Interesting, but I wonder if the speed of sound is the reason. Speed does not equal frequency. The speed of sound is the speed with which a sound propogates through the air - e.g. the delay in hearing a thunderclap.
However, if the resonant frequency of a metal string changes with temperature, then we are onto something Actually, it has to change, because metal contracts with cold, and that puts more tension on the string. The question is if it is enough to change the tuning significantly.
Speed doesn't equal frequency, but velocity equals the wavelength times the frequency, and the frequency depends on the source of the wave, so if velocity increases, the wavelength has to decrease. This would make it go sharp in heat though, not in cold, and it'd have to be an extreme difference for it to even matter, that's where I ****ed up. I also thought I was checking the tuning quickly enough for nothing to heat up or cool down, but I guess the metal conducts heat so quickly that I didn't remove them from the equation at all. I guess now the thing to check is the tuning at different time intervals in the cold.
I'll try it again, this time with a timer for 3 minutes in the cold. If it's even sharper than it is as soon as I get into the cold, then I'll know that something's losing heat and shrinking(probably the strings, because then they would go sharp; if the wood were losing heat, it would contract and the guitar would go flat, unless maple contracts faster and the neck tension changes, but the wood would take longer than a few minutes to get cold).
Edit: Yup, not only did it go sharper over time, each string's pitch changed at a different rate, in order of size. It has to be the strings getting looser or tighter with temperature then.