I would suspect flame adjustment should compensate for all other variables; more O2 in atmosphere is compensated for by adjusting the valve on the torch? I never thought that hard about the process
I believe a reason for the popularity of electric arc welding is the elimination of a subjective variable; flame adjustment... I may be all wet... but in addition to the size of the flame is it's mixture and characteristic of carbonizing or reducing, which I could only very loosely compare to electrode "stick-out" in GMAW. The flame is of different temperatures according to where the measurement is taken.
How severe were the ambient temperature variations? I would not be surprised if the temperature issue related to the temp of the steel being welded, not the air. As its temp lowers, its tendency to draw heat away from the weld increases, as do other issues in the HeatAffectedZone and indeed the entire weldment. Fortunately, with a torch you can adjust the temperature of steel quite conveniently. I find it much more troublesome to weld in winter cold.
What is your fuel gas?
Does absolute pressure in a tank vary with the barometer? If so how would that relate to [components of] the fuel gas' critical pressure?
Is this an anylitical exercise? Curiosity is valid, but there probably is a reason few (if any) welders concern themselves with these kinds of details. If the issue of primary concern is the weld it's probably OK to ignore the data you collected so far and focus on the nature of the difficulties/deficiencies in your welds... what's going on wrong?
Regards,
D