Aviation Law Monitor : Aviation Lawyer & Attorney : Michael Danko Law Firm : California Airplane & Helicopter Crashes : Aviation Law Monitor

Icing or pilot error?

Last April, the NTSB released the data from Flight 3407's FDR.  I blogged about that here.  Despite wide spread speculation that icing brought down the aircraft, it looked to me like pilot error -- not weather --  was to blame. 

Then, in May, the NTSB released an animation derived from the aircraft's flight data recorder, its cockpit voice recorder, and ATC transcripts.  I blogged about that here.  The animation, like the raw data from the FDR, made a strong case for pilot error.  From the animation, it appeared to me that an inattentive pilot allowed the aircraft to get slower and slower, until it became dangerously close to the speed at which the aircraft would stop flying altogether and simply fall from the sky.  Then, when the critical moment came, the pilot pulled back on the control yoke instead of pushing it forward, thereby inducing an aerodynamic stall.

The NTSB made public its official probable cause finding at a hearing yesterday.  No surprises to anyone who has studied the data.  According to an article in today's Buffalo News, the NTSB summed it up as follows: 

The plane got so slow that the "stick shaker" — a device that helps to prevent stalls — activated. But Renslow [the pilot] mistakenly pulled back on the plane's controls at that point, which is exactly the opposite of what he should have done.

In total, Renslow pulled back on the controls three times in response to the stick shaker and "stick pusher," forcing the nose upward. That caused and then exacerbated the stall.

It's almost unimaginable that a professional pilot would make the series of mistakes that the pilot did in this case.  Even a new student pilot would know better.  But that's what he did.

The NTSB played its animation for those who attended the hearing.  The animation shows the pilot's errors mount.  The activation of the "stick shaker" is depicted 2 minutes and 8 seconds into the animation. The shaking control yoke was a final warning to the pilot that he must immediately push the yoke forward.  But instead of pushing forward, the pilot pulled back. Three times.  After the third time, the aircraft stalled and crashed. 

There were countless points at which this aircraft could have been saved but, inexplicably, the pilot failed to take appropriate action.  

• Email This
• Print
• Comments (1)
• Trackbacks
• Share Link

I blogged about Scene Systems' animation of Flight 1549's landing in the Hudson here back in March.  Great effort, but I noted that it would take hundreds more hours of work before it could be used in court.  That's because it did not appear that the animation accounted for and synchronized all the available data for the flight.  For example, the flight path depicted in the animation could not have been true to the information from the flight data recorder, because the flight data recorder had not yet been downloaded and made available by the NTSB.  As a result, Scene System's finished product involved too much guesswork to ever be shown to a jury.

Just for fun, Kas Osterbuhr of Exosphere3d in Denver has been working on perfecting an animation ever since.  He emailed me the link late last night.  Kas, whose firm creates animations for use in court, explained to me that his animation is pretty much technically perfect.

Among the datasets utilized are: audio transcripts and recordings, digital flight data recorder, raw radar data, NEXRAD weather, witness statements, satellite imagery, elevation maps and several of the NTSB reports published in the docket. . .All aspects of this animation are based on actual data, whether from the NTSB docket or otherwise. The entire 3D reconstruction is built into a single environment where every piece of information can be aligned in position and on a timeline.

Tons of work went into this animation and it shows.  Aviation accident animations don't get any better than this.

One question, Kas.  The animation depicts flames coming from the aircraft's engines at certain times.  On what data is this based and what would happen if the judge ultimately determined that that evidence for this aspect of the animation is insufficient to allow it to be shown to a jury?  

November 9 Update: Kas' response is in the comments.

• Email This
• Print
• Comments (2)
• Trackbacks
• Share Link

NTSB Chairman Deborah Hersman's recent testimony before congress concerning the mid-air collision over the Hudson raises more questions than it answers.  She stated that  the Teterboro controller instructed the Piper pilot to switch to frequency 127.85 to contact the Newark controller.  But before leaving the Teterboro frequency, according to Hersman, the pilot read back to the controller "127.87,"  which was wrong.  Thereafter, the pilot was in contact with neither Teterboro nor Newark, and so neither facility could warn him of the impending collision. Hersman's remarks are here.

Hersman's implication is that the Teterboro controller failed to correct the pilot, and so the controller contributed to the pilot's getting "lost in the hertz" (out of radio contact) at a crucial moment.  However, the animation that the NTSB released on the same day that Hersman testified does not appear to back Hersman up.  It just doesn't sound as though the pilot read back "127.87" as Hersman states.  You can listen to the audio yourself beginning at minute 2:25. 

• Email This
• Print
• Comments
• Trackbacks
• Share Link

Two months ago, Scene Systems -- a litigation support firm -- released its animation of Flight 1549's crash into the Hudson. I posted here that, in all likelihood, the animation would not be admissible in court. The legal objection would be that the animation "lacked foundation." For example, without information from the Airbus' black boxes, Scene Systems couldn't confirm the aircraft's flight path or guarantee that the Air Traffic Control audio was properly synchronized to the aircraft's path of travel.  Therefore, the animation involved too much guesswork to be shown to a jury.

The National Transportation Safety Board has now released its own animation. Having retrieved the black bloxes, the NTSB was able to plot accurately the Airbus' position, speed, and altitude at each point along the aircraft's short flight.  The NTSB then properly synchronized the Air Traffic Control audio to the aircraft's flight path.

The only audio on the NTSB's animation is the radio transmissions between the crew and Air Traffic Control. As is typical, the NTSB did not make public the audio of the cockpit conversation between the captain and the first officer. The NTSB did, however, prepare a written transcript of that conversation. The NTSB superimposed the transcript on the animation. (HOT-1 is the pilot, HOT-2 is the first officer.)

Would this animation be admissible in court?  While Scene System's animation would not pass legal muster, the NTSB's work probably would. 

• Email This
• Print
• Comments (1)
• Trackbacks
• Share Link

I blogged here on whether it was icing that caused the crash of Flight 3407, or whether the pilot simply pulled back on the yoke when he should have pushed forward.  The NTSB's animation, using data gathered from the aircraft's black boxes, makes a strong case for the latter. 

The video is 2 minutes 39 seconds long.  Watch the airspeed drop dangerously low by 2:04 and the stick shaker activate at 2:07.  The pilot should have immediately pushed the yoke forward, which would have pointed the nose down and allowed the aircraft to regain airspeed.  Instead, he pulls the yoke back.

• Email This
• Print
• Comments
• Trackbacks
• Share Link

Scene System's animation of the crash of US Airways Flight 1549 is a viral hit.  The litigation support firm combined available ATC audio tapes, flight track information, and an on-scene photograph into a great recreation.  This is the exactly the type of animation used in court to help juries understand the details of an aviation accident.  

But would this particular animation be admissible in a lawsuit?  Probably not. It incorporates too much guesswork.  For example, Scene System overlays the animation with audio from Air Traffic Control tapes.  Are the movements and positions of the aircraft properly synchronized with the audio? To do that right, you'd most likely need information from the Flight Data Recorder , which isn't yet available. Without that data, the animation is objectionable as "lacking foundation."  It's safe to say that, before it could be shown in court, the animation would require hundreds more hours of work and refinement. 

Of course, Scene Systems wasn't out to produce a recreation that was admissible in court. It was just trying to show the type of product it is capable of. And it did that very nicely. 

• Email This
• Print
• Comments (2)
• Trackbacks
• Share Link