Thanks for the update with more information!
I think(?) swimming through the overflow during times of flood is probably the most likely source of GSF.
I personally have watched fish swim upstream in surface water flows where the water level was about 1/2 as deep as the fish's dimensions from dorsal fin to ventral fins. I do not know what compels them to swim upstream, but they work hard to do so.
I have also seen the salmon working up the fish ladders in Seattle (through the awesome viewing windows) against a pretty strong flow, so I suspect GSF could also swim up an outlet pipe if it was a direct connection from creek flows up and into a pond. (At least in that case, I understood why the salmon were so strongly compelled to swim upstream!)
It sounds like you modified your outflow pipe to create an "air gap" that invasive fish can no longer breach. Maybe you should try your Rosy Reds and SMB experiment again!
