NTSB zeroes in on Buffalo crash cause

The evidence is mounting that icing — and the crew’s reaction to it — played the main part in last week’s tragic Colgan Air crash in Buffalo.

After viewing a video from NASA, one wonders how much the crew — or most other pilots for that matter — knew about the existing investigations into what happens when the tails of turboprop aircraft ice up.

The spooky part of the video is at 15:40, when the test plane stalled (“stalling” in an airplane is the absence of lift). The pilots only recovered by retracing flaps, which allow planes to slow down, and descend without picking up airspeed.

Compare that to the National Transportation Safety Board’s timetable of when things started to go wrong for the plane in Buffalo:


The NTSB has said problems for the 74-seat Bombardier Dash 8 Q400 occurred when the pilots lowered the landing gear and tried to set the wing flaps to slow the aircraft for landing.

The video adds an additional part to the equation: The actions a pilot must take for a “tail stall,” are nearly the opposite of the actions he/she must take for a wing stall, the much more common type of stall.

How much time did the pilots have to take those actions once the plane was (apparently) stalling? Five seconds.

  • steve

    Are civilian pilots taught how to deal with this situation in their training before carrying passengers ? Are military pilots taught how to deal with this ? If military piliots are, and civilian plots aren’t, I want an ex-military pilot in the cockpit. And I just heard the plane was on autopilot even when it was icing up. This sounds criminal.

  • bigalmn

    I too wonder how much training pilots are given, but this plan was suppose to have the latest equipment. I don’t know anything about icing equipment, but you would think with today’s sophistication in detection and electronics, there would be some way to detect the icing and automatically start some type of element that heats the wings to prevent icing.

    This plane may have hit a unique spot where it iced up and others later or earlier did not. It sounds like the experts don’t know what to tell pilots so how do pilots really know what to do.

  • Bob Collins

    It depends what you mean by “civilian pilots.” The pilot certification system in the United States is multi-tiered and, technically, all pilots are “civilian.”

    For the average general aviation pilot — flying small airplanes in the daytime under visual flight rules — we’re trained to recognize stalls and recover from them quickly. In fact, it’s one of the first things we learn. For us, though, we practice under conditions where we’re up to about 3,000 feet and we know the stall is coming, which much different than the ones that take you by surprise.

    Then pilots can add “ratings” above that. A commercial pilot license, for example, allows you to fly for hire, or become a flight instructor.

    And instrument rating allows you to fly in instrument conditions are fly under instrument flight rules in good conditions etc.

    And an air transport pilot certificate allows one to fly left seat in a commercial jet. Then there are ratings for the type of aircraft you fly as well.

    So the training is VERY rigorous.

    Specific aircraft — a high-tail turboprop is a fine example — have specific flight characteristics that presumably the airline flying them are aware of and incorporate into their training.

    Colgan, as I understand it, was aware of the NASA recommendations not to fly autopilot in icing conditions and made it company policy. But it wasn’t required to.

    An autopilot is very valuable tool. It prevents the workload from overwhelming the flight crew. But, as we’ve heard in the last few hours, it compensates for changes in the flight envelope…more power to maintain a set altitude, or a different pitch to achieve a certain airspeed and if a flight crew isn’t monitoring the situation, it won’t occur to them to ask themselves why the AP is making a particular adjustment.

    The other thing we’re taught is if you do something and it results in a flight dynamic that’s unacceptable, UNDO what you just did in a hurry. Now, here that could be setting the flaps or even turning the AP off.

    But, again, you’re talking about humans and in an unexpected situation, you have to (a) recognize what’s happening (b) understand why what’s happening is happening and (c) know what to do about it.

    In this case that all had to be done in about 5 seconds. That’s asking a lot. My guess is as they were falling, they really didn’t know for sure exactly why.

    The way the NTSB described things yesterday, the plane lurched up, the left wing dropped and it lurched down, and then the right wing dropped. Just guessing — and that’s all I’m doing… the lurch up sharpened the stall, the left wing dropped (it has no left), he/she stomped on the right rudder (fearing a spin, which sounds like what was about to happen, you apply opposite rudder),and then the opposite rudder make the right wing drop. But, like I said, that’s a mere guess.

  • JohnnyZoom

    Bob, if you have read the novel Airframe, I’d love to hear your take on how accurate the technical portrayal of flying is. Or not.

    That book’s scenario may be unrelated to this one. But it painted a quite detailed and impressive picture for me of the flying process, and of the management of that process. I’d love to know where on the accuracy spectrum, from Apollo 13 to The DaVinci Code, it lies.

  • bsimon

    Bob writes

    “For the average general aviation pilot — flying small airplanes in the daytime under visual flight rules — we’re trained to recognize stalls and recover from them quickly. In fact, it’s one of the first things we learn. For us, though, we practice under conditions where we’re up to about 3,000 feet and we know the stall is coming, which much different than the ones that take you by surprise.”

    To add to Bob’s comment, stall training is under good conditions. Typically its probably a nice summer day. Skies are probably clear, you can see both the ground and the horizon. The flying surfaces are clear of snow, ice, or anything else except perhaps a few bug carcasses. You think “I’m going to practice some stalls,” pull back on the stick while dropping power – and voila! You’re stalled. Ease up on the stick, add some power and voila! You’re flying. The horizon is still right there in front, the ground down below, all where it should be.

    Now, take away the horizon, take away visual contact with the ground and put the aircraft into what they call ‘unusual attitude’. You’re under full power and a wing drops. Your butt tells you which one drops, and if you have the ability to focus some gauges on the panel will verify what your butt tells you. But once you correct, it is difficult to determine where to stop – and not overcorrect. The butt is not well calibrated for such maneuvers without a visual reference to level (like the horizon). It can also be difficult to tell if you’re going up or down, if you can’t see the ground.

    This is basically like holding your breath underwater & having a friend spin you forwards, back, left & right. Without opening your eyes, can you tell which end is up?

  • skyblu5555@aol.com

    Does anyone know whether the plane’s software was functioning properly or the speed senors were operating correctly; that is, did the pilots even receive correct information upon which to make the required instrument adjustments? All the focus seems to be either on the icing of the wing flaps or the failure of the pilots…while we know the Air France Air Bus plane traveling from Brazil to Paris crashed in the midst of a storm over the Atlantic because the speed sensor on the plane wasn’t working and the plane had been scheduled for a speed sensor replacement.

    Please indicate if you have heard anyting about NTSB investigation into the onboard software, hardware, and peripherals of the onboard flight system.

    Thanks.