Aviation Mishap Investigations: Other Conclusions

One of the subjects of this blog will be Major Commercial Aviation Mishap Investigations and others and alternative analyses, conclusions and recommendations that could be reached. The sole purpose of seeking and reaching these alternative analyses, conclusions and recommendations is to improve commercial aviation safety closer toward the point that it is mishap free.
Investigations can give us an insight into a mechanical chain of events. NTSB and other major mishap investigations often accomplish this goal. Yet investigations can also give us insight into human thinking, human nature and human factors. Human factors is just another expression for the human condition, and since mishaps are the subject, the part of the human condition most often observed is that of “human error.” This further, perhaps deeper insight might give us a window into ways to prevent mishaps from ever reoccurring.

It may be possible to reexamine the facts of major aviation mishap disasters for the purpose of drawing conclusions which would lead to recommendations which could make aviation closer to being mishap free than it is today. This may seem a daunting task, taking into account the complexities of aviation disasters. This may seem an unreachable task, taking into account all of the very talented people who even today pursue aviation mishap investigations formally for the NTSB and other groups. This may seem a redundant task considering the wide range of party status granted by the NTSB to interested groups.

But it may be possible that new analysis, conclusions and recommendations can be reached, though however daunting the task.

Essentially, that is the purpose of this blog. The free expression of ideas may lead the commercial aviation community towards a safer future.

43 thoughts on “Aviation Mishap Investigations: Other Conclusions

  1. Paul Miller

    That is a good question. Training must cover both aspects:individual proficiency in procedural knowledge and ability to magnify by teamwork.

  2. Bill

    First, let me ask you one question. When you train or are evaluated in a simulator, are you alone in the simulator responding to situations, or are you with other potential crew members or people being evaluated?

    How can flight simulations accurately evaluate how the flight crew “would interact” during real-time emergencies?

  3. Paul Miller

    Yes, they were low, it was night, they were in instrument meteorological conditions, they had ice building up and they were slowing to approach speed.
    Many things have to go right including procedures, equipment, training, handling, weather and crew resource management. How many of these went wrong?

  4. Bill

    It looks like addequate training is becoming an issue, as is fatigue.

    I still question the ability of anyone to adequately respond or react to that kind of an emergency, at that low of an altitude, during an approach, in that kind of rapidly changing weather situation. It seems as if it was a recipe for a disaster ready to happen.

    They were near 1000 feet, weren’t they?

  5. Paul Miller

    If the data was recorded it could also be sent to the ground by a data link digitally. That would be useful and usable for current wx obs and for wx awareness by ATC.

  6. Paul Miller

    I always thought that if I could transmit pix from the air by way of TV or now data, that the forecasters would have a better idea of what is out there plus the intensity of convective wx.

  7. Bill

    As a Flight Forecaster, one of the tools I found immensely useful was a good comprehensive pilot debriefing of the weather they actually experienced during a flight, in comparison to the weather the flight crew was briefed on, on completion of their flight…to be used for flight forecast verification/critique.

    Pilot or flight crew feedback is essential if forecasts are to improve and meet flying needs.

  8. Bill

    I first discovered the fine-line principle many years ago, in Florida. Although, fine-lines were a well known phenomena, no one ever taught putting 2 and 2 together.

    One night, while I was tracking a severe air mass thunderstorm, after putting out a warning our outlying facilities, I saw a very distinct fine-line moving in the direction of one of them. I decided to track the fine-line. It was moving something near 60 knots. That’s when I had my epiphamy!

    I called the facility and told them that in 15 minutes, they were going to experience a 60 knot gust. They thought I was a little bit loopy, but they acknowledged the info and hungup.

    About 30 minutes later, the forecaster called me back asking with amazement, “How did you know!” I told him my magic radar told me so. LOL

    He went on to describe how one of his windows blew out, and all the debris that had blown about and the tree branches that had falled.

  9. Bill

    …That same “fine-line” principle can also apply to the onset or afternoon regression of a sea-breeze. They can also, very often, appear as fine-lines on radar and give an indication of the general movement of any associated weather (e.g., afternoon thunderstorms).

  10. Bill

    Microburst gusts are a result of a down-rush of rain and cold air from a thunderstorm. The air in a microburst is cooler and saturated, and as such, is detectable on radar. The leading edge of a microburst along the surface of the ground shows up as a radar “fine-line”. The speed of progression of the “fine-line”, therefore, is the speed of the gust you will experience, from the direction from which the fine-line is moving.

    The only problem with radar detection on the ground, in close proximity of an approaching fine-line, is it being masked by “ground-clutter”.

  11. Bill

    In the Kenner accident, at between 95 – 150 feet on takeoff, and experiencing a microburst simultaneously, does allow much time for a situational evaluation and response. Under those conditions, it would have been better to wait and live another day, rather than to proceed under suspect conditions.

    It looks like it was a foot-race against “Mother Nature”, that “Mother Nature” won.

    Here’s an article from The Boston Globe:

    Globe Editorial

    Bird-brained at the FAA


  12. Paul Miller

    AA 1420 at Little Rock could have been avoided if the crew had been assisted by dispatch and diverted to a safe airport. Also holding could have been recommended, if a divert was not the best solution. Remember that the crew was tired, but more over, the crew was task saturated. They needed some help by ATC, Dispatch and NWS. Instead they were given the impression that the landing was the optimal operational path, when in fact, landing in the wind-field of the convective activity adjacent to the airport was the least favorable of options.

  13. Paul Miller

    Microburst recovery training was begun in late 1983 at United Airlines with microburst wind field data from the National Center for Atmospheric Research NCAR. Prior to that crews were trained to keep a minimum of “takeoff safety speed,” or what is known as “V2,” a required perfomance parameter in engine failure on take off procedure training. At Kenner the flight crew was attempting to maintain V2, instead of max power and 15 degree nose up attitude, while respecting the stick shaker or stall warning.

    The case for good training could not have been made any better than by the crew working together in USAir 1549 ditching in NYC. I say crew, because the pilot flying and the pilot monitoring worked well as a team, while the cabin crew and the flight crew also worked well together. They are a good example.

  14. Bill

    That USAir ditch in New York was a perfect example of why experience and training is so crucial to aviation safety. That wasn’t just fortunate happenstance. It looked like he was doing it for a training film on, “How to ditch a plane.”

    If I were him, I’d take the “payday” and retire. He fulfilled his destiny. LOL

  15. Bill

    …And the July 9, 1982 Pan Am flight 759 crash in Kenner, LA — on takeoff during a thunderstorm. They only got 150 feet off the ground before a microburst from an thunderstorm to the East (from over New Orleans) knocked them out of the sky.

    Tragic as it was, it brought attention to and discussion of microbursts to the fore, which resulted in wind shear detection and alert systems improvements.

    That thunderstorm was more than likely a result of that “Heat Island” phenomena — given the time of year, time of day (4PM), and the direction it was moving from.

  16. Paul Miller

    This was very critical in several mishaps. USAir 1016 was one very recent example. But there are many others as well. AA 1420 @ Little Rock, Jun 1999, comes to mind.

  17. Bill

    I agree with you wholeheartedly about the need for better interactions and attention between ATC and the weather folks.

    I think that more attention is paid to your descent and what’s happening once you touch the ground, than that transition phase between them — your approach (…what you indicate as being between 2000 feet and touchdown).

    That same phase is when you begin entry into the Friction Layer (…as a mentioned above, between approximately 200 feet and the surface). That’s where most of the surface-based meteorological phenomena takes place and can effect your airworthiness and safety the most.

    Although you spend just a very brief time in the Friction Layer, that time is the most critical time of your flight — when your most vulnerable in your transition from airborne to groundborne and visa-versa on takeoff.

  18. Bill

    Are the people who make these FAA regulations actually pilots, or are these regulations reviewed by anyone who actually has any flying experience?

    Does the flying community have any critical input prior to these regulations being made "the law of the land"?

    What influence does the NTSB have with the FAA — what is the relationship, and how are the NTSB's recommendations handled to assure they're considered for incorporation when regulations are being formulated or amended?

    What Congressional oversight is there to address issues of FAA/NTSB policies and procedures and regulations, and to address grievances?

    How often are their regulations audited for relevancy, contemporary contextual accuracy, technical accuracy for newly introduced aircraft — new aircraft designs, capabilities, crew demands, workloads, skills & requirements, and training … etc; and new ground tracking and monitoring capabilities?

    It would seem to me, from what your saying, there's a vast disconnect between the realities of flying and the regulations.

  19. Paul Miller

    Let’s discuss the July 2, 1995, Charlotte NC USAir 1016 mishap. One of the missing bits of data, therefore missing finding of fact, conclusion and recommendation concerned the lack of doppler weather radar at the Charlotte Douglas Airport.

    Hmmmm, why was that?

    Let me list the following parties:
    1. City of Charlotte
    2. Charlotte Douglas airport management
    3. USAir
    4. FAA
    5. NWS

    Of these five parties, none had been able to put a Doppler weather radar into operation at the Charlotte Douglas Airport since Dr Fugita’s micro burst analysis of the 1973 Kenner LA crash. A full twenty years had gone by.

    Of interest, at least two local TV news stations in Charlotte had put Doppler weather radar units into operation.

    So with all of their resources and safety mandates, none of these 5 parties was able to do in 20 years what two TV stations did in the pace of a few years.

    So home TV viewers had availability to Doppler radar, whilst the flight crew and ATC crew did not have.

    Yet the NTSB held the flight crew to a very high standard, essentially holding them largely accountable for the mishap.

    How is it that every day, flight crew are, under penalty of law and regulation, required to comply totally with FAA control, and then when a mishap occurs, suddenly held pretty much totally responsible for the mishap?

    Did or did not FAA guide the airplane into the weather, clear them to land and not warn them?

  20. Paul Miller

    I would like to see current real time convective weather tied more directly into ATC procedures. I do not like being directed through lines of convective weather in order to get to the runway, and most especially when there are other clear quadrants around the airport.

    Closer coordination between weather office and ATC in the approach and departure environment is needed for safety purposes. Now ATC and flight-crew receive time delayed data, often just a special or “the last hourly ob.”

    Instead, I want and need to know “what is happening now in my flightpath.”

    I can handle more at 4000 feet than I can below 2000 feet. Why? First at 4000 feet the plane is still clean and although slowed, still at high energy. At 2000 feet and below especially below 1000 feet, the plane is configured with flaps, slats and landing gear hanging out, at very slow speeds, not much more than minimum, and there for at a very low energy state. I am vulnerable, very vulnerable to convective weather. Same on takeoff, til I get my gear and flaps up and a little altitude.
    I think that the ATC and metro guys should be in the same room.

  21. Bill

    I guess what I’m trying to point out is, when it comes to weather briefings and safety, quality far exceeds and outweighs expediency as a priority.

    The people who give you your weather should be able to analyze, evaluation, and make changes and decisions during rapidly changing or unforeseen hazardous situations; not just read what someone else tells them to read from some other separate and distant location.

    If pilots have doubts or uncertainties — ask questions and demand answers.

  22. Bill

    Another interesting, and more recently recognized phenomena, is the “Heat Island”.

    That’s a microclimatic phenomena where cities and heavily populated and developed metropolitan areas capture and radiate more solar heat than their surrounding, less populated and “greener” suburbs.

    Because of their distinctly higher temperature in relation to the surrounding areas, they can kickoff convective activity over the “Heat Island”, where none was forecasted, or kickoff convection sooner than forecasted, because the temperatures can be 5 or 10 or maybe more degrees warmer than the area around them.

    They can generate their own, localized weather such as localized severe thunderstorms that drift in random directions, if there’s not a strong steering flow.

    They could also cause directional, and possibly wind speed, wind shear and gusts if there’s an airport in the nearby vicinity.

    That alone, or in concert with other local phenomena, such as sea breezes, can greatly alter and effect a general terminal forecast generated at a distant central forecast office and adapted or extracted for a specific airport near a “Heat Island”.

    …as an example for the need for forecasting attentiveness and pilot alertness.

  23. Bill

    Another interesting phenomena I’ve experienced — especially with younger, less experienced pilots is — some of them tend to be a little too overconfident of either their abilities or their aircraft’s abilities to handle serious weather situations.

    As a flight forecaster, I often worried about pilots who slighted their weather briefs or made some passing comment like, “Oh, I can fly around it [over it, under it…]” or “We can handle it” or “That’s okay. I need to get my flight time in.” Some even treated the weather brief as some kind of bothersome ritual they had to tolerate to fly. It wasn’t until their first really scary bad weather experience, that they started spending more time paying attention and asking questions and were a lot less jocular.

    Extending that, to the corporate world, how does that play out with non-flying flight managers/supervisors (…for lack of a better, more knowledgeable term — LOL).

  24. Bill

    I can't really say. I've been out of the system too long to give you an accurate opinion of what current procedures are.

    But, I can say, from experience, that many times these situations are local meteorological events, influenced by local topography and geography (sea breezes, Foehn winds, katabatic winds (downslope winds off of mountains), recent frontal passages, etc.

    Most of the weather briefings you receive are focused on enroute weather. Not too much attention is paid to landing and takeoff weather — probably because it only covers a relatively brief period of time of the overall flight.

    Unless some obviously severe weather is taking place, or immanent during takeoff/landing, all you get are "wind-altimeter setting-duty runway."

    Upper wind soundings are usually taken every 6 hours; or every 12 hours (00Z & 12Z at some locations. At some locations, not at the airport, but 10 or 20 miles away or further. Not only might they not be current for a rapidly changing weather event, they might not be for your destination airport. Winds for an airport without a colocated weather office, often use the closest upper-air location, if it isn't too far away, or the extrapolate the winds between two of the closest upperwind observation sites.

    That may work sufficiently enough for stable, unchanging weather, but, in a rapidly changing or fast moving situation, it might not be sufficient enough for a landing or takeoff — the two most crucial times of a flight, where there's the least amount of time to respond to sudden change, due to altitude and the business taking place in the cockpit.

    The only exception would be, at an airport collocated with a weather office that operated a Doppler wind sensing instrument to get their upper winds — which, I believe, is constant.

    The only other method, I'm aware of, is PIREPS. Hopefully, in the case of an unexpected phenomenon, he didn't experience a problem or crash his plane to get it — like the New York accident, which received a report of icing from a PIREP just before they crashed.

    With centralized weather forecasting offices — often many miles from the airport they make forecasts, warnings, and advisories for, they also might not be accurate enough during crucial situation — not fully attuned to local effects. That could be especially true if the dispatcher has limited weather skills as just being a "Flight Briefer" — not specifically skilled with the talents or tools of a fully qualified weather "Flight (Aviation) Forecaster."

    In marginal, potentially hazardous, or rapidly changing weather situations, I'd ask a lot of questions — on the ground during your prelaunch briefings, and in the air before landing, leaving yourself enough time to make decisions in advance for marginal calls, or conditions that deteriorate worse or faster than forecasted. Get enough information in advance to allow yourself a time margin for making decisions whether to procede or not.

    …You know, "Thunk Ahead". LOL

    Personally, I'd get all the information on any questionable weather I could. Based on the level of certainty expessed in the forecasts/warning and my confidence in the experience and professionalism of the briefer. Then I'd determine what my level of confidence was in the briefing. If my confidence level wasn't very high — there were uncertainties — and the weather I was encountering wasn't as advertised, I'd find another way.

    In legal terms, if it were my life and the safety of my plane and crew were at stake, I would rather go with "a preponderance of evidence," rather than, "beyond a reasonable doubt." Waiting to go "beyond a reasonable doubt," might be too late.

    …With a preponderance of evidence things aren't going well, reasonable's good enough for me. I wouldn't go beyond that. LOL

  25. Paul Miller

    Bill, I agree. Somehow, and I do not know how, all of this is lost between Metro class and the dispatchers desk. The only time we seem to recall the hazardous weather issues are during NTSB investigations. Why is that???

  26. Bill

    I agree with you on TRAINING and the focus. But, I think that they shouldn’t be limited to advising on just “convective” SIGMETS/AIRMETS.

    Although convective activity can introduce severe problems, phenomena such as low level CAT and clear-air wind shear (wind speed and directional shear), and the potential for icing are equally as hazardous on approaches and takeoffs.

    Much of the lower-level shear and turbulance occurs in the Friction Layer — within 200 feet or so of the surface — just before touchdown and immediately during takeoff.

    Things like density altitude could also become an issue, especially during extreme heat events, which are becoming evermore common.

  27. Paul Miller

    Human factors appears to be the issue here and in most of the mishaps. The best defense against human error (human factors) is TRAINING: lots of it and with a great deal of repetition.

    We may have transitioned from training to checking in many FAA supervised flight “training” programs. Modern airplanes are exceptionally complex, beyond your wildest imaginations. There is only one way for anyone to be proficient in the many modes of the various computer controlled systems for auto flight. That is training.

    Yet while MORE training is needed, we see that trend moving to LESS training. And the time in the training scenarios is often focused on CHECKING the airman versus TRAINING the airman.

    Approach and landing mishaps are common because so much complexity is occurring in this phase- so much navigation is occurring and so much communicating is occurring. Aircraft configuration is changing quickly and dramatically and so many parameters are changing. The only way to get and remain proficient is training, and lots of it.
    Also, we still do not optimize the pilot-dispatcher relationship, required by FAR Part121. Dispatchers could be actively advising the flight crew of current convective weather dangers and suggest alternatives, diversions, holding or other safe procedures. Much more can be done here and should be done here. After all, the flight crew is in partnership with the airline in safety.

  28. Bill

    For example, a 400 foot drop at 30,000 ft in turbulence is heart thumping, but not necessarily life-threatening, in most cases, unless there’s other air traffic nearby. That same 400 foot drop at low-level could be disastrous, because there’s so little time to respond to it with corrective measures.

  29. Bill

    The ground and ground obstructions are something you aren’t too concerned about, flying at 30,000 or 40,000 feet. Shouldn’t they be the paramount and deciding factor and concern when exercising procedures at lower levels, on approach?

    Planes don’t fly well in the ground. (A layman’s observation) LOL

  30. Bill

    Is it really reasonable to expect a pilot — a human being — to respond to both, an unexpected natural phenomenon as well as unexpected idiosyncrasies of a computer program and automated mechanical functions, simultaneously, at low-level during an approach?

    Wouldn’t “hands-on” save the precious little time there is to react at low-level at relatively high speeds over the ground and ground obstructions, to “feel the aircraft” and to respond to it?

  31. Bill

    Point well taken!

    Like insurance, it seems the problems always arise for situations that are unexpected and that you aren’t covered for. LOL

    Honing one’s skills is always on ongoing evolution, no matter how perfect you think you are. Mother Nature and the laws of physics are always testing — even if the FAA’s not.

    As the adage goes: “Always expect the unexpected” […and prepare for it]

    …That’s the one that will get you if you aren’t ready for it — a point proven by the last four serious aircraft accidents.

    All four — FedEx in Japan; Delta in the UK; the Amsterdam crash; and the Continental in NY, all occurred during approach. Another point of interest.

    Wouldn’t a discussion on the use of autopilot on approach be interesting, in light of these accidents…?

  32. Paul Miller

    The crew of the 737 may have had experience in routine ops, but may have had minimal training to the seat. So in terms of “non-normal experience” they may have been low time.

  33. Bill

    [From BBC News]


    ‘High risk’ of plane fault repeat
    Experts have warned there is a “high probability” that a fault which caused a British Airways jet to crash-land at Heathrow could hit other Boeing 777s.

    US air accident investigators called for a component to be redesigned after a Delta Air Lines plane reportedly encountered a similar problem.

    Manufacturers Rolls-Royce say the new part should be ready within the year.

    It comes after tests proved a build-up of ice in the engine was the most likely cause of the Heathrow crash.

    The Boeing 777, with 152 people on board, crashed in January 2008, causing one serious injury.

    The captain and co-pilot were praised for averting a major disaster….

  34. Bill

    It seems it was a tragedy of errors. The pilot’s radio altimeter was reading erroneously — the altimeter which apparently is read and used by auto-throttle, unless it is manually switched by the pilot to the secondary altimeter. When their actual altitude was 1950 ft above the ground, the faulty altimeter was reading -7 or -8 ft, according to the article.

    And, apparently because of lack of experience flying that particular model of plane (a third pilot was onboard in the jump-seat checking out the flight crew), they failed to check the discrepency in altimeters in a timely fashion and take corrective action. They failed to respond until they got the stall warning and the “shake-stick” feature activated. Sound familiar?

    …Based on info from the Seattle Times article.

  35. Paul Miller

    I have not seen the accident report yet, but I heard that a major, common and as of today unresolved safety issue was involved. In Europe and many other places outside of the US, the transition level, that is the flight level at which the cockpit altimeter is switched from standard, 1013 hphg (QNE) to locally adjusted (QNH) is often very low. In many countries the flight level is below 10,000 feet and in some cases at and about 4000 feet.
    If through human error the crew did not reset the altimeter to QNH and the local pressure was lower than standard, then the plane would register a higher distance above the ground than it actually was. If the QNH was higher than standard the plane would think that it was higher than it was.
    If let’s say, the plane believes that it is landing on the runway, but is still several hundred feet in the air, and the plane is on auto coupled, with the auto pilot and auto throttle, then the plane would attempt to close its throttles in anticipation of a landing touchdown on the runway. The plane would then slow to a stall, while the autopilot attempted to remain airborne by raising the angle of attack, through a higher nose attitude. Eventually, the wings would stall, and the plane would fall under its own weight. This may be a rough estimation of the events which occurred. The crew could have and I will toss out here, should have intervened, by disconnecting the auto coupled approach and hand flown, hand manipulated the throttles. In this way the crew could have maintained conrol of the plane safely and landed manually. Again this is just conjecture at this time.
    The problem of low altimeter resets has been a major concern of flight crews for along time, because the flow for resetting is normally done at much higher altitudes.

Leave a Reply