Deadly Deception! Unraveling the Mystery of Atlas Air Flight 3591.

- This sequence of incidents lasted a total of 31 seconds, approximately the same time it takes to drink a large glass of water. During these 31 seconds, years of delayed legislation, a misleading resume, personal shortcomings, and an incredible chain of events came together in the worst possible way. Stay tuned. - 100, 50, 40, 30, 20, 10. - Atlas Air Flight

3591

was scheduled as a domestic cargo

flight

from Miami International Airport to George Bush Airport in Houston, United States. The

flight

took place on February 23, 2019, but the complicated circumstances that led to what ultimately happened to the aircraft had begun more than a decade earlier. Because in response to several accidents and incidents that occurred during the 1990s, caused in part by pilot performance problems, a law called the Pilot Records Improvement Act or PRIA was implemented in the United States in 1996.

This law forced all airlines to review the training and employment history of any pilot they wanted to hire. And the background check had to cover at least five years prior to the start of the new job. The idea of ​​this law was to highlight any deficiencies in training that may have been noted during the pilot's initial training or with their previous employers. But it soon became clear that the law was missing some important details. After the horrific crash of Colgan Air Flight 3407 in February 2009, where the pilot made some very basic handling errors, it was discovered that he had been struggling during his initial training and had suffered multiple checkride failures.

Now this was exactly what PRIA was designed to detect, but the background check failed to do so. This led the National Transportation Safety Board to recommend that the PRIA should be modified by implementing something they call the Pilot Records Database or PRD. In this database, the training information of all pilots should be automatically uploaded and easily searched with the help of the pilot's license number, making the background check easier to perform and much more accurate. The United States Congress approved this suggestion and directed the FAA to implement the PRD within a reasonable time frame. But that period became increasingly longer without the implementation of the law being any closer.

Finally, the deadline for implementation was set as April 30, 2017. But unfortunately, that date passed without anything happening. 40 days after this deadline expired, Atlas Air hired the pilot who was to become the pilot in this story as first officer on its Boeing 767 fleet. So why is it important? Well, the first officer was 44 years old at that time and had a total flight experience of just over 5,000 hours. He had a very problematic training history behind him and, for example, he had resigned from two of his previous airlines, Air Wisconsin and CommuteAir, because he was unable to complete their initial training.

But in his application for Atlas Air, he conveniently did not include those airlines on his resume, stating that instead he had been studying for a college degree during those years, where his employment had a notable gap. This was partially true. He had been studying, but he had also failed his initial training at both airlines during that time. And one of those jobs was within the five-year scope of the PRIA legislation. He revealed that he failed an oral exam with Trans State Airlines, but did not reveal the subsequent failure to update his ERJ145 air transport license. And then it also failed a line check that wasn't mentioned.

But what perhaps should have been the biggest red flag in his training history on his resume, which, by the way, he also did not disclose, was his unsatisfactory attempt to promote to captain on ERJ175 with Mesa Airlines, which occurred just a year ago . A few months earlier he was doing his interview with Atlas Air. These training records later discovered showed several very worrying trends, including that when he was faced with a situation he was not expecting, he would sometimes react by simply starting to press random buttons and make movements on the flight controls to be seen. . like doing something instead of taking the time to properly assess the situation with his colleague.

And that's worth remembering. Now, all of this pointed towards an individual with personality traits unsuitable for a commercial pilot. But he was able to mask these deficiencies in interviews mainly thanks to PRIA legislation that relied on the applicant's honesty. It was very difficult for the airline to find records that the pilot did not voluntarily disclose and that was exactly why the pilot records database was designed to classify. But there was, however, a note in the pilot's paperwork that should have caught the attention of Atlas Air's recruiting team. That note came from his employment at Mesa Airlines and said the pilot would return as a first officer after he completed his command refresher course.

This allowed the co-pilot to pass the interview, which, by the way, did not include a simulator evaluation. And then he began training on the Boeing 767 with Atlas Air in the summer of 2017. Immediately, there were signs of trouble as he struggled to understand the aircraft's performance calculations and systems during his ground school training. He received additional training for this and eventually passed his oral exam, allowing him to continue to the fixed base part of the type rating. The type of specific training that pilots must undergo typically involves a ground school course and then fixed-base simulator training that is largely dedicated to learning scan flows and normal procedures.

And finally, a full flight portion in a moving simulator where emergency and driving scenarios are practiced. In fixed base training, his difficulties continued as he could not understand the normal procedures, which led him to undergo more training before he was allowed to enter the full flying portion. But only after a few sessions throughout the flight did his simulation partner get tired and complain to the instructors that the first officer was holding him back. This led to him pausing his entire flight simulator training and then restarting it from the beginning a few weeks later with a different partner.

Unfortunately, after starting again, training was interrupted by the effects of a hurricane that lasted several days. And after that, he failed his first type rating skill test on the 767. The examiner's stated reason for his failure was unsatisfactory performance in crew resource management, threat and error management, approaches that they are not about precision, sharp turns and just general judgment. However, the type rating examiner said that the first officer was very nervous during the check. And after another retraining session, he finally passed the type rating course and began his flight training in the real aircraft. So should this lackluster performance have raised any red flags?

Well, the airline's training director had reviewed the first officer's file, but concluded that there was no need to subject him to specific performance monitoring, in part because of the hurricane disruption that had disrupted the formation of him. At the time, it was thought that this interruption might have been the reason for his failure in the checkride. Now, after these initial problems, recurring training and monitoring over the next two years had gone quite well with satisfactory results until the day of flight

3591

. That was the story of the first officer, but what then of the captain? Well, the captain was 60 years old and had been working for Atlas since 2015.

He was quite experienced, with just under 11,200 hours of total time and 1,252 of those hours were flown on the Boeing 767. His training record was much better, but he had been under performance observation after a failed training day while still a first officer in 2015. This performance observation protocol was completed approximately a year and a half later with satisfactory results. And then he succeeded in his command refresh training in August 2018. Now, it must be said very clearly here that any person can suffer a single failure in training for a variety of reasons. Things like personal circumstances can make anyone have a bad day.

So having failed a single test doesn't necessarily mean anything. But repeated failures are a completely different matter. In any case, these were the two pilots who were scheduled to fly Atlas Air Flight 3591 together on February 23, 2019. They had flown together earlier the day before, during which they had been flying a night service. But this flight was scheduled as a nice local departure at 11:00am with a flight time of only 1 hour and 45 minutes. So in other words, it was potentially a pretty relaxing day. The pilots met before the flight and began reviewing pre-flight documentation while the cargo they were going to transport was loaded onto the plane.

There were no particular NOTAMs that were planned to affect them, but the weather around Houston seemed a bit challenging. A cold front was forecast to pass through the area from the northwest. And when colder air moves toward a warmer, moister air mass, it gets stuck under the warmer air and forces it upward. This can cause heavy rain and even thunderstorms to form. But the good news here was that the worst weather should have passed by the time they planned to land. They discussed this and then decided on the final fuel they would carry and then walked towards the plane.

Everything was being loaded according to standard operating procedures and the weight and balance of the aircraft were within limits. On the way to the plane, the crew was joined by a third pilot who wanted to ride with them in the jump seat to Houston. Obviously, this wasn't a problem as there were plenty of jump seats available in the 767 cabin, so he could join us. The co-pilot was going to be the pilot of the flight, so, once the captain verified that the technical condition of the aircraft was completely free of faults, the captain went outside to complete the tour.

And the co-pilot began to work on the cockpit configuration and prepared for departure. The plane they were going to fly was a 27-year-old Boeing 767-300, which two years before the flight had been converted into a freighter. It was in good condition and, as I mentioned before, had no recorded technical problems. Once the first officer was done with his cabin setup, he looked at the time on his nice big wristwatch and saw that they were getting close to their scheduled departure time of 11:00 and reported it to the captain. The final checklist was completed and the plane backed up, started its engines and got underway for takeoff.

At 11:08, Atlas Air flight 3591 took off normally from runway 09 in Miami. The initial climb and cruise sections of this flight were uneventful and the aircraft stabilized at its planned cruise level of 400 or 40,000 feet. As I mentioned earlier, this was a relatively short flight, so as they began to approach the top of the descent, the captain left the active ATC frequency to listen and record the latest weather for Houston. ATC information indicated that they could wait for one of the west-facing runways, most likely Runway 26 Left or 27, which would give them the fastest taxi route to the stop. But the actual landing strip would be given to the crew later, when they were in contact with the approach ATC sector.

The weather was fine, with some light to moderate rain in the area and some light westerly winds. Once this information was received, the copilot handed over the controls to the captain and began configuring the FMC for the planned arrival and approach. They talked about using the automatic brake in position 2 and landing flaps of 25 degrees. And with winds reported at 320 degrees at 14 knots, the crew decided to use a flight speed of 135 knots for landing. While they were discussing this, air traffic control came in and informed them that after they had passed the an-am point called GIRLY, they could continue on the LINKK ONE arrival route towards Houston.

And this coincided very well with what they had prepared to do. The problem, however, was that the cold front that had been predicted to pass was now between the plane and the airport. The front line was filled with active storm clouds and this would likely force the pilots to have to request radar vectors further ahead away from the arrival route to ensure they did not have to cross those storms. And this will be very important very soon. At 12:07, the Houston controller called the aircraft and gave them their first descent clearance to levelof flight 350. This was quickly read by the captain and then the aircraft began to descend.

Over the next few minutes, the pilots discussed various speed and altitude restrictions at waypoints. And there seemed to be some confusion as to how these restrictions should be interpreted. Sometimes a given restriction only applies if the plane follows a specific arrival, but not if it follows another arrival. And this discussion took quite a while during the initial descent. Air traffic control continued to grant more descent clearances. And at 12:12, the co-pilot handed over the controls to the captain and began his approach briefing. And speaking of briefings, here comes one from my sponsor. I just want to tell you about a fantastic streaming platform that I'm sure most aviation enthusiasts will love: Curiosity Stream, today's sponsor.

For a truly affordable price, Curiosity Stream gives you access to thousands of award-winning documentaries and series anytime, anywhere. With Curiosity Stream, you'll be able to dive into topics like the evolution of the jet engine, the secrets of the SR-71 Blackbird, and the story of the Wright brothers' first flight. You can also stream thousands of other programs, covering everything from science and technology to history, culture and sports. With new content added every week, I find that there is always something new to discover on the platform. I recently watched an incredible documentary that takes you on a journey through the history of flight, featuring some rare archival footage and interviews with aviation pioneers.

So if you're ready to take your love of aviation to new heights, head to Curiousstream.com/mentourpilot or scan this QR code on the screen here. If you use the promo code mentourpilot you will save 25% on your purchase, making it great value for money. Thanks Curiosity Stream. Now, let's continue the story. As the plane descended toward Houston, the crew was told to switch radios to Houston Center on the 133.8 frequency. The captain read this new frequency again, but waited to call the next controller until the first officer finished his approach briefing, took the controls, and the descent checklist was completed.

The new controller confirmed the arrival route and the captain asked what runway they could expect. Now the controller responded that he didn't have that information and that seemed to bother the captain a bit who was now trying to plan his descent to the airport. Knowing which runway to expect is very important, as approaching different runways will mean a different number of runway miles and that will have an effect on how fast a plane needs to descend. Descent planning in large aircraft is all about energy management. The more miles you have to fly, the slower you have to descend to get rid of the energy.

But if you plan on having lots of runway miles and suddenly get a shortcut to a closer runway, you may have to descend much faster, and you'll even have to use the speed brake or even the flaps to achieve it. The other problem with not knowing what runway and approach to expect is that pilot monitoring might have to go to the FMC CDU at a later stage and reprogram it for a different approach than they had previously reported. If that happens, this will take the pilot away from his task of monitoring the progress of the flight, and this will soon become important.

For the next few minutes, the pilots discussed the different speeds that would be used during the descent. The captain wanted to fly faster, as we often do, but the co-pilot pointed out that there were speed restrictions as part of the arrival route. Therefore, increasing speed might not be a good idea since they would have to slow down soon anyway. These are all pretty standard discussions that you would hear in any cabin. And it looks like the CRM was working pretty well at this stage, with just some standard banter. At 12:30:13 the aircraft was instructed to transfer to the Houston approach controller.

Now, they would finally be told which landing strip they would get. The controller welcomed them to the new frequency and then gave them the name of the current ATIS information as well as the altimeter pressure setting. He then told the crew to fly the published transition route for Runway 26 Left. As soon as the captain heard this, he dove into the FMC CDU and began setting up the final part of the route toward the approach. Once this was done, the crew adjusted their altimeters to the current pressure and the captain began activating the approach checklist. When that checklist was completed, the first officer started complaining that he didn't think the autopilot's virtual navigation or VNAV was doing a great job.

He expected him to stay a little higher than he currently was and to start slowing down sooner. And these comments led to a discussion among the pilots about how little they trusted the automatics. VNAV was probably doing what it could at the time. But every time the pilots updated the route or set another speed restriction, VNAV would have to reprogram its calculated descent path, causing fluctuations in the way it flew. At 12:34:08 the approach controller called the crew and warned them of a line of heavy precipitation looming just ahead of them. This was the cold front they had seen in the planning stage.

And the controller told them to let him know if they needed vectors to avoid. Basically, turning around the storms in front of them. And it is at this point that the actions that would eventually lead to the accident really begin. Because while the crew was watching their weather radar, trying to decide which path to take to avoid the storms, they were now handed over to the next approach controller. When the captain checked in, the new controller wanted to know if they needed vectors to the east of the weather front, which would give them a significantly longer route, or if they had other intentions.

The captain told the controller to stay alert. And now, suddenly, the first officer makes a statement about his instruments. It was not possible to hear on the cockpit voice recorder exactly what the problems were with his instruments, but the captain responded with a quick laugh and then the first officer handed the controls to the captain in an attempt to try. and quickly solve any problem. As soon as the captain took the controls, the first officer walked over and pressed the electronic flight instrument switch on the right side of his instrument panel. This switch was used to change the source used for instrumentation on your side from a primary source to an alternative.

Apparently it was a well known fix on the Boeing 767 that if the displays started to malfunction for any reason, pressing the EFI switch momentarily would reset the fault and then the switch could return to the normal position again with everything fixed. While the first officer did this, they continued to discuss what would be the best route for them to avoid the storms. The first officer was manning the radios while he was trying to fix his instruments and now asked the controller if they could turn west to avoid the clouds that way. This would give them the shortest route to fly and would probably save them quite a few minutes.

The controller came back and said the only problem with that vector was that there was a lot of outbound traffic in that sector that they would have to avoid. The pilots discussed this and then replied that they were fine going east because they had plenty of fuel on board anyway. But the controller who had heard his first request had now figured out a way for the aircraft to take that westward vector. But it would require them to maintain a higher descent rate, up to 3,000 feet, to stay clear of departing traffic. The first officer read this and the captain set 3000 feet on the mode control panel and selected grade change as the descent mode for the autopilot and autothrottle.

The EFI button had already been pressed and apparently fixed the problem because the captain laughed briefly and the first officer responded that EFI always fixes the problem. They were now instructed to turn left on a heading of 270 degrees. And after putting it on, the captain returned the controls to the co-pilot. As soon as the first officer took control, he remembered what the controller had said about maintaining a high descent speed, so he reached out and extended the speed brakes to flight position E10. This raised the wing spoilers, creating more drag and therefore increased descent speed. Now, whenever the speed brake was used, standard operating procedure at Atlas Air was for the pilot flying to keep a hand on it, as a reminder to put it away when it was no longer needed.

The speed brake is located on the left side of the throttle quadrant. So in order for the first officer to hold it, he needed to reach under it and pass the thrust levers. The captain was now back in the FMC CDU, trying to connect the route in the most sensible way to obtain descent guidance from the VNAV. And as I mentioned before, this also meant that he probably wasn't looking at the instruments at the moment. He asked if he should connect the route at a point called GRIEG on the approach and the co-pilot agreed and at the same time requested that flaps 1 be extended to further increase drag and descent speed.

Flap 1 was selected and the route to GRIEG was executed. And the captain then called LNAV available, which was not the right thing to do here since the plane was on a straighter course and should not follow that new route that he had created. But the first officer still attempted to engage the LNAV, but since the plane was not on an intercept course at that point, he did not engage. This caused a bit of confusion before they both realized the mistake they had made. And this is a small thing, but it shows some level of degraded situational awareness here, possibly from the added stress of avoiding storms, getting out of traffic, and planning the route to the approach.

The plane was now descending about 7,000 feet and had just entered some clouds associated with the front that were about to pass. This meant that the pilots were now flying in instrument conditions without any external reference and at the same time some light to moderate turbulence began to shake the aircraft. At 12:38:31, as the aircraft descended to 6,300 feet, a distinctive click was heard on the cockpit voice recorder. Immediately after this click, the thrust levers began to move forward from their inactive position toward near full thrust and the nose of the aircraft began to slowly pitch upward from an attitude of about minus 1 degree to 4 degrees nose forward. above.

So what was happening? Well, do you remember that big wristwatch the first officer wore? It is very likely that since he was holding the speed brake lever with his left hand when the turbulence began, that wristwatch came into contact with the TO/GA button, which on the 767 is located at the bottom of the levers. of thrust. That click heard on the cockpit voice recorder was therefore likely the TO/GA button being activated by mistake. Now, if the toggle switch were activated, it would send signals to the autopilot to maintain its current trajectory, maintain airspeed, and begin climbing at at least 2,000 feet per minute.

In any case, it is likely that none of the pilots noticed the TO/GA activation at this stage because none of them mentioned anything. Instead, the first officer began retracting the speed brake and moving it to the down and arm position, possibly in reaction to the movement of the forward thrust levers, as that would have been the reaction in the event the aircraft stalled. was leveling in the authorized area. altitude. Now, to understand what will happen during the next 31 seconds, we need to examine the way the human body reacts to certain accelerations and movements. We humans basically use three different ways to determine orientation and movement within a space.

The first way is with the help of the vestibular system, which is located in the inner ear, where the otolithic organs detect direction and accelerations. The second is the nerves in the skin, muscles and joints that sense their position based on gravity, general sensation and sounds. And finally, we use our vision, which uses visual information to make sense of all those other feelings. The problem is that the vestibular system cannot distinguish between the acceleration caused by gravity and other types of accelerations, such as that of an airplane that accelerates or maneuvers, for example. And once the visual cues disappear, as in this case, when the plane was inside a cloud, a type of sensory illusion known as illusionsomatogravica may begin to affect pilots.

And it can happen very quickly if we are not careful. This illusion is caused by the effect of thrust added to the plane, which caused a sudden acceleration. That acceleration would have forced the pilots to ease back into their seats. And at the same time, the plane began to pitch, causing exactly the same type of sensation. Within the vestibular system, both accelerations would have been added, and without a visual reference point, the first officer's brain would have interpreted these two accelerations as a much steeper and more sudden rise. In reality, this is exactly the same type of illusion that we use in full flight simulators to simulate accelerations and decelerations.

Now, the correct way to stop the progress of this illusion would have been to concentrate on the artificial horizon in front of him, which would have accurately told him that the pitch was much lower than what he was feeling. But remember, this pilot had a history of reacting immediately under stress without taking the time to carefully evaluate the situation in which he found himself. So what happened now was that the first officer began to bank the plane forward toward the ground. The autopilot was still activated, but the force of the co-pilot's descent command was enough to override it and the aircraft began an accelerated descent.

This then led to the next problem, which was a decrease in the plane's vertical acceleration and a rapid increase in forward longitudinal acceleration, which likely only made the somatogravic illusion even worse, causing a sensation of falling backwards. A very unpleasant feeling. This feeling is known as a reversal illusion, since the feeling is completely opposite to what is actually happening. This is probably why the first officer now yelled "Oh" and "whoa" as he began to lean forward even further. A sudden shock like this can trigger a fight-or-flight response in the human brain, which will encourage a strong physical reaction but severely hamper the brain's ability to accurately assess what's really going on.

This often leads to impulsive and incorrect actions known as the startle response, which is exactly what happened here. As the pitch was decreasing, the aircraft's descent rate continued to increase rapidly. An owl warning was activated in the cockpit, likely caused by the autopilot warning that manual input was overriding it. And he was followed by the first officer who yelled, "Where's my speed?" and "We're stalling!" Now, there was nothing in this scenario that should have made the first officer believe that the plane was stalled. The speed was well above the stall onset speed and was increasing rapidly.

There was no stick shaker and the red and black barber poles on the speed screen were nowhere near the speed they were flying at. The only possible explanation for this call was that the first officer felt the plane tilt violently and therefore possibly stall. Again, a quick scan of the instrument would have clearly shown what was going on and there is no indication that there was anything wrong with his instrument at the time. Now, the combined forces and acceleration caused by the sudden pitch forward, combined with the increasing acceleration of the aircraft, would have caused a force vector acting on the first officer's inner ear similar to a vertical drop with the attitude of the aircraft at the level of the wings. position.

Probably a truly terrifying feeling. Now, if the co-pilot had thought that he was stalling, the correct initial action would have been, in fact, to unload the wings by tilting them forward. But the continuation of that procedure should have been to recover the airplane to wing level once the indications of stall, shaking or shaking, had disappeared. But in this situation, there was no real indication of loss other than what was in the first officer's head. All the things that have happened up to this point, after the inadvertent activation of TO/GA, only took about 13 seconds, which is about the same time it took me to say this sentence.

Everything happened very quickly. But what happens then with the captain? What was he doing and why didn't he intervene? Well, when the first officer inadvertently activated the TO/GA button, the captain was still working the route in the FMC CDU, and had also just received a message from air traffic control that he was responding to. This probably meant that during those first crucial 13 seconds, he did not monitor the instruments and did not notice the extreme inputs that the first officer was suddenly making. When the owl warning sounded, he would have looked up, trying to make sense of what was happening, but the same sensory stimuli the first officer was experiencing would have affected him as well, probably slowing down his assessment and leading him into some situation. . real initial confusion.

If he had been paying attention to the instruments when all this started, he probably would have been able to take the controls and disengage the automatics much sooner. But from the position he was in now, everything would have been a sudden chaos and very difficult to assess correctly initially. The fact that he was with his head down when this all started would also have explained why he didn't realize that the flight mode annunciator had gone into roll mode as soon as the TO/GA button was accidentally pressed. In any case, at the time 12:38:51 seconds, the plane had entered a terrifying dive with an attitude of 46.1 degrees pitch down, and with that, a rapidly accelerating speed exceeding 359 knots.

This is above the airplane's maximum speed and about 100 knots faster than the maximum speed with flaps 1 extended. And just to give you a reference of how serious the situation of this aircraft was, a situation of alteration of the pitch of the aircraft, from which the pilots are trained to always try to recover, is defined as the aircraft that exceeds 25 degrees of nose-up pitch or 10 degrees pitch. below. In this case, the airplane had a nose attitude 36 degrees lower than what would be considered a rollover, descending 3,500 feet with a vertical speed of about 9,000 feet per minute. Here, the captain must have slightly gained orientation from him because some pitching force was recorded coming from the left control column.

And he was heard shouting, "What's going on?" The thrust lever was now also temporarily moved to idle and then raised again, but we don't know which of the pilots actually did that. But the fact that the captain now began to try to stop was not accompanied by a formal order from my controls, which would have been the correct thing to do. Instead, the two pilots were now pushing and pulling in opposite directions. The first officer could not answer the captain's question about what was happening. Instead, he continued making panicked calls, which he continued to do for the rest of the flight.

Now, the reason we know that the two pilots were working against each other at this point was because the flight data recorder began recording different angles of the left and right elevator at the rear of the horizontal stabilizer. Because there is a function that, in the event that one of the controls becomes stuck, will allow the other control to override the clutch and control the elevators separately to allow control of the aircraft despite the jam. This requires quite a bit of strength, and that's exactly what was happening now. Unfortunately, the captain's contrary orders were not enough to stop the descent that the first officer had already set in motion.

Vertical speed increased to more than 9,000 feet per minute as the plane descended 3,000 feet. So why didn't the captain take the controls? Well, in a rapidly developing emergency like this, it would have been very difficult for the captain to really understand what was happening and what was causing it. He would have verified that the co-pilot was not incapacitated and that he was still manipulating the controls. And because of that, it would have been very difficult for him to judge whether the first officer was trying to solve the problem or whether he was actually causing it. Intense shock and surprise work against logic and clear decision making.

And for the captain to go from programming the FMC to understanding that the first officer was suddenly pushing the plane's nose toward the ground in just a few seconds was probably too much to ask. So his delayed response was probably due to a combination of startle, confusion, and shock. In any case, the two pilots continued to input opposite controls for a total of about 10 agonizing seconds before the pilot in the jump seat yelled, "Get up!" This probably happened when the plane passed about 3,000 feet, which corresponded with the cloud base, and below the clouds, visibility was good.

This would have made instantly clear to both pilots a terrifying picture of their attitude and descent trajectory. And from this point, the first officer's control inputs changed to the maximum. Together with the captain's input, they managed to raise the nose to about 16 degrees below the horizon, causing a wing loading of around 4G while pulling at full force and with the overspeed warning blaring behind them. in the cabin. Unfortunately, all of this was too little, too late. There simply wasn't enough altitude to recover the plane. And at 12:39:03, just 31 seconds after the TO/GA button was inadvertently pressed, the plane crashed into the mudflats outside Trinity Bay with a vertical speed of 12,850 feet per minute and an airspeed total of 434. knots.

The plane was immediately destroyed and scattered over an area larger than 12 acres of mostly shallow water and muddy swamps. And the three pilots died immediately. Now, some of you may be wondering: why didn't any GPWS warning sound in the cockpit? With that kind of track record, that warning should have been triggered early and could have snapped the first officer out of the somatogravic illusion of him. Well, it turns out that the aircraft was equipped with an upgraded Honeywell GPWS system that was working fine during the accident. But the problem was that the angle and extreme speed the plane had as it approached the ground caused the computer to flag the radio altitude data being fed into the system as unrealistic.

That caused the computer to ignore the radio altimeter for three seconds, preventing the system from issuing a warning. But even if it had worked, it wouldn't have made any difference. At that point it was too late. The rescue effort very quickly became a recovery operation, as it was clear that there would be no survivors after this accident. The plane had been reduced to millions of very small pieces of debris that were scattered over a huge area. And it took specialized machines several weeks to recover most of it and take it to a hangar, where the pieces were placed as precisely as possible.

Both the cockpit voice recorder and the flight data recorder were eventually found. And although the audio quality of the voice recorder was not very good, a picture quickly emerged of what had caused this accident. The National Transportation Safety Board concluded that the accident had been caused by the inadvertent activation of the aircraft's TO/GA button, which then caused a change in pitch and acceleration that was initially undetected. These changes caused the first officer, who was the flight pilot, to experience a somatogravic delusion and, due to his documented personality weaknesses in stress management and situational awareness and simply general pilot suitability, he was unable to recover from the illusion and instead pushed the plane into a dive from which they could not recover.

The final report was highly critical of hiring procedures at Atlas Air, which had been unaware of the first officer's command upgrade failure just months before recruiting. But, above all, criticism fell heavily on the fact that the FAA did not implement the database of pilot records at the time of the accident. The first officer had clearly withheld some of his training history at two other airlines. And if this database had existed, a simple search of his license number would have revealed both the flaws and his attempt to hide them. And that probably would have prevented him from being hired in the first place.

This was an accident that really should not have happened. And it highlights how important it is for airlines to carefully vet their pilots and for important safety-critical legislation to be implemented as quickly as possible. The pilot records database has since been implemented. And this accident also led to someinvestigations into the possibility of implementing the most advanced military GPWS system on civil aircraft. A system like that would have been able to predict the plane's trajectory sooner, which probably would have been useful. Now, watch this video below. Or if you are interested in seeing more incidents and accident videos, check out this playlist.

Please consider supporting my work by sending a big thank you, purchasing awesome products, or join my fantastic Patreon team. All support is greatly appreciated. Have an absolutely fantastic day and see you next time, bye.