The Continental Flight 3407 crash on approach to Buffalo Niagara International Airport (KBUF) on Thursday has been quite thoroughly overcovered by the media.  Even highly professional direct and regular briefings from the NTSB haven’t staunched the blabbering and misinformation as journalists seek to fill out their quota of column inches or TV minutes.  I’m sure that the Internet is also rife with uninformed and idle speculation.

These are not negative observations, that is just the nature of technical catastrophes.  Not very many people are even close in expertise to the NTSB accident investigators.  They have a rare kind of training, experience, and mindset that would put any of the fictional CSI characters to shame.  So as the rest of us try to cope with our thoughts about the tragedy, we will wonder why it happened and try to understand.  The NTSB personnel will painstakingly gather the evidence and compare various theories against the facts.  They will exercise the limits of technology to solve the puzzle.  We get to sift through the misquotes and oddball statements of the media.

In my previous career, I enjoyed free rein to work with a wide variety of aviation technology.  In my last position, I was lead for stability and control for development of the Gulfstream G-5 (now the G500).  Earlier, I worked on analysis of four aircraft incidents and accidents of KC-135 aircraft while at Boeing.  I realize that writing yet another non-official theory for the Flight 3407 accident may not benefit the investigation (I doubt they’ll hear about this blog post), I’ll feel better airing a couple of ideas and maybe you will find them somewhat insightful.

The initial NTSB reports on the flight are puzzling in several ways.  It is clear that conditions were ripe for icing and that is a concern.  However, it is said that no other pilots reported severe icing.  To me that is a kind of red herring, because the special conditions for icing can be somewhat localized and transitory.  Icing can even vary considerably with relatively small changes in altitude.  Accumulation can happen suddenly and fast.  I also wonder how much icing was reported, even if it was not severe.  Some articles said that the crew reported icing on Flight 3407.

Another odd aspect of this accident is the initial severe pitch up of the aircraft, prior to the pitch down and roll.  It takes a lot of pitch moment to disturb an aircraft that way, and most of the “common” failure modes result in an initial pitch down.  Another mode, stall/spin, is usually not associated with large pitch rates nose up, the entry to stall is more gradual.  I would really like to see control deflections during the event if that is available from the flight recorder.  That would clarify a lot of things.  Unfortunately, the NTSB briefing talked about listening for the pitch trim motor on the cockpit voice recorder.  I’m guessing that the flight recorder does not have data for control deflections or the trim deflections, or they wouldn’t have had to guess about use of pitch trim from the CVR.

Any aircraft that experiences a sudden pitch change, especially nose up, will do so because of an equally sudden change of configuration at the horizontal tail.  In this case, with suspected icing conditions, icing of the horizontal tail has to be considered.  The textbook case is a situation where ice builds up during cruise or descent, reducing the effectiveness of the horizontal tail.  The ice builds up on the leading edge of the horizontal tail, greatly reducing the ability to hold the nose of the aircraft up at the correct attitude.  The horizontal tail has to produce a downward force to properly balance other forces and moments on the aircraft, and greater nose up trim than normal is needed with ice on the tail.  As the aircraft approaches the airport to land, the crew slows down the aircraft and deploys flaps and landing gear.  The flaps and landing gear create nose down pitch moments that overcome the capability of the iced up horizontal tail.  The aircraft dives and impacts the ground before the crew can regain control.

Flight 3407 crashed soon after deployment of the flaps and landing gear, too low to recover from a textbook case of severe horizontal tail icing.  But it initially pitched up, not down.  This is inconsistent with the textbook scenario.

Even worse for any theories based on icing: the crew activated the anti-icing boots on the aircraft soon after departure and that system was on during the accident, according to the NTSB.

It is important to remember that each accident or incident can be unique.  There is always a cause, but it doesn’t have to come from a textbook.  Worse, a theory can fit all the facts you have, but if you can’t validate it against the facts (eliminate other good theories), it may always be just a possible cause.  If you haven’t seen this in action, watch a few episodes of Mythbusters on the Discovery Channel.

Given what I’ve heard about Flight 3407, here is a possible cause.  There was significant icing at some point in the flight, providing accumulation of ice on the horizontal tail despite the anti-icing system.  Since the boots are fed from bleed air on the engines, I have to wonder about the flow rates at the tail when the engines are spooled down for descent.  The autopilot system automatically compensated by adding nose up trim as the ice accumulated.  It is possible that ice also accumulated on the control surface balance areas of the horizontal tail, which are probably not protected by an anti-ice system.  Due to the boots and other factors, the leading edge ice on the horizontal tail was not stable.  Configuration changes prior to landing caused the leading edge ice to break off the horizontal tail soon after deployment of the flaps and landing gear.  The sudden additional downforce as the tail regained its effectiveness caused the aircraft to pitch up and subsequently stall violently.  If ice accumulated on the control surfaces of the horizontal tail, the situation would be worse as the control system could have been jammed.

-- Walter Lounsbery, 2-16-2009

Q-400 Flight Evaluation