This thread has always intrigued me and I always look forward to reading the input from the experts. Hopefully they will chime in and critique my ramblings below.

I have found the following web site informative, and I reference it occasionally to brush up on the ins and outs of dissolved oxygen:

Can anyone comment on the possibility that it is OK if there is not open water exposed to the atmosphere on this lake? In other words, let’s assume that the percent O2 that is entering the water column via the diffuser is 21%, and that relatively little oxygen gets absorbed as it rises. If the air from the diffusers gets trapped under the ice, and spreads between the water/ice interface, won’t there still be diffusion of O2 into the water? Eventually, I would think, the air would make its way to the edges or fractures in the ice and escape, but if the diffusers continue to run, the air tracking along this ice/water interface should be at 21% O2 and diffusion into the lake water would occur. If this was happening across a surface area even larger than a hole in the ice (created by a diffuser) it would not be inconceivable that there is even greater surface area of air/water interface in this scenario compared to one with a smallish hole in the ice. One can even take it further. Let’s say the weight of the ice on that trapped air increases the partial pressure enough that there is even greater diffusion of O2 across the water surface (see link above). If that occurs, then it might, paradoxically, be a detriment to have a hole in the ice. Thoughts?