The speed of sound inside the barrel matters to performance independently of starting pressure.
The speed of sound determines the speed of expansion of a volume of gas. The molecules can't move away from one another any faster than their individual speeds due to temperature (which is the speed of sound). This is one of the reasons air gun efficiency drops as pellet speeds approach the speed of sound at normal temperatures - like the ones inside the barrel and tank. The speed of sound is much higher in a hot gas, like the one produced by gunpowder.
So maybe it's just that: the same amount of reg pressure pushes less on the pellet on a cold day, because the gas expansion is slightly less able to catch up with the pellet as the pellet approaches the same absolute final speed in the barrel. The change in the speed of sound is less than 0.5% between 0 and 20 deg. C - 1085 fps vs 1124 fps, but the change of the difference between this and the pellet (average) final speed is larger (assuming it matters). Maybe it's like having a slightly shorter effective barrel length, on a colder day.
There are ways around this problem, if it is a problem, on the air gun design side, but I think people mostly stumble on it by chance and not necessarily by design intent: