Integrated Salmon Hatcheries Can Pose Less Genetic Risk to Wild Populations Than Segregated Programs, Given Imperfect Implementation.

Hatchery programs can provide fishery and conservation benefits, but can also inadvertently threaten wild populations through genetic and ecological interactions. Two common, and non-mutually exclusive, strategies for mitigating the genetic impacts of hatchery programs on wild populations are reducing the number of hatchery-origin fish spawning in the wild and integrating wild-origin individuals into the hatchery broodstock. We compared the robustness of these two strategies to imperfect implementation (variation around target proportions of hatchery-origin spawners in the wild and wild-origin brood stock) using a quantitative population genetic model. Simulations revealed that incorporating wild-origin broodstock was more robust to both short- and long-term implementation errors compared to minimizing hatchery-origin spawners in the wild. Furthermore, relatively low levels of hatchery integration were required to achieve most of the increase in robustness, provided that the average proportion of hatchery-origin spawners was correspondingly low. We checked these findings against empirically observed levels of implementation error by parametrizing the model using data from hatchery programs in Washington and Oregon. These findings suggest that integrated hatchery programs can pose a smaller genetic risk to wild populations than segregated programs, given realistic levels of implementation error.