Adding to Roy's and Mike's excellent responses...

"O. Sami Saydjari" wrote in message
...
This question is a question on physcial phenomena, NOT on regulation (so
I am starting a new thread).

As I understand it, icing happens between +2C and -10C. Assuming this
is correct, I have three questions.

Bear in mind that the upper and lower limits are for different reasons.
Above the upper limit, no part of your airframe should be cold enough to
freeze the water. Below the lower limit, the water has already frozen, so
it doesn't form ice on you. In cumuliform cloud, supercooled water persists
well below -10 degC.

(a) Does that mean one is relatively safe if the surface temperature is
below -10C (and there is no temperature inversion, meaning that the
temperature is known to decrease as one goes up in altitude)?

Such surface temps often happen because of radiative cooling of the surface.
It's likely that there *will* be a temperature inversion as the air mass
doesn't usually cool that much, except in places that stay cold for very
long periods. It's always worth checking the temps aloft too.

(b) Is icing from 0C to +2C a possibility only when your aircraft skin
is colder than 0C (probably because you are descending from altitude)?
Or is there some other condition where this is possible?

In principle, even in equilibrium, parts of your wing where there's low
pressure can be below zero while your OAT probe is above zero. The
magnitude of this effect depends on wing loading, so it tends to be much
greater for big jets than GA aircraft.

Julian Scarfe