Is there a place with sample images of clouds with high water content that are likely to produce ice?
It would be useful to know what to look for other than making the experience over time oneself.
I’m no meteorologist but I fear it is not as simple as that. What mainly produces airframe ice is supercooled water. This is water which had been cooled below freezing point but could not condensate on any condensation core into ice.
This means the most likely icing will occur in humid air which has been cooled very recently. The longer the air is below freezing the more likely all the supercooled droplets will have found some core to condensate on.
As a result the worst are probably convective clouds. If you fly through the tops of building cumulus clouds below freezing level you will very likely pick up ice. If you fly in a solid more fog like cloud there will usually not be much ice.
I knew about the building cumulus and I actually had a chance to see the effect already.
Basic instinct tells me that a dark looking cloud is probably full of water. But it might not be the case or is it?
I believe dark looking means it is thick and thus only little light gets through it from above.
CBs and TCUs I will avoid no matter what. But I’m wondering what else is out there that I should avoid or at least be prepared for? Given a TKS wet wing aircraft.
Unfortunately other than isolated TCU or CBs you won’t easily see much else. From what I have experienced it can be very hard to guess whether you will pick up much ice. In my trip last week that I wrote up you would think I would have accumulated a lot of ice but it never arrived.
a dark looking cloud is probably full of water
As I remember, the cloud looks dark when its water drops are at their largest, at the end of their growth process, just before coming down.
Whether that says anything about humidity, i.e. percentage water to air, I don’t know.
As I’ve written before many times, my greatest ever ice collection encounter was in smooth stratus, base 1500ft, tops 4000ft, and I got about 30mm in 5 minutes.
Obviously the cloud was loaded with water but there was not a shred of convection in it, and it was a nice-weather day.
I have never collected anywhere near that much ice as fast, and I have been in some convective cloud (TCUs, and even briefly in freezing rain, but never a CB).
This tells me that looking at clouds is not helpful for this purpose. I keep out of them, except when the OAT is above 0C (or is about to be and I am descending anyway, or some similar escape route).
This means the most likely icing will occur in humid air which has been cooled very recently. The longer the air is below freezing the more likely all the supercooled droplets will have found some core to condensate on.
That’s an interesting theory. I don’t know either way, but is it true? It might offer a way to have a go at forecasting likely icing conditions. OTOH does it really take that long (minutes or hours) for SLDs to freeze up? I thought they can stay as SLDs for ages. There is always some dust in the air.
Hi Stephan,
you probably mean, can you tell the difference between a supercooled-liquid cloud and an ice cloud by looking at it? Offhand I could only tell you about radar or polarized lidar 
but I can try to find some experts.
I do know that a glory is a reliable sign of at least a liquid top (probably useful warning if you’re planning to descend through it, but I don’t have any practical experience with that because of the FAA “forecast icing is known icing” doctrine) and that a sun dog is a reliable sign of ice (but presence of ice does not mean absence of supercooled liquid!). Of course, if the cloud is thin enough that you can see a sun dog, it probably wasn’t much of an icing hazard to begin with.
because of the FAA “forecast icing is known icing” doctrine
That doctrine ended years ago.
There was a brief period of it, and some other stuff at other times (e.g. IMC below 0C = known icing) but all that has been removed because it didn’t make any sense.
In any case, a “forecast ice = known ice” position implies that a specific weather forecasting service(s) must be referenced. For non US flight, there are usually no weather services which could map onto any such US legislation. It would be a bit like the FAA demanding that every N-reg pilot carries the US VFR sectionals…
Hi Peter,
yes, Sebastian’s statement is often true. If there are ice crystals in a cloud that also contains supercooled liquid, the ice crystals will grow at the expense of the liquid drops because that state is energetically favored (latent heat). The time scale for glaciation by that process is hours.
However, sometimes it doesn’t happen that way for various reasons, and the supercooled liquid can remain for much longer. One reason is that ice nuclei are quite rare. A typical number over the continents for condensation nuclei for liquid is O(a few 100)/cm^3 (so when the relative humidity goes to 100%, you’re basically guaranteed to form a liquid-water cloud), but a typical number for ice nuclei is less than 1/cm^3 (sometimes much less, like 10^-5/cm^3). The reason for this discrepancy is that ice is much choosier about what makes a good condensation nucleus (some types of dust, pollen, bacteria, maybe soot, and these things are typically so large that they sediment out of the atmosphere fairly rapidly so are never present in high concentration for long), whereas liquid will condense onto anything that’s hygroscopic. So the cloud can be so starved for ice nuclei that it stays liquid. Another reason a cloud can retain liquid water for a long time is dynamics. If the liquid water can be replenished from some source and the ice can be gotten rid of quickly enough (by precipitation), then a mixed-phase cloud can be quite persistent. Stratiform clouds in the Arctic often remain mixed-phase in this way for days. These types of clouds usually have the liquid water at the top and a nice fat inversion above that, which sounds a lot like the cloud Peter iced up in.
As far as predicting icing hazard, there is a great tutorial here that summarizes what produces and maintains the supercooled liquid. (In general COMET is a fantastic resource for going beyond the Mickey-Mouse meteorology from the PPL textbooks…)
Oh, one side point: AFAIK SLD specifically refers to large droplets, i.e., freezing drizzle and freezing rain, and the reason people (meaning the weather services) care about the large ones is that they can strike the airfoils far from the stagnation points, outside the protected areas, so they’re a hazard even with certified de/anti-ice.
Hi Peter,
last I heard, the “IMC below 0C = known icing” crap is gone, but the “forecast ice = known ice” doctrine stands, so if there is an airmet for icing, non-FIKI aircraft have to stay out of the volume covered by that airmet (which is formulated as “geographical area in cloud between freezing level and level”). But I could be wrong about that (would love to be wrong about that, actually). Yes, outside the US the doctrine is probably pretty meaningless.
