I work as a machinist and do a lot of aero space work were we need to hold very tight tolerances. I have literally made parts that were within tolerance at 78-82°F that were no longer within tolerance at 68.5°F. based on this i would think the increase in velocity is caused by the bore size becoming smaller and increasing pressures when cold. Based on my research the linear thermal expansion coefficient of 416ss is 6.4 and going from 68°F to 0°F would cause the bore diameter to shrink by .0002 I have worked a lot with 15-5, 13-8 and 17-4 PH stainless and based on my experience with these materials I would think the bore size would change by more than the .0002
I'm not a metallurgist, but this has been covered before by some "in the know".
Iirc they came to the conclusion that a cold bore to cold bullet will not create extra pressure as they did calculations that both lead and copper (bullet) will shrink more than the steel bore diameter will so theoretically you should have larger clearances/less pressure between bullet and bore when cold-- even so, iirc there are larger tolerances in the actual machining of bores and bullets than the cold Temps would change it anyway.
It comes down to other variables such as "bullet weld", powder temp instability, primer temp stability, powder lot variances, etc.
OP said he thinks he figured this out being a seating depth issue anyway.
Problem is there are so many variables that come into play its hard to prove which one is affecting the velocity, so testing and backing up with multiple testing data sessions is necessary to know what happens to your rifle/ammo combo in cold vs hot temps.
Some people have seen re26 spike above 85 degrees, others have not-- is that due to lot variances or other variables? Since there is no real practical way for us to chemically test powders we just jave to test and record data for our combo