The ice contaminated tailplane stall phenomena remains remarkably misunderstood, partly because the aerodynamics involved are not often taught in primary flying.
The center of lift on the wing generally resides somewhat aft of the center of gravity. This generates a pitch down moment that can often be quite powerful. The horizontal stabilizer exists to create a balance force; thus the stabilizer must generate a lifting force in the downward direction. It operates as an inverted wing. The relationship between these pitching forces forms the basis for longitudinal stability. From the accident investigation standpoint, the main spars will almost always fail in a downward direction, due to negative G, following the inflight separation of the tail structures. Indeed, the sound of the main spars failing has been documented in both witness reports from the ground and on the CVR. It sounds like a large caliber rifle shot.
Thus, if the stabilizer is ice contaminated and actually stalls, the nose is going to pitch down. A stall, of course, is not the same thing as the inflight separation of the tail. Although stalled, lift is always still being produced to a limited extent, and the pitch down is not likely to exceed the structural capacity of the main spars. The occupants, however, may involuntarily test their seat belts.
That said, the typical ice contaminated tailplane stall rarely reaches this point. During the Twin Otter work, the tail was fully stalled only accidentally. What happens prior to the stall is an initial separation of the flow over a portion of the stabilizer. This was well documented by NASA through the use of flow visualization tufts. The flow separation seriously degrades the aerodynamic balance of the elevator, long before the tail actually stalls. When the aerodynamic balance is altered, the elevator control forces may lighten, leading the pilot to displace the control much farther than he intends for a given attitude ajustment. This gross displacement often leads to a PIO in pitch, which almost inevitably exacerbates the problem. Eventually, the elevator (or aileron, as in the ATR scenarios) will “snatch”, or violently displace into the region of low pressure. At the tail, this would a trailing edge down displacement, creating a nose-down pitch for the airplane. The tail is not yet stalled, and the elevator is fully functional aerodynamically. But the control force balance has been destroyed.
Control snatch is only possible with an aerodynamically balanced control surface. A fully powered, irreversible surface is not susceptible to this. (Reversible/irreversible refers to the ability to make the control wheel turn by manually moving the control surface from outside the airplane. If you can do this, the control is reversible. If you can’t, perhaps because there is an actuator between you and the control wheel, it is irreversible.)
The response to elevator snatch is to muscle it back into a trailing edge up position, which re-cambers the stabilizer and causes the flow to re-attach. This can require considerable force; NASA demonstrated 170 pounds of pull force required on the Twin Otter. The next step is to retract the flaps one notch, which reduces the downwash angle off the main wing and thus reduces the AoA at the tail, restoring normal pressure distribution and thus control force and stability.
If you have a powered control, it won’t lighten or snatch because the force used to position it far exceeds the airloads under nearly all circumstances. If the stabilizer were to stall due to ice accretion, a powered control allows the pilot to easily re-position the elevator to a camber that un-stalls the surface.
Thus, the Q400 does not have the snatch problem since it uses a powered elevator. The DC-9/MD-80, on the other hand, has experienced this type of problem, since it uses a spring/servo/geared/flipper/thing-a-ma-jig but, eventually, manually operated elevator.
The Twin Otter work also went a long way toward supporting the new certification requirements involving a 0.5 g pushover maneuver to investigate tailplane stall susceptibility during certification. The Q400 met these requirements; many, many older designs have not.
ICTS has been identified in accidents and incidents involving the Jetsream 31, DC-4, Convair 240 through 580 series, Viscount, YS-11, DC-9/MD80, DHC-6, Saab 340 (prior to mod) and others. There is no requirement that it be a T-tail design. It is fair to say that it could be a problem for most designs, if equipped with reversible flight controls and enough flap deflection. The question is whether the designer has properly optimized the tail area and stall margins with the flight control design and flap configuration. Most have; in cases where they have not, this has often been corrected through modification or flap restrictions. You should be able to get educated on this with respect to your airplane during training.
For a good look at the possibility of the flight crew misinterpreting a shaker/pusher event as an ICTS events, look at the Atlantic Coast accident at Clumbus, Ohio in 1993:
http://libraryonline.erau.edu/online...s/AAR94-07.pdf
There is no doubt in my mind that this crew did mistake the shaker and pusher for ICTS. Needless to say, if they had correctly interpreted their new-fangled tape display airspeed, they would have improved their mental model substantially.
I can currently identify 18 events involving ice-contaminated stalls of the main wing in which clear evidence exists of the pilot pulling the nose up in response to stall indications. There is a good case to be made in each situation that the stall indications could be mis-interpreted due to the mental model operative at the time. So the Colgan captain’s response to the shaker/pusher does not, in itself, indicate that he believed the elevator had snatched, although that remains a possibility.
On the other hand, the first officer of a Mid-Pac YS-11 at West Lafayette, Indiana in 1990 did actually retract the flaps without the captain’s command. As it turned out, this was the correct thing to do, so no one made an issue out of his “insubordinate” initiative. As the NASA video was far in the future at that time, I suspect that he reacted just as the Colgan first officer did…by re-configuring to the last known point of safety. The Roselawn crew attempted this as well, but were outmaneuvered by the automation.
To date, tailplane stall events have never been documented without a) ice accretion, and b) full landing flaps. There has not been a documented case of ICTS in transport airplanes for quite a few years, but who knows why that is. The manufacturer’s procedures must be followed, whether it be the Twin Otter’s flap restrictions in icing or the MD80’s tail deice requirement before landing flap extension.