Skip

Way too close for comfort
May 15, 2014 1:49 PM   Subscribe

Two Weeks Ago, I Almost Died in the Deadliest Plane Crash Ever How two jetliners nearly collided over the Pacific, why no one knows about it, and what it means for safety oversight aboard airplanes
posted by gottabefunky (67 comments total) 28 users marked this as a favorite

 
(did he manage to write that many words about possible collisions without mentioning GOL 1907 or did I miss something?)
posted by effbot at 2:11 PM on May 15 [2 favorites]


GOL 1907 is particularly relevant because one of the aircraft in that collision had its transponder inadvertently turned off. Functioning transponders are essential to TCAS, the collision avoidance system that warned the crew of this author's flight in time to prevent a disaster.
posted by zachlipton at 2:17 PM on May 15 [2 favorites]


Well, I don't think he set out to write a history of air collisions...

I started out thinking that the author was writing this article just because he was personally freaked out over what was, after all, an example of safety systems working properly. But I found his arguments ultimately pretty persuasive.
posted by showbiz_liz at 2:17 PM on May 15 [4 favorites]


I think this article overstates the extent of non-transparency in aviation. There is little chance this incident wouldn't have been reported, even if it hadn't had turned up in summary reports in that week or two before the article.

Aviation is self-policed in the way that medicine is self-policed, in so much as doctors oversee doctors and the system, largely to its benefit I think. And I think its one of the strongest "systems" at least as its developed in North American and Europe, between the manufacturers and controllers and pilots. Aviation has developed a strong culture base, one of the strongest in many ways, as noted by the New Yorker's medical writer in his book "The Checklist Manifesto"

That isn't to say that there's lots of room for modernisation. The gaps tracking over long ocean passages, the dated black box tech, the cultural problems in cockpits full of certain mindsets and overused to automation, and the looming issues of uncontrolled drone traffic and congestion are clearly areas to address. To date, I think its a system that does react in the right ways overall.

But he had a good fright, and not an everyday one.
posted by C.A.S. at 2:18 PM on May 15 [9 favorites]




I thought that was a great article. Complete with a grossly misleading title, exaggerated stock photography, and failing to address even its own question, namely, "why nobody knows about it," with anything remotely approaching an answer. But at least he managed to compare a potential mid-air collision with Tenerife, which happened on a runway.
posted by phaedon at 2:24 PM on May 15 [26 favorites]


There have been two midair collisions over Grand Canyon. The first collision, in 1956, killed 128 people aboard two airliners and resulted in the creation of the FAA. It crashed in a remote, difficult-to-reach part of the canyon, and some crash debris remains there today. The site was recently designated a national historic landmark.

The second collision was between two sightseeing aircraft in 1986 and resulted in changes to flight rules over the canyon itself.

I recall reading about a major effort in the 1990s to improve air safety by identifying and learning from near-disasters. (I think maybe Al Gore was somehow involved? Or maybe I'm just getting old and conflating everything that happened during the Clinton years.) I can't seem to find anything about it online, though.
posted by compartment at 2:25 PM on May 15 [3 favorites]


From Joe Sharkey's account: We would have a reunion each year and report on how we used our time.

Just chilling.
posted by roomthreeseventeen at 2:29 PM on May 15


Did he manage to write that many words about possible collisions without mentioning Wayfarer 515 or did I miss something?
posted by univac at 2:35 PM on May 15


He's vastly overstated the broken state of the system of incident reporting, and overstated the issue of lack of "oversight" of the safety system, based on his incredibly thrilling and personally terrifying sample size of one.

That being said, those US Airways pilots are going to have a tough time explaining why they were on a eastbound altitude heading west, that is as obvious and bad as trying to get on a highway on the wrong side of the road. And their coworkers, managers will be very aware.
posted by C.A.S. at 2:47 PM on May 15 [5 favorites]


Could be an mis-set altimeter. I was surprised to learn that GPS is not allowed to be used for height determination (really?)
posted by RobotVoodooPower at 2:53 PM on May 15 [1 favorite]


The article didn't really seem to have any basis for the fear-mongering in it. The systems worked, nobody responsible for the working systems was particularly surprised or panicked about the fact that their working systems worked. I'm not sure what more the writer really expected.
posted by jacquilynne at 2:53 PM on May 15 [3 favorites]


The article didn't really seem to have any basis for the fear-mongering in it. The systems worked, nobody responsible for the working systems was particularly surprised or panicked about the fact that their working systems worked. I'm not sure what more the writer really expected.

The very last system worked. Near misses have to be taken seriously and studied carefully for safety lessons because you won't always be so lucky.
posted by zachlipton at 2:55 PM on May 15 [6 favorites]


/thinks about reading this...

/looks at already booked plane ticket

/runs screaming from the room.
posted by the_royal_we at 2:56 PM on May 15 [12 favorites]


C.A.S., he's allowed to report about an incident he was personally involved in, and we're allowed to use inductive reasoning to guess, if it happened to this one person who luckily happens to be a writer in a position to tell us about it, that there are probably other near-misses that aren't being reported.

Anyway, if he's overstating the brokenness of the system, this is your chance to contradict him with more precise details. When the author writes:

The FAA even admitted that my initial information, the random phone call from a passenger, was “essential to [their] fact-finding.” Without the basic information I provided to them, they would not, by their own admission, have been able to connect the dots when the ATO began asking questions.

...that seems to me to indicate a somewhat dysfunctional system.
posted by JHarris at 2:56 PM on May 15 [8 favorites]


I was surprised to learn that GPS is not allowed to be used for height determination (really?)

Apparently barometric altimeters are not that accurate due to air temperature differences, but they are precise. So it's ok as long as everyone uses one to measure their altitude. If some planes use GPS, and other use a barometer, than there's a higher chance of collision.
posted by smackfu at 3:02 PM on May 15 [6 favorites]


I was surprised to learn that GPS is not allowed to be used for height determination (really?)

Almost any calibrated altimeter will be more stable at reading altitude than a GPS.

The FAA even admitted that my initial information, the random phone call from a passenger, was “essential to [their] fact-finding.” Without the basic information I provided to them, they would not, by their own admission, have been able to connect the dots when the ATO began asking questions.

...that seems to me to indicate a somewhat dysfunctional system.


Even if there are regulations requiring the airline to report these things, the FAA may consider customer reporting valuable too, because it creates an environment where airlines can't expect to get away with violating those regulations.
posted by aubilenon at 3:05 PM on May 15 [1 favorite]


"Modern air travel is such a raw miracle of technology it would almost certainly be the first of our achievements to awe generations before us."

Maybe the first time I've heard that- my first thought is always the web- but he's got a good point.
posted by stinkfoot at 3:06 PM on May 15


I dunno, Radio is pretty magical.
posted by aubilenon at 3:11 PM on May 15 [7 favorites]


I skimmed through the article but I didn't see any discussion about why the planes got so close in the first place. Isn't there some sort of scheduling or coordination of plane routes and altitudes to ensure they don't hit each other? Or do they basically just wing it and rely on the collision avoidance devices?
posted by pravit at 3:23 PM on May 15 [1 favorite]


I agree that the systemic breakdown is concerning, especially considering the small envelope allowed before a warning turns into national news and the lack of reporting of such an incident. I could see a couple pilots shrug it off because it doesn't seem like a big deal to them/they don't have time/they're afraid of being reprimanded, but that close call might contain data valuable to improving the system.

However, I'm also wary of an overly sensitive system. The safety record of US carriers is pretty absurd if you calculate the number of fatal accidents. If that safety bubble expands a mile in each direction, might the pilots ignore a potentially critical warning? Alarm fatigue is a big deal.
posted by Turkey Glue at 3:25 PM on May 15 [1 favorite]


Yeah, that's an impressively tall stack of assumptions and scare quotes. I wish he could have found out some facts before writing the article.
posted by ftm at 3:47 PM on May 15 [2 favorites]


This dude seems to be unable to read the sarcasm as the FAA official rolled his eyes and assured him that his pant-wetting panicky phone harassment was absolutely vital to reporting on this terrifying incident where two planes flew vaguely near each other once.

I hope he never drives down a freeway; you're continually within much less than 20 seconds of a fatal collision.
posted by Homeboy Trouble at 3:59 PM on May 15 [4 favorites]


my first thought is always the web

but that's just a CB-radio with an Etch-a-sketch stuck to it
posted by hap_hazard at 4:05 PM on May 15 [12 favorites]


I hope he never drives down a freeway; you're continually within much less than 20 seconds of a fatal collision.

Big difference in the amount of control a driver has and the amount a plane passenger has.
posted by roomthreeseventeen at 4:07 PM on May 15


This dude seems to be unable to read the sarcasm as the FAA official rolled his eyes and assured him that his pant-wetting panicky phone harassment was absolutely vital to reporting on this terrifying incident where two planes flew vaguely near each other once.

The fasten seatbelt sign wasn't even on. They're very, very lucky that there were no injuries. All it would have taken would be a parent with a lap baby on board, or someone walking to the bathroom.
posted by the young rope-rider at 4:22 PM on May 15 [1 favorite]


Big difference in the amount of control a driver has and the amount a plane passenger has.

Or that one is under the control of an experienced professional, and they let any goddamned crazy person drive a car.
posted by empath at 4:25 PM on May 15 [12 favorites]


The fasten seatbelt sign wasn't even on. They're very, very lucky that there were no injuries. All it would have taken would be a parent with a lap baby on board, or someone walking to the bathroom.

Assuming the maneuver was as violent as the author says. Given the sketchiness of the rest of the article, I'd like to hear the airline's side of things.
posted by cardboard at 4:33 PM on May 15


Fair point, although even "typical" turbulence can cause injuries.

One of the reasons I did not pursue freelance writing/journalism is because I didn't want to live my life thinking about how every single important moment could be spun and sold to someone else.
posted by the young rope-rider at 4:35 PM on May 15 [1 favorite]


I hope he never drives down a freeway; you're continually within much less than 20 seconds of a fatal collision.

That's an erroneous comparison; the highway system is designed such that cars routinely travel in opposite directions 20 feet apart. The air traffic control system is designed for airplanes to fly in vertically-situated lanes, and in this case a plane wasn't in its lane. The article compares it to meeting a car head-on and having to swerve violently.

Again, I'm not sure it's worth panicking over one anecdote with cherry-picked data, but it's still a valid response to a scary situation nobody wants to experience.
posted by Turkey Glue at 4:42 PM on May 15


JHarris

Of course he's allowed to report on his incident, and the aftermath.

I'd say its not up to me to offer precise arguments for why he's wrong, but up to him as a journalist to offer more detail as to why he's right that the system (of oversight and incident reporting) is 'broken". He is not even convincing that his personal pants wetting day wasn't reported, only that within the two weeks since his incident he hasn't proven that his scare wasn't reported.

United confirmed it, US Airways "no commented" (essentially a confirmation), and he had no proof that the FAA or ATO had NOT created an incident investigation, indeed it appears they have.

But, if I'm counter arguing, I would say that this is a serious incident, it will be seen as a serious incident within both airlines, the ATO, and the FAA, and there will be an incident report. I used to read weekly incident reports for general aviation pilots most of my life growing up. Those don't get reported in the newspaper unless they make a good headline, but they are taken seriously. An investigation is just that. The airline managements, the FAA, the ATO, the flight crews, and the pilots union all get involved.

They just don't get released within 14 days of the actual incident, prior to a full report, to someone calling up for pull quotes.
posted by C.A.S. at 4:46 PM on May 15 [3 favorites]


Aviation is self-policed in the way that medicine is self-policed, in so much as doctors oversee doctors and the system, largely to its benefit I think.

Oh my god, I am never flying again.

-M.D.
posted by hobo gitano de queretaro at 6:31 PM on May 15 [5 favorites]


These kinds of close calls have actually become much more common with the advent of GPS (which has only been used in commercial planes for a few years). With inertial navigation, the pilot would try to keep to hte assigned airway, but over the ocean you'd be constantly off course by hundreds of meters in some random direction because of navigation errors. With GPS, you're flying on your assigned course within a few meters ... and so is everyone else. Hence collisions are much more likely. There are ways to avoid this but when they fail, GPS makes a collision far more likely.
posted by miyabo at 7:24 PM on May 15 [2 favorites]


I have to fly cross country tomorrow. Thanks. Now I'm off to watch Flight.
posted by slogger at 7:55 PM on May 15


Patrick Smith often talks about the PEF that air crews are familiar with (Passenger Exaggeration Factor) wherein passengers will recall barrel rolls, other planes being so near that the other passengers were all clearly visible and so forth. I flew once with a coworker who travels much less than I and as we landed in the rain she swore the noticeable thump as we touched down could only mean the crew had lost control of the plane. I offered up that it may have been because surface tension, it is a thing, and a harder landing gives you a better chance of avoiding having your 155,000-pound 737 hydroplaning.
posted by ricochet biscuit at 8:12 PM on May 15 [5 favorites]


Did anyone catch the video before it was pulled off YT ?
posted by GhostRider at 8:17 PM on May 15


FWIW I was in a somewhat similar situation on a flight I took a few years ago. The aircraft was climbing out of the airport when the pilot suddenly throttled back and leveled out. About 2 seconds later I saw another airliner flash by out of my window. It was easily close enough to see the whites of the other passengers eyes and all that.

Well, that might be a bit exaggerated--it's really hard to judge distance up there with no frame of reference. But comparing it other planes on the ground after we landed my best guess was 1/4 or 1/3 mile away and maybe 100 feet below our altitude. So, closer that you're supposed to be, I'm pretty sure.

When I mentioned this to one of my airline/ex-military pilot friends, he basically waved it off. This one was moving (quickly) across our visual field. "The one you're really worried about," he said, "is the little dot that isn't moving at all."
posted by flug at 8:52 PM on May 15 [3 favorites]


About ten years ago I was taking off from London on a flight to New York. We were gaining altitude when I felt a sharp deceleration. I was pushed forward in my seat. I seem to remember the engine noise changing, but that might be my imagination because it all happened very quickly.

I just happened to be looking to my right, across the aisle and out the starboard windows of the cabin. Just then, another plane flashed by astonishingly quickly. Its course was about perpendicular to our own, but because we were flying perpendicular to it it seemed to zip away behind us. It was a twin engine aircraft with wings set high on the fuselage. I got the impression of blue and grey markings, and was reminded of those planes made by Antonov.

This is all vague, because I only saw it for less than a second. But it was very, very close. I could see rivets on the plane. If I had time, I could have easily have read the markings on the fuselage.

When I got off the plane I considered asking the captain what the deal was with that other plane on take-off, but the pilots weren't there to see us off as they sometimes are and... well I didn't want to make a fuss and I couldn't be sure it wasn't purely my imagination. Nobody else seemed to have noticed anything.

A while later somebody told me about some official website that listed aircraft incidents, including near misses (defined as coming, I think, within a mile of another plane). There was nothing on the list that could possibly be the plane I was flying on.

It's still a mystery to me. Could there have been a near miss that just wasn't reported?

On preview: flug? You too!?
posted by Dreadnought at 9:03 PM on May 15


I'd complete forgotten that I posted this (previously) in 2009 about the GOL 1907 mid-air collision over Brazil. It is very relevant to this topic and still interesting.

The point from that article is particularly relevant to the near-collision that is the subject of this post is the fact that the sky really is a big place. The near-misses that Dreadnought and I just outlined happened near airports where there is a natural concentration of aircraft and where ATC is (supposedly!) organizing things to avoid those sorts of near collisions.

But over a giant empty space like the Pacific Ocean there really is a nearly infinitely small chance that two aircraft will happen to be in the same place at the same time--to the point that it would actually be safer to just let aircraft proceed at whatever random, altitude, speed, and direction they like, as opposed to assigning specific altitudes and flightpaths, which tends to concentrate all traffic into just a few relatively congested lanes.

FWIW I've spent some time in flight simulators trying to deliberately crash one a/c into another. It is h-a-r-d. Just randomly crashing into another aircraft in an airspace as large as the Pacific Ocean is a really, really, REALLY unlikely event.
posted by flug at 9:46 PM on May 15 [4 favorites]


We were gaining altitude when I felt a sharp deceleration. I was pushed forward in my seat. I seem to remember the engine noise changing

This is what is frightening to me about these near misses. The pilots are obviously paying attention, seeing a potential situation, and had about 2 seconds of reaction time before the potential for a collision was over and so they stopped reacting.

So they aren't getting anything near the 10-15 seconds of reaction time outlined in posts above. They aren't getting enough advance warning to be able to take any action at all that might avoid a collision.

This can't be the way ATC in the area immediately around airports is supposed to work . . .
posted by flug at 10:09 PM on May 15


I hope he never drives down a freeway; you're continually within much less than 20 seconds of a fatal collision.

C'mon now, you can stop any car from almost any speed in 20 seconds. If you stop a jetliner from cruising speed in 20 seconds, everyone is dead.
posted by roquetuen at 12:04 AM on May 16


C'mon now, you can stop any car from almost any speed in 20 seconds. If you stop a jetliner from cruising speed in 20 seconds, everyone is dead.

Well yeah, but at 500 mph, you can move a pretty nice distance in 20 seconds. You only need to move, what, 200 feet to avoid a plane? You can move 4km in 20 seconds at 500mph.
posted by empath at 12:27 AM on May 16


Side note, after doing some googling, I am particularly concerned about the fact that United loaded 289 passengers onto an aircraft with only 213 seats. That doesn't sound very safe at all, collision or no collision.
posted by ftm at 5:30 AM on May 16 [1 favorite]


From one of the captions: A composite image of aircraft taking off shows the density of airport traffic

No, it doesn't.
posted by smackfu at 5:46 AM on May 16 [4 favorites]


> I was surprised to learn that GPS is not allowed to be used for height determination (really?)

In addition to the reasons mentioned above, the DoD can skew GPS without warning. GPS operates 2 signals from the satellites: an unencrypted civilian channel, and an encrypted US military channel. Up until the Clinton Administration, the "fix" one got on the civilian channel was low-resolution (accurate to around 300 feet), but they hit a switch somewhere and gave the civilians the same accuracy that the military uses. However, they can not reduce the resolution, but they can command the satellites to add as much or as little error as they want, and allegedly they can do this over targeted regions of the planet without affecting other planets. Great if you're in the Pacific, but not so great if you're over the Black Sea these days. In short, the GPS doesn't fail safe when you're an airliner.
posted by Sunburnt at 5:51 AM on May 16 [1 favorite]


> You can move 4km in 20 seconds at 500mph.

The plane was at 500 kts, about 575 mph, meaning 20 seconds is good for over 5km or around 3.2 mi.

If course, the plane can only go that far in a small range of evasive directions.
posted by Sunburnt at 6:07 AM on May 16


It is hard enough putting along at 90 knots in a small plane to see another airplane. Seriously - the sky is very, very large and the airplanes are very, very small and moving quickly. Even when controllers tell you approximately where traffic is in the sky, it's not uncommon to not see them, ever.

Above 18,000 feet, airliners set the altimer to 29.95 inches of mercury - that's the standard they use so that everyone is on the same page relative to atmospheric conditions. GPS altitude is not actually as useful or accurate in terms of the performance data that airplanes work with.

As an airplane descends, controlled airports will give them altimeter readings to use so that their altimeters again are reading correctly relative to sea level.

I know how this is going to sound, but 600 feet in a few seconds might be really alarming, but it's not exactly an uncontrolled freefall into hell, regardless what it feels like. The author's prose just comes off a bit overdramatic to me.
posted by Thistledown at 6:22 AM on May 16 [1 favorite]


Wow, giant image of tenerife in the middle of the article, that's trashy.
posted by kiltedtaco at 6:32 AM on May 16


This is exactly why we have TCAS - and TCAS works, very well. It's saved thousands of lives.

Clearly, either ATC or one of the flight crews messed up (or there was a navigational equipment failure, but you can guess which of these three options is least likely). This happens, and while there are humans in the system it will happen again. Hence TCAS, and why pilots are trained to respond to TCAS alerts immediately: move stick first, ask questions later.

Yes, there could have been a terrible accident. But the systems designed to prevent it did their job.

I feel exactly the same way about flying now as I did before that story - it's fabulously safe, and my next flight will be fabulously safe also. Next year, it will be safer still, if the lessons from this story are properly taken on board by the FAA and the airlines.
posted by Devonian at 6:39 AM on May 16 [2 favorites]


cardboard: "Assuming the maneuver was as violent as the author says."

If the entire cabin isn't wet with drinks and no flight attendants or unbelted passengers have broken bones, it's not that abrupt a maneuver.

Seriously, though, TCAS did what it was supposed to do. The vast majority of the time, TCAS isn't needed, but it is there for those times when the normal systems fail to maintain separation. I'm glad there is at least some investigation going on, though. Better not to have to rely on the last resort system if at all possible.

In reality, there's a lot more to be concerned about at takeoff and landing than anything that happens at cruise. Mid air collisions get so much press and make people hyperventilate precisely because of their rarity. Severe turbulence causes worse incidents than the one reported here almost weekly.
posted by wierdo at 8:57 AM on May 16


Side note, after doing some googling, I am particularly concerned about the fact that United loaded 289 passengers onto an aircraft with only 213 seats
I wondered about that too. I assume it was to justify his "ZOMG, almost the deadliest crash in history" claim. US Air flies 757-200s to Hawaii, so they hold 190 passengers max, so the max passenger load should have been 403 plus crew and lap children, making his "estimated 590 deaths, one of them mine" pretty weak. It does make you wonder about the accuracy of the rest of his report.

However, he does have a point that ACAS RA alerts probably should be manditory report events. It is a last ditch system and if two planes get close enough that it is invoked, something has gone very wrong.
posted by Lame_username at 9:11 AM on May 16 [1 favorite]


If it was really as violent and sudden as the author said, and the passengers were truly in freefall, wouldn't there have been injuries.

The fact that he didn't interview any of the other passengers or crew, makes his description of the events is more than a little suspect to me. I'm not saying the author is lying, but I think it's entirely possible that his memories are clouded by fear and emotions.
posted by inertia at 9:18 AM on May 16 [1 favorite]


I think it's entirely possible that his memories are clouded by fear and emotions.

Well, sure. But look: he provides an independent graph of altitude ("FlightAware data on United 1205 between Kona and Los Angeles") that clearly shows a blip at the right time, corresponding to a sharp dive and then return to level flight.

What I'd like to know is, if he had two candidate US Airways flights for the close call, why didn't he pull up the same FlightAware graphs for both of them and check for a corresponding blip in the other direction. It makes me a bit suspicious, because that would have been easy...

(Here's the tracklog for his flight - note the high (I assume averaged over a minute) descent rate (-300, -120 feet per second?) at 7:15-7:16PM EDT.)
posted by RedOrGreen at 11:01 AM on May 16 [1 favorite]


When the plane was landing though, the descent rate was much greater, up to -2,220.

I don't have any sort of knowledge about this stuff, but I just am very skeptical of the author's telling of what happened here, partially about the descent being a weightless inducing nosedive.

This kind of article about the incident which says;

The FAA says the required separation between aircraft in the airspace where the incident occurred is either 5 miles laterally or 1,000 feet vertically.

Preliminary data shows that the aircraft were 8 miles apart when the alert came. The United plane was told to descend. On the next radar hit 12 seconds later, the aircraft were 5.3 miles apart laterally and 800 feet apart vertically.


Well that just doesn't get as many pageviews as I NEARLY DIED IN A MID AIR COLLISION....
posted by inertia at 11:40 AM on May 16 [4 favorites]


I don't know what the limits are, but at certain altitudes, they operate under something called RVSM - reduced vertical separation minimums - but it requires specific equipment (assuming specific autopilot functions) to work.

I saw this posted by an airline pilot about the article...excerpted:

...To be very frank, the system didn't fail here, the system worked. First, somebody made a mistake, second, nobody caught that mistake, third a safety system stepped in to warn the pilots that a mistake was made, and fourth the undesired aircraft state of having a mid-air was avoided.

...the pilot likely followed the TCAS guidance with some amount of vigor, though it's unnecessary. I think you have something like 7 seconds to respond to an RA, so while you should certainly rotate the aircraft smartly into the guidance, there's generally no need to shove the nose over or yank it back.

As for this writer saying that the FAA isn't involved and that airlines simply self police these issues, this is completely false, and the writer misunderstands ASAP programs (or does not know that they exist). The FAA is a party to ASAP reports, and this would have likely been a mandatory ASAP.

So put simply, to the outside world, this may LOOK LIKE DISASTER. To those that deal with it every day, it's an uncomfortable screw up that will likely result in some retraining for whoever screwed up, once they figure out who that person is. Either a controller gave a bad altitude, or a pilot went to a wrong altitude. Both happen every day. At my last shop, I believe the majority of ASAPs were filed due to vertical deviations, where the pilot was at the wrong altitude.
"
posted by Thistledown at 11:45 AM on May 16 [4 favorites]


My points precisely, Thistledown.

The author wants to extrapolate from this incident a set of conclusions totally unsupported, argued without producing significant evidence, and demonstrating a real lack of awareness of how the system operates.

Always good to get in touch with your potential mortality, however
posted by C.A.S. at 12:39 PM on May 16 [3 favorites]


they can command the satellites to add as much or as little error as they want, and allegedly they can do this over targeted regions of the planet without affecting other planets (emphasis mine)

I should hope not.
posted by one more dead town's last parade at 1:30 PM on May 16 [1 favorite]


CNN is running with it now, with ATC to blame in preliminary reports from the NTSB. (As reliable as a CNN preliminary report can be.)
posted by cardboard at 4:04 PM on May 16


So obviously people with way more knowledge than me about this have already considered it, so why don't we have something more advanced than TCAS? It seems incredible we can't share trajectories in real time and offer longer range predictions about possible collisions. Is this a bandwidth problem? Computation? Political? Cost?

Or is TCAS in fact just fine?
posted by odinsdream at 4:20 PM on May 16


TCAS is just fine. The issue is that TCAS is the last line of defense for the aircraft, so if it activates it indicates that all the other protective measures failed.

Air Traffic Control (ATC) assigned the same altitude to two aircraft heading in opposite directions, which shouldn't have been done, and then ATC didn't notice the two aircraft were on a collision course, which should have set off alarms if the aircraft were operating in an area with radar coverage.

As with accidents, the investigation will take a bit of time, but there should be recorded data on everything done by ATC and the two airplanes.

Onboard display of other aircraft is being developed in a system called ADS-B In, but that is several years away from widespread use.
posted by cardboard at 5:32 PM on May 16


That is kind of shocking, as altitudes are standardised for heading, i.e. westbound ones might be odd numbered, eastbound even, so that in theory they are a thousand feet apart at minimum.

I wonder why the US Airways crew didn't pick up on the error in assignment even if it came from ATC
posted by C.A.S. at 12:54 AM on May 17


so if it activates it indicates that all the other protective measures failed

And by this do you mean that ATC properly assigned altitude and headings? Are there other components or is it ATC then directly to TCAS? If so, that seems to be a big gap that, to an outsider, seems solvable.
posted by odinsdream at 7:17 AM on May 17


"Those US Airways pilots are going to have a tough time explaining why they were on a eastbound altitude heading west, that is as obvious and bad as trying to get on a highway on the wrong side of the road. "

Often enough, the controller will send you at whatever altitude is open, east or west. Above 18,000 feet (and below 60,000), it's controlled airspace, and the controller is responsible for making sure they only send one plane at a time into the same space separated by the "separation minimums" either horizontally OR vertically. In that airspace ("Class A"), it's 5 miles horizontally and 1,000 feet vertically. (Skipping RVSM details.)

As a pilot, all you are required to do is climb or descend, on course, to the assigned altitude at 500 feet per minute or greater. Once in a while, when they figure out that your rate of climb/descent isn't sufficient to maintain separation, they'll ask you to speed up the climb or descent to avoid a conflict, or turn you left or right.

"Could be an mis-set altimeter. I was surprised to learn that GPS is not allowed to be used for height determination (really?)"

As stated above, GPS altitude is surprisingly different from the barometric altimeter reading. My plane has a super duper certified GPS that will let me fly to within 200 feet of the ground while still in clouds on some approaches, yet when cruising at, say, 20,000 feet, my GPS and barometric altitude readings can differ by 1,000 feet or more. All that matters for separation is that all planes in an area use the same altimeter setting. All planes have an altimeter, but not all have GPS (or even an electrical system). Next time I get a chance to fly past Mount Shasta or Mount Rainier, I'm going to stay level with the top of the mountain and see how accurate the GPS is.

The article is clearly exaggerated, but more importantly, he never answers the big question he raises. What does he want us to do differently? I can't think of much that would actually statistically improve safety. Super duper TCAS? Maybe change the "East fly at an odd thousand feet, West at even" into four directions? It would cut into the available routes at a time when many routes are already crowded.

We should probably start converting two lane highways into separated highways first, because when you pass a car going the other way, the cars' closing speed can be >120 mph, and you are a fraction of a second away from a sharp left (right if you're British) swerve and death. Or just, y'know, wear seatbelts all the time.
posted by Hello Dad, I'm in Jail at 2:06 AM on May 18


Hello Dad, I'm in Jail: yet when cruising at, say, 20,000 feet, my GPS and barometric altitude readings can differ by 1,000 feet or more.

The key word there is "cruising" - as in "cruise missile." GPS equipment that allows highly accurate supersonic or near-sonic flight at high altitudes would also allow relatively cheap-and-easy construction of ICBMs. Which would be a Bad Thing. It's my understanding the errors for GPS at high altitudes and/or speeds are intentional.

Next time I get a chance to fly past Mount Shasta or Mount Rainier, I'm going to stay level with the top of the mountain and see how accurate the GPS is.

On a high-visibility day, I presume.
posted by IAmBroom at 1:16 PM on May 19


IAmBroom: "It's my understanding the errors for GPS at high altitudes and/or speeds are intentional."

No, a consumer GPS chip will stop supplying position information if the fix has an elevation greater than some number higher than that at which any commercial airliner flies (60kft, I believe, but could be as much as 100kft) and/or faster than 1500kph or so. I say and/or because different companies interpret the requirement differently and some will shut down if either parameter is exceeded while others only when both are exceeded.

The reason GPS and barometric altimeters report different altitudes is that they are measuring different things. (also, GPS altitude fixes are of lesser quality than lat/long due to the geometry of the solution, but that's not the point) GPS altitude measures height above the WGS84 geoid, which is only an approximation of the actual shape of Earth's surface. A barometric altimeter measures barometric pressure relative to "normal" sea level pressure at the standard temperature and pressure, which is rarely the actual ambient temperature or pressure.

GPS is optimized for accuracy. Barometric altimeters, like LORAN, are optimized for repeatability. GPS' repeatability is shit. LORAN will only tell you where you are to within a mile or two, but if you take a fix at a given point and then go back there later, you will get essentially the same reading. With GPS, if you do the same thing, you will know where you are to within a much lower margin of error, but you may not get the same reading at the same point at two different times.

Of course, it's a damned miracle GPS works at all, so don't think I'm complaining. ;)
posted by wierdo at 5:18 PM on May 19 [1 favorite]


Here's a sea captain's take (related by James Fallows in The Atlantic) on this "there we were, about to die" nonsense.
posted by phliar at 3:45 PM on May 20


But over a giant empty space like the Pacific Ocean there really is a nearly infinitely small chance that two aircraft will happen to be in the same place at the same time--to the point that it would actually be safer to just let aircraft proceed at whatever random, altitude, speed, and direction they like, as opposed to assigning specific altitudes and flightpaths, which tends to concentrate all traffic into just a few relatively congested lanes.

There is a lot of space, but the rest of your reasoning seems to assume that aircraft would be randomly distributed over the whole of it. There are a limited number of straight lines you could draw between, say, Narita and LAX (one, in fact). Presumably more planes would be in that corridor than 1700 miles south of Pitcairn.
posted by ricochet biscuit at 1:09 PM on May 21


« Older Burger perverts welcome   |   Two recent studies support... Newer »


This thread has been archived and is closed to new comments



Post