Salmon hatcheries also artificially boost the quantity of salmon in the stream.
If a salmon hatchery released salmon at the base of a dam, when the fish return and the dam was now gone, they'd just keep going.
However, there's more to it than this, because dammed rivers lacking salmon hatcheries have seen salmon runs start once the dams are removed.
I don't think the old adage that salmon will only return to their original spawning grounds is the whole story.
Some percentage will just enter a different stream. Straying could be a genetic strategy, imperfect behavior, accidental, or some mix. And they're not all necessarily distinct; e.g. the genetic strategy might simply work by reducing accuracy in locating the original spawning ground.
Consider that even before humans streams and rivers would naturally be dammed, diverted, or otherwise change in a way that made it more difficult or impossible to reach the original spawning ground. What would be interesting is if the ratio of various phenotypes, like those that effect straying, has changed in response to the ecological upset caused by humans.
Source: I used to volunteer at one of those hatcheries raising endangered coho and releasing them in the spring. I spent a lot of time chilling in the bushes with NOAA scientists talking about fish.
Stocking can give it a faster kickstart though