The type of icing most likely in this/these incident(s) involved ice particle icing or glaciated/mixed phase icing. These relatively rare conditions are described in the report
The Ice Particle Threat to Engines in Flight (page 13).
Reports of rain streaming around the windshield are discussed and clearly answered as being due to the ice particles melting on the anti-iced surfaces. The conditions are outside of the conventional airframe icing boundaries, thus no airframe icing is seen. Also, note that the conditions are near, or above large Cb build-ups; the aircraft are not necessarily in Cbs.
I have encountered these conditions during test flights in 1997 (tests preceding those referenced in the report - USA 1998). On occasion small ice build-ups could be seen at the windshield edge or on the windscreen wipers - the windscreen was very wet.
Also, the IMC described in the report is often very thin cirrus cloud, which many pilots would declare as VMC due to the apparent good visibility.
The mechanism of pitot icing (AF447) is probably similar to TAT probe icing, or where the pitot is considered to be a very small scale engine. Ice particles, slowed by bends in airflow ducts, melt when in contact with heated surfaces, and the water acts as the glue for more ice particles to accumulate and freeze, which in time exceeds the heat-flow capability of the anti-icing structure; the report gives details – I urge all pilots to read the details.
Re “
What would explain that the air became warm and humid all of a sudden?”.
As a hypothesis, perhaps encountering an ice particle cloud (high ice water content) where the melting ice and changes in heat flow are beyond the instantaneous dynamics of the Air Cond packs. The report cites the possibility of very high Ice Water Content, as a guide, a mid value of 4gm/M^3 would require a stout umbrella at ground level if transformed to rain.
Re “
Maybe we as pilots have to modify the way we look at weather”. I agree.
Other sections of the report (page 17) discuss the shortcomings of radar in detecting ice particles and hence the inability to see the high altitude structures of Cbs, or new, emerging build-ups in the general cloud mass or anvil.
Defensive strategies include avoiding the core of large storms by a very wide margin, remaining upwind to avoid the anvil region, and being aware of new build-ups.
We must reconsider the way we look at very large storms; the judgment of size might be subjective, similarly ‘a very wide margin’ involves judgement and decision making processes.
My experiences indicate that the risk of encountering conditions favourable to malfunctions are present up to at least 30nm from a detectable Cb core, thus with a margin of safety, 50 nm should be considered – this consideration should also include turbulence. We need to focus more on planning ahead - rerouting, use of satellite pictures, and reasoned judgement of met forecasts – know before you go.
On two occasions during test flights, super, ‘super’ cells were seen – so large (area, height), so black, and so menacing, that ‘evil’ was a worthy term. These storms were not investigated during the tests and were avoided by 100 nm.