The only other thing I can think of is that somehow the fluctuation in the sump level changes the back pressure on the overflow somehow, thereby affecting how it's operating. However, that said, this might only make sense if 1) the outlet of the overflow is submerged (which I don't know if it is or isn't) and 2) the water level RISES in the sump, thereby putting more backpressure on the overflow. A level DROP would theoretically decrease the backpressure. So... OK I'm at loss now. Maybe if I saw the phenomenon in person I could observe other variables but using my imagination, I've hit the wall here.
