|
Getting your Trinity Audio player ready...
|
If you ask pilots of any experience level what makes a wing stall, I guarantee you’ll get a consistent answer: “Exceeding the critical angle of attack.” This, of course, is correct, and it is information that is ingrained into our flying brains from the first day of Private Pilot ground school. The logical follow-up question, “How do you recover from a stall?”, also yields a consistent and correct answer: “Lower the nose to decrease the angle of attack.” I can say these questions are answered consistently because I ask them of every student I’ve ever taught in university ground schools, from Private Pilot to Instructor Pilot, and they always give me the correct responses. This says nothing of my teaching, rather, it points to the fact that pilots of all stripes know this intuitively. A wing stalls when it exceeds its critical angle of attack (AOA), and to recover from a stall one must decrease the AOA. If this is intuitive, why then do so many pilots kill themselves by doing the exact opposite at the one moment in their lives when all the stall training they’ve undertaken actually counted?
Exceeding the critical angle of attack is what causes a wing to stall.
Loss of control inflight (LOCI) due to an aerodynamic stall/spin continues to be the number one killer of general aviation (GA) pilots. This statistic does not discriminate; pilots of all ages and experience levels have fallen prey to the LOCI monster. It’s a great mystery given the amount of stall training and evaluation pilots undergo in their quest for freedom aloft, and it begs the question: is something missing from our training? I’ve scratched my head over this one for years, even as I’ve taught spins to instructor pilot candidates, imparting on them the tremendous responsibility they will bear to ensure their future students never commit such a deadly error. Under what circumstances are pilots finding themselves in uncontrolled stalls or spins from which they are unable to apply their training in order to recover?
Distractions are the leading cause of stall/spin accidents.
Advisory Circular 61-67C Stall and Spin Awareness Training attempts to answer this by pointing to National Transportation Safety Board (NTSB) statistics indicating that distractions are the leading culprit of stall/spin accidents. As far back as 1980 the FAA announced its policy to incorporate distractions during the performance of flight test maneuvers. So there’s the answer – distractions. But that’s a vague term, and it doesn’t do enough to explain how a pilot loses control. Dropping a pencil in the cockpit, looking for traffic, jotting down a clearance, or even losing an engine on takeoff should never lead to an unrecoverable stall. This answer – distractions – doesn’t go far enough to explain what’s really going on; therefore, I began to search for something more specific, and I may have found it.
I think the answer may lie deeper in the realm of psychology, and I can point to two different tragedies that provide a clue; one involves a pilot, the other a police officer. In my Aviation Safety class at Minnesota State University, Mankato we discuss Colgan Flight 3407 in great detail. This crash of a regional airliner in February 2009, killing 49 onboard and one on the ground, ignited a flurry of new regulations ordered from on high by Congress and President Obama when he signed the Airline Safety and FAA Act of 2010. Examples include the requirement for first officers of Part 121 airlines to hold an Airline Transport Pilot (ATP) certificate, specific requirements for more rigorous ATP training, the requirement for airlines to review training records when hiring new pilots, and on and on. Because so many of these new regulations affect the careers of my students, I felt it was important to take a deeper look into what really happened on that fateful Colgan flight. What I found shocked me.
In summary, the captain stalled the airplane while intercepting the localizer for the ILS 23 approach into Buffalo, NY. How he stalled it is the shocker, but it may also provide the clue that explains all the other LOCI accidents mentioned above. I will forgo the details (anyone can access the full report online) and just take you to the moment when the captain responded to an artificial stall warning called the Stick Shaker. This system is used on larger airplanes to mimic the stall buffet at speeds generally five to eight knots above the stall. The stick shaker on this flight caught the captain by surprise because it activated much sooner than expected, and the captain responded by pulling back on the yoke. Let me repeat that. When warned of an impending stall, the captain pulled. This captain, holding an ATP certificate, would’ve undergone countless stall training throughout his career, from Private Pilot to his current level, knowing intuitively that a push is required to lower the AOA. Why did he pull?
The captain pitched up 19 degrees, even while the airspeed decayed rapidly.
Herein lies the clue. The NTSB report of this accident contains a detailed transcript of conversations and other sounds picked up by the Cockpit Voice Recorder (CVR), with each sound time-stamped to the tenth of a second. It also adds elements of flight control inputs gleaned from the Flight Data Recorder (FDR). At 2216 and 27.4 seconds the stick shaker activated. At 2216 and 27.8 seconds the control column moved aft. That’s 0.4 seconds. In other words, this pilot responded literally in a split-second to the stick shaker. The artificial stall warning startled him, and without thinking his muscle memory kicked into gear and he pulled on the yoke. Why? Because pulling generally gets us away from the ground and all the dangers associated with that. I think this is analogous to someone stomping on their car’s brake pedal at the first indication of a skid when that is often the worst thing to do. But the brake pedal usually brings us to our happy place; it stops the car and the danger is averted. The foot goes there automatically, without any connection to the brain. In an airplane pulling back on the yoke or stick brings us to our happy place, further from the ground and further from danger. This is also something we learn from day one of pilot training, and for the Colgan captain it was perhaps this muscle memory that overcame the appropriate response of pushing to reduce the AOA. This pilot may have committed something called “Action Error”, and that leads me to the story of a police officer.
The reader might remember the tragic shooting and killing of a young man in April 2021 by a police officer who had mistaken her firearm for a taser. This occurred in Brooklyn Center, MN less than a year after the death of George Floyd. Racial tensions were high, so this occurrence and the subsequent trial of the officer involved made front-page news. My interest in this tragedy piqued when I heard the defense arguments offered by the police officer’s attorney. In the heat of the moment, when the young man jumped back in his car to escape arrest by speeding away, the officer yelled “Taser!”, intending to use non-lethal force to stop him. But muscle memory kicked in and her hand reached automatically for her firearm. This action error, as explained by an expert witness on her behalf, occurs in high stress situations, and the hands, unconnected from the brain, will revert to the predominant action for which they’ve trained. The witness even pointed to aviation accidents, or mistakes in surgery, as prime examples. In the officer’s case the majority of her weapons training involved her firearm versus her taser, so that’s where her hand went.
In flying, as much as we practice stalls, the predominant behavior our hands learn is that pulling equals climbing. On every flight, for example, we always pull at rotation speed to get away from the ground. The captain of Colgan 3407, when faced with an impending stall while IFR at night and relatively close to the ground, suffered a disconnect between his brain and hands and pulled on that yoke for all his might. In fact, he even pulled against the Stick Pusher, another safety system designed to help a pilot lower the nose in a stall. The stick pusher, in this case, activated three times, and at each successive activation the FDR indicates a stronger and stronger pull by the captain, culminating in 160 pounds of aft force applied during the third activation. That’s a lot of panicked yanking on the yoke to get away from the ground that was rapidly rising to meet them.
This could provide the clue as to why so many other pilots may have done the precise wrong thing when entering a stall near the ground, pulling rather than pushing while falling to their deaths. In my search for the answer I took a deep dive into the NTSB database, looking for all general aviation accidents that resulted in a fatality due to an aerodynamic stall/spin. My goal was to find the smoking gun, a common trigger event such as the Colgan stick shaker that caused these pilots to go against their training and misapply the proper stall recovery procedure. The problem with GA accidents is the scarcity of information available for the NTSB to analyze. An average GA aircraft isn’t equipped with a FDR that can provide details such as control inputs and control force applications. Oftentimes the only information available is the wreckage itself, or eyewitness testimonies. Finding the smoking gun proved to be futile. Nevertheless, the NTSB does a fine job of piecing together all available information, usually culminating in a detailed eight to ten-page report. This information may provide enough clues to guide future thinking regarding stall training.
In analyzing these reports I’ve discovered a few commonalities that point to my theory. A review of the 100 most recent fatal stall/spin accidents since the time of this writing (going back to 2017) reveals that 90% resulted from a stall that occurred less than 1000 feet above ground level, and most of these (81%) were in the vicinity of an airport. Additionally, the weather almost never played a factor with 94% occurring in Day VMC and with excellent visibility and high ceilings or clear skies. Experience levels of the pilots committing these fatal errors were all over the map. This was not a discriminator; something else was going on. I believe these pilots, distracted by something in relation to the runway or airport environment (an engine failure after takeoff, for example), stalled their airplane, and once the stall occurred, they were caught by surprise and pulled all the more. In other words, the initial stall wasn’t the culprit; rather, it’s what happened immediately after. Because they were so low to the ground (<1000’) an element of stress kicked in and they committed a similar action error as identified in the trial of the police officer, and in the case of Colgan 3407. With the ground rushing up to meet them, they pulled with all their might to get away from it when a push was necessary to recover. It is counter-intuitive to push when near the ground, but a sacrifice of altitude is necessary to get the airplane under control.
This theory may not apply to all LOCI accidents resulting in a fatal stall or spin crash, but given what we know about the Colgan flight it is reasonable to assume that other pilots behaved the same way. That is, they ignored their training and pulled. And that leads back to the question posed earlier: is there something missing from our training? According to the Private Pilot Airman Certification Standards (ACS), pilots are evaluated on their ability to “configure the airplane, establish a pitch attitude that will induce a stall, and execute a stall recovery”, among other things. In other words, stalls are predetermined maneuvers as specified by the evaluator. There is no element of surprise to evaluate a pilot’s muscle-memory response to a startle-stall. Of course, this is something that would be difficult, if not impossible, to train and evaluate safely. Technology could be used, such as programming a flight simulator to stall at an unexpected moment while close to the ground, but in a flight simulator the fear element is absent (there’s no fear of actually dying.) Such is the case also when practicing stalls in an airplane at a safe altitude (minimum 1500’ AGL.) And it’s difficult to surprise a pilot with a stall in an airplane. Without those two elements being present – surprise and fear – action error may never surface.
The point is not to supplant today’s method of stall training, which is significantly valuable. Pilots must learn as a primary skill to recognize the behavioral characteristics of their aircraft before, during, and after a stall, and to recover from stalls with the airplane under control. I am merely suggesting that we explore new methods to augment existing stall training and evaluation. Can we do more on a checkride than just evaluate pre-canned power-on and power-off stalls? This would require additional research and testing by FAA working groups and steering committees, but something has to be done to reduce the still significant number of fatal LOCI accidents.
In the meantime, nothing prevents individual pilots from seeking additional skills in aircraft handling. In fact, I believe any training that involves maneuvering an airplane outside of normal parameters – aerobatics, spins, upset recoveries – is extremely valuable. The airplane used for this specialty training may have different handling characteristics than what most GA pilots fly, but there is direct application when considering the confidence gained as a skilled aviator. Upside down is upside down, whether in a Piper Cub or Pitts Special. Pilots should get comfortable getting uncomfortable so that when an abnormal situation occurs – a surprise stall low to the ground, for example – they do not panic. Such training should also incorporate muscle-memory: learning easy steps that can be taken home and practiced on the living room sofa (i.e. chair flying) so that action error never sneaks up on an unwary pilot.
Pilots should get comfortable getting uncomfortable so that when an abnormal situation occurs.
Regardless of whether or not one invests in the specialty training suggested above, every pilot should take their stall training very seriously. Stalls should never be square-filled items on a checklist – complete them and then move on to the next maneuver – and they should always be trained for more than just meeting the standards set forth in the ACS. With a qualified instructor pilot, and at a safe altitude, stalls can be practiced in many different scenarios. Slow speed acceleration stalls, secondary stalls, deep stalls, trim stalls, cross-controlled stalls (with an appropriate aircraft) and many more are all valuable when training and building muscle memory. If every pilot heeded the above suggestions, I’m convinced LOCI statistics would drop significantly. This is how not to lose control, according to what we can learn from Colgan 3407.
- How Not to Lose Control: What We Can Learn from Colgan 3407 - June 5, 2023
Great article, Jeff. FYI, you’ll find “begs the question” in Roget’s under “avoid” or “sidestep”.
I have nothing to do with aviation BUT YOU ARE 100% CORRECT. The appropriate physical movement of the yoke (design) must come naturally to the pilot in a time of stress, through may be audio warning or/and new design for the yoke to match the natural humans thinking while death is knocking the door. MODIFY YOKE MOVEMENT DESIGN.
There is always something about the Colgan flight conclusion that bugged me. I remember that time and tail plane icing was a big thing. Colgan 3407 was operating in icing conditions was I believe was accumulating ice. What is the recovery procedure for tail plane icing? The opposite of wing stall — pull on the yoke and retract the flaps. I wonder if the Captain was reacting to his recent training. I haven’t seen this elsewhere so I’m probably nuts but still ……
It was discovered that the crew had received required training on tailplane icing and I remember at that time that the buzz was that we were recovering from stalls in icing incorrectly.
Only certain types of aircraft were susceptible to tailplane stalls but the FAA had a one size fits all mentality which was subsequently found to be dangerous.
https://www.faa.gov/documentLibrary/media/Notice/N_8900.267.pdf
There is a lot in this accident that doesn’t get talked about much. The captain also had experience in the Jetsream, which had an accident or two due to tail plane icing. (Hibbing MN being one). His reaction had elements of a tail plane icing issue. The sad part is the Q400 probably would have flown right out of it if they added power and kept on the rudder.
Thanks for your comment, Kurt. The NTSB concluded that this captain was not responding to a tail stall, and I agree with that conclusion. You’re correct that the recovery procedure, according to the NASA tail stall video, is the opposite of a wing stall (i.e. Pull vs. Push; raise the flaps vs. keeping them in place; etc.) But I highly doubt that in the 0.4 seconds following the stick shaker activation that this pilot thought to himself – the airplane is stalling, we’ve accumulated ice, and over six months ago I watched a video that tells me to pull on the yoke rather than push, so I think I’ll pull, and I’ll even continue to pull as the stick pusher fights me to get the nose down. This captain panicked and yanked on the yoke, and then kept on yanking. The first officer, on the other hand, may have been applying those procedures when she raised the flaps. She had seen that video only one or two months prior to the accident.
The NTSB got this one wrong. Clear indications that BOTH pilots responded with a tail stall. The fact that the stick pusher was NEVER part of any simulator training event and that they had watched the now-replaced FAA icing video prior are clear indications. Stall training is not conducted in a realistic manner in general. This is not a matter of surprise, although that could be part of it. This is more that the entire way we train stalls for both large and small airplanes is NOTHING like what the lead up to how an actual accidental stall occurs. Pilots are not recognizing the stalls due to cognitive bias taking over when the actual scenario does not match their training. Consider the way that recognition primed decision making can go wrong, when a situation is similar, but not quite the same, as the one encountered a person can react to the wrong thing.
Do I recall correctly that both pilots were exhausted because of long flights to get to the airport for their flight? That must have compounded the problem. They weren’t thinking clearly.
I have wondered the same thing myself and discussed this with a retired airline captain who was of the same opinion. If you search on YouTube you will find videos from NASA that show tailplane icing tests in a DHC-6. They clearly show the pilot pulling on the yoke to unstall the horizontal stabilizer. The Colgan first officer retracted the flaps which is consistent with the tailplane icing training.
Kurt, I agree with you (and Duane, below). In my opinion, that is EXACTLY What was happening. For several years prior to that incident, most (if not all) the 141 and 142 commercial and airline training centers were POUNDING “Tailplane Icing” scenarios in their simulator training programs, going so far as to show the FAA’s “Tailplane Icing” video at least once during every Initial and Recurrent Class. It was shown in every one of the classes I attended. I think this crew was the eventual outcome (the first “unintended consequence”) of what I do believe was a sincere attempt by all the training facilities to remind pilots there are more than one type of icing events to consider before deciding on how to recover from an icing-related incipient stall condition. I think this largely inexperienced crew (by their own admission if you listen to the CVR made public by the NTSB) DID fall back on their training (of Tailplane Stall) by 1. The Captain’s action to pull the Yoke Aft, and 2. The F.O.’s immediate Un-commanded and Un-announced retraction of the Flaps, because their training scenarios had been almost exclusively Tailplane Stall related. In effect, that is ALL they had trained in the simulator outside of the basic “Maneuver-based stall recovery” exercise during the first sim session of their various training events. And it came back to bite everybody in the butt, in this case. Right after that incident, I noticed that many Training facilities stopped showing the FAA’s “Tailplane Icing” video. Coincidence? I don’t think so…
Great article. I think most pilots don’t spend enough time flying and maneuvering just above stall speed. I fly jets, rotorcraft, and gliders. I believe flying gliders has made me a better pilot than anything else. After coming off the tow plane, we can struggle for a long time a few hundred feet above a ridge, making figure eights while flying 5-10 knots above stall speed, with a wind affecting the shape of the turns. Battling for altitude and maneuvering just above stall speed really teaches you to not pull on the stick even when you really want to!
Agreed, Richard (and great article Jeff). I’m still a novice pilot < 120 hours & 35 of those were 30 years ago when I had to put my dream on-hold for a long while & restarted the training in late 2019. Just a couple quick thoughts/suggestions again from a novice perspective: 1. Students need to understand the "why we do this" for various flight maneuvers – specifically for this article it's "why do we do/practice power-off & power-on stalls", suggest that instructors "check-in" with their students on the "why" at various times so that they remember the why in case they forgot or are overwhelmed with the current topic or lesson. 2. I'm struggling a bit with a combo cross-wind landings and a new-to-me, narrower runway than what I had at my training airport. I recently acquired a wonderful-to-me old 172G, a gentle old airplane that I really want to keep nice and not tear it up with a terrible landing. So I'm becoming very familiar with go-arounds :) My point here is that go-arounds are excellent training opportunities for stall avoidance. The airplane is configured to land, it wants to land, and suddenly you're forcing it to climb. power-off, carb-heat on, with flaps, nose-up trim, you apply full power & carb heat off, that nose even in a tired old 172G wants to reach for the sky, so the pilot really needs to get that situation under control, and they learn the feel of a mushy airplane, slow speeds, getting that nose down to help get the airspeed up, etc.. Perhaps instructors could incorporate more go-arounds during training (random points "deer on the runway!" when the student is about to land, that kind of thing) is what I'm suggesting, is all. Again just comments from a novice's perspective, thanks.
I flew hang gliders for about a decade – I was already an airplane pilot when I started. When you’re learning to fly a hang glider, on later low-level training flights (before going to the top of a mountain) you will sooner or later make the mistake of getting slow while still 10 or so feet in the air. At that point, as the glider begins to accelerate toward the ground, every instinct demands that you keep the nose up. The impact on arrival isn’t dangerous – that’s the point of limiting the start height – but it’s quite a shock nonetheless: it’s a hard landing where your body is the landing gear. After just one or two of those experiences, and the immediate feedback from your instructor (“get the nose down”) you actually do get the nose down, and the glider picks up just enough airspeed to allow you to stop the descent by flaring as you reach the ground. The resulting absence of pain is almost a spiritual experience.
This is shockingly effective at altering the startle response.
To this day, my response to being startled in an aircraft is to get the nose down – now. This seemed like a good reaction when I flew sailplanes, and although it can surprise passengers (and instructors) because if I get startled they get startled (!), it seems like a generally sensible one in airplanes too. Could it cause trouble? Theoretically, near the ground, maybe, but my hang gliding experience also taught me to flare before impact so I don’t think so. I think it’s more likely to keep me out of trouble: for example, I probably don’t need to remember to use rudder to keep the wings level in a stall because if I’m startled by a wing drop I will push – and the ailerons will work fine.
I don’t know if this startle response is common to all hang glider pilots. I don’t know how you train it – sure, you could startle students and rap their knuckles if they don’t push, but we’re trying to improve student retention. Still, I have to think that training pilots to push when startled would be a good idea.
I was one of the accident investigation party members for the Colgan 3409 accident and you neglected to mention that the first officer is on CVR announcing the uncommanded retraction of the flaps during the go around. This rapid loss of lift insured that the stall was unrecoverable.
You beat me to that point, John, I wonder why the NTSB mentioned it but never put it in the summary. Also, they theorized that it was her training as a CFI that made her “muscle memory” react that way. So now nearly every pilot has to become a CFI for 1,000 – 1,500 hours…..
Another point missed in this article was that at the time, training was only accomplished for “approach to stall”, not “full stall prevention” which is the standard now. The concept for recovery was completely different – wing is still flying, so power out of the situation. I was told this was from the days when they trained in the aircraft (early days of 707’s, DC-8’s, etc very few, if any simulators), therefore, safety in doing the maneuver was more important than how it was done, and the PTS at the time gave acceptable altitude loss targets which we were trained not to exceed, so no diving down to recover.
Thanks for your comment John. I’d be interested to hear more about your participation in the accident investigation. I completely agree that the First Officer’s raising of the flaps was contributory to this accident, but that wasn’t that point I was trying to make. (“I will forgo the details (anyone can access the full report online) and just take you to the moment when the captain responded to an artificial stall warning called the Stick Shaker.”) In trying to solve General Aviation stall/spin crashes, where we don’t have the luxury of a FDR or CVR, can we learn from this Captain’s split-second panic reaction to the stick shaker? (0.4 seconds.) Is that what many GA pilots are doing when they kill themselves, and is there a way to train against that?
The same way that pilots are told to react to a tailplane icing stall….
I have a theory, based on nothing but my own uneducated thinking. In stressful situations, many humans have an instinctive reaction of assuming the fetal position, curling up into a ball until danger passes. For a pilot strapped into his seat and holding the stick/yoke, that instinctive reaction will result in pulling on the controls. So, danger threatens, and a pilot reacts by pulling. Take that theory for whatever it’s worth.
Peter, I think your theory is spot on! (Similar to my analogy of stomping on the car brake in a skid.) How do we train pilots against that?
Have them maneuver the aircraft at altitude, close to the stall, for a few hours, until their startle response converts. It’s no guarantee – glider pilots do a whole lot of that, and sometimes they spin in too – but it might help.
Go fly. Go fly slow. Listen/feel. Understand muscle movements may be full deflection inputs. Do recoveries visually and on instruments. If you know somebody with an aerobatic plane go fly in all attitudes, lots of nothing but sky or nothing but ground out the front. Understand that sometimes all it takes is a slight unload (push) and when you jam in the power your feet should be moving too. Do it enough and you’ll realize it’s no big deal…and if you tickle an edge inadvertently some time in the future it’ll be a yawn to fix.
I am a new private pilot but I agree with Dave Lewis, it’s important to make the maneuvers scenario based instead of checking the box on the elements of a maneuver. If you go into a power off stall maneuver as if you are on final approach, and act like you are on short final when you stall, you are a lot more likely to take the proper actions when you are faced with the situation in real life. This may not have direct application to the case study in the article, but it could be helpful in GA.
“Muscle memory” is the PC term for panic. Choose life, stay proficient or stay on the ground.
Instinct, which I consider to be a reaction done without thinking, needs to be developed with training to deal with an event so the correct instinctive reaction is chosen. This a skill that fades, thus recurring training is needed. Having a real life experience to reinforce the training embeds the correct instinct.
Several months after receiving my private ticket I was on short final with the wind straight down the runway. Crossing a line of tall pine trees, the airplane suddenly sank toward the ground with the stall warning blaring. I would have to say pushing the nose down with the ground rushing towards me was the hardest, and most certainly non-instinctive reaction I’ve ever done while flying. Thankfully, the training I had received overcame any incorrect ‘instinctive’ behavior.
In the 42+ years of flying since that event, pushing the nose down (and adding power) is done whenever needed or anticipated without any confliction (as opposed to the first time).
When I took an intro aerobatics course some 43 years ago, two powerful lessons stand out, which I still remember well.
The first lesson: Although I was already a CFII and always taught my students that “an airplane can stall in any attitude at any airspeed”, it’s a hard concept to teach or learn in the average trainer. My aerobatics instructor emphasized it this way: we were recovering after a multi-turn spin in a vertical dive (he used the term “competition recovery”) with the airspeed climbing toward 100 knots when he told me to “pull too hard” on the stick. So we’re vertically diving, and suddenly the wing stalled. “Now recover from the stall”—so I pushed us over semi-inverted in order to get the wing flying again—hard to do when the ground is rushing up at 8 or 9000 fpm! Then of course I recovered from the dive by more gently pulling to prevent the secondary stall. Good lesson.
The other lesson: the first time I rolled inverted in level flight, I had an enormous instinct to pull on the stick—which promptly started us diving toward the ground. I learned to roll and push—helped some by some nose down trim before starting the roll—but it was still hard to overcome that instinct. I think I’d developed that instinct from doing steep turns when we were required to do them up to 60° bank, which required a healthy pull to maintain altitude.
Incidentally, that last lesson became a life saver when I was later caught in the wake turbulence of a 737 unexpectedly. The airliner was actually below my airplane, but the mountain wave effect caused its wake turbulence to rise instead of fall. Suddenly the TR182 was rolling to the right onto its back—not all the way, but somewhat past vertically, to which I responded with hard left rudder and pushing on the yoke. The airplane righted, and we flew on to Denver, but the memory is still quite vivid.
Missing from the discussion of the Colgan crash is the question of the fitness of the Captain. Given his demonstrated incompetence in several failed checkrides throughout his career, it seems obvious that he was deficient in several essential areas and, importantly, failed to disclose (covered up) these previous failures in presenting his resume to hiring authorities at Colgan! If a complete resume (warts and all) had been presented, would he have been hired or even offered an interview?
This whole sordid tragedy is a near perfect illustration of Chaos Theory where a single change at the beginning of a process leads to a dramatically different outcome!
Never get in a hurry with an airplane. Gravity always wins!
Interesting article. Is there any data to support your theory that pilots who have completed upset recovery training have a lower incidence of GA stall\spins in the pattern? Certainly, aerobatics and upset recovery training can only improve piloting skills, but I’ve seen many authors make this same recommendation. Any actual data?
Mark, I am not aware of specific data regarding GA pilots. That’s a good hypothesis for future research, although I think it would be difficult to do given that UPRT is not required for GA, nor is there any standard for how it’s taught. Likewise for aerobatics. One thing I do know; however, is that every single certificated GA pilot who stalled/spun their airplane into the ground had received stall training and had been examined on that training according to the PTS or ACS. If, after all these years, LOCI still continues to be the number one killer of GA pilots, then something has to change regarding how we train. There may be researchable data in the future as the European Aviation Safety Agency (EASA) now requires “on aircraft” UPRT for obtaining a commercial pilot license. I wonder if Europe will see a future drop in LOCI because of this. I live in Minnesota and have taught all four of my kids how to drive. We spent many driving sessions on an ice and snow covered parking lot, skidding the car around and having a ball “playing” in a safe environment. This can only help them should they ever enter an unplanned skid on the roadways. Why not apply the same to pilots? (The military does this, by the way. My first checkride as a student pilot in the Air Force included loops, rolls, and spins.)
There’s something missing in the 20-20 hindsight of this discussion – there was a total lack of situational awareness contributing to the incorrect pitch control response since “intercepting the localizer” is nowhere near enough to the ground to say that fear of imminent ground contact by itself was the major factor for the fatal mistake.
Hi Jeff:
Another confirmation data point. A couple years ago in recurrent sim for my corporate flying job, I was having a particularly bad session which resulted in me pulling it into a stall with the shaker and continuing until I “crashed “. In spite of 40 years and 10K hours of flying. Chalk it up to lack of coffee, a bad nights sleep in the hotel at training, or whatever. My brain was definitely “disengaged” from my actions at that moment. A poignant lesson for me as I fly, operationally.
Jeff,
Good piece. I’ve had my own hypothesis for some time, arguing that the initial startle response may be due less to a fear of the ground, but rather to a fear of appearing “out-of-compliance”. But it’s pure hypothesis. Your point about the difficulty in extracting information from NTSB reports is spot on…we could discuss that at length. That said, in research done into inflight icing over the past 25 years, I have identified 22 cases in which the initial response to a stall indication (shaker, buffet, etc.) was a confirmed application of back pressure. In those cases without a DFDR, the only reason I know this is because the pilot survived and documented it, so obviously he/she eventually stopped pulling and started pushing. That research was FAA-funded and can be found here:
https://www.tc.faa.gov/its/worldpac/techrpt/tc14-44.pdf
Additionally, I recently penned an article on Linked In regarding the whole stall recovery issue and particularly, the interpretation of the Colgan accident. That article is here:
https://www.linkedin.com/pulse/did-we-get-colgan-right-steven-green-fraes/?trackingId=hn4vU1k3Tc%2BDqNJoFQyM3Q%3D%3D
I agree we should look at training for an answer. I would start with how we have always approached stall training. I think there is a fundamental flaw in that initial stall training. The methodology is to push the wing till it stalls. So the student is nervously sitting there waiting for that indication that a full stall as been achieved before pushing the nose down. The student is taught to ignore all the warnings of an impending stall. Ignore the stall warning, ignore the change in engine pitch, ignore the airspeed indicator, ignore all sensory information about the impending stall – until the plane stops flying only then can you push the nose down. That is some significant muscle memory training at at early stage. DON’t PUSH THE NOSE DOWN UNTIL THE FULL STALL!!! Maybe we should make it a fail if the student doesn’t immediately react to the impending stall indicators and actually waits to stall the wing. You pass by not stalling. We teach what to do to recover from a problem instead of teaching how to not cause the problem.
Certainly a valuable input. My recollection is that approach has been tried. For several years prior to Colgan recognition of incipient stall and minimization of altitude loss was taught. 47 years ago during my initial training, a stall was allowed to develop until aerodynamic buffet, regardless of artificial indications. Then, probably about 15 years ago, in the sim, instructions were to add power and to not decrease pitch attitude so much as to lose altitude. After Colgan, now we’re back to “breaking” the stall by lowering the nose.
I’m not a CFI, just a corporate line pilot so maybe instructors can more precisely comment on the changes in stall training over the years.
Paul, I agree with you 100%. From my earliest training flights I wondered why in the hell am I still pulling on the yoke with the stall horn blaring! Negative training at it’s worst. Too much training emphasis on stall recovery and too little on stall avoidance. Especially since the most common stall, base to final turn, is essentially unrecoverable any way.
As an instructor, i have used the Colgan accident in presentations. One thing that I did not see in your article was in the NTSB presentation was that there appeared to be a ground proximity warning alert to “Pull Up, Pull Up” at about the same time the stick shaker activated. I wonder if that did not assist in his miss handling of the aircraft? A similar accident several years ago on a fatal go around I think also was aided by outside events, when the tower said he needed to go around with left traffic, and the pilot’s first response was a quick pull up and steep turn to the left, and the aircraft (still dirty) rolled over on its back and crashed. After 65+ years it is still valid to “fly the airplane to the crash!!”
Your article reminded me of the only thing in my life that was similar to a panic and LOC that have cost a lot of pilots their lives. A bad car accident. I had a 1970 Camaro that I had set up with impeccable handling (for the time period) with Koni shocks, sway bars, etc. I loved practicing driving fast on curvy roads and was totally into figuring out how to come to a complete stop or avoid an accident. I knew locking up the brakes was a big no as we didn’t have antilock brakes back then. I had successfully avoided some close calls where I almost hit a cow, etc but had always done what I had planned. I knew my car was uncontrollable in a skid. I was only 18 and coming down a highway around a fairly tight corner and was doing the speed limit. There was a car on the right shoulder as I came out of the corner that decided to do a U turn right in front of me. Instead of using all the skills I had learned during constant practicing, I panicked, locked up the brakes and skidded right into her as she was doing her U turn. No one killed but both cars totaled. I relived that over and over and knew for sure that I could have steered around her to the right onto the grass and back onto the road if I had not gone into a panic skid. I hadn’t related this to flying until your article and totally see why good pilots still auger into the ground on gorgeous days.
It seems to me that when a pilot crashes after stalling the plane the stall occurs quickly and unexpectedly. A long time ago while practicing touch and go’s with a student we hit a downdraft at the end of the runway and only 200-300 feet up. All of a sudden I found myself in an airplane not only not climbing anymore but actually starting to descend. I pulled up a little more hoping to arrest the sink but quickly triggered the stall horn which I didn’t expect and it caught me by surprise. Everything was happening quickly and took me a few seconds to absorb the situation, but my muscle memory kicked in and I pushed the nose down. I honestly thought we were going to end up in the street as the TV antennas were not too far below but we slowly started climbing away from the ground.
I have always questioned the jet simulator training I’ve been doing for the last several decades regarding stalls. There have been quite a few corporate jets stalling and crashing during circling or maneuvering for landings and this is where in my opinion stalls should at least be practiced.
This is an insightful and thought provoking article and a very important read in aviation literature. It begs the question, where does critical thought end and rote action begin. There was an obvious disconnect between the split second actions of this captain and copilot and what would be the normal reaction if time allowed for critical thought and action…..Studying this topic is important and although the loss of life and property are always tragic, in this instance it has and can be a catalyst for future in-depth analysis…..great job.
In a Private Pilot check ride, practicing stalls is usually telling the prospective pilot to display a power-on/off stall and show how to recover from it. A variation might be for the Examiner to instruct the candidate to not recover until requested, so the amount of penetration into the stall is an unknown. But even then, the candidate knows that a stall is coming, lots of time to mentally prepare. As discussed in the article, how do we make a stall unrecognizable event until until an immediate response is needed- even on an Instrument ride, where spatial illusion is simulated under the hood, we know what might be coming.
It’s like simulating a vacuum failure on the Instrument ride- the examiner covers the AI (steam gauges), and we are prepared to follow whatever action we were taught to mitigate the loss of an instrument. We all know that when such an event happens in reality, it’s nothing like that, and the event only comes initially evident when there’s differential information from different instruments conveying the same information.
So, how do we create a genuine “surprise event” while flying at a safe altitude, that might teach the correct procedures instinctively rather than by rote? I am not an ATP, but have been in a B75/76 full motion sim when pilots were being taught recovery from a near-inverted event (“wake turbulence” was the issue in the sim). Talk about getting scared and surprised- on the initial event these two ATP pilots didn’t know what was coming- I nearly required a change of underpants, and the terror experienced by the pilots was obvious. Will economics ever allow full-motion simulation to be used for GA training and testing?
I seemto recall that this Captain had already been
previously let go by two other companies for failure to fly to the standards the companies required.
Can you comment on that?
Great article. I, too have thought long and hard about the Colgan accident.
First, great comments by all. And good points regarding other factors present during the event.
Those other factors notwithstanding, Jeff’s question regarding stall training in general is a valid one.
May I suggest we begin to teach, when the aircraft stalls, or if startled, the first step is do nothing for a count of three. That’s right do nothing PHYSICALLY for three seconds. During that time, analyze.
In most cases (not sure if this would apply in an inverted flat spin), if in a normal category airplane, the attitude will trend towards a recovery.
In all aircraft I’ve flown from Cessnas and Pipers to Falcons and Gulfstreams (both conventional and fly-by-wire), when stalled, the aircraft begins to recover on its own. There is no reason to hurry. The pilot really doesn’t have to do anything except let go of the yoke/stick.
In fact a colleague who flew FA-18’s stated, if they found themselves unable to recover a loss of control situation, before punching out, let go of the controls to see if the aircraft might recover itself.
Like a lot of folks stated above, how we were trained has contributed to these accidents. I suggest training students that stalls are only problematic if we make them so. Teach them that stall accidents are not usually caused by pilots not acting, but rather by very quickly acting incorrectly.
Let’s take our hang glider colleague’s comments above to heart and translate his experience/training and apply it to power aircraft syllabus.
From “Stick and Rudder” by Wolfgang Langewiesche, page 33:
“And THAT is the real danger of stalling: this faulty reaction to the stall, rather than the stall itself. It is quite rare that a pilot is killed simply because he stalled. But it happens with tragic monotony that a pilot is killed because, stalled when he did not expect it, he either fails to recognize the stall for what it is, or fails to control that impulsive desire to haul back on the stick”.
And on page 247:
“…, your first reaction must be to get the stick forward. … You must trust yourself to do this even if you should be close to the ground. And that’s a large order.”
In my experience, the best advice for pilots young and old is to read Langewiesche’s “Stick and Rudder”. Had the Air France AF 447 and the Colgan pilots read Langewiesche, their passengers would still be alive …
This inept crew led to Congress changing the hourly requirements to become an airline pilot. It is not the hours in the log book that makes a proficient pilot, it is the quality of the training received. Towing banners or flight instructing to build time does not compare to building time in the right seat of an airliner under the command of an experienced captain. ALPA loves the new rules that has created a shortage of airline pilots that increases the pilot’s value. As a retired airline captain with 9 jet airliner type ratings, I totally disagree with the 1500 hour requirement.
Well maybe. It would have been insightful to know the transition time you had from certificate to certificate to airline copilot to place this comment perspective.
I had about 400 hours and a new commercial and instrument when I got hired by TWA. A nice captain let me do the ATC communications when when I was a 707 flight engineer, since I had no real time instrument experience. In 1968 the ATP required 1200 hours. I had 1270 .hours when I was type rated on the 4 engine Convair 880 in captain training. TWA
Captain training was up or out. You made it or you were terminated.Glen, I believe 1000 hours in the first officer seat of an airliner is better training and experience than 1000 hours of instructing or towing banners.
Although you have received many good responses I feel I have to add mine. I believe your theory is correct. I was trying to find a way to land our Scout towplane short on a downward sloping runway. The runway sloped downward on the approach end when landing into the prevailing west wind. I discovered the Scout would not sink enough to match the slope. A day or two later I flew my RV-6 to the field and tried to get the proper sink rate. I thought I was above stall speed, but quickly learned I was not. At about fifty feet I stalled and the RV dropped like a rock. Without thinking my instinct took over and overpowered my training and previous experience. I yanked the stick back before my brain kicked in. My first thought was the RV is toast. I have only a theory as to why it did not hit the runway. It stopped short by about a foot. I think then I had recovered enough to add power to fly down the runway, pull up, and go around. My belief is that the stall was close enough to the ground that downward momentum had not fully developed and the plane and I were saved by ground effect. I do not know if that is correct, but that was the last time I tried that. It was super frightening and I am still in disbelief that I yanked the stick aft. At fifty feet the closeness of the ground was not in question. It was right there.
From what i remember this pilot and copilot were carrying on a conversation on approach that had nothing to do with flying. I’d say lack of attention to what is going on including aircraft speed was a contributor to this crash. Pay attention to takeoff and landing as you do know this is not automatic and anything might occur. You just cannot be relaxed in these configurations!
Finally, Glen brought up what I think is the biggest factor in this accident. No one was flying the plane, all this talk of a stall, when there would be no stall if someone was on the gauges. Have you ever ridden with someone in a car or plane that has to face you when they talk? I really believe this “pilot” was face to face with a co-pilot having a conversation,and at the risk of flaming, it was a female,and may have caused more inattention.
Wonderful article that leads me to my military helicopter flight trainer. As a medevac pilot flying low and slow in the deadman’s curve very close to the ground I understand the human fear factor of striking the ground to be a high distract or. The military rigorous training is such that we pilots fly close to the ground often. Over time this reduces the distract of the fear associated with striking the ground. We practice nap of the earth maneuvers and emergency procedures at such altitudes. When I transitioned to fixed wing platforms, my training also transferred. I was also fortunate to have fixed wing instruction with combat vet CFIs that afforded me low flying in fixed wing and practice emergency procedures at very low altitudes under safe conditions. Flying close to the ground with a CFI that has a proper plan for such training May in fact be an additional opportunity to reduce the awful problem the GA community has with stalls. Stay safe and always fly the the experienced CFIs.
Jeff, while I am no NTSB accident investigator it is reasonable to suspect Colgan 3407 was likely an attempt to overcome a tail stall in Icing conditions gone bad. That said, your premise of incorrect reflex actions in stall situations low and slow is spot on. If I were to reflect on flight training I believe the emphasis has always been stabilized approach and energy management. While I don’t disagree, anyone who has an instrument rating understands how easy it is to become cognitively overloaded. During those critical moments when seconds count you have to have simple reflex actions. For me its push and add power (I would not do well in a tail stall unless I was talking it through and anticipating like Colgan 3407). I believe many of the stalls that occur in GA are from failure to prevent going below DMMS in the first place. The second is the insistence in training to go around if anything becomes unstable. I agree with this conceptually (especially early on) but then it doesn’t foster the stick and rudder skills needed to deal with the unexpected. I don’t know what the answer is but I suspect having the simple mantra of stall warning push-power will serve the GA community well.
This is a very interesting article, particularly because I had not heard of that airliner accident. It reminds me of the Swedish bizjet being flown by the PiC when his attitude indicator suddenly showed a sharp pitch up; he pushed immediately to counter the apparent pitch change and 90 seconds later he and his copilot were dead, from FL300 or thereabouts.
I used to teach on various fighter aircraft, including F16 (MacDill AFB), Tornado GR1 and little Grobs. I also taught Spatial Disorientation at the Royal Air Force’s Centre of Aviation Medicine. From all the time teaching cognitive failure and distraction were key factors in human failures. For instance SD is less of a problem unless the pilot is distracted, then it becomes a killer. Cognitive failure is serious when at a critical stage of flight, such as when close to the ground, or flying in degraded visibility etc . . . In other words, it’s not normally just one failure.
We have to stop using the term, “muscle memory.” This just distracts us, as exemplified in the above discussions.
Memories are in the brain, not the muscles, and memories are the result of conditioning or in this case, training. Electricity may be in the wire, but the source is the generator.