Comparisons of life-history traits among chinook salmon populations revealed regional differences in many traits. The definition of geographic regions which contained populations with similar life-history attributes was an important step in the establishment of tentative ESU boundaries. The following discussion includes information on anthropogenic changes in habitat quality, stock transfers, and artificial propagation efforts. The impacts of these activities on genetic integrity, abundance, and other potential risks to chinook salmon populations are discussed in later sections in more detail and are included here only to the extent that these activities may have altered the expression of life-history traits in presumptive native populations.
Chinook salmon are found in most of the rivers in this region. WDF et al. (1993) recognizes 27 distinct stocks of chinook salmon: 8 spring-run, 4 summer-, and 15 summer/fall- and fall-run stocks. The existence of an additional five spring-run stocks has been disputed among different management agencies (WDF et al. 1993). The Skagit River and its tributaries--the Baker, Sauk, Suiattle, and Cascade Rivers--constitute what was historically the predominant system in Puget Sound containing naturally spawning populations (WDF et al. 1993). Spring-run chinook salmon are present in the North and South Fork Nooksack Rivers, the Skagit River Basin, the White, and the Dungeness Rivers (WDF et al. 1993). Spring-run populations in the Stillaguamish, Skokomish, Dosewallips, and Elwha Rivers are thought to be extinct (Nehlsen et al. 1991). Summer-run chinook salmon are present in the Upper Skagit and Lower Sauk Rivers in addition to the Stilliguamish and Snohomish Rivers (WDF et al. 1993). Fall-run stocks (also identified by management agencies as summer/fall runs in Puget Sound) are found throughout the region in all major river systems. The artificial propagation of fall-run stocks is widespread throughout this region. Summer/fall chinook salmon transfers between watersheds within and outside the region have been commonplace throughout this century; thus, the purity of naturally spawning stocks varies from river to river. Captive broodstock/recovery programs for spring-run chinook salmon have been undertaken on the White River (Appleby and Keown 1994), and the Dungeness River (Smith and Sele 1995b). Supplementation programs currently exist for spring-run chinook salmon on North Fork Nooksack River and summer-run chinook salmon on the Stillaguamish and Skagit Rivers (Marshall et al. 1995, Fuss and Ashbrook 1995). Hatchery programs also release Suiattle River spring-run chinook salmon and Snohomish River (Wallace River) summer-run chinook salmon (Marshall et al. 1995, Fuss and Ashbrook 1995). The potential impacts of artificial propagation and rearing programs (especially delayed-release programs) on the expression of life-history traits were taken into account when comparing the characteristics of each stock.
Adult spring-run chinook salmon in the Puget Sound typically return to freshwater in April and May (Table 1) and spawn in August and September (Fig. 10) (Orrell 1976, WDF et al. 1993). Adults migrate to the upper portions of their respective river systems and hold in pools until they mature. In contrast, summer-run fish begin their freshwater migration in June and July and spawn in September, while summer/fall-run chinook salmon begin to return in August and spawn from late September through January (WDF et al. 1993). Studies with radio-tagged fish in the Skagit River indicated that river-entry time was not an accurate predictor of spawning time or location (SCC 1995). In rivers with an overlap in spawning time, temporal runs on the same river system maintain a certain amount of reproductive isolation through geographic separation. For example, an 18-km river section (at river kilometer (RKm) 35-53) of poor spawning habitat separates the spawning areas for summer and spring runs on the Sauk River (Williams et al. 1975).
The majority of Puget Sound fish emigrate to the ocean as subyearlings. Many of the rivers have well-developed estuaries that are important rearing areas for emigrating ocean-type smolts. Puget Sound stocks also tend to have relatively large eggs, with average diameter being greater than 8.0 mm, which may be an adaptation for their subyearling smolting strategy. In contrast, the Suiattle and South Fork Nooksack Rivers have been characterized as producing a majority of yearling smolts (Fig. 11) (Marshall et al. 1995). Analysis of scales from adults returning to the South Fork Nooksack in 1994 and 1995 indicated that 69.1% of the fish had emigrated as yearlings (WDFW 1995); however, analysis of adults returning in 1980-85 showed only 16.4% of the fish had emigrated as yearlings and 75% of these were hatchery fish (WDFW, unpublished). The reason for this difference is unknown. Glacially influenced conditions on the Suiattle River may be responsible for limiting juvenile growth, delaying smolting, and producing a higher proportion of 4- and 5-year-olds compared to other chinook salmon stocks in Puget Sound, which mature predominantly as 3- and 4-year-olds (Fig. 12). Puget Sound stocks exhibit a similarity in marine distribution based on CWT recoveries in ocean fisheries. Tagged fish have been primarily captured in Canadian coastal and Puget Sound waters (Fig. 13). Marine recoveries of CWTs from Nooksack River spring-run chinook salmon have occurred to a lesser extent in the Puget Sound fishery than in other Puget Sound stocks. This may be due to the geographical position of the Nooksack River Basin at the northern end of Puget Sound and/or the allocation of effort by fishers in Puget Sound. Additionally, Elwha River summer/fall chinook salmon CWT recoveries in Alaska and Puget Sound appear to be intermediate in their frequencies between Puget Sound stocks and Washington coast stocks.
Anthropogenic activities have limited the access to historical spawning grounds and altered downstream flow and thermal conditions. Water diversion and hydroelectric dams haveprevented access to portions of several rivers. Furthermore, the construction of Cushman Dam on the North Fork Skokomish River may have resulted in a residualized population of chinook salmon in Lake Cushman. Watershed development and activities throughout Puget Sound, Hood Canal, and Strait of Juan de Fuca regions have resulted in increased sedimentation, higher water temperatures, decreased large woody debris (LWD) recruitment, decreased gravel recruitment, a reduction in river pools and spawning areas, and a loss of estuarine rearing areas (Bishop and Morgan 1996). These impacts on the spawning and rearing environment may also have had an impact on the expression of many life-history traits and masked or exaggerated the distinctiveness of many stocks.
Life-history similarities--emigration timing, age at maturation, and ocean migration--among spring-, summer-, and fall-run chinook salmon may be related to the relatively recent deglaciation (10,000 b.p.) of the Puget Sound region. It is unclear when suitable freshwater habitats for chinook salmon became available in the Puget Sound area following deglaciation (Busack and Marshall 1995). However, chinook salmon in Oregon coastal rivers, which were not glaciated, also show little differentiation in life-history characteristics, except for run timing. The life history exhibited may instead represent an optimized strategy for stocks in the Puget Sound area regardless of run timing or simply the homogenization of stocks due to artificial propagation.
Fall-, summer-, and spring-run chinook salmon are found in this region. Rivers in this region tend to be short with low gradients near the coast. These low gradient areas are preferred spawning sites for chinook salmon. The relatively small size of the rivers limits the amount of spawning habitat available and minimizes the likelihood of spatial separation of run times. The Chehalis and Umpqua Rivers are physically much larger than any of the other basins, although they do not maintain proportionately larger chinook salmon runs. WDF et al. (1993) recognized 2 spring-run, 4 summer-run, 4 spring/summer-run, and 23 fall-run "stocks" on the Washington coast. According to the Oregon Department of Fish and Wildlife (ODFW), the Oregon coast from the mouth of the Columbia River to Cape Blanco contains 11 spring-run, 1 summer-run, and 33 fall-run populations (Kostow 1995). Peak spawning periods for spring, spring/summer, and summer-run populations occur from mid-September to early October which is somewhat later than in Puget Sound and the Strait of Juan de Fuca. Peak river-entry times for spring- and summer-run stocks range from May to August. In general, populations considered spring, spring/summer, and summer runs return to the river at an immature stage and hold in the river for an extended period before spawning. In contrast, fall-run fish enter freshwater at an advanced stage of maturation. Peak spawning periods for coastal fall runs occur from late-October to early-December, with a tendency for later spawning in more southerly rivers. The existence of multiple runs on many of the smaller coastal river systems is associated with low summer flows that physically limit access or result in high summer water temperatures in the lower river reaches (Nicholas and Hankin 1988).
Chinook salmon from the Washington and Oregon coasts emigrate to saltwater primarily as subyearlings and utilize the productive estuary and coastal areas as rearing habitat. The limited size of many coastal watersheds mandates the reliance on extended estuary or coastal rearing by juveniles. Furthermore, high summer water temperatures and related low flows may be responsible for early emigration. Chinook salmon from coastal populations (ocean-type) tend to have much larger eggs than inland, predominantly stream-type, populations (Rich 1920b , Nicholas and Hankin 1988, Lister 1990). Larger eggs result in larger juveniles and may enable an earlier and more successful emigration to marine rearing habitat (Fowler 1972, Kreeger 1995). The Washington and Oregon coasts contain numerous large estuary areas: Grays Harbor, Willapa Bay, Tillamook Bay, Coos Bay, Winchester Bay (Umpqua R.), and Yaquina Bay. Emigrating juveniles from rivers without well-developed estuary systems may undertake coastal migrations to estuary feeding areas or find sufficient rearing habitat in coastal areas, but it is unclear which strategy they undertake. Coastal chinook salmon from this region also mature at a later age than stocks from Puget Sound, the lower Columbia River and southern Oregon coastal areas (Nicholas and Hankin 1988, SCC 1995, QFD 1995, WDFW 1995). The majority of the runs are composed of 4- and 5-year-old fish, with a small proportion of 6-year-olds. The numerically large populations of chinook salmon on smaller coastal rivers may create competition for mates and select for larger (older) male chinook salmon (Roni and Quinn 1995).
Marine recoveries of CWTs indicate a similar ocean migration distribution for Washington and northern Oregon coastal stocks. The majority of the recoveries are from 4- and 5-year-old fish in British Columbia and Alaska fisheries. This is a more northerly oceanic distribution than is observed for Puget Sound, Lower Columbia River, and Southern Oregon and California stocks. A proportion of fish from stocks in the vicinity of Cape Blanco tend to exhibit a "north-and-south" migration pattern, with a proportion of recoveries occurring in Oregon and California coastal waters (Nicholas and Hankin 1988). The existence of a transition zone in migratory patterns may be due to natural and/or anthropogenic factors. CWT ocean recoveries of Umpqua River spring-run chinook salmon, specifically Rock Creek Hatchery fish, show a north and south distribution. The mouth of the Umpqua River is almost 100 km north of Cape Blanco; however, the Umpqua River has received transfers of Rogue River spring-run chinook salmon, a south migrating stock, during rebuilding programs over the past decades. The north-south migratory pattern may be the result of hybridization of Rogue and Umpqua River stocks. Differences in age and oceanic migration pattern between the Washington and Oregon coast and neighboring regions were among the most pronounced of any life-history comparisons.
The coastal drainages south of Cape Blanco are dominated by the Rogue, Klamath, and Eel Rivers. The Chetco, Smith, Mad, Mattole, and Russian Rivers and Redwood Creek are smaller systems that contain sizable populations of fall-run chinook salmon ( Campbell and Moyle 1990, ODFW 1995). Presently, spring runs are found in the Rogue, Klamath, and Trinity Rivers; additionally, a vestigial spring run may still exist on the Smith River (Campbell and Moyle 1990, USFS 1995). Historically, fall-run chinook salmon were predominant in most coastal river systems south to the Ventura River; however, their current distribution only extends to the Russian River (Healey 1991). There have also been spawning fall-run chinook salmon reported in small rivers draining into San Francisco Bay (Nielsen et al. 1994).
Chinook salmon populations south of Cape Blanco all exhibit an ocean-type life history. The majority of fish emigrate to the ocean as subyearlings, although yearling smolts can constitute up to approximately a fifth of outmigrants from the Klamath River Basin, and to a lesser proportion in the Rogue River Basin; however, the proportion of fish which smolted as subyearling vs. yearling varies from year to year (Snyder 1931, Schluchter and Lichatowich 1977, Nicholas and Hankin 1988, Barnhart 1995). This fluctuation in age at smoltification is more characteristic of an ocean-type life history. Furthermore, the low flows, high temperatures, and barrier bars that develop in smaller coastal rivers during the summer months would favor an ocean-type (subyearling smolt) life history (Kostow 1995).
Run timing for spring-run chinook salmon in this area typically begins in March and continues through July, with peak migration occurring in May and June. Spawning begins in late August and can continue through October, with a peak in September. Historically, spring-run spawning areas were located in the river headwaters (generally above 400 m). Run timing for fall-run chinook salmon varies depending on the size of the river. Adult Rogue, Upper Klamath, and Eel River fall chinook salmon return to freshwater in August and September and spawn in late October and early November (Stone 1897, Snyder 1931, Nicholas and Hankin 1988, Barnhart 1995). In other coastal rivers and the lower reaches of the Klamath River, fall-run freshwater entry begins later in October, with peak spawning in late November and December--often extending into January (Leidy and Leidy 1984, Nicholas and Hankin 1988, Barnhart 1995). Late-fall or "snow" chinook salmon from Blue Creek, on the lower Klamath River, were described as resembling the fall-run fish from the Smith River in run and spawning timing, as well as the degree of sexual maturation at the time of river entry (Snyder 1931).
Populations in this region are readily distinguished from more northerly coastal populations by their oceanic migration patterns. Recoveries of CWTs in ocean fisheries occur primarily off the Oregon and California coasts. The majority of the spring and fall runs are composed of 3- and 4-year-old fish, with a small proportion of 5-year-olds (Snyder 1931, Kutkuhn 1963, Nicholas and Hankin 1988, Barnhart 1995). Analysis of scales from "late-fall run" fish returning to the lower Klamath River indicated that there was a higher proportion of 5-year-old fish (up to 51%) compared with spring- or fall-run fish returning to the upper Klamath River (Snyder 1931). In general, fish from coastal populations south of Cape Blanco mature earlier than those to the north.
Other morphological and physiological differences between geographic regions have been observed. McGregor (1923a) and Snyder (1931) described significant differences between Klamath and Sacramento River fish in gill arch and pyloric caeca counts, in addition to body size and fecundity. Dorsal fin ray, anal fin ray, and branchiostegal counts for the Klamath River chinook salmon were significantly lower than for Columbia River ocean- or stream-type chinook salmon stocks (Snyder 1931, Schreck et al. 1986). Rich (1920b) found that coastal stocks from the Umpqua and Rogue Rivers had larger eggs than Columbia River stocks. Egg diameters for fall-run chinook salmon on the Klamath River averaged 9 mm (Snyder 1931), which is similar to ranges presented by Nicholas and Hankin (1988) for Oregon coast chinook salmon but much larger than for populations in the Sacramento River (see California Central Valley section). Furthermore, data collected by McGregor (1922, 1923b) indicated that for a given length, Sacramento River fish have a higher average fecundity and smaller egg size than fish from the Klamath River. While coastal populations south of Cape Blanco are substantially different from those to the north, there is some finer scale differentiation between shorter coastal system and the two larger river basins, the Rogue and Klamath Rivers.
Agricultural, logging, and mining activities, in combination with periodic flood events (e.g. 1955, 1969), have affected all of the coastal river systems to some degree. Mining activities have also caused severe habitat degradation. The construction of dams on the Rogue, Klamath, and Eel River Basins has restricted the distribution and potentially altered the life history of chinook salmon, especially spring-run fish that historically utilized upstream habitat. Lost Creek Dam (RKm 253) eliminated one-third of the spawning habitat of spring-run chinook salmon in the Rogue River (Kostow 1995). Additionally, changes in river flow and temperature have allowed fall-run chinook salmon to spawn in more upstream locations and increased the opportunities for interbreeding between fall and spring runs (ODFW 1990). Similarly, dam construction on the Klamath River Basin has eliminated much of the spawning habitat for spring-run fish and increased the potential for interbreeding between spring and fall runs. Fish passage to the upper Klamath River was blocked at Klamath Falls by the Link River hydroelectric dam in 1895. Several dams have subsequently been constructed on the mainstem Klamath River. Historically, the largest spring-run population in the Klamath River Basin was in the Shasta River; however, this population was extirpated in the early 1930s as a result of land use practices and water diversion dams. Since 1962, the upper limit to anadromous migration has been the Iron Gate Dam (RKm 306). Additionally, the Lewiston water diversion dam (RKm 249) on the Trinity River has prevented access of spring-run chinook salmon to their historical spawning grounds on the East Fork, Stuart Fork, and Upper Trinity River and Coffee Creek (Campbell and Moyle 1990). Hatchery-reared smolts, especially yearling smolts, from mitigation hatcheries on the Klamath River (Iron Gate Hatchery) and Trinity River (Trinity River Hatchery) have probably altered age of maturation and smoltification estimates derived from the scales of unmarked returning adults. The life-history attributes of coastal chinook salmon south of Cape Blanco are quite distinct from those to the north, in the Upper Klamath River Basin, and those in the Central Valley. These differences exist in spite of artificial propagation and the loss of ecologically distinct spawning and rearing habitat areas.
The Sacramento and San Joaquin Rivers and their tributaries contain several different groups of chinook salmon based on run timing and habitat utilization. Historically, spring-run fish were predominant throughout the Central Valley, occupying the upper and middle reaches (450-1,600 m in elevation) of the San Joaquin, American, Yuba, Feather, Sacramento, McCloud, and Pit Rivers, with smaller populations in most other tributaries with sufficient cold-water flow to maintain spring-run adults through the summer prior to spawning (Stone 1874, Rutter 1904, Clark 1929). Winter-run populations historically utilized the upper watersheds (450-900 m in elevation) of the upper Sacramento, Pit, McCloud, and Calaveras Rivers and were not as numerous as the spring or fall runs (Slater 1963, Reynolds et al. 1993). Fall and late-fall runs spawn in the lower reaches (60-600 m) of most rivers and streams in the Central Valley (Clark 1929, Hallock and Fry 1967, Reynolds et al. 1993). Fall-run chinook salmon are currently the most numerous of the runs in the Central Valley. Habitat degradation due to dams, water diversions, and placer mining, as well as past and present land-use practices have severely reduced the range and number of spring- and winter-run chinook salmon and to a lesser extent fall and late-fall runs (Clark 1929, Needham et al. 1940, Reynolds et al. 1993, Fisher 1994).
Central Valley chinook salmon exhibit an ocean-type life history. Large numbers of fry have been observed emigrating during the winter and spring (Rutter 1904, Rich 1920a, Calkins et al. 1940, Kjelson et al. 1982, Gard 1995). High summer water temperatures in the lower Sacramento River (temperatures in the Sacramento-San Joaquin Delta can exceed 22C) present a thermal barrier to up- and downstream migration and may be partially responsible for the evolution of the fry migration life history (Rich 1920a, Kjelson et al. 1982). Water withdrawals for agricultural and municipal purposes, have occasionally been of a sufficient magnitude to result in reverse flows in the lower San Joaquin River.
Age estimates from scales of returning adults in 1919 and 1921 indicated that 89% of the fish had emigrated as subyearlings (Clark 1929). Scale samples in Clark's study were from returning adults taken below the confluence of the Sacramento and San Joaquin Rivers. Scale samples were made throughout the year during the course of the in-river fishing season (there were two closures during early June to early July and late September to early November) and would have included all of the runs. Calkins et al. (1940) sampled both the fall and spring runs on the upper Sacramento River and determined that the proportion of adults that emigrated as subyearlings in both runs was 90%. Gard (1995) stated that the majority of smolts from all four runs on the upper Sacramento River currently emigrate as subyearlings. The emigration of spring, fall, and late-fall runs is completed prior to high summer temperatures in the lower river, while winter-run emigration does not begin until after the summer temperatures have started to diminish in August (Fig. 14). In contrast, Fisher (1994) suggested that a large proportion of late-fall and spring-run juveniles emigrate as yearlings, the average length for late-fall-run and spring-run smolts being 160 and 115 mm, respectively. Using scales from returning adults, Calkins et al. (1940) estimated that the average size of subyearling fall- and spring-run smolts at the time of ocean entrance was 88 and 83 mm, respectively. Emigrating juveniles sampled in the upper Sacramento River are, on average, less than 70 mm in length (Gard 1995). Vast numbers of fry (<70 mm) were observed rearing in the Sacramento-San Joaquin River estuary, but relatively few larger smolts were found in the late spring or fall (Kjelson et al. 1982). Fry tend to remain in the estuary for an extended period of almost 2 months (Kjelson et al. 1982). The tendency for fish to emigrate as fry appears to be characteristic of this region and is linked to summer water conditions (low flow and high temperatures).
As with the timing of smolt emigration, the timing of the adult return migration and spawning is dictated by high summer temperatures. Fall- and late-fall runs enter freshwater at an advanced stage of maturity and move quickly to their spawning sites. The return migration does not begin until late August or September (fall run) or December (late-fall run) after summer temperatures have declined (Hallock and Fry 1967). Fall-run and late-fall-run chinook salmon peak spawning occurs in late October and early February, respectively (Fisher 1994). Winter-run and spring-run fish enter freshwater well in advance of spawning. Winter-run adults historically would have migrated upstream at a time of high river flows in late November through January and held in upriver areas until spawning sometime in April-July (Slater 1963, Fisher 1994). The eggs deposited would have developed during the summer months in the cold headwaters of the Sacramento, Pit, McCloud, and Calaveras Rivers. Fry would then emigrate in the fall after temperatures in the lower river had cooled. The migration of the spring run began in March and April, later than the winter run, when river flows were still sufficient for these fish to gain access to the cool, spring- and snow-fed upper reaches of rivers. Spawning did not typically start until late August (lasting through early October), and fry did not emigrate until river flows had risen in early winter. Winter- and spring-run fish no longer have access to the vast majority of their historical spawning and juvenile rearing grounds, but their migration and spawning timing still reflect the appropriate timetable to utilize these areas.
Estimates of the age at maturation for Central Valley stocks differ between studies; this may be due to differences in scale pattern interpretation, or there may have been a shift to younger spawners. Fish gill-netted in 1919 and 1921 below the confluence of the Sacramento and San Joaquin Rivers were primarily 4 years old (46.5%), with 5- and 3-year olds comprising 32.5 and 17.0% of the spawners, respectively. The use of fish collected in gill nets introduces a considerable bias; differences observed in the percentage of 5-year-olds between 1919 and 1921 (24.0% vs. 41.0%), was thought to be due to a change in the gill-net mesh size from 14 cm to 19 cm. Additionally, the large mesh size would potentially explain the low incidence, 1.1%, of 2-year-old fish in 1921. Rich (1921) estimated females caught in the troll fishery off Monterey Bay in 1918 would mature in their third or fourth year. The predominant age classes among returning fall- and spring-run adults sampled at Redding in 1939 were 3- and 4-year-old fish (Calkins et al. 1940). Furthermore, the incidence of 2-year-old males (jacks) was 8.8 and 27.3% for the spring- and fall-run fish, respectively. Five- and 6-year old fish contributed less than 5% of the return for both runs (Calkins et al. 1940). Near the turn of the century, Rutter (1904) observed large numbers of small male "grilse" (jacks) in Battle Creek, a tributary to the upper Sacramento River. Samples taken from the McCloud River from 1909-12 suggested that approximately 10% of the males matured as 2-year olds without leaving freshwater (Rich 1920a). The mean age composition for fall-run chinook salmon from the upper Sacramento River, for the 1973-79 brood years, was 24, 57, 19, and <1% for 2-, 3-, 4-, and 5-year-olds, respectively (Reisenbichler 1986). Hallock and Fisher (1985) estimated that for winter-run chinook salmon, 3-year-old returning adults constituted the majority of returning fish (67%), with 2-year-old and 4-year-old fish representing the remainder of the age classes (25 and 8%, respectively). More recently, Fisher (1994) estimated that the 3-year-old age class was predominant among all runs, being 77, 57, 91, and 87% of each run for fall-, late-fall-, winter-, and spring-runs, respectively. The age structure of fish from the San Joaquin River Basin appears to be much younger than that of the Sacramento River (Neillands 1995). Up to 30% of the returning adults in the Merced and Tuolumne Rivers are 2 years of age; this includes a number of 2-year-old females, "Jills," which are not normally observed in other river systems. The younger age of maturation is probably related to warmer water temperatures in the San Joaquin River rather than being genetically influenced, given the genetic similarity between Sacramento and San Joaquin River fall-runs. Furthermore, analysis of chinook salmon age structure in the San Joaquin River is complicated by the influence of river flow on the survival of emigrating juveniles. During extreme drought years, there has been a near failure of the corresponding year class of smolts. It has yet to be determined whether the shift toward a younger age structure in the Central Valley during this century is environmentally-mediated, due to the selective harvest of older (larger) adults, or reflects an underlying genetic change.
Sacramento River chinook salmon reproductive traits are very different from coastal California and the Klamath River populations. Information on Sacramento River chinook salmon eggs sizes is limited. Page (1888) estimated the average egg diameter was 6.7 mm for eggs collected at the Baird NFH on the McCloud River. The average egg diameter for winter-run eggs in 1992 was 6.91 mm (USFWS 1996a). Quinn and Bloomberg (1992) found that chinook salmon in New Zealand (from Sacramento River transplants in 1901-07) have considerably smaller eggs, (0.17 g), relative to coastal stocks in British Columbia, (0.47 g). The fecundity of Central Valley females was also considerably higher for a given body size than for females from the Klamath River (Snyder 1931).
Historically, low summer flows and associated high temperatures have been major factors in determining the life-history characteristics for each of the four runs in the Central Valley. Winter- and spring-run adults utilized colder mountain streams to provide a suitable holding, incubation, and fry-rearing environment during months when the environment on the lower river was inhospitable. Fall- and late-fall-run fish delayed the adult return migration and spawning until temperatures had declined to acceptable levels. Differences in habitat utilization provided a spatial separation between runs in addition to temporal differences. The duration of freshwater rearing appears to have been minimized to allow emigration to estuarine rearing habitat before temperatures rose to deleterious levels.
Anthropogenic activities have primarily affected the spring and winter runs. Placer mining in the 1800s destroyed spawning and rearing habitats either directly or through increased sedimentation. Mine wastes still affect water quality. Water diversion and hydroelectric dams have limited or prevented access to most of the upriver areas that were historically utilized by spring and winter runs (Clark 1929). Agricultural and municipal water withdrawals have reduced river flows and increased temperatures during the critical summer months, or in some cases even reversed river flows (Reynolds et al. 1993). Changes in the thermal and water flow profiles for Central Valley rivers have presumably subjected chinook salmon to strong selective forces. The degree to which current life-history traits reflect predevelopment characteristics is largely unknown, especially since most of the habitat degradation occurred before chinook salmon studies were undertaken late in the nineteenth century.
One consequence of dam construction has been alteration of the river thermal profile. The completion of Shasta Dam (RKm 505) in 1944 eliminated access to the McCloud, Pit, and Upper Sacramento Rivers. However, water subsequently released from Shasta Dam has had a more uniform, cooler, thermal regime, 12-15C, than prior to dam construction (Moffett 1949). This cool water provided new spawning habitat for spring- and winter-run adults attempting to migrate to their historical spawning grounds. The released water was also significantly warmer than historical levels during the autumn and winter, thereby accelerating egg development and fry emergence (Moffett 1949). Accelerated embryonic development may effect subsequent smolt emigration timing and reduce estuarine survival. Additionally, dam construction has eliminated the spatial and temporal barriers that once separated the fall run from the spring run and increase the potential for hybridization. The expected loss of spawning habitat above Shasta Dam led to efforts to salvage fall- and spring-run adults destined for the upper Sacramento River (Calkins et al. 1940). In a program that paralleled the GCFMP recovery effort, fish were intercepted at Balls Ferry (RKm 446) or Keswick Dam (RKm 486) and transferred to the Coleman NFH for spawning, to Deer Creek (RKm 353) for natural spawning (spring run only), or allowed to remain in the Sacramento River (primarily fall run) to spawn naturally. The primary criteria for separating spring and fall runs was a late June cut-off date that varied from year to year (Moffett 1949). In all, some 15,972 "spring-run" chinook salmon were hauled to Deer Creek from 1941-46. A considerable proportion of transferred fish died shortly after transfer to Deer Creek because of high water temperatures (Moffett 1949). There was no provision in the plan to identify winter-run adults, and a number were incidentally hauled to Deer Creek (Slater 1963). The absence of baseline information on spring-run fish from the mainstem Sacramento River and Deer Creek prevents any estimate of the impact of these fish transfers, nor is there any information for estimating potential interbreeding between winter and spring runs. The loss of spring-run spawning habitat in the headwater areas has eliminated the spatial separation that once maintained the genetic isolation between spring- and fall-run populations, and a certain amount of mixing has probably occurred in both hatchery and naturally spawning populations (Fisher 1994). Stock transfers and high straying rates may have resulted in the loss of distinctive life-history characteristics between fall-run populations. Perhaps because fall-run fish utilize mainstem areas and rear in freshwater for a limited period, there has been little selective pressure for geographic adaptation within the Central Valley. Alternatively, local extinctions and recolonizations due to natural drought cycles may have prevented distinct populations from forming among fall-run chinook salmon. Nevertheless, differences in the life-history traits of winter, spring, fall, and late-fall runs are still apparent in spite of massive changes in their spawning and rearing habitat, and these differences underscore the distinctiveness of these stocks.