Here's BPA's 2003 assessment.
LINK CONCLUSIONS
1. Fish from old hatchery stock consistently have very low fitness (usually less than 50%
that of wild fish) when breeding in the wild. The fact that Hold x W crosses consistently
produce fewer offspring than W x W crosses (Table 4) suggests that having Hold breeders
in a system might lower the fitness of the wild population. Whether the surviving wildborn
offspring of such crosses re-establish “wild” levels of fitness after one full
generation of selection in nature remains to be tested.
2. Fish from new, conservation hatchery stock have fitness that is about equal to that of
wild fish (less than wild in two years, greater than in the third year). The same pattern is
apparent whether one examines the relative fitness of individual parents or that of pairs
that left at least one offspring. The similar fitnesses Hnew x W and W x W pairs, suggests
that having Hnew fish in the system is probably not obviously dragging down the fitness of
the wild population for genetic reasons (as might have been expected under some models;
e.g. Lynch and O’Hely, 2001). Thus, the conservation hatchery program appears to have
added a demographic boost to the population without having obvious negative genetic
consequences - at least in regards the effects of domestication selection and mutation
accumulation that should occur in the hatchery. We have not yet conducted a formal
analysis of the effect of the hatchery program on the effective size of the wild population
(e.g. Ryman et al., 1995), but the high levels of microsatellite diversity we still observe in
both runs suggest that reduced effective size is not a problem.
3. The surprisingly large number of missing parents, and the fact that most missing
parents are fathers (Fig. 3), suggests that precocious parr or resident trout are obtaining
matings that produce anadromous offspring. Alternate explanations for offspring that
lack both parents include a large number of unclipped hatchery fish or wild strays
entering the system.
Future Work
We hope to continue genotyping fish through the rest of this decade. Additional
questions we plan to address include:
(1) Do F1 progeny (born in the wild) of Hnew x W, Hnew x Hnew and W x W winter run
parents differ in their production of F2 progeny?
We know from our current analyses that all three types of matings occur on the spawning
ground, and that all three types of mating produce offspring that return to spawn as
adults. F2 offspring of those winter F1s that spawned in the late 1990s are now returning
(see Fig. 2). If we continue sampling through the end of the decade we will have a large
number of returned F2s from multiple brood years with which to test the relative fitness
of different types of F1s (Fig. 2). Given the apparently high fitness of Hnew hatchery fish,
our expectation is that the three types of wild-born F2’s will have similar fitnesses
.
(2) Selection to maintain the difference between summer and winter runs:
What is the rate of hybridization between the runs? What are the phenotypes (run time,
size, freshwater residency) and actual fitnesses of any hybrids?
(3) Selection on measurable phenotypic traits:
We can use standard selection gradient analysis (Lande and Arnold, 1983) to analyze
fitness as a function of body size, run time, age and freshwater residency (known from
scales), after controlling for hatchery/wild genetic background.
(4) Quantitative genetic parameter estimation:
From our pedigrees we can estimate the heritabilities of, and genetic correlations among
any measurable phenotypic traits. We can also estimate the average breeding value for
each trait in individuals of HxH and WxW genetic background, in order to test whether
genetic changes in the hatchery, and subsequent mating with wild fish, could be changing
phenotypic distributions in the wild population (Ford, 2001).
(5) Parental contributions of resident, non-anadromous fish
We sample all potential breeding adults passed over the dam, and we know from our
ground truthing experiments the expected rate of mismatching owing to experimental
error. Therefore, unassigned offspring are either wild strays from out of the basin, or
were parented by resident fish (non-anadromous O. mykiss, or precocious parr). We will
use likelihood methods (Rannala and Mountain, 1997) to attempt to determine the most
likely source of missing parents (of offspring that only match to a single known parent),
and whether fish lacking both parents are most likely to be Hood River wild, Hood River
hatchery (unclipped) or immigrants from adjacent steelhead populations. Because we
sample all anadromous parents, the Hood River is an ideal system in which to ask
questions about the rate of parentage from resident fish and about the sources of those
fish.
(6) Effective size estimation
From the pedigrees we can obtain direct estimates of the effective size (Ne) of each
population over time. These data will be used to estimate the impact of hatchery
programs on the effective size of the wild population and to provide basic parameter
estimates such as the variance in family sizes (number of returning adults) for hatchery
broodstock, for H fish in the wild, and for W fish in the wild. These are important
parameters that are unknown for most populations and can be very useful for estimating
Ne and the effects of supplementation in other steelhead populations (e.g. sensu Ryman et
al., 1995). We can also use our system to evaluate the accuracy of indirect methods for
estimating effective size (e.g. Waples, 2002; Anderson et al., 2000). If the indirect
methods give very different values from the pedigree-based estimates, then we can ask
what assumptions of the indirect methods cause the difference. Note that because of our
ability to sample all potential anadromous parents, we can take into account the
contributions of non-anadromous, resident fish in our calculations.
Previous Topic
Index
