Wild Chinook Research Articles

Influence of Juvenile Size on the Age at Maturity of Individually Marked Wild Chinook Salmon

AbstractThe age at maturity for an organism represents a trade‐off between survival and reproductive fitness and is controlled by both genetic factors and environmental conditions. In the case of Pacific salmon Oncorhynchus spp., most of our understanding of the importance of various drivers in determining age at maturity has relied on information from cohort analyses using average size and age conditions or from hatchery fish that might not reflect observed patterns in wild fish. I used 14 years of adult return data from individually passive integrated transponder (PIT)–tagged wild Chinook salmon O. tshawytscha from the Snake River to examine the influence of juvenile size on subsequent age at maturity. I also included information on the natal watershed as a surrogate for genetic influence. I found a significant, positive relationship between juvenile length and the proportion of adult salmon returning at age 4 versus age 5, but the effect of juvenile length varied among the three watersheds. This suggests that local environmental conditions supporting juvenile growth largely control the ultimate age composition for a given brood of this stock of salmon but that genetic differences also exist.

Two Alternative Juvenile Life History Types for Fall Chinook Salmon in the Snake River Basin

Fall Chinook salmon Oncorhynchus tshawytscha in the Snake River basin were listed under the Endangered Species Act in 1992. At the time of listing, it was assumed that fall Chinook salmon juveniles in the Snake River basin adhered strictly to an ocean-type life history characterized by saltwater entry at age 0 and first-year wintering in the ocean. Research showed, however, that some fall Chinook salmon juveniles in the Snake River basin spent their first winter in a reservoir and resumed seaward movement the following spring at age 1 (hereafter, reservoir-type juveniles). We collected wild and hatchery ocean-type fall Chinook salmon juveniles in 1997 and wild and hatchery reservoir-type juveniles in 1998 to assess the condition of the reservoir-type juveniles at the onset of seaward movement. The ocean-type juveniles averaged 112–139 mm fork length, and the reservoir-type juveniles averaged 222–224 mm fork length. The large size of the reservoir-type juveniles suggested a high potential for survival to salt water and subsequent return to freshwater. Scale pattern analyses of the fall Chinook salmon spawners we collected during 1998– 2003 supported this point. Of the spawners sampled, an overall average of 41% of the wild fish and 51% of the hatchery fish had been reservoir-type juveniles. Males that had been reservoir-type juveniles often returned as small “minijacks” (wild, 16% of total; hatchery, 40% of total), but 84% of the wild males, 60% of the hatchery males, and 100% of the wild and hatchery females that had been reservoir-type juveniles returned at ages and fork lengths commonly observed in populations of Chinook salmon. We conclude that fall Chinook salmon in the Snake River basin exhibit two alternative juvenile life histories, namely ocean-type and reservoir-type.

Relating results of chronic toxicity responses to population‐level effects: Modeling effects on wild chinook salmon populations

Standard toxicity tests assess the physiological responses of individual organisms to exposure to toxic substances under controlled conditions. Time and space restrictions often prohibit the assessment of population-level responses to a toxic substance. Compounds affecting various toxicity endpoints, such as growth, fecundity, behavior, or immune function, alter different demographic traits and produce different impacts on the population. Chronic effects of immune suppression, reproductive impairment, and growth reduction were examined using life history models for Chinook salmon (Oncorhynchus tshawytscha). Modeled immune suppression acted through reductions in age-specific survival, with first- and second-year survival producing the greatest changes in the population growth rate (lamda). A 10% reduction in various reproductive parameters all produced a similar lamda, but different sensitivity and stable age distributions. Growth reduction models incorporated effects to both survival and reproduction and produced additive effects. Overall, model output indicated that for Chinook salmon, alteration of first-year survival has the greatest relative impact on lamda. Results support the importance of linking toxicity endpoints to the demographic traits that they influence and help generate toxicity tests that are more relevant for the species. Life history modeling provides a useful tool to develop testable hypotheses regarding specific and comparative population-level impacts.

Comparison of maturation timing, egg size and fecundity between hatchery lines of chinook salmon and their wild donor stocks

Increasing concern has been expressed about the genetic effects of cultured salmonid fishes on natural populations. Avoidance of extreme negative outcomes was one reason for the establishment of a genetic management policy for the State of Alaska. However, domestication within the hatchery may still cause divergence from the wild donor population. This divergence could potentially lead to adverse impacts on wild stocks through straying and introgression. This study examines potential domestication in two Alaskan chinook salmon stocks. The Little Port Walter (LPW) Hatchery Chickamin River stock resulted from a small collection of wild broodstock in 1976. The LPW Unuk stock was founded with a larger number of individuals in 1976 and has had subsequent infusion of wild gametes. These lines have been maintained at LPW through ocean ranching of tagged smolts. Comparisons are made between the hatchery lines, progeny of wild chinook collected from the Chickamin and Unuk Rivers, and hybrids between the hatchery and wild groups. Mature ocean‐ranched female chinook salmon returning to the facility were periodically graded for ripeness and spawned. Body size and meristic measurements were collected from these mature spawners. Maturation timing, fecundity, and individual egg size of these fourth generation hatchery fish are compared with that of offspring of wild fish from the same donor stock. Stock of origin is confirmed for all spawners and offspring using microsatellite DNA analysis.

Evidence of a Lunar Gravitation Cue on Timing of Estuarine Entry by Pacific Salmon Smolts

Results from a 4-year tagging study of the migration behavior of Pacific salmon smolts indicate an effect of lunar gravitation on the time of saltwater entry at an artificial tidal barrier. We used passive integrated transponders to tag several populations of juvenile Chinook salmon Oncorhynchus tshawytscha, coho salmon O. kisutch, and sockeye salmon O. nerka, including hatchery and wild Chinook salmon. We discovered through exploratory data analyses that the temporal variation in the date of initiation of passage at the barrier by the majority of the Chinook salmon populations studied was most closely related to the date of lunar apogee, followed by the date of quarter moon. Both phenomena are associated with weaker net lunar gravitation effects on water bodies than the full or new moon, which showed no consistent effect. Using available data we could not discern any effect of other environmental variables, including water temperature, nocturnal illumination, salinity, and lake level. Lunar apogee occurred on average 1 d prior to the onset of substantial passage by Chinook salmon smolts. The phenomenon was consistent in all 4 years, even though the lunar apogee cycle advanced annually by 7 d. An intermediate relationship was noted for coho salmon smolts and the weakest relationship for sockeye salmon smolts.

Statistical analysis of the relationship among environmental variables, inter-annual variability and smolt trap efficiency of salmonids in the Tucannon River

Abstract Spring and fall chinook (Oncorhynchus tshawytscha), steelhead (Oncorhynchus mykiss) and hatchery spring chinook in the Tucannon River, Washington, USA are listed as “threatened” under the Endangered Species Act of 1973. Restoration and management of both species can be facilitated by understanding how biotic and abiotic factors affect their smolt trap efficiency. In this paper, we examine the effects of the rate of water flow, water temperature, the level of staff gauge, debris load, and Secchi disk readings on their weekly smolt trap efficiency from 1998 to 2003 using a generalized linear model (GLM) with a binomial response (link function – logit). The nonlinear relationships between the smolt trap efficiency and abiotic variables are also analyzed using a generalized additive model (GAM) with a binomial response (link function – logit). Both GLM and GAM analyses showed that the trap efficiency varied among years for fall chinook, steelhead, and wild spring chinook, but not for hatchery spring chinook, and that the level of staff gauge and the rate of water flow were the most important factors altering trap efficiency. The partial residuals from GAM analyses were used to determine the optimal number of sampled fish with a known efficiency and to detect possibly misleading results from GLM analyses.

Blood chemistry correlates of nutritional condition, tissue damage, and stress in migrating juvenile chinook salmon (Oncorhynchus tshawytscha)

We used factor analysis to examine the correlation structure of six multivariate blood chemistry data sets for migrating hatchery and wild juvenile chinook salmon (Oncorhynchus tshawytscha). Fish were sampled (1998–2002) from juvenile fish bypass systems at dams or (one data set) from fish transport barges on the Snake and Columbia rivers. Analyses were performed to determine which blood chemistry analytes covaried, to facilitate interpretation of the data sets, and to provide insight into controlling physiological mechanisms. Four underlying factors were derived from the analyses: (i) a nutritional factor composed of total protein, cholesterol, calcium, and alkaline phosphatase, (ii) a tissue damage factor composed of the enzymes alanine aminotransferase, aspartate aminotransferase, and creatine kinase, (iii) a lipid metabolism factor composed of triacylglycerol lipase and triglycerides, and (iv) a stress factor composed of cortisol, glucose, Na+, and Cl–. Although causal mechanisms cannot be directly inferred from our analyses, findings of published research provide tenable causal mechanisms for the observed structure. The consistency of the correlation structure among data sets suggests that composite (latent) variables may be more reliable indicators of some physiological responses than changes in individual variables.

Hooking Mortality by Anatomical Location and Its Use in Estimating Mortality of Spring Chinook Salmon Caught and Released in a River Sport Fishery

AbstractWe estimated the hooking mortality of spring Chinook salmon Oncorhynchus tshawytscha that were caught and released to determine whether selective fishing on hatchery Chinook salmon would reduce harvest mortality of wild fish in a sport fishery in the lower Willamette River, Oregon. Hooking mortality in the fishery was estimated from hooking mortality rates for each of five anatomical locations (jaw, 2.3%; tongue, 17.8%; eye, 0.0%; gills, 81.6%; and esophagus–stomach, 67.3%) and from the frequency of these anatomical locations in the sport fishery (jaw, 81.5%; tongue, 5.1%; eye, 0.4%; gills, 5.1%; and esophagus–stomach, 7.8%). Mortality rates by anatomical location were estimated from recaptures of 869 tagged fish that were experimentally angled and of 825 tagged controls that were trapped in a nearby fishway. Anatomical hook locations in the lower Willamette River sport fishery were determined with creel surveys. We estimated hooking mortality rates of 12.2% for wild Chinook salmon caught and released in the sport fishery and 3.2% for the entire run of wild Chinook salmon based on a mean encounter rate of 26%. Hook location was the primary factor affecting recapture of hooked fish, but fish length, gear type, bleeding, and the elapsed time to unhook fish were also significant factors. A selective sport fishery in the lower Willamette River can be used to reduce harvest mortality on runs of wild Chinook salmon while maintaining fishing opportunity on hatchery Chinook salmon. The effect of selective fisheries for Chinook salmon in other rivers would depend on the frequency distribution of anatomical hook locations and on river‐specific encounter rates.

Central administration of corticotropin-releasing hormone alters downstream movement in an artificial stream in juvenile chinook salmon ( Oncorhynchus tshawytscha)

We evaluated the effect of corticotropin-releasing hormone (CRH) on spatial distribution and downstream movement in an artificial stream in juvenile Chinook salmon ( Oncorhynchus tshawytscha) during the period when the fish were able to tolerate seawater. An intracerebroventricular (ICV) injection of CRH (500 ng) to hatchery fish significantly increased the proportion of fish that were distributed downstream of a mid-stream release site. A second group of hatchery fish were given ICV injections of saline (control) or CRH (500 ng) and released near the top of the stream. The time taken to enter a trap at the lower end of the stream was recorded. In all cases the groups given CRH had a higher proportion of fish that did not enter the trap within 60 min of release. However, in those fish that did enter the trap, treatment with CRH increased the speed of downstream movement to this point relative to control fish. Wild sub-yearling Chinook salmon were captured during their downstream migration to the estuary and given ICV injections of saline or CRH (500 ng) either 2, 3, or 7 days after transport from the river. As with hatchery fish, a significantly higher proportion of wild fish that were administered CRH did not enter the trap at the lower end of the stream. The mean time of passage for control fish decreased on each successive day (day 2 > day 3 > day 7). In contrast, the mean passage time of the wild fish that were given CRH was relatively constant through time, and was only significantly faster than control fish on day 2. The current study provides evidence that CRH alters the downstream movement of juvenile Chinook in a simulated stream environment, and produces behavioral effects similar to those of juvenile salmonids that are stressed during their downstream migration.

Genetic Structure of Wild Chinook Salmon Populations of Southeast Alaska and Northern British Columbia

Genetic Structure of Wild Chinook Salmon Populations of Southeast Alaska and Northern British Columbia

Interactions between natural and hatchery chinook salmon parr in a laboratory stream channel

Pre-smolt (fry and parr) hatchery chinook salmon Oncorhynchus tshawytscha are increasingly being used in supplementation projects to augment natural but diminished salmon populations, but little is known of the behavioral interactions between sympatric hatchery and wild chinook salmon populations and how those interactions may affect success of a supplementation program. We used experiments conducted in a laboratory stream channel to document behavioral interactions that occur when hatchery salmon are stocked to streams containing existing natural chinook salmon juveniles. We found that natural fish aggression was suppressed when hatchery fish were larger than natural fish in the spring and summer, and the aggressiveness of natural fish increased when paired with similar sized hatchery fish during fall. Types of micro-habitat used by natural chinook salmon were mainly affected by hatchery fish during the fall trials when more natural fish were found in open pool areas, and fewer used cobble cover when hatchery fish were present. Observed behavioral shifts by natural chinook salmon parr could increase energy expenditures and exposure to predators when hatchery fish are present, but there were few consistent effects from hatchery salmon on the growth or movement patterns of natural chinook salmon during this study. Relative fish size, increased stream densities, and the behavior of the hatchery salmon appeared to be a major factors influencing observed changes in the behavior of juvenile natural chinook salmon in the stream sections.

Influence of Flow and Temperature on Survival of Wild Subyearling Fall Chinook Salmon in the Snake River

Summer flow augmentation to increase the survival of wild subyearling fall chinook salmon Oncorhynchus tshawytscha is implemented annually to mitigate for the development of the hydropower system in the Snake River basin, but the efficacy of this practice has been disputed. We studied some of the factors affecting survival of wild subyearling fall chinook salmon from capture, tagging, and release in the free-flowing Snake River to the tailrace of the first dam encountered by smolts en route to the sea. We then assessed the effects of summer flow augmentation on survival to the tailrace of this dam. We tagged and released 5,030 wild juvenile fall chinook salmon in the free-flowing Snake River from 1998 to 2000. We separated these tagged fish into four sequential within-year release groups termed cohorts (N = 12). Survival probability estimates (mean ± SE) to the tailrace of the dam for the 12 cohorts when summer flow augmentation was implemented ranged from 36% ± 4% to 88% ± 5%. We fit an ordinary least-squares multiple regression model from indices of flow and temperature that explained 92% (N = 12; P < 0.0001) of the observed variability in cohort survival. Survival generally increased with increasing flow and decreased with increasing temperature. We used the regression model to predict cohort survival for flow and temperature conditions observed when summer flow augmentation was implemented and for approximated flow and temperature conditions had the summer flow augmentation not been implemented. Survival of all cohorts was predicted to be higher when flow was augmented than when flow was not augmented because summer flow augmentation increased the flow levels and decreased the temperatures fish were exposed to as they moved seaward. We conclude that summer flow augmentation increases the survival of young fall chinook salmon.

Migrational Behavior and Seaward Movement of Wild Subyearling Fall Chinook Salmon in the Snake River

Abstract Flow augmentation increases flow and decreases temperature in reservoirs in the lower Snake River during the seaward migration of wild subyearling fall chinook salmon Oncorhynchus tshawytscha. A study of the migrational behavior and seaward movement of wild subyearling fall chinook salmon in the Snake River was necessary to help understand the efficacy of flow augmentation. We studied fall chinook salmon in the Snake River during 1992–2001. After analyzing mark–recapture data, we deduced that fall chinook salmon passed through at least four migrational phases, including (1) discontinuous downstream dispersal along the shorelines of the free-flowing river, (2) abrupt and mostly continuous downstream dispersal offshore in the free-flowing river, (3) passive, discontinuous downstream dispersal offshore in the first reservoir encountered en route to the sea, and (4) active and mostly continuous seaward migration. We used ordinary-least-squares multiple regression to test the effects of flow (m3/s), t...

Growth of Wild Subyearling Fall Chinook Salmon in the Snake River

Abstract Growth is an important determinant of life history development for juvenile anadromous salmonids. We collected juvenile fall chinook salmon Oncorhynchus tshawytscha in two reaches of the Snake River to describe growth in fork length (mm/d) and to test for a relation between growth and water temperature. Growth rate during shoreline rearing was significantly higher (P = 0.003) for parr in the warmer of these two reaches (grand means = 1.2 ± 0.04 and 1.0 ± 0.04 mm/d). Because smolts from the two reaches share a common, relatively warm downstream migration route, growth rates were similar between smolts from the two reaches (P = 0.18; grand means = 1.3 ± 0.04 and 1.4 ± 0.04 mm/d). By pooling data across reaches and life stages, we found that growth rate generally increased as water temperature increased (N = 17, r 2 = 0.62, P = 0.0002). The growth rates we observed were probably lower than for fall chinook salmon in a historical rearing area now inaccessible because of dams, but they were still rapi...

Interspecific Effects of Artifically Propagated Fish: an Additional Conservation Risk for Salmon

Abstract: For more than 120 years, hatcheries have released enormous numbers of Pacific salmon to compensate for numerous human insults to their populations, yet the ecological effects of this massive effort are poorly understood. We tested the hypothesis that hatchery‐reared steelhead salmon ( Oncorhynchus mykiss) released into the Snake River Basin negatively affect the survival of wild Snake River steelhead and chinook ( O. tshawytscha) salmon. Because climatic conditions can influence salmon survival, we included an index of the El Niño–Southern Oscillation ( ENSO) as a covariate in our analyses. Based on time series of hatchery releases and rates of smolt‐to‐adult survival, we demonstrate that the survival of wild chinook salmon is negatively associated with hatchery releases of steelhead. The state of the ( ENSO) did not affect the strength of this relationship. We observed no relationship between survival of wild steelhead and steelhead hatchery releases. Our results suggest that industrial‐scale production of hatchery fish may hinder the recovery of some threatened salmonids and that the potential interspecific impact of hatcheries must be considered as agencies begin the process of hatchery reform.

Juvenile Life History of Wild Fall Chinook Salmon in the Snake and Clearwater Rivers

Dam construction in the 1950s and 1960s blocked passage to the historical spawning area of Snake River fall chinook salmon Oncorhynchus tshawytscha. We obtained water temperature data and collected juvenile fall chinook salmon in three present-day spawning areas from 1992 to 2000 to investigate the relation between water temperature and juvenile life history events. We used historical water temperatures and the literature to depict juvenile life history in the historical spawning area. Water temperatures in the three present-day spawning areas differed significantly from winter to spring, when eggs were incubating (P ≤ 0.0001), as well as during spring, when juveniles were rearing and starting seaward migration (P ≤ 0.0001). When water temperatures were warmer, the timing of most life stages was generally earlier. The life stages included fry emergence (r2 = 0.85, N = 14, P < 0.0001), growth to parr size (r2 = 0.94, N = 15, P < 0.0001), and smolt emigration (r2 = 0.93, N = 14, P < 0.0001). The percentage of parr that overwintered in freshwater and migrated seaward the next spring increased when spring water temperatures decreased (r2 = 0.40, N = 12, P = 0.02). As the historical spawning area was warmer than present-day spawning areas, fall chinook salmon juvenile life history progressed on an earlier time schedule. We conclude that dam construction changed juvenile fall chinook salmon life history in the Snake River basin by shifting production to areas with relatively cool water temperatures and comparatively lower growth opportunity.

Use of tissue and sediment‐based threshold concentrations of polychlorinated biphenyls (PCBs) to protect juvenile salmonids listed under the US Endangered Species Act

Abstract Under the Endangered Species Act, the National Marine Fisheries Service has authority to protect listed species from any adverse actions that may jeopardize the population's ability to recover and increase to sustainable levels. Listed salmon species in the northwest United States are known to travel through urban areas in their migration from river to ocean. Species such as the chinook salmon (Oncorhynchus tshawytscha) often spend several weeks in these urban estuaries where they can be highly exposed to urban‐related contaminants that reside in the sediments and accumulate in their prey species. The concern is that these contaminants are bioaccumulated to levels that may impact the ability of individual salmon to grow and mature normally. This paper provides a framework for determining the tissue and sediment concentrations of polychlorinated biphenyls (PCBs) that are likely protective against adverse effects in listed salmonid species. The relevant ecotoxicological literature was examined and 15 studies were selected that met the pre‐established criteria outlined here. For each study, the lowest tissue concentration (residue) of total PCBs associated with a biological response was selected. The tissue concentration associated with the 10th percentile of these 15 studies was chosen to represent the residue effect threshold (RET) above which wild juvenile salmonids would be expected to exhibit adverse sublethal effects from accumulated PCBs. This value (2.4 μg PCBs g−1 lipid) is expressed in terms of the lipid‐normalized concentration because of the large effect lipid can have on the expressed toxicity and the substantial variability in lipid content observed in salmonids over their life cycle. A sediment concentration that is expected to produce the RET was then determined using the biota‐sediment accumulation factor approach. The sediment effect threshold, which varies with the total organic carbon content in sediment, is the level above which adverse effects may be expected in juvenile salmonids due to accumulation of PCBs from environmental exposure. Bioaccumulation of PCBs was examined in one river system as a model for determining an appropriate bioaccumulation factor for wild juvenile chinook salmon. Evaluation of exposure to potentially deleterious concentrations of PCBs based on tissue residues is the preferred approach; however, the sediment effect threshold may also be used in cases where bioaccumulation has been characterized in an estuary. The threshold values presented here are intended as interim guidelines that should be modified as more data become available. Additionally, because of the uncertainty around many of the factors and assumptions that comprise the single threshold effect values, it is recommended that future studies be employed to help determine a range of acceptable values that would afford protection under various environmental and biological conditions. Published in 2002 by John Wiley &amp; Sons, Ltd.

Preliminary examination of contaminant loadings in farmed salmon, wild salmon and commercial salmon feed

This pilot study examined five commercial salmon feeds, four farmed salmon (one Atlantic, three chinooks) and four wild salmon (one chinook, one chum, two sockeyes) from the Pacific Coast for PCBs (112 congeners), polybrominated diphenylethers (PBDEs – 41 congeners), 25 organochlorine pesticides (OPs), 20 polycyclic aromatic hydrocarbons (PAHs), and methyl and inorganic mercury. The farmed salmon showed consistently higher levels of PCBs, PBDEs, OPs (except toxaphene) than the wild salmon. The mean concentrations in pg/g were 51,216 vs 5302 for total PCBs; 2668 vs 178 for total PBDEs; 41,796 vs 12,164 for total OPs (except toxaphene). The farmed salmon levels are likely a consequence of the elevated level of contamination found in the commercial salmon feed (mean concentrations in pg/g were 65,535 for total PCBs; 1889 for total BPDEs; 48,124 for total OPs except toxaphene). Except for a single high wild chinook value, PAHs were highest in the feed samples followed by the farmed fish and the three other wild fish. The Bio-Oregon-1996 feed of hatchery origin showed a level of PAHs ten times higher than any other feed. The genotoxic implications of such a high PAH level are considered for juvenile chinook salmon. Toxaphene and methyl mercury concentrations were not notably different between the wild and farmed salmon. There was no clear low contaminant brand of salmon feed. The human health implications of eating farmed salmon are considered from the perspective of the current WHO and Health Canada (2000) tolerable daily intake (TDI) values for PCBs. Based on a TDI of 1 pg TEQ/kg bw/day, this analysis indicated a safety concern for individuals who on a regular weekly basis consume farmed salmon produced from contaminated feed.

Growth and Long-Range Dispersal by Wild Subyearling Spring and Summer Chinook Salmon in the Snake River Basin

Abstract Wild juvenile spring and summer chinook salmon Oncorhynchus tshawytscha typically pass downstream in the Snake River as yearling smolts when migrating seaward. A small number of these fish disperse downstream from natal streams into the Snake River as subyearlings and then continue moving seaward. We estimated that the naturally produced subyearling spring and summer chinook salmon that dispersed into the Snake River in 1993, 1994, and 1997 originated from both wild (67%) and naturalized hatchery (33%) stocks. These fish were large in terms of mean fork length (range, 78–87 mm) during shoreline rearing. While moving seaward, they grew rapidly (range of means, 1.0–1.3 mm/d) to yearling smolt fork lengths (range of means, 122–128 mm). The estimated dispersal distances from natal streams to the first two dams encountered during seaward movement ranged from 172 to 810 km. Although we could not demonstrate seawater entry, we believe that the wild spring and summer chinook salmon had met the growth and...

The road to extinction is paved with good intentions: negative association of fish hatcheries with threatened salmon

Hatchery programmes for supplementing depleted populations of fish are undergoing a worldwide expansion and have provoked concern about their ramifications for populations of wild fish. In particular, Pacific salmon are artificially propagated in enormous numbers in order to compensate for numerous human insults to their populations, yet the ecological impacts of this massive hatchery effort are poorly understood. Here we test the hypothesis that massive numbers of hatchery-raised chinook salmon reduce the marine survival of wild Snake River spring chinook, a threatened species in the USA. Based on a unique 25-year time-series, we demonstrated a strong, negative relationship between the survival of chinook salmon and the number of hatchery fish released, particularly during years of poor ocean conditions. Our results suggest that hatchery programmes that produce increasingly higher numbers of fish may hinder the recovery of depleted wild populations.