Growing up in the 1970s and 1980s, fatal crashes involving airliners in the U.S. weren’t an uncommon thing – at least, not as I remember it. It seems like the rate was one or two catastrophic crashes per year, in which most or all of the passengers or crew were lost.
Setting aside crashes due to terrorism or intentional acts by the crew, it now seems like such crashes, at least here in the U.S., are very uncommon. As far as I’m able to determine from a few minutes of Googling, there’ve been only 3 fatal crashes involving airliners in the U.S. in the past 10 years:
The Asiana crash at SFO in 2013, which was due to pilot error
A Delta Connection crash at the Lexington, KY airport in 2006 (also apparently due to pilot error)
I know that we have a number of posters here who are pilots, or otherwise involved in commercial aviation. Is what we’re seeing a sign of better safety in commercial flight? Are airliners themselves less prone to mechanical failure now? Or, are the airlines getting better at maintenance? Or, is this long “safe” period just a statistical anomaly? (Or, is it just that my recollection of the frequency of accidents a few decades ago is off?)
This is worldwide, but this would seem to indicate that the skies have gotten much safer. That doesn’t even include the increase in passenger miles flown today which is dramatically higher than at the peak for fatalities (1972).
Excellent point, as well, on crashes declining at the same time that passenger miles has increased. From this chart by the U.S. Department of Transportation, billions of passenger-miles on U.S. domestic commercial flights:
1960: 31
1965: 53
1970: 108
1975: 120
1980: 191
1985: 276
1990: 346
1995: 404
2000: 516
2005: 584
2010: 565 (decline from a height in 2008, due to the recession)
2014: 608
I was told by pilot friends that much of this is due to improvements in weather prediction & tracking – many airline crashes involve bad weather, but now the pilots are better informed about storm locations & movement, and can divert their planes to avoid the bad weather.
He said, only half joking, that the increase in airline safety is due to Seymour Cray & his supercomputers. (The labs where he did his initial supercomputer work are located within sight of the runways here at Minneapolis-St Paul airport.)
Air traffic controllers and pilots are humans. Sometimes shit happens. That’s why there’s a lot of redundancy built into the protocols. (Read-backs, phonetic alphabets, standardized vocabulary, etc.) It looks like in this case the controller noticed something was amiss pretty quickly and worked hard to fix it.
I’m really interested to find out what happened friedo.
“The Taiwan-bound jetliner appeared to clear the 5,713-foot peak of Mt. Wilson by no more than 800 feet, according to website data cited by The Times. However, broadcast towers rise an additional 400 feet from the summit, potentially reducing the clearance.”
The ATC did not direct the airliner towards the Mt. Wilson. She instructed them to “turn left heading 180”. She should have told them to turn right but the heading was fine. For some reason the crew turned to the north (speculation: maybe they turned to 018, wouldn’t be the first time that kind of error occurred.) After that she tries to sort the situation out. Lack of standard phraseology (“What are you doing? turn southbound now”) and english second language issues may have played a part in why they didn’t turn south more promptly.
To the OP: Air travel becomes increasingly safe because whenever there is an incident there is an investigation, and, if necessary, procedures are changed to try and prevent a reoccurrence. This process is happening continually and ideally improves safely. Sometimes it just means change for the sake of change, but overall the results speak for themselves.
Note that due to the way the radio scanners pick up ATC transmissions the actual instruction to the EVA aircraft is missing, however other news articles confirm that the instruction was to turn left heading 180 and that is what EVA read back.
So Richard Pearse, did the air traffic controller mistakenly route a wide-body jet with 353 people aboard toward Mt. Wilson? I listened to your link but I am a mere grounded mortal. There were sounds of confusion. Help defeat ignorance amoungst us Dopers. Please create a new thread if necessary.
Starting about 30 years ago there was the recognition that accidents were becoming fewer. So learning from accidents, while fine and all, wasn’t good enough. We had to start learning from almost-accidents too. Which led to this: Aviation Safety Reporting System - Wikipedia (ASRS pronounced as 4 letters: A-ess-Are-ess). Which at first had little uptake due to trust issues. But after that slow startup the process now is well-trusted and heavily used by all aspects of aviation: pilots, maintenance, and ATC.
Starting about 15 years ago there was a recognition that ASRS was nice, but some issues were really specific to individual airlines. Either goof-prone policies, or perhaps just mission-based exposure. e.g an airline that operated heavily into airports in mountainous terrain might have a much larger exposure to those risks than the industry as a whole. So the ASRS reports on mountain-related problems might never rise above the industry-wide noise floor even though this particular carrier was having many mountain-related close calls.
This led to the creation of similar projects at the individual airline level. Negotiated between the carrier and the pilot union and blessed by the FAA. As with ASRS, the basic deal is for the crew to report any anomaly or screw-up. In exchange for the data gained, the FAA & airline management agree to not prosecute the reporter. There’s not a good wiki on this since it’s unique to each airline and such internal procedures.
FAA has also created a similar program within the ATC world for their people. Which helps highlight local issues which, like the mountain example above, might be endemic in one facility, but not big enough to move the national ASRS needle.
Of course, just like with ASRS, wilful disregard of law or procedure is still punishable. But overall it’s the recognition that to err is both A) human, and B) not a crime. The more errors we can detect and analyze, the more we can design the opportunity for error out of the system.
Starting about 10 years ago as aircraft with sophisticated onboard data recorders became the norm, this was extended to gathering and analyzing 100% of the data from 100% of the flights. Flight operations quality assurance - Wikipedia (“FOQA” pronounced FOE-kwah). Every single flight is downloaded and processed within a few days. The whole thing goes into a data base and standard queries are run looking for marginal behaviors. We get reports monthly on how the crew force as a whole is doing against various goals to improve. And as with ASRS, being highlighted as near or outside the envelope isn’t a crime (absent intent).
FAA ATC has a similar program that analyses radar data, including ground control radar, looking for close calls.
Through statistical analysis, FOQA has found defects in approach or airspace design, bugs in ATC or aircraft software, FAA or airline procedures that look good on paper but are too unreliable in practice or are widely misunderstood and hence misused, etc.
If something looks anomalous enough the relevant union staff will call the crew or controllers asking for more details which narratives also go into the database for company and FAA analysis.
The payoff from that is we also get monthly news letters on lessons learned from these events. “There I was …” stories for the little glitches that happen every day and used to be water-cooler conversation between individuals if even that. Decades ago we’d pore over the three or four annual accident reports looking for insights. Now we get a couple dozen lessons learned per month. Each lesson learned is much smaller, but since a mishap is almost always the end result of a chain of small errors, the more we can plant the seeds of small error avoidance the more chains will be broken earlier.
Collectively, all these things are driving risk out of the system design, and also driving individual variation out of performance. If we stay right in the center of the envelope all the time we’re real safe. When operating closer to the edge has become normalized either through systemic failure, culture, or through a resources / requirements mismatch, safety is degraded. In a data-driven world, having the data to see this drift happening before somebody gets hurt is a game-changer.
All of which results in the near-zero accident rate you’re seeing in big jet US aviation. Note I’m not suggesting this is a US-only effort. I’m just not up on other countries’ details. Although most major ICAO nations are somewhere down this same road.
'm just
Thanks, LSLGuy, for the authoritative input as always.
I just wanted to add a clarification about the above incidents. All three were pilot error, and the second one (Continental Connection, which was Colgan Air 3407) was a particularly egregious one where a captain who had failed three check rides inexplicably did exactly the wrong thing in response to a stall – not once, but twice, the second time overriding the plane’s automatic stall recovery system.
I have to concur that air travel is getting dramatically safer for all the reasons already noted, including technology, training, and the aircraft themselves. I have the impression that one of the most dangerous periods in aviation history – depending on how you measure it – was the early jet age defined by the Boeing 707 and Douglas DC-8, because it combined a rapid growth in air travel with new and relatively unproven aircraft. In fact that’s true even if one goes back a little earlier, to the British Comet (which I believe was the first passenger jet in commercial service). The early Comets had a serious design defect where metal fatigue could cause sudden catastrophic in-flight hull failure.
But the 707 and DC-8 had their share of problems, for many reasons not necessarily all related to the aircraft themselves (after all, a 707 functioned for a while as Air Force One). But there were many spectacular accidents, like the famous Air France flight 007 which was returning with the members of the Atlanta Art Association after a month-long visit to Paris and a celebration of the new Jet Age. On June 3, 1962, it rolled down the runway at Orly airport and simply failed to leave the ground and became the worst single-aircraft accident of the time. In November 1963, Air Canada, one of the world’s safest airlines, had one of its DC-8s crash shortly after takeoff from Montreal, killing all aboard. No definite cause was determined (there were no flight recorders then) but it was thought to be a problem with the pitch trim system, and three months later, an Eastern Airlines DC-8 crashed due to a pitch trim problem. In 1970 another AC DC-8 crashed on approach to Toronto, although that appeared to be pilot error in accidentally deploying spoilers on approach.
These kinds of things don’t seem to happen any more, or at least are vanishingly rare. It’s interesting to look at the raw stats for accidents with the older aircraft compared to newer ones. The stats may be a bit skewed due to time in service but probably mitigated by the fact that the oldest ones are pretty much all retired. Also obviously the number of aircraft produced will affect the numbers, but among the common passenger aircraft it’s quite dramatic how much lower the accident numbers are for the newer ones.
To take the above two aircraft as a benchmark:
[ul]
[li]As of October 2015, the DC-8 had been involved in 146 incidents, including 83 hull-loss accidents[/li][li]As of September 2015, the 707 has been in a total of 246 major aviation occurrences, including 172 hull-loss occurrences[/li][/ul]
Now compare to newer planes, with the 747 so old that it represents a kind of transition point:
[ul]
[li]As of May 2016, the 747 [in service since 1970] has been involved in 132 aviation occurrences, including 60 hull-loss accidents[/li][li]As of October 2015, the Boeing 767 [in service since 1982] has been in 45 aviation occurrences, including 15 hull-loss accidents [/li][li]As of October 2016, the 777 [in service since 1995] has been involved in 18 aviation accidents and incidents, including 6 hull-losses[/li][/ul]
