Experts seemed to have missed one critical aspect of the Jeju Air flight 2216 crash video: the aircraft didn’t slow down as it slid along the runway. That video was unlike any I’ve ever seen. I’ve viewed numerous gear up landing videos, and in all cases, those aircraft slowed down relatively quickly after touchdown. But even to the eye, the Jeju Air jet doesn’t appear to slow much it all while it’s on the runway.
Last week, in episode 362 of the Aviation News Talk podcast, I talked about this anomaly, and my analysis of the video to determine the aircraft speed at various points in the video. Since then, we’ve learned that the flight data recorder and cockpit voice recorder didn’t contain the final four minutes of the flight, so much of what we’ll know about the flight will come from a video shot by a restaurant owner.
In accidents, speed is critical. If you’re going to hit something, you want to do it at the slowest possible speed to minimize injuries and deaths. Had flight 2216 slowed normally while it was on the runway, and not gone roaring off the end of the runway, lives could have been saved.
A few days after the accident, I began wondering about the aircraft’s speed as it went off the end of the runway. So I imported the video into video editing software and began examining it frame by frame. That let me estimate the aircraft speed at a number of points along the runway and in the grass overrun. What I found was surprising: Flight 2216 exited the runway at substantially the same speed it was at when it touched down on the runway! During its almost 14 seconds of sliding down the runway on its belly and two engines, it traveled 2/3rd of a nautical mile and yet it was moving at about a constant speed the entire time it was on the runway.
To figure this out, I looked up the length of a Boeing 737-800, and most sources said it’s 129.5 feet long. Then I looked for objects in the video, such as trees, the terminal building and a tower, where you could clearly see both the nose and the tail pass the point. Then I counted the number of video frames between the nose and tail passing each point, to determine how many fractions of a second it took.
The video appears to be shot at 23.98 frames per second. So 12 frames would be half a second. Then it’s easy to calculate the speed at which a 129-foot-long aircraft would have to be moving if, for example, there’s a ½ second time difference between when the nose and the tail pass the same object.
But there is some inaccuracy, probably on the order of +/- 10% in measuring the speed this way. For example, if it took 10 frames for the airplane to pass an object, if it actually passed the object half a frame early or half a frame late, we could be off by as much as 1 frame, which would be 10% of the measurement. So what’s important about the speed numbers I’ll give you is not the absolute number—because it could be off by as much as 10%—but by the trend of the speeds, as the aircraft moved down the runway. And that trend…. is what is surprising.
I made the first measurement at the point where the tail touched the runway, and created a small cloud of smoke. That speed measurement is the least accurate, as it’s the farthest away from the camera and the aircraft is a small target and it’s not perpendicular to the camera. It was the only measurement where it could have reasonably been either 11 or perhaps 12 frames. Adding and subtracting 10% from both of those measurements gives a speed range from 137 to 183 knots, with an average of 160 knots, so let’s say the aircraft first touched down at 160 knots. That of course is very fast for a 737, but you’d expect it to be faster, as the gear and flaps were up.
As the aircraft passed in front of the terminal building, it took 10 frames to pass, which is 183 knots, with an error range of 164 to 201 knots. As it passed the tower, it took 11 frames, which is 167 knots with a range of 150/183 knots. As it passed another tower, it was 10 frames for 183 knots. It then passed a sign in 10 frames, which would be 183 knots. This was probably the most accurate measurement, since the plane was nearly perpendicular to the camera. It then went off the end of the runway, and when it was about a plane length into the grass, it measured 12 frames or 153 knots with an error range of 137 to 168 knots.
So, the numbers bounce around a little bit because of the error in this measurement technique. And while experts commented on the aircraft landing fast, they have missed this point.
I can only think of three reasons why the engines might be high power as the aircraft slid down the runway.
1) Somehow the signals between the thrust levers and the engines were interrupted, and the crew couldn’t bring the engines back to idle.
2) The crew tried to activate the thrust reversers, but instead of providing reverse thrust, they provided forward thrust, and
3) The crew forgot to lower the gear, touched down on the runway…and then decided to try to go around.
And if you’re thinking, that is crazy, no pilot would ever do that. Remember that this aircraft touched down about halfway down the runway at high speed and the pilots could have thought going around would be the best course of action. And believe it or not…if they did, they would not be the first crew to have tried to do that.
On May 22, 2020, Pakistan International Airlines Flight 8303, an Airbus A320, was on a domestic flight from Lahore to Karachi. During its initial landing attempt, the aircraft touched down without the landing gear deployed, resulting in a belly landing that caused significant damage to both engines. The pilots initiated a go-around, successfully got the airliner back into the air, and travelled a few kilometers. But then the damaged engines failed, leading to a crash in a residential area near the airport. 97 of the 99 people on board died, along with one person on the ground.
Toward the end of the video, one can see that the thrust reverser is deployed on the #2 engine. It’s difficult to tell if the #1 engine thrust reverser was deployed. And we don’t know if the pilots tried to deploy the thrust reversers immediately after touchdown, or perhaps much later sometime later during the slide down the runway.
An alternate hypothesis is that the aircraft touched down at high speed, and did decelerate normally as it was sliding down the runway. But the data don’t support that theory. As I mentioned, the least accurate speed measurement was the one when the aircraft first touched down. If the aircraft was at the high end of the error range, it touched down at 183 knots. But the last speed I measured, when the aircraft was still on the very end of the runway was 183 knots, with an error range of 164 to 201 knots. And that measurement should have been the most accurate, since it was directly opposite the camera location. So even if the aircraft touched down at 183 knots, the data show it was still doing 183 knots at the end of the runway, and the only significant deceleration occurred after the aircraft was in the grass.
Many pundits have suggested that the berm at the end of the runway that supported the ILS’s localizer antennas was a key factor in this accident, but I disagree. Speed was the key factor, and speed kills. Even if there had been unlimited grass available, at the speed the aircraft left the runway, it would almost certainly have broken up anyway, though perhaps with less loss of life.
But imagine the possible outcome if, after touchdown, the crew had been able to bring the engines to idle, and the aircraft ran off the runway and hit the berm at 50 knots instead of 150 knots. 150 knots is three times 50 knots, so the energy involved in a crash at 150 knots is going to be nine times higher than it would be at 50 knots, since energy increases by the square of the velocity. That’s tantamount to the difference between a car hitting an object at 90 mph versus 30 mph. Speed kills, and the high speed at which this aircraft exited the runway is what killed nearly everyone on board.
The only way for this aircraft to have exited the runway at essentially the same speed at which it touched down is if the engines were developing high power. Hopefully investigators can figure out why there was so much forward thrust when the aircraft came firing off the end of the runway. We first broke this story on the Aviation News Talk podcast. If you’re not already an Aviation News Talk subscriber, go to aviationnewstalk.com to sign up for the free podcast.
Comments