Two Decembers later the same thought bothered me again, and this is what trickery to do for measuring temperature of a cold night when external temperature this is at 1.5 Celsius or about there?
Handheld multimeter lucking cold junction adjustment for thermocouple compensation and practically this is a dead end.
But the answer and solution came to me by thinking that
Zero Celsius calibration this is equal to a K type thermocouple shoved in to the snow.
Therefore workaround suggested this is leaving outside a glass of water for an hour and then measure water temperature indoor .... this workaround will increase measuring accuracy because of all reasons mentioned at my post above.
Second interesting testing by shoving a thermocouple in to the snow, this is inspection of multimeter calibration.
Today I do own three U1272A, and Zero Celsius calibration after testing all three, it is measured at +/- 1 Celsius.
First meter measurement: minus 0,5
Second meter measurement: 0,0
Third meter measurement: Plus 1,0 Celsius
Naturally any measured difference when comparing identical multimeter, this having to do with precision of calibration adjustment when this was performed at factory or elsewhere.
From the other hand any calibration report this serves and the purpose of consultant for us when we seek ultimate measuring accuracy, in a way we should mathematically compensate any detected deviation marked at calibration report from our own measurement.