I've long believed and oft expressed that terminal harvest is the answer to the management issues mixed stock fisheries present. Several of the challenges have been addressed in this thread, but I think there are ways to work around each of them. As usual, getting enough momentum in the state and federal governments to make such changes would probably be the hardest obstacle to overcome.

First, a couple of caveats/assumptions:
1. All hatchery fish are intended for harvest and re-seeding future generations.
2. Some component of the wild salmon population should be available for tribal and sport harvest in the absence of open ocean, mixed stock, commercial fisheries.

I think the scenario I would most favor would be one where, as Carcassman mentioned, we make some decisions to manage streams with few or no wild fish strictly for harvest and build hatcheries in the lower reaches of those streams. These hatcheries should be owned and operated by seafood processing companies and/or treaty tribes, with federal and state government footing some percentage of the construction and production costs (to compensate for any hatchery fish that get harvested in sport fisheries). Ideally, these streams would be located in coastal areas, to ensure the best quality product (from a fat content perspective) possible. If a stream is managed strictly for hatchery production, it can be assumed that every returning fish is available for spawning the next generation or harvest. Any fish not in marketable condition could be used in nutrient enhancement programs in streams being managed for wild stocks (with a per x pounds government incentive available to companies who choose to participate).

The next step would be to cease all hatchery operations on streams with wild populations, and manage those streams strictly for wild fish. Spend some of the money saved by closing the hatcheries on habitat improvements. In the absence of mixed stock fisheries, my hypothesis is that we would see much larger numbers of wild salmon returning to our rivers, and they would provide meaningful in-river opportunities for sport and tribal fisheries. Clearly, there would still be the problems of managing the tribal gillnet fisheries and sport fisheries so that struggling stocks don't get overfished, but I think that would be a much more surmountable task than trying to effectively estimate encounter rates in mixed stock, open ocean fisheries.

To be clear, I am proposing effectively ending open ocean, commercial fisheries as we know them today. Obviously, that carries a significant human impact, as it would eliminate a lot of jobs. Ideally, I think the scenario I described could also create a lot of new jobs (at the terminal hatchery sites), so that might absorb some of that impact. Ceasing hatchery operations on rivers shifting to a wild fish management paradigm would also land a lot of state hatchery workers on the unemployed list. Perhaps the new hatcheries could employ some of those workers as operations specialists and managers - they would clearly be well-qualified for those positions.

Such a shift would also change the nature of tribal and sport fishing. In the beginning, that might not all look very good, but assuming the theory that more wild fish would be available holds true, both sport and tribal interests should end up with better, less-complicated fishing opportunities. One sport fishery that would become difficult to justify would be the open ocean fishery. Certainly, losing that fishery would seem unacceptable to a lot of sport fishers, but if the goal is to eliminate mixed stock fisheries, it would be hard to argue that fishery should remain open.

Again, I see plenty of issues here, but I think the potential benefits of moving to a system of terminal harvest far outweigh the negatives, where the very opposite appears true under the status quo.