Icing should not be possible below -15C in cloud where there is no vertical lifting.
If there is lifting (convection) the theoretical floor is -40C and below that supercooled water cannot exist.
The worst temperature for collecting ice is -5C and I have proven that countless times 
In fact I can prove that on just about any flight in IMC around that temperature.
There is a graph out there somewhere (I can’t find it but I know bookworm had it) showing -5C to -6C as the worst place to be.
In general, -3C to -8C is where things really start to happen and the quicker you can get through that range (going up or down) the better.
Below -15C I have never had any ice whatsoever but then I would not be flying in convective wx in heavy IMC in the first place.
There is a small aerodynamic heating effect but at piston GA speeds it is small – of the order of 1-3C – and not worth much, especially as when in IMC in which convection could exist you won’t be flying above Va, will you
I seem to recall that Timothy’s magazine (maybe one of the later issues which are under “embargo” for a period) contains a seriously scary “near death experience” article by a well known pilot who always previously used to write in a certain “private” forum that ice is much less of a problem than is made out to be! My view is that it is only a matter of time before such pilots stop writing that kind of stuff – either because they eventually got really scared or because they are dead. And a fair number of old-timer IR holders are dead (including one who used to share my hangar, and didn’t want his rubber boots fixed), embedded in various terrain, though to be fair they would all have been OK had they had oxygen and actually filed an IFR flight…
I find it very hard to know when icing will exist. Plenty of times I go through a layer at -5 and get nothing even when in it for some time. I think that it important to have a plan B to exit any icing but don’t be terrified of it. It is a meteorological phenomenon that sometimes leads to irrational fear. It is a serious thing and can be very dangerous.
Clearly the most dangerous situation is icing with subzero levels down to the surface. Unless you have proper anti and deicing equipment that is not a place you should be.
especially as when in IMC in which convection could exist you won’t be flying above Va, will you
Umm, probably! Unless I have strong evidence of convection or feel it.
The question about icing and irrational fear it generates, leads us to find personnal rules to follow (and indeed ease our fears…)
It depends on the pilot – and of course on the airplane equipments: anti-ice, de-ice, SE or ME, turbo or not (climbing fast and high), all that should be taken into account.
Peter seems (among other subjects) to have established solid rules of behaving to cope with potential icing situations and actual icing conditions.
He is SE, NA, and prop-deiced.
I would like to ask him a question (as he thought about upgrading his airplane): how full TKS would change your personnal rules about icing ?
Tks !
It’s a very good question, and those who want to create trouble for me 
will point to some past postings of mine where I said that full TKS does not make much sense without a turbo (= a FL250 ceiling).
Full TKS gives you a lot more leeway for climbing up through icing layers, but it doesn’t give you much of an advantage for penetrating fronts, especially warm fronts (many miles of dense wet IMC, tops FL200-250).
It’s a very useful step and should not degrade the overall availability of the aircraft, unlike a turbo which seems to give an almost universal significant extra downtime. Full TKS is 50kg down, a TB21 (turbo) with full TKS is 100kg down which makes it a 2-seater (for me and just one Size 8 girl ) in any practical travelling situation.
Probably what stops me doing it is the lack of installation facilities I would trust to not screw my aircraft up, and the inability to refill it where I am based (hangar rules).
Many thanks Peter. Not so long ago there was a discussion in another thread about icing.
AFAIK and one can read there one is safe below -15C. This AOPA-film describes a possible danger from 0 to -20C.
This contradiction cause some uncertainty in dealing with ice. What is really true ?
It is a gradual scale – no hard limits.
I have always got ice between about -1C and -8C. You could bank on it if you stay in IMC for long enough. It might take 30 mins or it might take 30 seconds before it starts.
Below -8C it is less likely / can take longer to start. But you will still get it.
Below -15C, in totally stratus cloud, at piston GA speeds (i.e. no significant local cooling effects due to local pressure drops) I would say icing is almost impossible. I have never seen a single tiny speck and I have flown in IMC for maybe 2hrs, like that.
In convective cloud (vertical currents) icing is possible down to -40C. That situation is somewhat less applicable to light GA because if you do find yourself in serious convective wx while in large-extent IMC then you might have a “very interesting” flight One normally tries to avoid that scenario.
Below -40C supercooled water cannot exist, supposedly.
The above obviously assumes that your OAT probe is not 5-10C out If it is, you may as well give up… You need to check yours, with a known accurate thermometer. I have a PT100 one, accurate to 0.1C.
Below -15C, in totally stratus cloud, at piston GA speeds (i.e. no significant local cooling effects due to local pressure drops) I would say icing is almost impossible. I have never seen a single tiny speck and I have flown in IMC for maybe 2hrs, like that.
I saw some yesterday at -18. not a huge amount but it was there. TAS of 210 so not super quick.
Over the last few days I’ve done some armchair flying looking at the DWD (German met office) icing forecast and autorouter.eu GRAMET. That makes me ask two questions.
1) It appears that for DWD every cloud at below 0C temperature makes them show light icing. In cases where areas and levels were shown as moderate or even severe the GRAMET shows icing as well. When planning a flight in a non FIKI TKS aircraft should I wait it out or can I trust GRAMET and the interpretation of vertical soundings (skew T) more than DWD’s icing forecast?
2) What about planning and performing a flight well over – more likely – or under an area with icing as depicted on a GRAMET. Imagine a situation of DWD showing light icing all over the middle of Germany from 2500ft to FL80 and GRAMET only shows a certain layer within that area as icing. I understand that in the case of an emergency like engine trouble I would be descending into the icing area but other than that staying on top of it while enroute should be fine.
I use three sources:
Gramet
The RocketRoute briefing pack
The Jeppesen icing forecast
Of these, the RocketRoute pack is usually the most conservative and I have found it often predicting ice that isn’t there. The combination of the Gramet and Jepp site seems usually to be pretty accurate.
[URL moved to a clickable link since some URLs don’t work properly if pasted-in alone]
