Drift ALWAYS occurs (though it may be minimal in rare cases and not need correction) when you use two different digital recorders with different internal clocks that are not synced together.
But drift isn't always from two different clocks. Drift also can occur due to the difference between the speed of electrons (constant practically speaking) and the speed of sound, which varies.
The speed of sound is dependent on temperature and humidity, so the inherent time lag between sources (in milliseconds) could be different early on in a set in a cold, uncrowded club compared to later in the set with a croweded, hot, sweaty, humid club -- leading to drift later in the set even if the clocking is the same and the the two sources were synced at the outset of the recording. Generally, this probably isn't worth accounting for since it would have too great of an effect, but the OCD might want to re-sync periodically throughout the show or time-stretch (or better yet figure out exactly where the club is heating up and accelerating the speed of sound and time stretch only during that period).
You're right that changing environmental conditions will affect how sound travels in space, and thus the AUD/mic feed.
In my experience, however, the amount of that change has always been negligible and I have never felt a need to correct for it - particularly when the SBD and AUD tracks I recorded were clock-synced when captured. So, for practical purposes, I believe it is a reasonable operating principle that drift between two non-synced digitally recorded sources will be constant over time, due to the different internal clocks, and that environmental factors are negligible.
One example where there was significant environmental change, but no noticeable drift from those environmental factors, was a PBS set I recorded at Joshua Tree Music Festival last May. The set began a half hour before sunset and ended in the dark. The air temp. was probably ~ 100 deg. and there were not too many people at the beginning of the set. By the end of the set, air temp. had probably dropped at least 25 deg. to the mid 70s. A lot more people also filled in the concert bowl area as the sun dropped. Humidity was pretty constant - bone dry. I shifted and aligned the SBD tracks to the AUD tracks, but found no noticeable drift over the course of the recording.
http://www.archive.org/details/pbs2009-05-16.matrix.flac16I could see where you might get something more noticeable in conditions like the summer in the SF bay area, when warm and drier daytime conditions change to more humid and colder temperatures as the fog rolls in with the sunset. But, I think you'd need to have something like significant change in environmental factors coupled with an Allman Bros.-like 3+ hour long set before drift from any environmental factors becomes noticeable.