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

NOAA Tech Memo NMFS F/NWC-200: Snake River Chinook Salmon (cont):


Distribution and Abundance

Historically, spring and/or summer chinook salmon spawned in virtually all accessible and suitable habitat in the Snake River upstream from its confluence with the Columbia River (Evermann 1896; Fulton 1968). Evermann (1896) observed spring-run salmon spawning as far upstream as Rock Creek, a tributary of the Snake River just downstream from Auger Falls and more than 1,442 km from the sea.

Human activities have substantially reduced the amount of suitable spawning habitat in the Snake River (Fig. 1). Even prior to hydroelectric development, many small tributary habitats were lost or severely damaged by construction and operation of irrigation dams and diversions; inundation of spawning areas by impoundments; and siltation and pollution from sewage, farming, logging, and mining (Fulton 1968). More recently, the construction of hydroelectric and water storage dams without adequate provisions for adult and juvenile passage in the upper Snake River has precluded the use of all spawning areas upstream from Hells Canyon Dam.

The Snake River contains five principal subbasins that produce spring- and/or summer-run chinook salmon (CBFWA 1990) (Fig. 2). Three of the five subbasins (Clearwater, Grande Ronde, and Salmon Rivers) are large, complex systems composed of several smaller tributaries which are further composed of many small streams. For example, the Middle Fork of the Salmon River is a tributary of the Salmon River subbasin that is 171 km long and contains 28 streams that produce spring- and/or summer-run chinook salmon (Mallet 1974). In contrast, the two other principal subbasins (Tucannon and Imnaha Rivers) are small systems in which the majority of salmon production is in the main rivers themselves. In addition to the five major subbasins, three small streams (Asotin, Granite, and Sheep Creeks) that enter the Snake River between Lower Granite and Hells Canyon Dams provide small spawning and rearing areas (CBFWA 1990).

The historical size of the Snake River spring and summer chinook salmon population is difficult to estimate. Chapman (1986) provided estimates of chinook salmon abundance for the entire Columbia River during the late 1800s but did not attempt to partition the Snake River runs. For the years 1881 to 1895, Chapman estimated a combined return of 2.5 to 3.0 million adult fish for spring and summer chinook salmon runs in the Columbia River. Historically, it is estimated that the Salmon River alone produced 39 and 45% of the Columbia River spring and summer chinook salmon adults, respectively (CBFWA 1990). Fulton (1968) estimated that 44% of all Columbia River spring and summer chinook salmon entered the Salmon River. By combining the above estimates and considering other production areas in addition to the Salmon River, the total production of the Snake River was probably in excess of 1.5 million spring and summer chinook salmon for some years during the late 1800s.

By the mid-1900s, the abundance of adult spring and summer chinook salmon had greatly declined. Fulton (1968) estimated an average of 125,000 adults per year entered Snake River tributaries from 1950 through 1960. Raymond (1988) estimated that the combined annual returns averaged 100,000 wild fish from 1964 through 1968, adjusting for fish removed by the river fisheries below McNary Dam in Zones 1-6. In another analysis, the average run of Snake River fish over McNary Dam from 1954 through 1961 and over Ice Harbor Dam from 1962 through 1969 was reported to be 90,919 fish (CBFWA 1990).

Since the 1960s, counts of spring and summer chinook salmon adults have declined considerably at Snake River dams (USACE 1989). Counts at Ice Harbor Dam declined steadily from an average of 58,798 fish in 1962 through 1970 to a low of 11,855 fish in 1979. Over the next 9 years, counts gradually increased and reached a peak of 42,184 fish in 1988. In 1989 and 1990, counts dropped sharply again to 21,244 and 26,524 fish, respectively. These counts, although illustrative of population trends for all fish, are not indicative of the abundance of wild fish in the population, because adult counts at dams have been confounded by hatchery-reared fish since 1967. Unfortunately, counts at dams cannot be reliably separated into hatchery and wild components.

The annual abundance of wild fish passing the uppermost dam on the lower Snake River since 1967 can be estimated by two methods, both of which are subject to bias. The first method is to subtract the returns to all hatcheries from the count at the dam. This method is appealing in its simplicity, but it does not account for potentially large differential mortalities after dam passage. The second method entails establishing an expansion factor based on the relationship between adult counts at the uppermost dam on the lower Snake River and redd counts in index areas prior to hatchery influence (J. Williams) (2). The annual abundance at the uppermost dam can then be estimated after 1967 by multiplying the annual redd counts by the expansion factor. The weakness of this method is that there are only 6 years (1962-67) of dam counts when only wild fish were present to establish the relationship. Also, 5 of the 6 years represent relatively high dam and redd counts. This calls into question the accuracy of extrapolating to situations in which abundance is low. Even so, this method would likely provide the better estimate of the number of wild fish passing the uppermost dam because the dam and redd count indexes used would be temporally consistent.

The expansion factor (EF) is given by:

Expansion factor equation

Using this method, the estimated number of wild adult spring and summer chinook salmon passing over Lower Granite Dam averaged 9,674 fish from 1980 through 1990 with a low count of 3,343 fish in 1980 and a high count of 21,870 fish in 1988.

Redd counts in index areas provide the best indicator of trends and status of the population of wild spring and summer chinook salmon in the Snake River Basin; counts used in this review are detailed by subbasin in the Appendix 1 Table (White and Cochnauer 1989; CBFWA 1990; M. White (3); K. Peterson (4)). Redd counts are available since 1957 from all areas except the Grande Ronde River, for which enumeration began in 1964. Therefore, we provide two perspectives of the abundance of redds over time--one beginning in 1957 excluding the Grande Ronde River and the other beginning in 1964 including the Grande Ronde River. Redd counts in the Clearwater River were excluded from all analyses because the current population was derived from hatchery outplantings of nonindigenous fish (see Stock Histories section).

Trends in abundance of redds are similar for both time series (Fig. 3). Redd counts have declined sharply over the last 33 years. In 1957, over 13,000 redds were counted in index areas excluding the Grande Ronde River (Fig. 4). By 1964 and including the Grande Ronde River, the annual count in index areas was 8,542 redds. Over the next 16 years, annual counts in all areas declined steadily, reaching a minimum of 620 redds in 1980. Annual counts increased gradually over the next 8 years, reaching a peak of 3,395 redds in 1988. In 1989 and 1990, counts dropped again to 1,008 and 1,224 redds, respectively.

The abundance of wild Snake River spring and summer chinook salmon has declined more at the mouth of the Columbia River than the redd trends indicate (Chapman et al. 1991). Prior to curtailment in the mid-1970s, the in-river fisheries in the Columbia River below McNary Dam harvested 20 to 88% of these fish annually (Raymond 1988). Therefore, any analysis of population decline using redd counts provides a conservative approximation of the actual decline in abundance of adults.

In the near term, we are pessimistic concerning the expected abundance of Snake River spring and summer chinook salmon. Based upon the lowest return on record of jack (precocious male salmon that return after 1 year of ocean residence) spring and summer chinook salmon to Lower Granite Dam in 1990 (357 compared to 2,451 in 1989), we expect adult and redd counts to drop considerably over the next 2 years. There is a strong possibility that, over the next few years, we may witness record low returns of wild spring and summer chinook salmon adults to the Snake River.

Life History Characteristics


Adult chinook salmon migrating upstream past Bonneville Dam from March through May, June through July, and August through October are categorized as spring-, summer-, and fall-run fish, respectively (Burner 1951). In general, the habitats utilized for spawning and early juvenile rearing are different among the three forms (Chapman et al. 1991). In both rivers, spring chinook salmon tend to use small, higher elevation streams (headwaters), and fall chinook salmon tend to use large, lower elevation streams or main-stem areas. Summer chinook salmon are more variable in their spawning habitats; in the Snake River, they inhabit small, high elevation tributaries typical of spring chinook salmon habitat, whereas in the upper (5) Columbia River they spawn in larger, lower elevation streams more characteristic of fall chinook salmon habitat. Differences are also evident in juvenile outmigration behavior. In both rivers, spring chinook salmon migrate swiftly to sea as yearling smolts, and fall chinook salmon move seaward slowly as subyearlings. Summer chinook salmon in the Snake River resemble spring-run fish in migrating as yearlings, but migrate as subyearlings in the upper Columbia River (Schreck et al. 1986).

Gilbert (1912) first categorized the two behavioral types and referred to those juveniles that migrate seaward as subyearlings as "ocean-type" chinook and those that migrate seaward as yearlings as "stream-type" chinook. A strong tendency toward one or the other types is found within most streams, with ocean-types dominating in the southern range from California through the coastal streams of Oregon and Washington and stream- types dominating in the northern range from British Columbia (excluding Vancouver Island) through Alaska and in the Yukon River (Taylor 1989). The Columbia River is located in the middle of the range and produces chinook salmon populations with the highest diversity in juvenile migrational behavior and timing. Some tributaries or areas produce only ocean-type juveniles (main-stem areas of the Columbia and Snake Rivers), some produce only stream- type juveniles (upper tributaries of the Columbia and Snake Rivers), and some produce both types (many tributaries of the Columbia River below the confluence of the Snake River). In both the Columbia and Snake Rivers, spring- and fall-run adults produce stream-type and ocean-type juveniles, respectively; however, in the upper Columbia River, summer-run adults produce ocean-type juveniles, whereas in the Snake River, they produce stream-type juveniles.

Life history information thus clearly indicates a strong affinity between summer- and fall-run fish in the upper Columbia River, and between spring- and summer-run fish in the Snake River. Genetic data (discussed below) support the hypothesis that these affinities correspond to ancestral relationships.

The relationship between Snake River spring and summer chinook salmon is more complex. Some streams in the Snake River are considered to have only spring-run fish (e.g., those in the Grande Ronde River), some only summer-run fish (e.g., those in the Imnaha and the South Fork of the Salmon Rivers), and some both forms (e.g., many streams in the Middle Fork of the Salmon River and upper reaches of the Salmon River). These designations persist in spite of the observation that some fish returning to "spring" chinook salmon streams may not pass Bonneville Dam until early June; conversely, some fish from streams having populations recognized as "summer" chinook salmon may pass upstream in late May (at Bonneville Dam, chinook salmon are called spring-run until 1 June and summer-run until 1 August). This has led to confusing appellations such as "late spring" or "early summer" fish.

Elevation appears to be the key factor influencing run/spawn timing. In most cases, spring chinook salmon spawn earlier and at higher elevations than summer chinook salmon. This is generally true whether spring and summer runs from the same stream or different streams are compared. Where the two forms co-exist, spring-run fish spawn earlier and in the upper ends of the tributaries, whereas summer-run fish spawn later and farther downstream. Spawning fish in both groups tend to use the upstream portions of their respective spawning areas first and the downstream portions last.

An obvious connection to elevation is water temperature, with higher elevations generally characterized by lower annual temperatures. Brannon (1987) showed that spawning times for Fraser River sockeye salmon were progressively earlier as the mean temperature of the incubation period decreased. Presumably this is an adaptive behavior, because post-spawning embryo development is retarded in cooler water, requiring more incubation time.

Two hypotheses can explain the presence of both spring and summer chinook salmon in some streams. The first hypothesis is that the two forms arose from a single colonization event by one of the forms. Subsequently, a slight shift in run-timing of some individuals in the population might have allowed expansion into habitat that could not be utilized by the original colonists. The result of this expansion might be a single population, with a cline in the frequency of genes controlling run-timing associated with the cline in stream elevation and incubation temperature. Alternatively, some degree of reproductive isolation between the two forms might develop following expansion into the new area.

The second hypothesis is that spring- and summer-run fish are two independent evolutionary units, and the reason both forms are sometimes found in the same stream is that two colonization events occurred. Under this hypothesis, habitat suitable for summer- run fish is unlikely to be adequate for spring-run fish (and vice versa); therefore, such habitat can only be colonized by fish of the appropriate run-time from another area.

Both hypotheses are consistent with the idea that environmental factors are important in determining time of spawning and, therefore, time of entry into fresh water. That is, "spring" chinook salmon return early and spawn early because the streams they spawn in are colder and the eggs require longer incubation time; furthermore, adverse weather conditions may reduce the success of individuals that spawn too late in the season. In this view, "summer" fish can afford to migrate upriver and spawn later in the season because their spawning locations, being typically at somewhat lower elevation, present less exacting requirements for spawn timing and embryo development. The two hypotheses differ in their predictions regarding the evolutionary relationships between the two forms. According to the first hypothesis, spring- and summer-run fish from the same stream would be more closely related to each other than either is to fish of the same run-time from other streams, whereas the second hypothesis leads to the opposite prediction. At present, there is insufficient information to determine which of these hypotheses is true. (It is also possible that the first hypothesis is true in some cases and the second hypothesis in others.)

Other Life History Characteristics

Detailed life history data (age at spawning, sex ratios, etc.) are plentiful for hatchery populations, but limited and inconsistent for wild populations. More data are also available for some subbasins and streams than others, and different types of data are available for different streams at different times. Moreover, most of this information was gathered during the past 30 years--after man's activities may have disrupted the structure of the populations. The expression of these characteristics by populations can be influenced by short- and long-term population disturbances, stochastic processes, and various sources of sampling error (e.g., sampling only certain segments of a population). The Columbia Basin Fish and Wildlife Authority identified this type of life history data in their "critical data gaps" sections of most of their subbasin production plans (CBFWA 1990). Nevertheless, limited information for seven important life history characteristics of wild fish are available and summarized in Appendix 1 List A.

Age at spawning and associated fecundity differ between the adults returning to the Middle Fork and main Salmon Rivers and all other areas where information is available. In these two areas, 3-ocean adults (especially females) with higher fecundity predominate, whereas 2-ocean adults with lower fecundity predominate in other areas. This is in spite of the fact that spring- and summer-run chinook salmon inhabit parts of both areas. This suggests that geography or other enviromental factors are more influential in determining age at return than run-timing.

In contrast, the outmigration timing of smolts at dams strongly aligns with adult run- timing. Recent studies by NMFS have shown that, in two consecutive years, smolts from summer-run only streams (Imnaha and South Fork of the Salmon Rivers) arrived at Lower Granite Dam much earlier than smolts from spring-run only or mixed streams (Matthews et al. 1990).

No apparent patterns or relationships were found in any of the other life history characteristics examined. Additional data of this kind will be critical to more precisely define the evolutionary relationship between Snake River spring and summer chinook salmon in the future.

Phenotypic Characteristics

Schreck et al. (1986) compared 29 phenotypic characters (meristic, body shape, size of fins, etc.) of wild and hatchery stocks of spring, summer, and fall chinook salmon in the Columbia River. There were significant differences among the stocks of chinook salmon for each of the characters. Between-year variation did not account for the differences among stocks of chinook salmon. Characteristics of geographically proximal stocks tended to be similar, regardless of time of freshwater entry. Based on phenotypic and genetic characteristics, these researchers found that spring chinook salmon stocks are more similar to stocks with different run-timing that originate on the same side of the Cascade Range than to other spring chinook salmon from the other side of the range. Spring chinook salmon from west of the Cascade Range were more similar to fall chinook salmon from the same or nearby streams; spring chinook salmon from the Salmon River had stronger affinities to summer chinook salmon from the same river than to spring chinook salmon from west of the Cascade Range. Similarly, two groups of summer chinook salmon were identified. Populations in the upper Columbia River aligned with fall chinook salmon stocks of the middle and lower Columbia River, whereas populations from the Salmon River aligned with spring chinook salmon stocks in Idaho. These affinities parallel the similarities between these groups in juvenile migration behavior and timing discussed above.

Stock Histories

Stock transfers from within or introductions from outside the Snake River were unreported prior to the mid-1900s. Since then, transfers have been extensive. Here, we briefly review the history of artificial propagation and summarize the outplantings of spring and summer chinook salmon in the Snake River.

Chapman et al. (1991) listed 24 facilities in the Snake River that have produced, held, or released various life stages of spring or summer chinook salmon since the early 1960s. Currently, major hatcheries producing spring or summer chinook salmon in the Snake River include Sawtooth, McCall, Rapid River, Pahsimeroi, Dworshak, Kooskia, Lyons Ferry, and Lookingglass. Satellite facilities for brood-stock capture, juvenile rearing or conditioning, and juvenile release are associated with most of the hatcheries. Two additional hatcheries, the Clearwater and Nez Perce Tribal Hatcheries, are planned for construction in the near future.

Stocks used in most hatcheries were derived from various exotic lineages, mixtures of exotic lineages, or mixtures of exotic and native lineages (Howell et al. 1985; CBFWA 1990; Chapman et al. 1991). However, the Tucannon River stock raised at Lyons Ferry Hatchery and the Imnaha River stock raised at Lookingglass Hatchery have had minimal exotic influence. Both stocks are released from the hatcheries back into their native streams. Stocks nonindigenous to the Snake River that were released from hatcheries or outplanted into various streams in the Snake River include Carson, South Santiam, Little White Salmon, Marion, Willamette, Klickitat, Cowlitz, and Leavenworth. Stocks that originated in the Snake River but were released into nonnative streams within the Basin include Rapid River, McCall, Sawtooth, Lookingglass, Pahsimeroi, Hayden Creek, and Imnaha.

Many millions of eggs, fry, or smolts as well as many adults have been released directly from hatcheries or placed into other streams or drainages over the last 30-40 years. These outplantings are summarized in Appendix 1 List B. A brief report for each principal subbasin follows.

The Tucannon River subbasin received only two small plantings of nonnative fish: 16,000 Klickitat stock and 10,500 Willamette stock spring chinook salmon fry, in 1962 and 1964, respectively.

The native runs of chinook salmon in the Clearwater River subbasin were nearly, if not totally, eliminated by hydropower development. In 1927, Island Power and Light Company built a dam on the river near its mouth at Lewiston, Idaho. From 1927 through 1940, inadequate adult fish passage in the dam's fish ladder virtually eliminated salmon runs into the basin (CBFWA 1990). Fulton (1968) stated the dam "prevented passage" during the 14-year period, but the area above the dam was subsequently made available to salmon by improvements to the fishway in 1940. He further stated that chinook salmon returning since then were from "re-stocking." Holmes (1961) provided a detailed record of fish passage at the dam. Spring and summer chinook salmon were observed during only 3 years prior to 1950, after which counts were conducted annually. Counts of 311 and 102 spring and/or summer chinook salmon were reported in 1928 and 1929, respectively. In 1938, only two fish were counted. When counting resumed in 1950, seven chinook salmon were observed passing the dam during the time period typical for spring- or summer-run fish. Some or all of these fish could have been from either restocking or straying (Chapman et al. 1991) (see discussion below). The dam was removed in 1973. Harpster Dam on the South Fork of the Clearwater River blocked chinook salmon runs into this tributary (CBFWA 1990). Finally, the construction of Dworshak Dam on the North Fork of the Clearwater River in the early 1970s eliminated this tributary from use by anadromous salmonids.

The first efforts at re-stocking the Clearwater River occurred from 1947 through 1953, with annual introductions of 100,000 eyed eggs from the headwaters of the Middle Fork of the Salmon River. Since then, millions of salmon of mixed and pure exotic lineages were released into various areas of the subbasin. Even if a few native salmon survived the hydropower dams, the massive outplantings of nonindigenous stocks presumably substantially altered, if not eliminated, the original gene pool. One member of the NMFS Technical Committee suggested that a remnant, indigenous stock may exist in one tributary of the Lochsa River. Recent electrophoretic data for this stock were inconclusive (see Genetics section below).

Hatchery influence began relatively recently in the Grande Ronde River subbasin, with the first release of smolts from Lookingglass Hatchery into Lookingglass Creek in 1980. Since then, four other streams (Big Canyon and Catherine Creeks and the main and upper Grande Ronde River areas) have received various outplantings from the hatchery in addition to annual releases into Lookingglass Creek. Principal stocks used were Lookingglass, Carson, and Rapid River. All streams contained some native fish before the outplantings, as indicated by the presence of redds (CBFWA 1990). Redd counts, which did not increase dramatically in any of the streams after the releases, suggest that, in general, the outplantings did not lead to large increases in the populations inhabiting the streams.

The Salmon River subbasin can be divided into the South Fork of the Salmon River; the Middle Fork of the Salmon River; the main river below Stanley, Idaho; and the main river above Stanley. We treat each of these areas separately.

The South Fork of the Salmon River is a native summer-run chinook salmon stream. Hatchery influence began in 1976 with the first planting of smolts from McCall Hatchery into the main river above the adult trapping facility near Cabin Creek. Since then, hatchery fish were outplanted as smolts in this area annually and as fry or smolts in other tributaries during some recent years. The McCall Hatchery stock was originally established from adults trapped at Little Goose and Lower Granite Dams during the mid- to late 1970s. The original gene pool was likely made up of native South Fork stock, with heavy influence from other summer-run streams and, perhaps, a small infusion of spring chinook salmon genes (Chapman et al. 1991). Since 1980, only adults returning to the trapping facility were used as brood stock.

The Middle Fork of the Salmon River received a single, small outplanting of nonindigenous fish. In 1975, 22,000 spring chinook salmon fry from Rapid River Hatchery were outplanted in Capehorn Creek, a small tributary of Marsh Creek at the upper end of the Middle Fork.

The initial outplanting of nonindigenous fish in the main Salmon River system below Stanley occurred in 1966 with the first smolt release from Rapid River Hatchery into Rapid River. Rapid River stock originated from mid-Snake River stocks above Hells Canyon, including the Weiser and Powder Rivers and Eagle Creek. Since 1966, millions of fry or smolts and many adults of various lineages were outplanted into 14 tributaries or areas of the main Salmon River. Numbers and time periods of outplantings varied by stream. Many of the outplants were into streams that contained relatively healthy populations of (or at least some) native fish, as indicated by previous redd counts. Stocks outplanted include, but were not limited to, Rapid River, Sawtooth, McCall, and Pahsimeroi.

The main Salmon River above Stanley contains seven streams or areas that have received hatchery outplants since 1968. Since then, the main river itself received over 10 million spring chinook salmon outplants of either fry or smolts. As with the lower main river, many of the releases were into areas that harbored relatively healthy populations of native fish. Stocks released were primarily Rapid River and Sawtooth.

Only two small releases of nonindigenous fish have occurred in the Imnaha River subbasin. In 1966, 119 adult spring chinook salmon were transferred into the river from the adult trap at Hells Canyon Dam. In 1984, the river received 4,258 spring chinook salmon smolts from Lookingglass Hatchery.


Natural straying (fish spawning in a nonnatal stream) of anadromous salmonids appears to be minimal for most species. "Wandering" as described by Chapman et al. (1991) can occur when conditions in home streams are detrimental or inhospitable to returning adults or when adults miss their home stream and are trapped above obstacles that preclude their return.

Two recent studies examined straying rates for hatchery-reared spring or summer chinook salmon in the Columbia River drainage. Fulton and Pearson (1981) documented a straying rate of 0.5% in an extensive experiment involving 12 separate releases of spring and summer chinook salmon in the mid-Columbia River. Quinn and Fresh (1984) examined straying of four brood years (1974-77) of spring chinook salmon in the Cowlitz River. Of those recovered, only 1.4% were found outside the Cowlitz River and only 0.2% actually spawned in a nonnatal river. The other fish returned to nonnatal hatcheries, but could have been either strays or "wanderers." Furthermore, the analysis showed that straying correlated positively with age at return and negatively with the number of returning salmon. Straying may be higher in older fish and when numbers returning are few.

Chapman et al. (1991) extensively reviewed coded-wire-tag recoveries from wild spring chinook salmon streams in Washington and wild spring and summer chinook salmon streams in Idaho. Although millions of tagged hatchery fish were released from nearby hatcheries over many years, no tags were found during carcass checks on any of the wild streams (6). Moreover, tagged fish from one hatchery rarely appeared at another hatchery, except where traps prevented possible wandering adults from leaving a hatchery once they entered. The only exceptions occurred in the Grande Ronde River during 1986 and 1987. About 60% of the releases of Lookingglass/Carson stock released into the main Grande Ronde River were recovered in wild fish areas of the Minam and Wenaha Rivers. The reasons for this apparent anomaly are unknown.

Studies of straying of wild spring/summer chinook salmon have not been conducted. However, we have no reason to believe they would be any higher (and, more likely, they would be lower) than for hatchery-reared fish.


Protein electrophoresis has been effectively used to study population structure in anadromous Pacific salmon since the early 1970s, and allele frequency information for Snake River spring chinook salmon has been available for over a decade (Milner et al. 1981). A number of more recent studies (Schreck et al. 1986; Utter et al. 1989; Winans 1989; Waples et al. 1991) have considerably expanded the geographic coverage, and development of additional genetic markers has increased the sensitivity of the technique. Significant findings of these genetic studies can be summarized as follows:

  1. On a broad scale, Columbia River populations can be grouped into three clusters (Fig. 5) a) spring- and summer-run fish from the Snake River and spring-run fish from the mid- to upper-Columbia River, b) spring chinook salmon from the Willamette River, and c) fall chinook salmon. The third cluster also includes some hatchery stocks of spring chinook salmon from the lower Columbia River and some upper Columbia River summer-run fish with life history patterns similar to fall-run fish.
  2. Fall chinook salmon are distinct genetically from spring-run fish in both the Snake and upper Columbia Rivers.
  3. Summer chinook salmon in the Snake River are genetically very similar to spring chinook salmon in that river. However, summer chinook salmon in the upper Columbia River are genetically very similar to fall chinook salmon in that river.
  4. As a group, Snake River spring and summer chinook salmon are characterized by relatively low levels of genetic variation. Winans (1989) found that heterozygosity values in Snake River spring and summer chinook salmon were about half as large as those in lower river stocks of similar run-timing. It has been suggested (Utter et al. 1989; Winans 1989) that these relatively low levels of genetic variation may reflect past bottlenecks in population size; however, other explanations cannot be ruled out. A more recent study (Waples et al. 1991) using more gene loci suggests that the difference in level of genetic variability between Snake River and lower Columbia River stocks may not be as great as previously thought.
  5. As a group, Snake River spring and summer chinook salmon also have been shown to be genetically distinct from other chinook salmon populations in North America, with two exceptions. One group is spring chinook salmon from the upper Columbia River. In recent genetic studies, this group is primarily represented by samples from hatcheries using Carson stock fish. This similarity may be due to the origin of the Carson stock, which was initiated to mitigate losses to upper Columbia River populations eradicated by construction of Grand Coulee Dam. Founding brood stock was collected at Bonneville Dam (Mullan 1987) and likely included some and possibly many Snake River fish. Subsequently, Carson stock has been extensively outplanted in the Columbia and Snake River Basins (Howell et al. 1985). According to Mullan (1987), the Wenatchee, Entiat, and Methow Rivers are the last remaining drainages in the upper Columbia River Basin with "wild" runs of spring chinook salmon, and over a million smolts of Carson stock hatchery fish are released annually into each of these rivers.
Utter et al. (1989) also found an unexpectedly high level of genetic similarity between Snake River spring and summer chinook salmon and samples from the Klamath River in California. The authors speculated that the apparent similarity was largely an artifact that would disappear as more genetic data became available. This has proved to be the case. Data collected more recently (NMFS and University of California at Davis, unpublished data) indicate that substantial allele frequency differences exist between the two groups at several gene loci not examined in the earlier studies.

Although early genetic studies demonstrated that fall chinook salmon in the Snake River are distinct from spring- and summer-run fish, relatively little was known until recently about relationships between the latter two forms. The study of Utter et al. (1989) included two samples each (one hatchery and one wild) of Snake River spring and summer chinook salmon. They found nonsignificant allele frequency differences between the two spring-run samples (Valley Creek and Rapid River Hatchery), as well as between the two summer-run samples (Johnson Creek and McCall Hatchery). Modest (but statistically significant) frequency differences were found between the combined spring- and combined summer-run samples. In a more recent study using substantially more gene loci, Waples et al. (1991) found highly significant allele frequency differences for every pairwise comparison of samples from 11 spring and summer chinook salmon populations in the Snake River Basin, including the four populations examined by Utter et al. (1989). This result presumably reflects the greater sensitivity in the latter analysis provided by the increased number of genetic characters examined.

The results obtained by Waples et al. (1991) demonstrate that some population subdivisions can occur at the level of individual streams. That is, the authors were able to reject the hypothesis that all samples (or any pair) were drawn from a single, random mating population. For example, in the South Fork of the Salmon River, the frequency of the variant ("83") allele at the gene locus ADA-1 was 0.154 in the Secesh River but only 0.015 in nearby Johnson Creek (Waples et al. 1991). It is highly improbable (P < 0.001) that both samples could have been drawn from the same population; furthermore, the two samples also differed significantly at 10 other gene loci. Thus, although the mouths of Johnson Creek and the Secesh River are close to each other in the same drainage, there is genetic evidence for restricted gene flow between the two populations.

For perspective, it should be noted that it is not inevitable, even using a large number of loci, that significant genetic differences will be found between samples. For example, allele frequency differences between spring chinook salmon from Carson, Leavenworth, and Little White Salmon Hatcheries are so minor that they can be attributed to random error in drawing the samples (NMFS and WDF, unpublished data). This result presumably reflects the frequent transfers of fish or eggs between these facilities. Nonsignificant tests comparing allele frequencies over all gene loci are also commonly found in comparisons of temporally-spaced samples from the same population. Waples et al. (1991) reported such a result for two samples from Rapid River Hatchery and two from McCall Hatchery.

In their study, Waples et al. (1991) found general agreement between groupings based on genetics and run-timing. For example, the Salmon River spring chinook salmon samples (Marsh Creek, Valley Creek, upper Salmon River, and Sawtooth Hatchery) shared a relatively high degree of genetic similarity, as did summer-run samples from the South Fork Salmon River (Johnson Creek, Secesh River, and McCall Hatchery) and the Imnaha River (a wild and a hatchery sample). However, these clusters also conform largely to geographic patterns, and in some cases substantial differences were found between fish of similar run- timing from different areas (e.g., between spring-run samples from the Salmon and Grande Ronde Rivers). Thus, it cannot be determined from available data whether geography or run-timing is more important to the genetic structure of Snake River spring and summer chinook salmon. Such a determination will require analysis of samples of spring and summer chinook salmon from streams where both forms occur.

Recent (unpublished) electrophoretic data gathered by the NMFS ongoing genetic monitoring program for Snake River chinook salmon and steelhead provide some additional insight into population structuring in the Grande Ronde Basin. This area is of interest because the Lostine River, a Grande Ronde tributary, was a relative outlier in the Waples et al. (1991) study, which was based on samples collected in 1989. Preliminary data show that a 1990 sample from the Minam River (a wild population) is genetically distinct from the 1989 samples, including the Lostine River. In contrast, a 1990 sample from Catherine Creek, another Grande Ronde tributary, is genetically more similar to samples from Carson Hatchery than it is to the Minam River or any of the other 1989 NMFS samples. This latter result presumably reflects the effects of repeated releases of fish from Carson stock and elsewhere into Catherine Creek in the last decade (see Stock Histories section).

Relatively little genetic information is available for chinook salmon from the Clearwater Basin. Waples (1990) found that Kooskia Hatchery, which has received fish from a variety of stocks, is genetically closest to samples from Carson stock spring chinook salmon hatcheries. In contrast, Red River, which has been heavily supplemented with Rapid River stock, is very similar genetically to spring chinook salmon samples from Rapid River Hatchery and the upper Salmon River. In 1989, William Miller of the U.S. Fish and Wildlife Service at Dworshak Hatchery provided NMFS with 11 adult and 19 juvenile chinook salmon taken from the White Sands Creek area of the upper Lochsa River. He suggested that genetic analysis might help resolve speculation that a remnant population of spring chinook salmon persists in the stream. Results of that analysis were inconclusive (Waples 1989). The possibility that genetic characteristics of the White Sands fish differ somewhat from those of other Snake River chinook salmon could not be ruled out, but such differences could not be convincingly demonstrated given the small number of individuals available for analysis.

Table of Contents