Two quick points.

First, spawning in a hatchery fails to take into account the selectivity in spawning in the wild. I.e., in a natural setting, the most robust and dominant males and females will seek to spawn with each other, which will have an obvious effect on the genome of their offspring. Hatcheries combine milt and roe more or less at random, which negates this key component of natural selection.

Second, survivability in downriver migration and at sea is only one component of a steelhead's adaptation and suitability to its environment. Fish born in the wild have to survive what is arguably the most dangerous phase of their lives as alevin and fry, whereas hatchery fish don't face the same hazards and "weeding out" process. Hatchery fish are thus given a leg up on survival that they haven't "earned," and therefore traits that are less desireable for the alevin/fry/smolt stages (which would be selected against in the wild) can easily be passed along by hatchery fish. Indeed, having large numbers of steelhead be born and bred in a hatchery will tend to select for traits that are advantageous in the hatchery environment but harmful in the wild -- a tendency to crowd, to be more surface-oriented (and therefore more subject to predation by birds, otters, humans), etc.