Sometimes the sizing of overflow connections requires more than just the velocity through the outlet nozzle. You could use the procedure suggested by JStephen and this would give you an outlet of about 3" ID for an allowable head of 4" above the centerline of the overflow nozzle. This would work well if the inflow is steady, and the overflow simply falls from the outlet nozzle through the air to an open drain.
The complications come in when you have varying inflow, and the overflow goes somewhere where you do not want entrained air. Another potential complication is the formation of a siphon.
Imagine that you have put a 3" outflow nozzle near the top of your tank with a pipe down the side of the tank discharging into an open drain. Now you start filling the tank. As the level reaches the bottom of the overflow nozzle water starts flowing out. But the rate of outflow is less than 85 gpm so the level in the tank continues to rise. As the level gets towards the top of the nozzle air will be entrained with the exiting water and you will have gurgling and maybe a bit of vibration in your pipe. This may or may not be a problem.
Once the level reaches the top of the nozzle no more air can be entrained. When the water level reaches approx 4" above the centerline of the outlet nozzle the outflow will match the inflow and you might think this is now a steady state - but it is not. What happens is that the pipe running down the side of the tank starts to fill with water and causes a siphon sucking the water down. At this stage the overflow rate jumps to around 300 gpm and the level in the tank drops rapidly until air is entrained once more and the siphon is broken. Now the level slowly increases and you get into a cycle of alternating high and low flowrates. Again, this may or may not be a problem.
The usual practice with overflow pipes is to size them so that the level in the tank never goes above the centerline of the outlet nozzle, and the piping system is designed to cope with (or prevent) air entrainment. Try to get a copy of the article "Designing piping for gravity flow" by P.D. Hills which was published in Chemical Engineering, Sept 5, 1983, pgs 111-114. This is the best reference I have seen for this type of piping.
Katmar Software
Engineering & Risk Analysis Software