Why is the standard classroom experiment demonstrating the GHG effect of CO2 wrong ?
classroom: A project and laboratory-based curriculum"
J. Geosci. Educ. 49, 274-279 (2001)
They don't seem to find much difference between closed and open containers. Ben is clearly right to stress the effect of convection/absence thereof. And this must the main effect (as it is for real greenhouses) in the experiment described as shown by another experiment using a heavy gas (Ar) for control without IR absorption bands.
a/ simple calculation show that the increase in temp due to CO2 IR absorption can't account for more than a 5% increase w.r.t. the normal temperature difference (with air) between the container and the environment.
b/ even with sealed containers, convective effects/absence thereof at the interface between the CO2 layer and air can't be ruled out. The important point appears to be the existence of an interface here and not the thickness of the air layer.
@FGR isn't there some contradiction between your points 2 and 5 ? Either CO2 does not settle at the bottom, or there is some stratification but not both.
Actually, this is really what happens. CO2 settles at the bottom and prevents convective cooling. The important point is the existence of a CO2-air interface where convection stops so that the differential warming has nothing to do with radiative GH effect but with the suppression of convection as pointed out by Ben.
This can be shown by replacing CO2 by Argon which has about the same molar mass but has no infrared absorption bands.
Monoatomic gases do not have the vibrational ddf which are excited by IR radiation in CO2. A simple model shows that large temperature rise found in the CO2 container cannot be explained by the IR CO2 absorption bands but can very well be explained by the suppression of convection.
Interestingly, if you keep monitoring the temperature long enough (without introducing more heavy gas) you find that after rising and reaching an almost steady state value, it suddenly drops with a discontinuous slope, corresponding to the moment when the level of the heavy gas falls below the temperature sensor, showing that it is really the existence of an interface which is at the root of the absence of convection. There is a real greenhouse effect here, but it has nothing to do with the radiative GH effect supposed to cause global warming.
Also, I don't see why you said or implied that CO2 would not stay at the bottom because it is warmer. It would have to be a lot warmer to compensate for the density difference (44/29)=(273.15+t)/(273.14+27)==>t=182 °C !
What is difficult is really to understand this convection mechanism and why it is inhibited by the interface and how in the end it overtakes. We should try to estimate the diffusion rate. I'll do that if I have time.
More basically my question was: given that this experiment does not appear credible, do you know a table-top experiment demonstrating the GHG properties of CO2 (apart from measuring its absorption lines directly ?)
For BA, I give it to Ben in spite of the numerous interesting points made by FGR, because he was clearly the first to spell the fundamental reason for the described experiment to be unrelated to radiative GHG effect, namely that the effect here is entirely due to (lack of) convective heat transport.