I know someone who disputes the idea that the Earth's atmospheric temperature would average -18 deg C in the absence of greenhouse gasses. He maintains that conductive heating would warm the air and convective transport mix it, and, since the gasses are not greenhouse gasses they would not lose the heat by radiating it away, and so the heat could build up.
The thought experiment runs thus: a planet in earth's orbit without water and an atomosphere of, say, oxygen and nitrogen.
It seems to me that being such a small percentage (15%) of the incident radiation striking the Earth (even if, in the absence of GHG, the incident radiation would be higher) the effect could not possibly build up because the conduction would work the other way at night and most of it would, in turn be radiated away into space.
In short, he is saying that a significant portion of the effective warming we see in the atmosphere is attributable to this effect, making the basis for calculating the greenhouse gas effect suspect.
I don't have the mathematics to estimate this effect on balance, and it seems to me that it must have been accounted for previously, but I do not know. I would particularly appreciate a reference to a paper or article that discusses this aspect of atmospheric warming.