I think this is the relevant posting from that thread back in 1999. Basically during active fermentation the CO2 will act like a blanket but not afterwards. I would assume though that as long as some CO2 is produced the bubbling of it through the airlock will eventually displace the O2 and not let any to re-enter. So the shake of a carboy to force some CO2 out of suspension is valid just not sure how much would be required.
Second point is that the CO2 and O2 do not separate out but wil mix in all but the most ideal situations, near 0 deg K perhaps?
Don
-------
This is not an "equlibrium problem". It's a good example of "pseudo-steady state". I don't expect that to mean much to most folks, but it realy says it all. Or in plain english......
There are a few things that you need to consider. 1) The foam layer at the top of the wine early in fermentation is nearly all CO2 (by volume), there is no air in the bubbles that comprise this foam. 2) On a volume basis the total amount of CO2 produced by the frementation of grape juice is many times larger than the fermenter head space, this would be true even if you used a plastic bucket that was only half full. 3) The system is not static you are continuously introducing CO2 at the surface, in order for O2 from the air to reach the surface of the foam layer it would have to pass through a continuously moving front of newly formed CO2. As long as there are no breezes accross the surface, this font of emerging CO2 is enough to keep the O2 concentration very low close to the surface. Thus the term "CO2 Blanket", but rather than being applied from above as with most blankets it is regenerated continuously from below. So even a loose fermenter cover is adaquate early in fermentation because it just has to prevent air cross air breezes, the emerging CO2 does the rest. But, even with a cross breeze, the foam layer would still offer a lot of protection. 4) Finaly, the situation discribed in 3) is stable for days because of 2). I knew I needed 2) for something.
Later in fermentation, the foam dissapears and CO2 generation slows enough so that O2 can reach the surface of the wine by diffusion (diffusion = net effect of random molecular motions, i.e. no breeze required) This is why its important to rack into a container with a small headspace and an air lock.
Finaly, everything in your discussion below would be true if the rate of CO2 production was low enough as to be neglected. In such a case, the entire head space would become uniform with the air above.