> Do we understand the mechanisms of this "genetic memory"

I don't think there is particular evidence for "genetic" memory here. The salmon were already further down river, they just kept swimming upstream. While most salmon do return to the place of their birth, a small percentage always stray, which is how salmon are able to colonize new habitats and survive things like ice ages.

>While most salmon do return to the place of their birth

I wonder to what degree that is even true. Like sure they probably return to the same rivers, but how far up the river they swim is likely unrelated to where they were actually born. If you extend that river further or introduce side streams that didn't exist when they were born, they're probably just as likely to end up in one of those places.

Exactly that. They also need the right kind of gravel to spawn in. The kind you find in mountain streams.

Glad they are doing well.