I think there has been enough evidence and science posted on this thread - over the past 33 pages or so - that some broad-scale assertions/assumptions can be made.
Every site has overlapping temporal and spatial scales - and potential and realized impacts that fluctuate - and so those potential and realized impacts vary. They vary annually, seasonally, weekly, daily, and hourly - dependent upon many complex interacting factors such as tidal flows, estuarine flow, regional weather and watershed-scale run-off, large-scale climatic processes like PDO/El Niño and La Niña/CC, migratory and holding behaviour of both juvenile and adult salmon - stocking densities and fish health of the cultured fish and the wild fish - and position of that particular site within this mix of interacting and synergistic variables.
If we average these impacts over time and geographic area - a trend develops - where Ford and Myers (2008) found: "reductions in survival or abundance are greater than 50%" on wild salmon from farmed salmon operations: http://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.0060033
A responsible regulator would take these findings with the seriousness they deserve...
So, your hypothesis is that if fish farms were removed, we could expect a 50% increase in the returns (of course this is 2008 article). Obviously, the next question would be to examine all the data over the last 30 years and see if the runs were 50% larger than they are today. That would help with first hypothesis - that is to prove the capacity is there. The second would be to examine the commercial and sport fishing takes and compare that to the data from 30 years ago. Did we diminish numbers so substantially because of over fishing that the runs are still trying to recover? We did have salmon run collapses in the 80's - any oldtimers want to talk about the collapse of salmon from the Georgia Strait or from Campbell River? Finally, we need to understand how the Fraser River sockeye run of 2010 was second largest in a century - this doesn't make any sense unless the data shows it should have been twice as big as it was. If we can match this data and it proves the statistical correlation you are looking for, then you would have a mighty fine argument.
By preforming statistical regressions on your data sets, we can actually test your ideas and prove the point. Again, AA, I am not actually commenting on the validity of all the claims - it is whether they are hyperbolic and don't prove what they are supposed to prove, namely that the source of our troubles for salmon are FF's. I actually worry that the real source of the problem is rising temperatures in our streams that are making them non-conducive for the spawners. We all talk about the sockeye circling out front waiting for a rainfall to cool things down so they can make there push upstream. What if that is the most significant problem? Closing all the FF's in the world won't change rainfall or ocean temperatures and if all they do is make us feel better, lets put that energy into something that works.