9.0 kg for the fall run in the Columbia River. Deschutes River fall-run chinook salmon are similarly described as having a small size for their age (Kostow 1995) which suggests some degree of relatedness with the extirpated Umatilla River fish.
The expression of fall-run life-history strategies in the Yakima River are potentially biased by changes in spawning and rearing habitat and introductions of non-native populations. The development of agricultural irrigation projects on the Yakima River during the last century has resulted in lower river flows, higher water temperatures, river eutrophication, and limited or impeded migration access (Davidson 1953, BPA et al. 1996). Several million "upriver brights" and smaller numbers of lower Columbia River fall-run hatchery chinook salmon have been released into the Yakima River (Howell et al. 1985, Hymer et al 1992b). The "upriver brights" stocks represent a composite of Columbia and Snake River populations and were generally founded by random samples of fall-run chinook salmon intercepted at a number of mainstem dams (Howell et al. 1985). The majority of these introductions on the Yakima River have occurred below Prosser Dam (RKm 76) and may be responsible for genetic and life-history differences between Marion Drain and lower Yakima River fall-run fish (Marshall et al. 1995). Water temperatures in the Yakima River have increased significantly, such that returning fall-run adults must delay river entry, and juveniles must emigrate from the river sooner than occurred historically (Watson7). Conditions above Prosser Dam are such that only in the Marion Drain (RKm 134), a 27-km long irrigation return water canal which is supplied with more thermally stable groundwater, is it possible for fall-run chinook salmon to naturally produce smolts in any number (BPA et al. 1996, Watson see footnote 7). It has been speculated that the Marion Drain fish are representative of "native" Yakima River fish (Marshall et al. 1995); if this is the case, then the phenotypic expression of their life-history traits (spawn timing, age at smoltification, age at maturation, size at maturation) may have been altered by the artificial environment in which they currently exist. For example, warmer winter temperatures and high stream productivity contribute to the production of large, 95 mm, outmigrating subyearling smolts in late April (Watson see footnote 7) which, in turn, result in the high incidence of 2-year-old mature males observed. The persistence of life-history differences among some populations of ocean-type chinook salmon in the Columbia River Basin, despite extensive stock transfers and geographic constriction of available habitat, is indicative of the significance of these traits.
Columbia River Stream Type--Stream-type chinook salmon in the Columbia River are represented by spring-run fish from the Klickitat River upriver to the accessible tributaries of the Columbia and Snake Rivers and summer-run fish in the Snake River Basin. With the exception of the Klickitat River, all of these rivers are located upriver from the historical location of Celilo Falls, near the present Dalles Dam.
In the Columbia Basin, the Klickitat, Deschutes, John Day, Yakima, Wenatchee, Entiat, and Methow Rivers contain "native" stream-type chinook salmon. Marshall et al. (1995) reported that the spring run on the Klickitat River has some genetic and life-history similarities to lower Columbia River (ocean-type) spring-runs. However, this run exhibits classical stream-type characteristics--yearling smolt migration and limited recoveries of CWTs from coastal fisheries (Howell et al. 1985, Hymer et al. 1992b, WDF et al. 1993). Scale samples taken from Klickitat River spring-run fish early in the 1900s (prior to extensive artificial propagation efforts) indicated a 1-year freshwater residence prior to emigration to the ocean (Rich 1920b). Transplants of Cowlitz and Willamette River spring-run chinook salmon to the Klickitat River (Howell et al. 1985) may be responsible for the few ocean recoveries of CWT-marked fish released from the Klickitat River Hatchery. Finally, vertebral counts from Klickitat River spring-run fish (average 71.3) clustered with stream-type (71-73 vertebrae) and not ocean-type populations (66-69 vertebrae) (Schreck et al. 1986).
Tributaries to the Snake River that contain "native" stream-type populations include the Tucannon, Grande Ronde, Imnaha, and Salmon Rivers. A stream-type run in Asotin Creek existed until recently, but may now be extinct (WDFW 1997a). In a previous status review, stream-type chinook salmon in the Clearwater River system were determined to have been introduced from a number of Snake River and Columbia River sources (see Appendix D) and were not considered for listing under the ESA (Matthews and Waples 1991). Stream-type fish in the Columbia River and Snake River Basins spawn across a large geographic area that encompasses several diverse ecosystems.
Stream-type fish remain in freshwater throughout their first year and sometimes second year following emergence (Healey 1991). Typically, stream-type chinook salmon undertake extensive offshore ocean migrations; therefore, few CWT-marked fish from stream-type stocks are recovered in coastal or high seas fisheries (Healey 1983, Howell et al. 1985, Olsen et al. 1992, Hymer et al. 1992b). Spring runs enter the Columbia River from March through mid-May, and summer runs from mid-May to mid-July (Galbreath 1966). Fish passing over Bonneville Dam (RKm 235) prior to 1 June are designated by the U.S. Army Corps of Engineers (USACE) as belonging to the spring-run, although there is considerable overlap (Galbreath 1966). The majority of stream-type fish mature at 4 years of age, with the exception of fish returning to the American and upper Salmon Rivers, which return predominantly as 5-year-olds. Fish ascend to the upper reaches of most river systems, and in some cases access to these areas is only possible during the high spring river flows from snowmelt and spring storms. The return migration and spawning timing for summer-run (stream-type) fish on the Snake River is somewhat later than, and in somewhat lower reaches than used by the spring runs, although this distinction is apparently not always clear (Chapman et al. 1991). The use of smaller tributaries for spawning and extended juvenile rearing by stream-type chinook salmon increases the potential for adaptation to local ecosystems through natural selection relative to ocean-type populations (which spawn in mainstem areas and migrate more quickly to the marine environment).
An important adaptation by stream-type chinook salmon in the Columbia and Snake River Basins is the early maturation of resident males (Gebhards 1960, Burck 1967, Mullan et al. 1992, Sankovich and Keefe 1996). These resident males may play a crucial role during years with low numbers of returning adults by ensuring returning females spawn successfully. The expression of this life-history trait may vary depending on the location and physical characteristics of each river, but the fact that all stream-type populations appear to express this trait is indicative of its importance. Additionally, stream-type females produce much smaller eggs, generally less than 8 mm in diameter, than Columbia River or coastal ocean-type females. Reductions in egg size are compensated for by increases in total egg number; however, perhaps due to the energetic costs of their extensive migrations and/or their prolonged residence in freshwater prior to spawning, the percentage of body weight devoted to gonads appears to be less in stream-type stocks than in coastal ocean-type stocks (Lister 1990, Bartlett 1995). Producing a greater number of smaller eggs may be an appropriate strategy to maximize long-term survival in response to the environmental fluctuations of high-altitude spawning habitats. Furthermore, large eggs may not be as important to stream-type fish, which smolt as yearlings.
Comparisons of chinook salmon populations in the Columbia River Basin indicated some morphological differences between life-history types (Schreck et al. 1986). Samples showed stream-type populations averaged 71.2-72.5 vertebrae, significantly more than the typical ocean-type population with 65.9-69.45 vertebrae, except for "fall-run" fish taken from the lower Yakima River (70.6 vertebrae). Electrophoretic analysis of these fish by Schreck et al. (1986) placed the lower Yakima River fall-run with Snake River stream-type populations, in contrast to subsequent studies by other researchers. When the lower Yakima River sample is excluded, there is a clear distinction in the average vertebral counts of ocean- and stream-type populations.
Stream-type chinook salmon spawn in rivers whose headwaters are located in one of three major mountain systems: the Cascade, Blue, and Rocky Mountains. The Salmon River lies in the Northern Rockies Ecoregion and spawning areas for stream-type fish are predominantly above 1,000 m and average approximately 1,500 m. The Grande Ronde and Imnaha Rivers, tributaries to the Snake River, originate in the Blue Mountains with spawning areas at approximately 1,000 m and higher. The John Day River, a tributary to the Columbia River, has its headwaters in the Strawberry Mountains and contains spawning areas on the North, Middle, and South Forks at approximately 1,000 m. Even prior to the construction of Pelton Dam, spawning areas for spring-run chinook salmon on the Deschutes River lay below 1,000 m (Nehlsen 1995). The Klickitat, Yakima, Wenatchee, Entiat, and Methow Rivers all contain stream-type spawning areas at relatively lower elevations, 500-1,000 m. Differences in elevation and geography are correlated with differences in temperature, rainfall, and productivity, with obvious impacts on salmon development rate, growth, and carrying capacity. Schreck et al. (1986) analyzed several aspects of spawning and rearing habitat for different rivers in the Columbia River Basin. Differences were most apparent between upper (Klickitat River and upstream) and lower Columbia River tributaries. There are two geographically-defined clusters of stream-type chinook salmon rivers: relatively low elevation rivers in the Columbia River Basin and the higher elevation rivers in the Snake River Basin.
Anthropogenic activities have significantly influenced the distribution of stream-type chinook salmon. Not included in this review is the spring run on the Wind River, which is a hatchery stock founded by intercepting spring-run fish at Bonneville Dam destined for upriver tributaries (Howell et al. 1985, Hymer et al. 1992b, Marshall et al. 1995). Stream-type chinook salmon on the Methow, Entiat, and Wenatchee Rivers were influenced by GCFMP transfers of fish destined for rivers above Rock Island Dam. River surveys undertaken prior to the onset of the GCFMP indicated that spring-run (stream-type) fish historically existed in the Wenatchee, Entiat, and Methow Rivers, but the run size had diminished considerably by the 1930s, and the run on the Entiat River may have been extirpated (Craig and Suomela 1941, Mullan 1987). Returning adults intercepted at Rock Island Dam each year prior to 9 July were classified as spring run and either transferred to spawning sites on the Wenatchee or Entiat River, or to hatcheries for spawning (Fish and Hanavan 1948). Hybridizations between late-returning stream-type (spring-run) and early-returning ocean-type (summer-run) fish probably occurred under this system (Chapman et al. 1991, Waknitz et al. 1995). Alternatively, Fish and Hanavan (1948) observed that presumptive spring-run fish transferred to impounded stream sections and allowed to naturally spawn all did so within the normal spawning period recorded for spring-run chinook salmon. Given the small size of the spring-run populations that existed on these rivers prior to the GCFMP, the majority of the fish intercepted at Rock Island Dam were probably destined for rivers above Grand Coulee Dam (Fish and Hanavan 1948, Chapman et al. 1991). Subsequent increases in run-size in the Wenatchee, Entiat, and Methow Rivers following the GCFMP suggest that introduced fish became established in these rivers (Mullan 1987).
The construction of the Hermiston Power and Light (1910) and Three Mile Dams (1914) on the Umatilla River and the Lewiston Dam (1927) on the Clearwater River were largely responsible for the extirpation of native stocks of stream-type chinook salmon on those systems (Olsen et al 1992, Keifer et al. 1992). Fish from a number of sources have since been used to reestablish stream-type chinook salmon stocks on the Umatilla and Clearwater Rivers. Certain spring-run chinook salmon stocks, such as the Carson NFH stock, have been widely transferred to rivers throughout the Columbia and Snake River Basins, and their integration into many local populations is likely.
Hydroelectric dams and/or irrigation diversions affect virtually every river containing stream-type chinook salmon (although irrigation effects are less significant in much of the Snake River Basin) and have produced changes in thermal regime, loss of spawning and rearing habitat, or direct mortality by stranding or upstream and downstream passage injury (Lindsay et al. 1989, Matthews and Waples 1991). Identifying regional life-history differences among stream-type populations is complicated by stock transfers and the difficulty in separating hatchery and naturally produced fish. Culture practices and differences in water conditions, primarily temperature, may alter the observed expression of numerous life-history traits, such as body size and age of smoltification and maturation.