Subscribe by Email

Sunday, July 13, 2014

Asiana 214 and NTSB Credibility

The National Transportation and Safety Board of the United States, NTSB, recently released it’s findings on the crash of AsianasFlight 214, which crashed on July 6, 2013 at San Francisco International Airport (SFO).  All professional aviators and their instructors as well as air traffic control specialists should read and study the report.  The information in it describes the enormous potential for and realization of unimagined failures that always exist in the complex relationship between humans and the machines they control. 

There is no question that the crew of Asiana 214 (humans) failed to effectively manage their Boeing 777 in the final moments of the flight.  The NTSB detailed the operational errors of the crew, the design flaws of the flight guidance system and some relevant issues related to their training.  The report also highlights numerous other contributory factors, but uncharacteristically for the NTSB, missed a very obvious one, air traffic control.  The NTSB has been the worlds “gold standard” in aviation accident investigation for decades.  Unfortunately, it seems the gold might have lost some of its glitter.

The probable cause of Asiana 214 focused on human failures, primarily by the crew and secondarily by designers and trainers, which were thoroughly examined in the investigation.  However, there are other human components of this story that should also be examined for their role in this fatal accident.   Specifically overlooked was the environment in which the flight operated.  The machine was a Boeing 777. The environment was the airspace near SFO. The humans are the engineers, instructors, air traffic controllers and pilots who design, train and operate them. The NTSB report on Asiana 214 surprisingly describes the contributory role of the aircraft and each of these other elements except air traffic control.  How can this report have such an obvious omission?  This is very perplexing.  Was it just a huge oversight or an intentional exclusion?   In light of the fact that there was speculation by the Board on other contributing factors, this oversight is even more astonishing. 

Currently, at large metropolitan airports where the air traffic volume is heavy, a flight crewsrole has evolved into that of a facilitator between air traffic control instructions and the airliners flight guidance system rather than that of aviator.  Like all pilots flying modern airplanes into highly controlled airports, including SFO, the crew of Asiana 214 was tasked with converting the commands (not suggestions) of ATC into a desired flight path and energy state by programming the aircrafts flight guidance system.  The Boeing 777 is both designed and its pilots trained to operate it with reference to the flight guidance system.  Desired parameters are entered and the computed information is shown to the pilots by reference to the flight directors on the primary flight display (PFD) or heads up display (HUD).  The pilots follow the flight director commands manually or by engaging the autopilot.  The flight directors are turned off in flight on only very rare occasions.

Because the Asiana crew was ineffective applying their approach clearance to the flight guidance system, their workload increased dramatically.  ATC’s instruction to maintain a high speed longer than normal increased task saturation as well.  A late landing clearance by SFO tower provided an additional distraction diverting the crew’s attention away from subtle flight guidance information in the final seconds before the crash.   

Although there are many aspects to the air traffic control handling of Asiana 214, I will limit the discussion of ATCs contributory role to the time period after the crew received their approach clearance.

From the NTSB history of the flight (italics): 

"According to recorded information, NORCAL stated “Asiana 214 heavy, San Francisco airport 9 to 10 oclock, one seven miles, do you have it in sight?” The flight responded they had the field in sight and NORCAL cleared them to fly a visual approach to runway 28L on an assigned heading of 310° to intercept final approach. At a point approximately 14 NM5 from the airport NORCAL instructed the flight to maintain 180 kts until 5 NM from the airport."

This command by ATC, as well as the inoperative electronic glide slope, set up a series of events that facilitated the crash.  It is not certain, but clearly reasonable to infer that if the crew had been allowed to slow the aircraft at their discretion, or if there had been an operable electronic glide slope, the unintended consequences of autothrottle “HOLD” would have occurred at a higher altitude or not at all.  How can the NTSB omit these relevant facts?

The NTSB found that the flight could have been completed successfully without the electronic glide slope.  That statement is not particularly probative since it is made without context. As the Asiana pilots pointed out in their submission, a Flight Safety Foundation (FSF) report on approach and landing accidents (ALA) stated that 75% of flights did not use or have a precision approach aid available.   The FSF further states that the accident risk is 5 times greater for commercial aircraft flying non-precision approaches.

In the hearings the NTSB conducted on the human factors of this accident, Dr.s Sarter and Abbott and Captain McKenney all testified to the interrelationships between parts of a safety system. As the “swiss cheese metaphor” for safety systems shows, the slices are only effective when used in combination.  Removal of one slice may or may not be significant depending on the make up of the other slices.  Why does one plane crash in the same conditions that allow many others to be successful?  That is why the role of ATC in this accident cannot just be brushed as inconsequential.   These are exactly the issues the NTSB is tasked with considering.

“The trainee captain stated that after arming and intercepting the localizer he set the DUYET intersection crossing altitude of 1,800 ft. MSL in the mode control panel (MCP) altitude window and began using the AFDS vertical speed (VS) mode to descend at 1000 feet per minute (FPM). The IP stated the arc on the navigation display (ND) showed they would be high at DUYET. The observer stated he thought they were a little higher than the normal profile when they were cleared for the approach, and they extended the landing gear earlier than normal because they needed to get down. The trainee captain set the command airspeed bug to 172 kts, and acknowledged when the observer commented this was below the assigned 180 kts.”

This emphasizes the fact that the 180 knot restriction was increasing the crew's task saturation by complicating energy management solutions. 

At the prompting of the IP, the trainee captain increased the command vertical speed to 1,500 FPM. The trainee captain commanded the flaps be set to 20° and requested the IP to reset the command altitude from 1,800 ft. to 3,000 ft., the missed approach altitude, as the flight approached 2,000 ft. MSL.

According to recorded data, at 1,600 ft. MSL the AFDS pitch mode changed to FLCH SPD, the throttles began to increase power, and the airplane pitch attitude began to increase; this was followed by autopilot disconnect and shortly thereafter a reduction in the thrust levers to the idle position, followed by a change of the autothrottle mode from THRUST to HOLD. The trainee captain stated he considered pressing the FLCH pushbutton to obtain a higher descent rate but he could not recall what he did for sure. He disconnected the autopilot and called out “manual flight.” None of the three pilots could recall the autothrottle status displayed on the flight mode annunciator (FMA). The IP stated he set the command airspeed to the approach speed of 137 kts. and turned both flight director (FD) switches off and then turned the right FD switch back on. Recorded data showed the left FD switch was turned off but the right FD switch remained on.

Pre selecting the missed approach altitude is common practice by B777 crews.  However, setting an altitude above the aircraft without the flight guidance capturing and maintaining either an altitude or glide slope caused the autopilot to initiate a climb. The pilot naturally and appropriately disconnected the autopilot and manually flew the aircraft. Unfortunately, that unusual set of circumstances caused the auto throttle mode to remain in "HOLD" and not respond to the commanded speed as the crew expected. 

“The trainee captain called for the flaps to be set to 30°, and after a delay due to the airspeed being in excess of the flap limit speed of 170 kts. the IP placed the flaps to 30°.”

Because ATC instructed Asiana 214 to maintain 180 knots until 5 miles, they would be 10 knots above the limit speed for flaps 30° and 43 knots above their target approach airspeed at approximately 1500AGL (3 degree glide path).  Without an electronic glideslope or a VNAV PATH, this complicated programming of the flight guidance system.  At 1000’ AGL, their mindset would likely have been slowing the airplane down quickly as they descended and closed on the runway.  The power needed to be reduced to idle to get below the limit speed for flaps 30 (170 knots) as well as slowing to the target approach speed of 137 knots.  This maneuver, set up by ATC requirements was clearly contributory to the accident.

Flight 214 was handed off to, but not acknowledged by, San Francisco Tower.  In combination with the ATC imposed energy management complication, lack of electronic glide slope and flight guidance mode confusion, this was another significant distraction. 

“The instructor pilot stated he saw 2 red and 2 white precision approach path indicator (PAPI) lights at 1,000 ft. MSL and speed was a little high. The observer saw the descent rate on the vertical speed indicator was in excess of 1,000 FPM as the flight descended through 1,000 ft. MSL, and he called out “sink rate” several times. The IP stated he heard this callout but they were still high and the descent rate was only 1,000 to 1,100 FPM.”

Slightly above the glide path and slowing, the trainee captain manually flew the aircraft toward what he thought would be a stable approach at 500’.   Unaware of the significance of the autothrottle “HOLD” mode, the aircraft slowed toward target speed.  Still not cleared to land and below 1000the crew once again queried San Francisco Tower and received landing clearance 9 seconds later.  The late prompted landing clearance was an ATC distraction that diverted the crews attention at a very critical time.    

The aircraft was on glide path with the idle power as it descended toward 500’.  At about 500the aircraft began to drift below the PAPI (glidepath).   As the pilot began to adjust the descent with pitch, the aircraft quickly slowed well below target speed.  By the time the crew was alerted, noticed the low speed and initiated a go around it was too late.

In any accident, there are many factors that intersect to create the causal conditions.  Failure to consider all of them results in an incomplete and ineffective investigation. The relevance air traffic control resources, procedures and practices might have had on this accident is not the depth of analysis we have come to expect from the NTSB.

It would be extremely naive to believe ineffective or improper use of the flight guidance system to comply with air traffic control instructions only affects a few crews.  Recognition by the NTSB that air traffic control practices have a profound impact on how crews operate the flight guidance systems of modern airliners would be more consistent with their role in aviation safety. 

Since the report on Asiana 214 was prepared under his leadership, I would hope that Acting NTSB Chairman Hart would be given the opportunity to give his perspective on this controversy during his confirmation process.


  1. I received comment from Christine Negroni, author of "Deadly Departure" and popular blogsite "Flying Lessons"

    Hi James,
    I tried to post a comment on this blog post, which was quite good, (the post, not necessarily the comment) but it disappeared into the ether when I tried to preview it. 

    Anyway here is what I intended to say. Feel free to post it as a comment if you like.

    The post was very interesting and you have convinced me that the NTSB did not do as expansive a job as it should have done in looking into all the contributing factors in the Asiana 214 event. As you describe it, it reminded me of the Comair "wrong runway" crash in Lexington, Kentucky. In that case the clearance to take off was issued while the plane was still at the gate, I'm thinking around 10 minutes prior to takeoff. There was no eyeballing of the runway for obstructions; other traffic or wildlife (a no- too far-fetched possibility considering the time of day - dawn - and the fact that this is horse country) Had the controller eyeballed the field for obstructions he would have seen the plane lined up on the shorter of the two runways. But he was alone in the tower and pre occupied with other work. 

    I don't know why in that case the board chose to concentrate entirely on the pilots non-sterile cockpit. Certainly, it was a hole in the cheese, but far from the only hole. 

    And of course, that's always the case as you point out.

    Thanks for taking time to review the report and do the analysis. Sorry the NTSB failed this time. 

    1. Having attended the Comair wrong runway hearing, it was clear then that the standing Board absolutely suffered the extreme of the "Monday morning quarterbacking" that the commenter below correctly notes.

      Its cause though was not at all Type A. Rather, it was the normal and predictable response that accurately reflects the human subconscious that fully internalizes that something terribly bad could have happened to them in the past or might happen to them in the future. Regardless of its cause, it is clearly a detriment to a sound safety system.

      This, along with incomplete post-accident procedures, inadequate investigative follow-up and inattention to relevant facts, yielded a politically suitable, yet functionally incomplete and therefore, useless finding.

      So it was with the Asiana 214 report.

      The more things change, the more they stay the same...

      Great post Captain Blaszczak.

  2. Excellent synopsis Jim! Unfortunately, the investigative processes at the NTSB are outdated. They rely on the manufacturer and the airline to paint a picture of what should happen in certain circumstances based upon the guidance they claim to provide crews for the conduct of their operation. But, there seems to be little reflection on the veracity of that guidance.

    For instance, wouldn't it be nice to know if any other crews also had trouble with automation issues on approaches without electronic glideslope and/or visual glidepath indicators? As you made mention, wouldn't it have been nice to know the prevalence of occurrence of high energy states created by poor vectoring/speed assignments by ATC? How many other crews also failed to make a stable approaches under these conditions? How many other crews have not been stable at 500' and chose to continue? How many pilots are unaware that the "wake up" speed protection of the autothrottles are inhibited under this rare, albeit critical condition?

    Here is my point. The NTSB wants to measure aircrew performance based upon the theoretical, not the reality. If someone were to hear that x% of crews have also found themselves high and fast at X airport, and/or X% of non-precision approaches were also incorrectly flown due to automation complications, and/or x% of crews have also been unstable at 500' and chose to continue the approach - I believe it may leave the impression that the problem may lean heavier on inadequacies of the system vice a couple pilots deemed "inadequate" or "mismanaging the situation."

    It would also call to question the NTSB's insistence on finding a "probable cause" and recognize that there is no such thing. Without a greater understanding of the prevalence of occurrence of certain conditions, the NTSB will continue to give the industry a reason not to change for the better. Why not just lay the primary blame on the pilot so we can continue to do things just the way they are?

    Finally, pilots are there own worst enemy in this process. Monday morning quarterbacking leaves everyone feeling that they would have never made that mistake. These comments are sickening. I have made hundreds of calls to pilots that have found themselves in "undesired aircraft states." It leaves me wondering how anyone can believe that it can't happen to them. Type A personality working against them.

  3. I could not agree more! Thank you for your thoughtful comment.

  4. Good article, and I agree. There is more to it, as I have outlined previously in some articles. In addition to this, the Capt was off the A320, so likely looking for that sight picture, plus A320 procedures, not looking for the thottles to move and expecting it would capture the speed (only Boeing's have this scenario of A/T on and no speed protection, to my knowledge). The check airman was likely sweating it until 500', where all looked like it was good, then might have just glanced at the chart for a moment to plan the turn off, as this student was quite competent and had an excellent record. I have written a few bits on this in various articles