Sensor Question...

I've heard that, in cold weather, noise is far less evident, with a CMOS sensor, at higher ISO's than when the temperature is warm.

Is this true and, if so, how's that work?

Sensor noise is caused in part by the sensor sites heating up during long exposures or by the random noise levels of the sites as amplified at high ISO settings (electron activities). Cooling reduces this noise, which is why astronomical sensors are frequently chilled with refrigerants up to and including liquid nitrogen. Now how much improvement you'd see between 75 F and 25 F I wouldn't care to venture a guess. Between 115F and 0 F you might notice a consistent improvement. Basically, think of noise as evaporating water from a pan. If the pan's at freezing point, you won't get much evaporation; as you heat it up, you energize the molecules and some start evaporating off. At a high enough temperature, you'll start seeing the steam coming off, and at boiling point, you'll really notice the activity and the water evaporating.