[sphelps]

I know what you're getting at but a single overflow still won't produce turbulent flow, it will still be laminar so a higher velocity would still be better. Ideally you would want the smallest "linear dimension" without entering the transition or turbulent flow regime, a Reynolds number less than 2000. A lower velocity will result in lower surface tension, and could result in surface build up in some areas. Kind of like a vacuum cleaner, a large area results in low velocity and it can't bring in particles at a distance, decrease the area and the suction is magnified and you can bring in particles from larger distances.

The main point here is why waist the space with two overflows when one is good, possibly even better. Same idea as pipeline design, you want laminar flow but why use 12" when 6" works? Bigger isn't always better.

In the case of a weir I was under the impression the idea was to primarily increase the water level of the river upstream. The reason the overflow length is so long, usually covering the full river width, is to minimize water level fluctuations upstream due to inconsistent flow rates. Never heard of anything relating to surface skimming, could be wrong though.