Gonna have to disagree with the Doc a bit. A well run hatchery program designed for harvest production does not need wild fish infusion. The broodstock needed of a large program is huge; well in excess of many of the existing wild stocks. The inbreeding of a wild run with a couple hundred spawners is way higher than say, Hoodsport chum to 20-50K spawners. You get in trouble when you have a small number of spawners. Period.

Plus, a well-run hatchery program selects for fish that as juveniles perform well in confinement and the well post-release. What survives spawns. But, if those hatchery fish spawn in the wild they bring hatchery adaptations to the wild, which is bad. You can't make one fish do both well. It is either wild or hatchery.

Hatcheries are significantly more efficient at converting water to fish. You need less water and less land to make fish there. Plus, since you convert more eggs to fish, you need fewer spawners (more harvestable). To use Hoodsport again, with Finch Creek and that few acres it produces more chum that any single WA river (management has something to do with that, too-but I digress). A Hoosdport-style operation on (say) the Skagit could put a bigger chum run there with lots of additional harvest. No ecosystem benefits but you could log/farm/develop the watershed.

If you want well managed fisheries on known numbers of fish you will either have to fish in the rivers and bays, where stocks are separated and updates can be done or you can invest in massive, daily, stock identification methods in marine waters (they used to do this with Fraser sockeye-mostly by scales) and do updates outside with daily changes in stock compositions. It would be much more complex from a people management perspective but could be done. Also cost a lot.