The Adverse Reproductive Consequences of Supplementing Natural Steelhead Populations in Oregon with Hatchery Fish
Mark W. Chilcote
Abstract: The proportion of wild fish in 12 mixed populations of hatchery and wild steelhead (Oncorhynchus mykiss) was evaluated for its relationship to mean and intrinsic measures of population productivity. The population mean of
ln(recruits/spawner) was used to represent mean productivity. Intrinsic productivity was represented by values for the Ricker a parameter as estimated from fits of spawner and recruit data. Significant regressions {p < 0.001} were found
between both measures of productivity and the proportion of wild fish in the spawning population {Pw}. The slopes of the two regressions were not significantly different {p = 0.55} and defined a relationship suggesting that a spawning
population comprised of equal numbers of hatchery and wild fish would produce 63% fewer recruits per spawner than one comprised entirely of wild fish. Study findings were not sensitive to likely levels of data error or confounded by extraneous habitat correlation with Pw. Population status assessments and conservation monitoring efforts should include Pw as a critical variable. For natural populations, removal rather than addition of hatchery fish may be the most effective strategy to improve productivity and resilience.
Between 1978 to 2000, depending on the steelhead population, regulations were implemented that made it illegal for anglers to keep wild fish. Based on information presented by Hooton (1987) and Reingold (1975) it was assumed that 10% of the wild steelhead caught and released under these new regulations died as a result of handling stress. Therefore, the fishery mortality rate for each population after implementation of the .wild release. regulations was calculated as 10% of the estimated harvest rate for the period prior to the regulation change.
Discussion
Natural productivity in 12 populations of Oregon steelhead was significantly influenced by 4 variables, one of which was the level of hatchery fish in the spawning population. It appeared the presence of hatchery fish depressed overall population productivity, reduced the number of recruits, and lowered the genetic fitness of wild fish. These negative effects were insensitive to the type of hatchery fish involved. Although the hatchery fish represented in five of the study populations were from hatchery broodstocks developed from the local wild population and managed in manner to avoid domestication, the advantages of this strategy as purported by (Cuenco et al. 1993, Flagg et al. 2000) were not apparent. Even if the acceptable statistical significance level for the hatchery broodstock variable was raised from 0.05 to 0.10, thereby enabling the inclusion of this fifth variable in the production regression model, its influence would have been minor (Table 2). For example, using this 5-variable model, the estimated reproductive success for .wild-type. hatchery fish would have been 0.335, whereas the reproductive success for .semi-domesticated. hatchery fish would have been 0.293. Overall, these results demonstrated that the use wild fish for hatchery broodstocks as a means to create hatchery fish that are reproductive equals of wild fish in the natural environment does not appear to be a promising endeavor.
The recruitment curves developed from the results of this study (Figure 3) illustrate that the number of naturally produced fish can be expected to decline as the presence of hatchery fish in the spawning population increases. It appears that naturally spawning hatchery fish, regardless of broodstock type, are a potential impairment to the subsequent production of recruits. When more than 10% of the naturally spawning population is comprised of hatchery fish, this impairment is not trivial.
There has been considerable interest concerning the use of various types of hatchery programs to help rebuild and restore depressed populations of wild fish (Waples 1991; Olney et al. 1994; Cuenco et al. 1993). Sometimes described as .supplementation. (Sterne 1995) this approach has both intuitive and theoretical appeal as reflected by Cuenco (1994) and Flagg et al. (1999).
However, based upon the results of a variety of simulations using the productivity model developed from the observations of this study, it appears that supplementation may be an ineffective tool for recovering depressed populations of wild fish. Such depressed populations appear to respond weakly to the addition of more spawners if they are hatchery fish (Figure 4). In addition, a byproduct of the supplementation strategy is a decline in the genetic fitness of wild fish. Although the magnitude of this fitness decline is not large when the level intervention is low (-3.9% when Ph = 0.10), it rapidly increases once the resulting proportion of hatchery fish becomes greater than 0.50 (Figure 6).
The results of this study suggest that naturally spawning hatchery fish, regardless of broodstock origin and quality, are ineffective at producing offspring that survive to adulthood. The observation that hatchery fish from .semi-domesticated. and .wild-type. hatchery broodstocks have essentially the same reproductive success makes a genetic explanation of this observation more difficult. It would seem that the .wild-type. hatchery fish should be genetically more similar to the local wild fish. Therefore, they should have better reproductive performance than hatchery fish from .semi-domesticated. broodstocks. However, as speculated by Reisenbichler and Rubin (1999), it is possible the genetic change that occurs as fish adapt to the hatchery environment may have more significance in terms of reproductive success in the natural environment than do the genetic shortcomings related to the geographic origin of the hatchery broodstock. Should this be the case, any genetic effect of broodstock origin may be difficult to detect.
An alternative genetic mechanism, unrelated to stock origin or selective changes, may be operating to create the reproductive differences between hatchery and wild fish. Genetic differences may arise from the common situation that returning hatchery fish are the offspring of substantially fewer parents than is the case for wild fish returning to the same basin. For example, approximately 160 fish are used annually as broodstock for the North Umpqua summer steelhead hatchery program. In contrast, the number of wild fish that spawn naturally in the North Umpqua basin is typically greater than 3,000 fish (Appendix 1). Therefore, the genetic base for the hatchery return is approximately 80 families, whereas for the wild fish it is roughly 1,500 families.
In summary, this study found that hatchery fish are poorly suited to reproduce under natural conditions and when allowed to do so have an adverse impact on the recruitment and productivity of natural steelhead populations. These results confirm the study supposition that hatchery fish are maladapted for reproductive survival in the natural environment. This confirmation is robust and applies to hatchery fish regardless of hatchery broodstock origin and various attempts to mimic the genetic and reproductive characteristics of wild fish. Further, it appears supplementation of depressed wild populations with hatchery spawners is an ineffective conservation strategy. Such efforts can be expected to yield only minor gains in the number of naturally produced recruits and cause a loss in the genetic fitness of wild fish. Therefore, the results of this study are consistent with the view that the most effective conservation role for hatcheries is one of impact avoidance, not direct intervention. It appears that limiting the proportion of hatchery fish in naturally spawning populations to less than 0.10 is an appropriate strategy to achieve this conservation role.