Midair collision at 37,000 feet
March 27, 2009 10:45 AM Subscribe
The sky is a really big place, right? So how did a Boeing 737 and a Legacy 600 private jet manage to collide head-on at 37,000 feet over the Amazon jungle in Brazil? William Langewiesche's detailed analysis of the 2006 crash--which killed all 154 aboard the 737--provides some answers.
It seems the sky just isn't as big as it used to be: "Until recently, head-on airplanes mistakenly assigned the same altitude and route by Air Traffic Control would almost certainly have passed some distance apart, due to the navigation slop inherent in their systems. But this is no longer true." The navigation systems of the two aircraft are so precise that when they met--at a closing speed of about a thousand miles per hour--the Legacy 600 was only 2 feet below and 30 feet to the left of the 737.
It seems the sky just isn't as big as it used to be: "Until recently, head-on airplanes mistakenly assigned the same altitude and route by Air Traffic Control would almost certainly have passed some distance apart, due to the navigation slop inherent in their systems. But this is no longer true." The navigation systems of the two aircraft are so precise that when they met--at a closing speed of about a thousand miles per hour--the Legacy 600 was only 2 feet below and 30 feet to the left of the 737.
- Another look at the incident from Salon's Patrick Smith
- Another analysis of an aircraft disaster by William Langewiesche, the 1999 crash of EgyptAir 990 (previously on MeFi)
Well, the other part of the problem is the tradition of limiting planes to just a few designated airway corridors, especially when we have all this fancy TCAS, GPS, and IRS/INS technology. Packing all the traffic onto fixed corridors just multiplies the chances a plane might occupy your X/Y/Z space.
posted by crapmatic at 11:13 AM on March 27, 2009
posted by crapmatic at 11:13 AM on March 27, 2009
Nice article.
Well, the other part of the problem is the tradition of limiting planes to just a few designated airway corridors, especially when we have all this fancy TCAS, GPS, and IRS/INS technology. Packing all the traffic onto fixed corridors just multiplies the chances a plane might occupy your X/Y/Z space.
Well, what's the alternative? Fly wherever you want? I also could see airway corridors with sufficiently randomized margins so that no two planes would be flying along the same exact route until the route is 'recycled' after a month or two ago. But that would require more details and more confusion when a plane has to be flown manually. Rather, like the article points out, it seems prudent to divide corridors in terms of altitude for direction -- odd altitudes like 37,000 for east, 36,000 for west. You're right, but there's no better alternative -- that's like saying the practice of having traffic lanes and thus packing all of the cars onto fixed lanes is to blame for having traffic accidents.
posted by suedehead at 11:35 AM on March 27, 2009 [1 favorite]
Well, the other part of the problem is the tradition of limiting planes to just a few designated airway corridors, especially when we have all this fancy TCAS, GPS, and IRS/INS technology. Packing all the traffic onto fixed corridors just multiplies the chances a plane might occupy your X/Y/Z space.
Well, what's the alternative? Fly wherever you want? I also could see airway corridors with sufficiently randomized margins so that no two planes would be flying along the same exact route until the route is 'recycled' after a month or two ago. But that would require more details and more confusion when a plane has to be flown manually. Rather, like the article points out, it seems prudent to divide corridors in terms of altitude for direction -- odd altitudes like 37,000 for east, 36,000 for west. You're right, but there's no better alternative -- that's like saying the practice of having traffic lanes and thus packing all of the cars onto fixed lanes is to blame for having traffic accidents.
posted by suedehead at 11:35 AM on March 27, 2009 [1 favorite]
William Langewiesche is the son of Wolfgang Langewiesche, who is the author of the seminal text Stick and Rudder.
posted by backseatpilot at 11:36 AM on March 27, 2009
posted by backseatpilot at 11:36 AM on March 27, 2009
Well, what's the alternative? Fly wherever you want?
Actually, yeah, RNAV.
posted by crapmatic at 11:38 AM on March 27, 2009
Actually, yeah, RNAV.
posted by crapmatic at 11:38 AM on March 27, 2009
The sky seemed big, until I had the opportunity to pilot a small plane over the L.A. basin. Granted, it's a high traffic area, but all of a sudden you realize that planes are everywhere, in all three dimensions, all of them moving swiftly (and "they're all coming right at us!"); but only able to change course with the slow, measured grace of oceanliners.
I gained tremendous respect for pilots and ATCs that day.
posted by malocchio at 11:50 AM on March 27, 2009 [1 favorite]
I gained tremendous respect for pilots and ATCs that day.
posted by malocchio at 11:50 AM on March 27, 2009 [1 favorite]
Also, wow, excellent article. He certainly has a gift for narrative. Interesting that he seems to harbor a lot of hate for the bizjet industry.
There are some very basic human factors work at play here (the issue with the warning lights and bells in the article is basically what I did at my last job). In hindsight, sure, the transponder failure was a catastrophic one, but what can you do? If every failure indication involves a large red blinking icon and klaxons in the cockpit, the pilot quickly learns to ignore all of them. Tiered warning systems are important to declutter the noise, and a transponder failure is not Catastrophic, with a capital C, in the safety-of-flight sense. Only issues that pose an imminent threat to the aircraft and its occupants really justify the kind of warnings that will make you jump out of your seat.
Modern systems and avionics have undoubtedly made the skies safer. There's a reason why you can take off from JFK even though you can see your hand in front of your face outside. However, it does seem to be leading towards pilots taking a backseat approach to flying (hah!). My takeaway from this article is that modern flight training needs to impress upon FLY THE AIRPLANE as Rule #1. You are not supposed to be learning how the radios work while you're in the air anymore. We have simulators, interactive software for your computer, and you can push every button and twist every knob before you ever step into the cockpit.
Maybe I'm biased, but I'm against all of the OMG BAN THE SMALL AIRPLANES!! rhetoric that I hear sometimes. If you want to carry more passengers to more places, we're going to have to start using smaller aircraft to ferry fewer people at a time to smaller, more remote aircraft. This is where equipment like TCAS comes in very handy. Pilots will still need to learn the very basics and stick to them, though. When you're in clear air, fly the airplane, see and avoid, and don't get distracted.
posted by backseatpilot at 12:32 PM on March 27, 2009
There are some very basic human factors work at play here (the issue with the warning lights and bells in the article is basically what I did at my last job). In hindsight, sure, the transponder failure was a catastrophic one, but what can you do? If every failure indication involves a large red blinking icon and klaxons in the cockpit, the pilot quickly learns to ignore all of them. Tiered warning systems are important to declutter the noise, and a transponder failure is not Catastrophic, with a capital C, in the safety-of-flight sense. Only issues that pose an imminent threat to the aircraft and its occupants really justify the kind of warnings that will make you jump out of your seat.
Modern systems and avionics have undoubtedly made the skies safer. There's a reason why you can take off from JFK even though you can see your hand in front of your face outside. However, it does seem to be leading towards pilots taking a backseat approach to flying (hah!). My takeaway from this article is that modern flight training needs to impress upon FLY THE AIRPLANE as Rule #1. You are not supposed to be learning how the radios work while you're in the air anymore. We have simulators, interactive software for your computer, and you can push every button and twist every knob before you ever step into the cockpit.
Maybe I'm biased, but I'm against all of the OMG BAN THE SMALL AIRPLANES!! rhetoric that I hear sometimes. If you want to carry more passengers to more places, we're going to have to start using smaller aircraft to ferry fewer people at a time to smaller, more remote aircraft. This is where equipment like TCAS comes in very handy. Pilots will still need to learn the very basics and stick to them, though. When you're in clear air, fly the airplane, see and avoid, and don't get distracted.
posted by backseatpilot at 12:32 PM on March 27, 2009
The author is also the guy who enraged the Fire Department of New York by writing that some of the fire fighters were busy looting stacks of jeans instead of rescuing people on 9/11. The fire fighters who apparently were accused (he didn't name them or their company but people could figure out which house he was talking about) because they were all killed. Tough issue.
posted by etaoin at 12:42 PM on March 27, 2009
posted by etaoin at 12:42 PM on March 27, 2009
"The fire fighters who apparently were accused (he didn't name them or their company but people could figure out which house he was talking about) because they were all killed."
Say... huh?
posted by bz at 1:23 PM on March 27, 2009
Say... huh?
posted by bz at 1:23 PM on March 27, 2009
See also this report from Wired about redesigning New York's airspace to maximize efficiency. Apparently, once planes take off, they have to stay six miles apart from eachother, which I'd think would make midair collisions in US territory virtually impossible.
posted by Rhaomi at 1:46 PM on March 27, 2009
posted by Rhaomi at 1:46 PM on March 27, 2009
Here is a New York Times article on the 9/11 firefighter looting claim and some later review of the evidence. While I have always found Langewiesche's writing captivating, his initial reporting on this alleged incident and his later responses to his critics cause me to seriously doubt his journalistic integrity and judgment.
posted by iknowizbirfmark at 2:30 PM on March 27, 2009
posted by iknowizbirfmark at 2:30 PM on March 27, 2009
These senseless deaths could have been easily prevented if only they had been wearing their piloting slippers.
posted by Rhomboid at 2:48 PM on March 27, 2009
posted by Rhomboid at 2:48 PM on March 27, 2009
A lot of these accidents have to do with anachronisms like combining precise digital instrumentation and GPS with relatively ancient ideas like "Pilots must be able to switch transponder to standby for ground operations".
There is almost no reason that a modern transponder with automatic ground standby ever needs to be manually turned off. If it ever does need to be turned off, it should be a drastic and blatantly obvious measure, like pulling a circuit breaker, not a button you can accidentally press on some obscure menu. Radio is another one. We've got a plane loaded chock full of advanced electronic equipment and computers, and 90% of the information you get about other traffic is being conveyed by voice on a half-duplex AM radio that can only listen to one or two frequencies at a time and that doesn't even have a way of determining when the receiver is out of range. Aviation is full of anachronisms. Some of them are outdated legal requirements that haven't caught up to technology. Others are demands by pilots who haven't caught up to technology. Much of it is just that people haven't even opened their eyes to the opportunities.
There still needs to be an overhaul of how aviation traffic is managed. A lot of work has been done to improve the air traffic system, but it's not an overhaul, just steady improvements. And it's obviously not enough. It's expensive and it's a headache for everyone involved to boot. Perhaps create a new set of flight rules for these high-speed, high-precision planes where the radar/ATC/navigation systems are more fully computerized and integrated with each other. Instrument flying can be left to those who are working with less accurate systems or (god forbid) watching their instruments and flying by hand. Computers can make decisions a hell of a lot faster than me as a pilot, and two computers talking to each other are infinitely faster at deciding what to do than me talking to ATC.
Unfortunately now is a really crappy time to be looking for solutions, as the global economy is hitting aviation especially hard. I don't know where the money or motivation to make those sorts of changes would come from when everyone is hunkering down into survival mode.
posted by cecilkorik at 3:02 PM on March 27, 2009 [1 favorite]
There is almost no reason that a modern transponder with automatic ground standby ever needs to be manually turned off. If it ever does need to be turned off, it should be a drastic and blatantly obvious measure, like pulling a circuit breaker, not a button you can accidentally press on some obscure menu. Radio is another one. We've got a plane loaded chock full of advanced electronic equipment and computers, and 90% of the information you get about other traffic is being conveyed by voice on a half-duplex AM radio that can only listen to one or two frequencies at a time and that doesn't even have a way of determining when the receiver is out of range. Aviation is full of anachronisms. Some of them are outdated legal requirements that haven't caught up to technology. Others are demands by pilots who haven't caught up to technology. Much of it is just that people haven't even opened their eyes to the opportunities.
There still needs to be an overhaul of how aviation traffic is managed. A lot of work has been done to improve the air traffic system, but it's not an overhaul, just steady improvements. And it's obviously not enough. It's expensive and it's a headache for everyone involved to boot. Perhaps create a new set of flight rules for these high-speed, high-precision planes where the radar/ATC/navigation systems are more fully computerized and integrated with each other. Instrument flying can be left to those who are working with less accurate systems or (god forbid) watching their instruments and flying by hand. Computers can make decisions a hell of a lot faster than me as a pilot, and two computers talking to each other are infinitely faster at deciding what to do than me talking to ATC.
Unfortunately now is a really crappy time to be looking for solutions, as the global economy is hitting aviation especially hard. I don't know where the money or motivation to make those sorts of changes would come from when everyone is hunkering down into survival mode.
posted by cecilkorik at 3:02 PM on March 27, 2009 [1 favorite]
Rhaomi: If you've got two planes heading directly at one another from 6 nautical miles apart, either the pilots or the controller would have about 15 seconds to somehow become aware of the situation and react appropriately. Not really very much time at all.
posted by cecilkorik at 3:11 PM on March 27, 2009
posted by cecilkorik at 3:11 PM on March 27, 2009
Well, what's the alternative? Fly wherever you want?
The post begins with the key phrase, which is the sky is a really big place. If we allow planes to be flown at random altitudes and headings, the chances of a mid-air collision are extremely small, but not quite zero. But it makes people uncomfortable when nothing but random chance is preventing a big accident.
So instead we have a system of air traffic control that constrains flight vertically to a few specific altitudes, greatly reducing one degree of freedom. The flightpath is then constrained horizontally to be along specific airways, reducing another degree of freedom. Instead of being randomly distributed throughout the whole volume of the sky, flights are all aiming for a few specific lines in the sky.
A system of positive control exists to keep them apart, and the resulting accident rate is almost as low as it would be if they were permitted to fly completely at random. And the better the navigation equipment and instrumentation becomes, the worse it gets, because flights are generally closer to those few lines in the sky than they would be if they followed them rather loosely instead.
None of this applies in the terminal environment of course, where everyone is aiming for a runway. You've got to have positive control there or many accidents would occur. But it's interesting that safety would most likely be improved by scrapping the air traffic control system in airspace that isn't particularly near an airport.
posted by FishBike at 3:33 PM on March 27, 2009
The post begins with the key phrase, which is the sky is a really big place. If we allow planes to be flown at random altitudes and headings, the chances of a mid-air collision are extremely small, but not quite zero. But it makes people uncomfortable when nothing but random chance is preventing a big accident.
So instead we have a system of air traffic control that constrains flight vertically to a few specific altitudes, greatly reducing one degree of freedom. The flightpath is then constrained horizontally to be along specific airways, reducing another degree of freedom. Instead of being randomly distributed throughout the whole volume of the sky, flights are all aiming for a few specific lines in the sky.
A system of positive control exists to keep them apart, and the resulting accident rate is almost as low as it would be if they were permitted to fly completely at random. And the better the navigation equipment and instrumentation becomes, the worse it gets, because flights are generally closer to those few lines in the sky than they would be if they followed them rather loosely instead.
None of this applies in the terminal environment of course, where everyone is aiming for a runway. You've got to have positive control there or many accidents would occur. But it's interesting that safety would most likely be improved by scrapping the air traffic control system in airspace that isn't particularly near an airport.
posted by FishBike at 3:33 PM on March 27, 2009
If we allow planes to be flown at random altitudes and headings, the chances of a mid-air collision are extremely small, but not quite zero. But it makes people uncomfortable when nothing but random chance is preventing a big accident.
So instead we have a system of air traffic control that constrains flight vertically to a few specific altitudes, greatly reducing one degree of freedom.
The problem is that commercial carriers don't want to fly at random altitudes - there is a narrow band that almost all flights will want to be in for reasons of fuel efficiency.
However, improvements are slowly being made to the way routes are determined and assigned.
posted by Urban Hermit at 4:22 PM on March 27, 2009
So instead we have a system of air traffic control that constrains flight vertically to a few specific altitudes, greatly reducing one degree of freedom.
The problem is that commercial carriers don't want to fly at random altitudes - there is a narrow band that almost all flights will want to be in for reasons of fuel efficiency.
However, improvements are slowly being made to the way routes are determined and assigned.
posted by Urban Hermit at 4:22 PM on March 27, 2009
The fire fighters who apparently were accused (he didn't name them or their company but people could figure out which house he was talking about) because they were all killed."
Say... huh?
Sorry, dropped a couple of words. I was trying to say that the firefighters who apparently the targets of the accusation couldn't respond because they had been killed when the second tower collapsed.
posted by etaoin at 4:49 PM on March 27, 2009
Say... huh?
Sorry, dropped a couple of words. I was trying to say that the firefighters who apparently the targets of the accusation couldn't respond because they had been killed when the second tower collapsed.
posted by etaoin at 4:49 PM on March 27, 2009
The single biggest problem seems to be the awesome accuracy of the modern avionics, keeping both planes within wingspan distance of their assigned locations. This assures that when two planes accidentally contrive to be in the same place at the same time, they will positively succeed. All it would have taken is 100 feet of error in either plane's navigation system and the accident would have been much less likely to happen. And the separations are kept as large as they are on the assumption that those errors are there. The ironic thing is that because of the error assumption flyways are wide, therefore there are relatively few of them, and the chance that two planes accidentally end up on the same one are increased.
Of course, we are raised to think that accuracy is a good thing, and staying on your assigned path is a good thing, and randomness is a bad thing. But in this case a little randomness would have saved 154 lives. Of course randomness isn't a guarantor of anything, except that if everyone flew wherever they wanted a certain nonzero collision rate would be guaranteed. That that rate might be very very low, possibly even lower than what we can expect from the occasional highly accurate fuckup, just creeps people right the hell out and I think you can rule out any chance of people rationally deciding between the two possibilities.
posted by localroger at 4:53 PM on March 27, 2009
Of course, we are raised to think that accuracy is a good thing, and staying on your assigned path is a good thing, and randomness is a bad thing. But in this case a little randomness would have saved 154 lives. Of course randomness isn't a guarantor of anything, except that if everyone flew wherever they wanted a certain nonzero collision rate would be guaranteed. That that rate might be very very low, possibly even lower than what we can expect from the occasional highly accurate fuckup, just creeps people right the hell out and I think you can rule out any chance of people rationally deciding between the two possibilities.
posted by localroger at 4:53 PM on March 27, 2009
The problem is that commercial carriers don't want to fly at random altitudes - there is a narrow band that almost all flights will want to be in for reasons of fuel efficiency.
I should have been clearer--I didn't mean random in the sense of rolling dice and picking a completely random altitude. I meant random in the sense of not being limited to specific intervals of 2,000 (or even 4,000) feet.
Instead let each plane fly at its exact optimum altitude, which varies enough with aircraft type, weight, and fuel prices that the chances of two nearby aircraft actually being at the same altitude would be quite small.
posted by FishBike at 5:05 PM on March 27, 2009
I should have been clearer--I didn't mean random in the sense of rolling dice and picking a completely random altitude. I meant random in the sense of not being limited to specific intervals of 2,000 (or even 4,000) feet.
Instead let each plane fly at its exact optimum altitude, which varies enough with aircraft type, weight, and fuel prices that the chances of two nearby aircraft actually being at the same altitude would be quite small.
posted by FishBike at 5:05 PM on March 27, 2009
It sounds like aircraft autopilots should take a cue from ethernet collision avoidance mechanisms and introduce a small amount of jitter or fuzziness in the initial flight paths to prevent these types of mid air collisions. There probably is a set of generalizable rules for routing any sort of traffic that get rediscovered in each system as deficiencies become apparent.
posted by euphorb at 8:30 PM on March 27, 2009 [1 favorite]
posted by euphorb at 8:30 PM on March 27, 2009 [1 favorite]
One thing I just can't over is that these guys are still flying professionally. Yes, the pilots are not solely at fault and a whole series of incidents and just plain bad luck conspired to lead to catastrophe on that day. Nevertheless, the cavalier attitude these pilots displayed in the (non-)operation of the aircraft is to my mind a major reason why events turned out the way they did. Flying an aircraft requires constant vigilance and situational awareness; if you can't deliver that you (and the public) are better off if you opt to drive a golf cart for a living.
posted by oxidizer at 8:59 AM on March 29, 2009
posted by oxidizer at 8:59 AM on March 29, 2009
« Older Somebody's watching you... | If FAS 157 started it will FAS 140 will keep the... Newer »
This thread has been archived and is closed to new comments
posted by NotMyselfRightNow at 10:52 AM on March 27, 2009