The genetics of a managed Atlantic salmon stock and implications for conservation.

Abstract

Numerous populations of wild Atlantic salmon have declined in recent years. The Atlantic salmon in Girnock Burn, an upland spate tributary of the River Dee, Scotland have been monitored intensely by government scientists since 1966. The burn is equipped with upstream and downstream traps, which have enabled monitoring of juveniles leaving the burn and adults returning to it since 1966. Recently, due to a decline in numbers of female returns, a supportive breeding program was instigated. Using microsatellite-based DNA profiling, this study exploited existing and novel tissue samples to investigate aspects of Atlantic salmon biology and conservation. A panel of up to 12, mainly highly polymorphic, microsatellite loci were employed to derive allele frequency data and to resolve parentage in egg, parr, smolt and anadromous adult samples taken between 1991 and 2004. Genotyping error was investigated and rectified where possible. Overall, the detected error was low (c.0.5%), providing confidence in subsequent population and parentage analyses. The error rate involved in estimating the age of salmon in Girnock Burn from scale readings was also estimated (c.2-8%). A study of the dynamics of natural spawning, based on the parentage of parr, confirmed that multiple matings by anadromous returns of both sexes were prevalent. Not all anadromous returns were apparently successful spawners; data from parr and existing redd samples failed to detect a contribution from 35% of males and 29% of females. An important aspect of the work was to determine the success of the supportive breeding program. Results showed that, in comparison to natural spawning, the program gave a more complete and even representation of adult spawners in offspring. In addition, there was no detectable difference in the output (number of smolts) of the two schemes when the number of eggs used in each was taken into account. The distribution of juvenile kin (parr aged 1+) within the burn was determined, which revealed clustering of full and half sib groups. This was found to impact on standard population genetic analyses. Adjacent samples (n = 50), each sampled over a c.1.5 km stretch of river were shown to exhibit significant allelic differentiation, while samples from individuals selected at random over a 7.5km stretch did not. Parentage analysis of adult returns showed that the number of returns likely to be philopatric was higher than would be predicted solely from physical tagging data. This was attributed to ‘leakage’ of the downstream parr/smolt trap. An initial investigation into the role of mature parr in adaptation of populations to the environment was made, although sire type (i.e. anadromous male or mature parr) was not found to affect survival in the freshwater environment in this case. More research into this aspect is warranted, particularly with the possible impact of predicted climate change on male parr maturity. A comparison of genetic diversity through time (measured by allelic richness) revealed no detectable change between 1991 and 2004. Estimates of the effective population size using different genetic (temporal) methods were associated with a large degree of uncertainty, and were surprisingly high (ranging from 595 to 1992) c.f. demographic based estimates (ranging from 95 to 144), which was likely to be due in part to violation of assumptions made in the calculations. These findings have highlighted a range of avenues for future lines of research, should aid in the management of Atlantic salmon within Girnock Burn and assist in the design of sampling regimes.