James Fallows

James Fallows is a national correspondent for The Atlantic and has written for the magazine since the late 1970s. He has reported extensively from outside the United States and once worked as President Carter's chief speechwriter. His latest book is China Airborne. More

James Fallows is based in Washington as a national correspondent for The Atlantic. He has worked for the magazine for nearly 30 years and in that time has also lived in Seattle, Berkeley, Austin, Tokyo, Kuala Lumpur, Shanghai, and Beijing. He was raised in Redlands, California, received his undergraduate degree in American history and literature from Harvard, and received a graduate degree in economics from Oxford as a Rhodes scholar. In addition to working for The Atlantic, he has spent two years as chief White House speechwriter for Jimmy Carter, two years as the editor of US News & World Report, and six months as a program designer at Microsoft. He is an instrument-rated private pilot. He is also now the chair in U.S. media at the U.S. Studies Centre at the University of Sydney, in Australia.

Fallows has been a finalist for the National Magazine Award five times and has won once; he has also won the American Book Award for nonfiction and a N.Y. Emmy award for the documentary series Doing Business in China. He was the founding chairman of the New America Foundation. His recent books Blind Into Baghdad (2006) and Postcards From Tomorrow Square (2009) are based on his writings for The Atlantic. His latest book is China Airborne. He is married to Deborah Fallows, author of the recent book Dreaming in Chinese. They have two married sons.

Fallows welcomes and frequently quotes from reader mail sent via the "Email" button below. Unless you specify otherwise, we consider any incoming mail available for possible quotation -- but not with the sender's real name unless you explicitly state that it may be used. If you are wondering why Fallows does not use a "Comments" field below his posts, please see previous explanations here and here.

James Fallows: Disasters

  • Today's 777s-in-Peril Update: Asiana 214, MH 370

    Lawyers offer their contention of why a pilot is not to blame for one accident. A former NTSB investigator directs blame at the pilot in another.

    How a plane looks just before touchdown into a strong crosswind. This is hard. The Asiana landing at SFO should have been easy. ( YouTube )

    1) Asiana 214. If you'd like to see professional pilots landing big airplanes, under difficult circumstances, with hundreds of lives at stake, watch any 20-second portion of this video below, taken at Dusseldorf airport during a very strong crosswind. Even the first 15 seconds will give you the idea. Or the elegant maneuver by an Air Berlin crew from 1:40 to 1:55.

    What you'll see in all cases are pilots executing the familiar "crab-into-kick" procedure for crosswind landings. First the plane "crabs" -- it approaches at an angle, to keep its direction of flight aligned with the runway. Then, just as the plane is about to touch down, the pilot "kicks" the rudder to align the airplane itself with the runway. That allows the plane to land without putting impossible cross-stress loads on the landing gear.

    Some of the landings in this clip are more precise than others, and some of that variation is beyond the pilots' control, depending on last-minute gusts and shifts in wind. But all of them show the proficiency expected of professional flight crews. 

    Watch a little of that, including artful landings of Boeing 777s, and then consider the claim from Asiana airlines' lawyers, as reported yesterday in the NYT, that autopilot software was somehow to blame for the crash of an Asiana 777 at SFO last year.

    Remember that this crash -- which killed three people, injured many more, destroyed the airplane, and shut down the airport for a time -- happened on a clear day, with light winds, in what would be considered the very most benign flying conditions. Remember that according to cockpit recordings, other members of the flight crew were warning the captain that he was mis-flying the approach and letting the plane get too "low and slow." And consider that in the two decades of the 777's operation, with many hundreds of thousands of landings worldwide by the more than 1000-strong airplane fleet, there appear to have been zero reported incidents of autopilots causing the plane to land short of the runway. You can read a 777 accident/incident list here. The one other episode involving auto-throttles and landing problems, in 2008, was traced to ice that obstructed the fuel system and kept the engines from responding properly. This was not a factor for the Asiana at SFO.

    Of course we shouldn't prejudge the legal process. And if you were a lawyer for Asiana, you'd probably try to push this "the autopilot made me do it" argument too. But, c'mon.

    [For the record, I am an admirer of NYT reporter Matthew Wald, but -- as he knows -- I disagree with the implication of the lead of his story. It was this, with emphasis added:

    While the world has been fixated on the disappearance of Malaysia Airlines Flight 370, the Korean carrier involved in the crash of a different Boeing 777, the Asiana flight into San Francisco last July, raised design issues on Monday that put another question mark over the model of jetliner.

    The Asiana and Malaysia episodes have absolutely nothing in common, and from my perspective nothing in the Asiana lawyers' filing "puts another question mark" over one of the world's most widely used and best-safety-record airplanes. To me, this connection is like saying that a car-jacking put "another question mark" over a certain model of car, if that car had also experienced carburetor problems.]

    2) Malaysia 370. No theory of the plane's disappearance makes sense. But I've mentioned several times that I thought the "Chris Goodfellow scenario" required few logical leaps than most. Goodfellow, a Canadian who now lives in Florida, has hypothesized the following sequence: a sudden inflight emergency, followed by a turn back toward airports in Malaysia, followed by a still-unexplained incapacitation of the crew, and a still-unexplained flight out over the ocean.

    If you would like to see an argued-out (rather than merely speculative) version of a contrary hypothesis, check out this on Leeham News and Comment. The item is based on an interview with Greg Feith, a former NTSB investigator, who argues (a) that the wreckage might well never be found, and (b) that the most likely scenarios, in his view, involve one of the pilots deliberately bringing the plane down. Sample:

    Feith believes there will be several plausible theories that all will point to a deliberate act by someone with intimate knowledge of flying the Boeing 777, most likely one of the pilots.

    Too many deliberate actions maneuvering the airplane and turning off communications systems occurred to have any plausible mechanical failure explanation. He completely discounts theories that a fire, either in the electronics bay or involving lithium-ion batteries being transported in a cargo bay, disabled the airplane.

    He also discounts a theory that there was a depressurization that incapacitated the pilots and allowed the 777 to meander over the skies of the Gulf of Thailand, Malaysia and the Strait of Malacca before turning south 3,000 miles over the Indian Ocean before running out of fuel.

    No one knows what happened, and it's possible that we may never, or not for a very long time, get conclusive evidence one way or another. But this article is worth considering as a strong counter to the inflight-emergency view.

    Previous post

  • Where to Get Your Next Fix of MH370 News

    Chris Goodfellow, who has offered the so-far least-unlikely explanation of what happened, does a 90-minute Q&A.

    Evolving search areas for Malaysia Air 370. ( via VOA )

    Executive summary of what you'll find below: If you're looking for more discussion of MH370, please swap the first 90-minutes of this "Google Plus Week" (embed below) for the next 90 minutes you would spend with cable news. It features the person who came up with the at-the-moment-least-implausible explanation of what happened, and its BS/insanity quotient is practically nil, in stark contrast to the normal cable level.

    Now the details.

    Yesterday I was on CNN's Reliable Sourceswith Brian Stelter (and Poynter's Andrew Beaujon), talking about the way CNN, in particular, has decided to go wall-to-wall in covering the missing airplane. Stelter pointed out that when CNN dealt with MH370, its ratings went up; when it didn't, the ratings went down. Therefore the network had quite rationally decided to make itself into the Missing Plane Channel for the foreseeable future, much as it became the Gulf War Channel in the early 1990s, in the period that first established its worldwide role.

    I replied: I understand this business logic, which fits the reality of the modern highly diversified news ecology. If you want to find out, right now, what's up with sports, or weather, or the stock market, or political trends, you know where to go. If you want to find out about the MH370 search, you know you can go to CNN.

    But my main complaint was that CNN had been so undiscriminating in filling these hours with nutso speculation -- black holes, "radar shadow," attack on Israel -- right alongside people who kept their discussion within the realm of the plausible. It's one thing to say that almost nothing is known about what happened with the plane. It's something else to have people gas on about things with no evidence to support them and with strong common-sense obstacles to being true.

    Political analogy: no one knows who will be elected president in 2016. But if we treated electoral handicapping the way cable news has often treated MH370, we'd have panelists speculating how Megyn Kelly might do against Stephen Colbert in the crucial swing states. ("Kelly will help the Republicans with the youth vote, and women, without in any way depressing interest from their traditional base of older white men. But Colbert, who is from South Carolina, could open new possibilities...") After all, you can't prove they won't be the nominees.

    Which brings us to Chris Goodfellow. Nearly three weeks ago, when most discussion concerned hijacking or pilot criminality, he offered on Google+ a different MH370 hypothesis. In essence it was: that some mid-flight emergency (probably a fire) had broken out on the plane; that the pilots had immediately turned back toward the nearest big airport, which was on an island off Malaysia; that for some still-unknown reason they had become incapacitated or disabled; that also for unknown reasons, possibly fire that disabled their radios, they had not been able to communicate; and that the plane, on autopilot, had flown on until it either ran out of fuel or crashed for another reason.

    Soon thereafter, I wrote that this was the first hypothesis that made face-value sense to me. Maybe things didn't play out this way --  but this scenario started out with the Occam's Razor advantage of requiring fewer assumptions or suspensions of probability than others.

    An item in Slate immediately and with great certainty declared that Goodfellow's scenario could not be true. Its author, Jeff Wise, became a regular on CNN making that same point -- and meanwhile promoting the hypothesis that the plane had landed in Central Asia. Eg, "the 777 is capable of landing on small airstrips and on relatively unimproved surfaces, such as packed dirt and dry lake beds. In such a scenario, the odds are good that, unless they were murdered, the passengers remain alive." (Also here. For the record, in his original anti-Goodfellow item Wise included me among people who he thought had been taken in by Goodfellow. He knows that I disagree with him.)

    This weekend Chris Goodfellow did a Q-and-A session on the "Google Plus Week" channel on YouTube. The next time you're looking for 90 minutes' worth of discussion of what could have happened, what we know and don't, and why the plane's reported cargo of lithium ion batteries deserves attention, I'd recommend this over any comparable time with cable news.

    It could turn out that Goodfellow's view is entirely wrong, but his pattern of thinking about the puzzle is systematic and worth hearing.

    Previous post

  • As the Search Goes On for the Plane That Disappeared, How Can So Many Flights Take Off and Land Safely Each Day?

    More on the details of designing for 1-in-a-billion accident rates.

    Representation of flights underway at any given moment. ( Flixxy.com )

    Earlier today, I quoted the longtime aviation writer J. Mac McClellan on the one-in-a-billion risk factor to which modern aircraft design is held. Someone familiar with such standards writes in:

    I'm a system safety engineer for a small-ish system supplier, so I'm pretty familiar with the 10^-9 standard. There are a number of issues with probabilistic risk assessment, but I think the history of the 1 in a billion standard is pretty interesting. This is an excerpt from the proposed rule change to FAA regulations regarding system design, referred to as the ARSENAL draft of 25.1309. [Excerpt begins:]

    The British Civil Airworthiness Requirements (BCAR) were the first to establish acceptable quantitative probability values for transport airplane systems.  The primary objective in establishing these guidelines was to ensure that the proliferation of critical systems would not increase the probability of a serious accident.  Historical evidence at the time indicated that the probability of a serious accident due to operational and airframe-related causes was approximately one (accident) per one million hours of flight.  Further, about 10 percent of the total accidents were attributed to failure conditions caused by the airplane’s systems. Consequently, it was determined that the probability of a serious accident from all such failure conditions should not be greater than one per 10 million flight hours, or “1 x 10 -7  per flight hour,” for a newly designed airplane.  Commensurately greater acceptable probabilities were established for less severe outcomes. 

    The difficulty with the 1 x 10 -7  per flight hour probability of a serious accident, as stipulated by the BCAR guideline, was that all the systems on the airplane must be collectively analyzed numerically before it was possible to determine whether the target had been met.   For this reason, the  (somewhat arbitrary) assumption that there would be no more than 100 failure conditions contributing to a catastrophe within any given transport category airplane type design was made. It apparently was also assumed that, by by regulating the frequency of less severe outcomes: 

     * only 'catastrophic failure conditions' would significantly contribute to the probability of catastrophe, and  

     * all contributing failure conditions could be foreseen. 

    Therefore, the targeted allowable average probability per flight hour of 1 x 10 -7  was apportioned equally among 100 catastrophic failure conditions, resulting in an allocation of not greater than 1 x 10 -9   to each.  The upper limit for the average probability per flight hour for catastrophic failure conditions became the familiar “1 x 10 -9 .”  Failure conditions having less severe effects could be relatively more likely to occur." [Excerpt ends.]   

    They basically worked backwards from the existing accident rate, made a few assumptions about contributions from complex systems and got us this number. There are a few questionable assumptions such as the number of catastrophic failure conditions. Thankfully, more goes into safety now than estimating probabilities such as human factors and common cause assessments. But it does point out that the standard was arbitrary to begin with, so changes in public perception may eventually change the standard.

    One minor correction on Mr McClellan's note. The 10^-9 standard is referred to as "extremely improbable" rather than just "improbable" and it is in terms of average probability per flight hour, not per average flight. See section 7c(1) of the arsenal draft of 25.1309 under "Probability Ranges."

    (1) Probability Ranges. 

          (i) Probable Failure Conditions are those having an Average Probability Per Flight Hour greater than of the order of 1 x 10-5 .  

          (ii) Remote Failure Conditions are those having an Average Probability Per Flight Hour of the order of 1x 10-5 or less, but greater than of the order of 1 x 10-7 . 

          (iii) Extremely Remote Failure Conditions are those having an Average Probability Per Flight Hour of the order of 1x 10-7 or less, but greater than of the order of 1 x 10-9. 

          (iv) Extremely Improbable Failure Conditions are those having an Average Probability Per Flight Hour of the order of 1x 10-9 or less. 

    I sent this to Mac McClellan, and he replies as shown below. (We added the photo, which is of a different engine failure from the one he mentions):

    Yes, the standard evolved over time and has some interesting twists. For example, passengers can be seated in a turbine engine rotor burst zone and would presumably be killed by a burst. A rotor burst energy is now treated as infinite and debris will pass through anything or anybody in the zone. However, no required crew can be located in the burst zone. The 10-9 standard doesn't necessarily apply to a passenger or passengers staying alive, but to the airplane and it's ability to reach a runway.

    Destroyed turbine engine from a Delta MD-88 with P&W engine, in 1996. (Wikipedia photo)

    Also, much of the historic data floating around is historic. Rules change constantly, and there was a big change after the DC-10 in Iowa where the center engine exploded. Douglas had installed triple hydraulic lines to the tail control surfaces but the lines were routed close together. The engine burst took out all three systems. Up to that time triplex was enough. After that triplex was only good enough when you could demonstrate that no single foreseeable event would take out all three.

    The 777 was certified under pretty current rules, as was its FBW which does meet the 10-9 through triple redundancy and several levels of computer participation. The final level is direct law where the cockpit controls command direct movement of a surface with no enhancement or protection for speed or CG [Center of Gravity] or other considerations.

    This is way more technical detail than most people will want to follow. But so much of this story, which continues to command interests, turns on precise technical details; and for those who are interested in the safety and redundant-design criteria of modern aircraft, this will be instructive.

    Previous post                                                                        Next post

  • When a 1-in-a-Billion Chance of Accident May Not Seem 'Safe Enough'

    The very safety of modern air travel makes accidents both intellectually more difficult and emotionally more disturbing.

    Navy pilots fly a P-8A Poseidon during a mission to assist in search and rescue operations for MH370. (US Navy)
    As the mystery about the fate and location of MH370 continues, and as theories come and go about what might have happened, here is a note from J. Mac McClellan, long-time editor of Flying magazine, about a phenomenon I've mentioned frequently. First-world commercial air travel has become so extremely safe that when something does go wrong, figuring it out can be a huge challenge -- which heightens the mystery and, for many people, the terror of these episodes, by making them seem so random. You're sitting there grumbling about the discomforts of modern flight -- and then, for no apparent reason, your plane is the one headed into the sea. McClellan writes: 

    As you probably know the FAA standard, and pretty much the global standard, for certifying critical components and systems is one in a billion probability of failure, or 10 to the minus 9th. The FAA calls this standard "improbable."

    That means in a transport category airplane [JF note: this includes airliners] the certification standard for a failure, or combination of failures, that would prevent the airplane from successfully landing on a runway must be one in a billion flights. Not hours, flights.

    I remember that when the 777 was introduced it was such a sales success and was expected to live such a long service life that some people speculated the fleet could actually make a billion flights. Of course, you don't need to make a billion flights to draw the magic short one-in-a-billion straw. But it is something to think about. Transport flying is now so safe that the long time standard of 10 to the minus 9th may not satisfy the public.

    I'm sure you are also tired of hearing about all of the things a transponder does that it really can't. Every comment on a transponder says it reports course and speed, but we know a transponder, even a Mode S as you have and the 777 has, reports only an identification code and Mode C pressure altitude. Course and speed all must be calculated by observation by radar. I guess the media and experts have mixed up what ADS-B does with what a transponder does. [For more on transponders, here; on for ADS-B, here.]

    Also odd that the 777 FBW [fly-by-wire, or electronic system for directing the airplane's control surfaces] system has escaped almost all speculation. It was the first for Boeing. And it was failure of the pitot input that put the FBW system into "direct law." [That is, it disabled the normal automated limits on "control inputs" the pilots could give to the airplane. In "normal law," which prevails within normal flight circumstances, the autopilot impedes or buffers any control input thought to be unsafe, for instance too sharp a turn or too steep a climb at too low an airspeed.] This handed the Air France crew an airplane that the computer could no longer control while expecting the humans to quickly diagnose a problem the computers couldn't. I'm not saying FBW has anything to do with 370 but it must be on the list of considerations.

    This is not speculation, simply some basic info that I haven't seen touched on during the endless TV interviews and such.

    The sobering point here is again that the very safety of modern air travel makes these episodes both intellectually and emotionally even more difficult.

    One other aspect of the drama is the national reactions and tensions it has highlighted -- of course in Malaysia and China, also in Australia, even in Israel. Thanks to many people writing in with on-scene reports of reactions in China and Malaysia. Will sort them out and report as I can.

    Previous post                                                                    Next post

  • Malaysia 370, Day 10: One Fanciful Hypothesis, and Another That Begins to Make Sense

    Did the airplane hide in a "radar shadow"? Probably not. Did the flight crew act like heroes? Possibly so.

    Tintin and a disappearing airplane ( Tintin Wikia )

    I rejoin the Internet after a day away to find no additional hard evidence about the fate of Malaysia Air flight 370, but a number of new rumors and possibilities. To run through a few:

    1) The "Radar Shadow" Hypothesis. Many, many readers have sent in links to a post early today by Keith Ledgerwood. He suggests that the Malaysian plane might have avoided radar detection by sneaking up on and deliberately flying right next to another 777, so that radar operators would see only a single blip from this ad-hoc formation flight.

    You can read the intriguing details for yourself, but the crucial points are:

    • The other plane, a Singapore Airlines flight en route to Spain, would not have known the Malaysia flight was right behind it, because its onboard collision-warning system (called TCAS) senses other aircraft by their transponder signals. Since MH 370 had its transponders turned off, the Singapore TCAS system would have nothing to work with -- and would get no warning from ground-based radar operators, who would not realize they were looking at two planes.
    • Meanwhile, MH 370 could creep very close to the Singapore plane without crashing into it, because the Singapore transponders were still working, and would broadcast its position to the Malaysian plane. (Plus, in the night sky the Malaysia pilots could see the other plane's green, red, and white navigation lights as it flew along ahead of them.)
    • After going as far as it wanted in the Singapore airplane's shadow, MH 370 could peel off at some point and head toward its intended destination.  

    Is this possible? At this point, when no normal expectations have panned out, I suppose almost any conjecture must be entertained.

    Is it likely? Or even plausible? Neither, in my view.

    Apart from the general rococo-ness of the plotting, this interpretation rests on a piece of evidence that I view in a very different way from what's implied in the post. Keith Ledgerwood notes that the two planes followed exactly the same course across a series of aerial way points ("intersections" with 5-letter names like IGREX and VAMPI) at very close to the same time. Isn't this suggestive of something strange?

    Actually, not. On many heavily traveled air corridors, planes will be sent along exactly the same sequence of way points at intervals of a few minutes. (If you listening to Air Traffic Control near a major airport, you'll hear one plane after another receive the same routing instructions.) I view it as routine rather than exceptional that planes might have crossed the same sequence of intersections.

    So maybe this will turn out to mean something -- and if so, all respect to Mr. Ledgerwood. My bet is that this will be another interesting-but-fanciful interpretation, and that the cause will prove to be something else. 

    2) The Pulau Langkawi possibility. Over the weekend Chris Goodfellow, an experienced pilot, offered via Google+ a very different sort of explanation. Far from carrying out an elaborate scheme, he says, the pilots may have been caught by surprise by an inflight fire, a major systems failure, or some other genuine emergency. At that point they called on the reflex nearly all pilots develop: the constantly updated awareness of where the nearest airport is, if they should suddenly need to get back to the ground. As he puts it:

    We old pilots were always drilled to always know the closest airport of safe harbor while in cruise. Airports behind us, airports abeam us and airports ahead of us. Always in our head. Always. Because if something happens you don't want to be thinking what are you going to do - you already know what you are going to do.

    When trouble arose, Goodfellow says, the pilots tried to head for what they knew to be the nearest very long runway, with an unobstructed over-water approach, on the Malaysian island of Pulau Langkawi. (Pulau means "island.") Here's the Google Earth idea of how the Langkawi runway might look in daylight, although the plane was of course approaching at night. That runway is 13,000 feet long -- enormous.

    But they never made it. Before getting the plane down, Goodfellow suggests, the pilots could have been incapacitated -- and the plane would fly on until it ran out of fuel. This view is notable for the light it casts on the MH 370 flight crew. Far from being villains, schemers, or the objects of a hijacking plan, he says they were in fact heroes, struggling until the last to save their aircraft, themselves, and the 237 other souls on board. Referring to the senior pilot, he says: 

    This pilot did all the right things. He was confronted by some major event onboard that made him make that immediate turn back to the closest safe airport.... 

    Fire in an aircraft demands one thing - you get the machine on the ground as soon as possible....

    Smart pilot. Just didn't have the time.

    Goodfellow says he is certain this is what happened: "No doubt in my mind."  I think there's doubt about everything concerning this flight. But his explanation makes better sense than anything else I've heard so far. (And he has updated it in light of developments since his original post.) It's one of the few that make me think, Yes, I could see things happening that way.

    3) Flight 714. Many readers have written in to say that the best fictional reference for the mystery of this plane is not Thunderball, nor You Only Live Twice, nor any other part of the James Bond oeuvre. Instead it's Tintin, as a reader in Los Angeles explained:

    I can go the Thunderball reference one better…the comparison I make is to the plot of the Tintin story “Flight 714”, in which a rich man’s jet is hijacked by part of the crew and crash landed on a deserted island in the java sea. 

    The numerous parallels are quite interesting…it’s a crew takeover, they drop out of sight of radar, it all takes place in the same general part of the world…and the scene in which they show how the plane lands (on a hastily constructed airstrip, which is then dismantled) could explain a lot.  Frankly at this point, you’d be better off reading Flight 714 than watching the cable news reports.

    Previous post                                                                       Next post

  • Malaysia 370 Update: Landing Strips, Cell Phones, and More

    Has the plane landed at a secret site? Probably not; here's why.

    Dr. No, who had a hidden island command center ( via Glamstruck )

    Update: See previous articles at the bottom of this post.

    The ongoing Malaysia 370 investigation coincides with my being in transit, with family, and away from the Internet most of each day. (Writing this from the passenger seat of a car on a four-hour drive, hoping that my TMobile hotspot via Samsung Galaxy III holds up.) Here is a quick update on some of the developments since the inflight dispatch yesterday:

    1) Derek Thompson sums up recent news for the Atlantic. You can see it here.

    2) Rupert Murdoch loses his mind. You can see it here. What's most amazing about the response below is that it happened before anything was known about the flight -- whether it had blown up, ditched in the sea, been hijacked, landed safely by mistake somewhere, etc.

    It's possible that the jihadist interpretation will turn out to be true. But the word "confirms," before anyone knew (or yet knows) what happened to the flight, from perhaps the single most powerful "journalistic" figure in the world is ... well, it "confirms" a lot of judgments about Murdoch.

    3) What about those cellphones?  We all know that cellphones can minutely track our movements as we walk or bike through cities or drive through the countryside. So why aren't they being used to track this flight?

    One answer: We don't know whether all the phones were seized and disabled, if this was a hijacking. Another: phones can track us in our normal life because we're operating right at ground level, and in places designed to offer phone coverage. At airliner-flight levels, 35,000 feet in the case of this plane, and at airliner speeds, there usually is no coverage. (Try to make a call from 30,000+ feet on your next cross-country flight.) At any altitude there is usually no coverage over open water or in remote, jungle, mountain, or desert areas, which describes most of the path of this flight. More in a good AP explainer here

    4) What about some other runway? Buried in our collective memory is the image from You Only Live Twice, or even Dr. No** (which I mention as an excuse to use the poster above),  or other fantasy movies of a hidden, secret runway that magically opens up just long enough for an airplane to land, and then disappears or is covered over again. Sadly I do not see such an image on the Internet right now.

    Based on the facts as now understood, it is conceivable that the plane, rather than crashing, was deliberately flown to some remote side. (In another Tweet, Rupert Murdoch said it would be somewhere similar to Osama bin Laden's Af-Pak hideout.) 

    The main challenge here is that a Boeing 777 is a big airplane, which needs a big, flat space on which to safely land. This Boeing technical manual suggests that in normal circumstances, you'd want 7,000 feet or more to land a plane full of passengers and have margin for error. Slate quoted a 777 pilot who said that if the plane was on fire (ie, the worst kind of in-flight emergency), he would try to put it down on anything above 5,000 feet. 

    WNYC has produced a map showing the 5,000-foot runways within conceivable flight range of the plane. Sample here:

    Theoretically possible landing sites, via WNYC

    Congrats for the work that went into this  -- and I mean to sound supportive rather than churlish in hoping that the next version of the map will have popups giving the names of the relevant airports, plus elevation and runway length. My guess is that only a small fraction of those shown would be suitable -- by terrain, location, elevation, and other factors -- as a deliberate diversion site. And even if all of them were feasible, it's a finite list. Most airports that big would have control towers; in that part of the world, many would be military-run; and spy satellites can easily pick out mile-long runways from above.

    My claim: if the plane had landed at a runway big enough to handle it, we would know that by now.

    5) What about the 45,000-foot altitude claim, and the 40,000-feet-per-minute descent? Reports since last night speculated that the plane had climbed very high, and then descended very fast, perhaps indicating: an incompetent/amateur pilot; a professional pilot bent on disorienting the passengers or destroying the plane; or something else strange. 

    To put this in perspective: in normal airline flights, you have rarely if ever been above 40,000 feet. Most airliners operate in the high-20s through the high-30s, in thousands of feet. Assuming that pressurization systems still worked, passengers wouldn't necessarily have noticed a difference at 45,000.

    They certainly would have noticed a 40,000-fpm descent. In normal airline flights, you've rarely if ever felt a descent of more than 2,000-fpm. Most of the time, airliners go down by 1,000 - 1,500 fpm. Descending 20 or 30 times that fast would mean that the plane was pointed more or less straight down, with engines running. 

    So if this happened, it would have been remarkable, and terrifying. And among the problems would be pulling out of the dive without subjecting the plane (and crew and passengers) to G-forces beyond what any of them were designed to tolerate.

    6) What about the Chinese role? There will be a lot more here, but for now, before we head into an area where my little hotspot will give out, here is a note from a reader making good points:

    Since you're one of the few people left who think of aviation as part and parcel of a national identity, the Chinese reaction has been fascinating as well.   

    1.  The highly responsible and flexible response by the Chinese leadership

    2.  The obvious panic by the public and family members who are not being kept in the loop and may (or may not) have easy access to information.

    3.  The inability of said Chinese leadership to 100% control their own people (the satellite leak,etc).

    The now default American security response (Terrorists!  Coming to get us!) is pretty weak as well.  Although one would hope the NSA can get away from Yahoo Chat for a few minutes to do something useful.

    Signal is flickering out. More when back on line.

    Previous posts on Malaysia 370:

    1. The first grim indications

    2. Why cable news should be on a 24-hour news delay

    3. How airplanes report their data while flying and more

    4. Why a better black box won't help and more

    ** My Atlantic colleague and movie guru Christopher Orr writes in to set the record straight:

    This is Thunderball, where the pilot's twin brother takes his place, hijacks the plane, and lands it on the bottom of the ocean. 

    Though the purpose is less clear in this case, given that MA 370 presumably wasn't carrying nukes.

    Previous post                                                                           Next post

  • Why Malaysia Airlines 370 Remains So Profoundly Mysterious, and Why a Better Black Box Wouldn't Help

    The absence of data about this flight is itself a significant data point.

    Track of Malaysia Air 370 on March 7 ( Flight Aware )

    (Please see update with the March 14 news.) Here is the heart of the mystery over what has happened to Malaysia Airlines flight 370:

    • If the airplane did keep on flying, presumably there would be evidence of that fact  -- at a minimum through "primary radar returns," blips on civilian or military radar screens showing that something was in the air even if the plane's transponder was not sending back specific identifying info.
    • If the airplane did not keep on flying, presumably there would be evidence of that -- through wreckage on the ground, oil slicks or debris in the sea, satellite detection of a flash or explosion at the relevant time. 

    As of now, six days later, there is no clear evidence of either type. Or other evidence to suggest difficulties with the weather (in contrast to Air France 447 -- and I'll have more on this soon), suspicious actions by passengers or attackers, problems with the flight crew, a pattern of failure with this kind of airframe, or any of the other usual components of the "accident chain" in aviation disasters. As I mentioned earlier, airline travel is now so amazingly safe that when something does go wrong, the cause usually turns out be some previously unforeseen triple-whammy combination of bad-luck factors. Air-safety experts refer to this as the "Swiss cheese" factor: the odd cases in which the holes in different slices of Swiss cheese happen to line up exactly, letting the improbable occur.

    But so far MAS 370 is in a category of its own, in the shortage of useful data and the mismatch of what is known with most imagined scenarios. This is a source of additional heartache for affected families, anxiety for some in the traveling public, and embarrassment for the Malaysian officials clumsily running the search. (As mentioned, I am a fan of Malaysia-the-country and of Malaysia Airlines, but Malaysian safety officials are looking bad.) Yet it is the frustrating reality. The closest comparison would be the crash of TWA flight 800 18 years ago. The absence of data is itself a surprising data point.

    Now, about one common pundit claim: If only we had better "black boxes," and more real-time streaming of black-box data, we'd be spared mysteries of this sort. Michael Planey, a Washington-area consultant who has worked for several airlines and did air-safety investigations for the Air Force, writes in to explain why this is a false hope. 

    I'm quoting his message in full detail, since in cases like this the details matter. If you don't want to deal with all the specifics, his main point is: the disappearance of this airplane remains profoundly mysterious, and would probably remain so even if one much-discussed "remedy" had been in place. I turn the floor over to Mr. Planey:

    Would realtime streaming of black box data end the mystery of what happened to MH370? Probably not.  Here’s why.

    As the search for Malaysia Airlines flight 370 continues in earnest, many have called for the implementation of realtime streaming of black-box data.  It is an understandable reaction to an inexplicable event:  that a modern airliner could simply vanish without a trace.  The thinking is that real-time black-box data would make it possible to locate the aircraft more quickly; to understand what had happened to the aircraft causing it to lose contact with air traffic control; to perhaps prevent an aircraft safety incident through monitoring of aircraft systems and highlighting suspect or anomalous data.  But is that really the case with this aircraft and this flight?  Unfortunately, I suspect not.

    The last loss of a commercial airliner in trans-oceanic flight was Air France Flight 447 on June 1, 2009.  In that case, some system failure reports and warnings were transmitted via ACARS [JF note: a data transmission system linking in-flight airplanes with ground stations] in the last moments before the aircraft crashed into the Atlantic.  This data was useful in the preliminary understanding of the event, but it was not enough data to paint the complete picture of the complex system failures and flight crew actions that led to the crash, nor prevent it from happening.

    In that case, the data transmission was of no particular use in locating the debris field.  Rather, traditional air traffic control and radar data was used to pinpoint the last known location of Flight 447 and the search began at that point.  The aircraft wreckage was located by the next day in the expected area.  In the current case of MH 370, the same type of location data is available, but the search has been fruitless.  This opens the up possibilities of the aircraft’s fate to scenarios where data-streaming would again be ineffective.

    Given that the Boeing 777-200 aircraft on this flight had been recently inspected and operated without incident over the prior ten days, there are no red flags leading to a likely cause of the disappearance.  Even though this aircraft was equipped with an ACARS system like the Air France flight, no relevant data transmissions were made.  This reasonably points to a thoroughly unforeseen, catastrophic event (such as TWA Flight 800) or perhaps a deliberate action such as hijacking, terrorist action or even flight crew suicide.  

    In the case of the immediate, catastrophic event, data streaming would likely cease at the moment of the event.  Either a complete loss of electrical power would disrupt the data stream or a mechanical break in the aircraft systems would prevent data transmission.  Further, if an aircraft was in an out-of-control attitude such as a steep dive, a spin or a hard roll, maintaining a direct link with a satellite would be nearly impossible, thus again breaking the data stream and rendering the system incapable.

    If the demise of MH370 is due to a deliberate action, realtime data-streaming is again unlikely to yield definitive answers.  If hijackers were sophisticated enough to completely cut-off all communications (radios, ACARS, transponder, ADS-B) then it would stand to reason that the data link would be cut off in the same manner.  Further, the detonation of a bomb would not show a prior indication of the event in the flight data-stream.  Perhaps, the very slight chance of aircraft depressurization or loss of fuel volume would be detected at the moment, but it is unlikely that such a signal could be successfully transmitted before the communications system was rendered useless.

    It is important to note that the “black-box” is actually a pair of boxes.  The Flight Data Recorder secures information from a host of flight systems and the flight management computer.  The Cockpit Voice Recorder captures the last 30 to 60 minutes of dialogue in the cockpit and adds significant context to the FDR data.  In the investigation of AF447, the CVR was critical to understanding why the flight crew took the actions they did, even as the data could show what those actions were.  Capture of both information streams would be necessary for a full picture of what was happening at the critical moment.

    If days of intensive air and sea search efforts have yielded no clues, it is hard to believe that the aircraft and its crew were capable of providing any more useful information at the time the aircraft disappeared.

    Previous post                                                                                 Next post 

  • Ukraine and Malaysia Airlines: Why Cable News Should Be on a 24-Hour Delay

    The best reactions to breaking news are rarely the first ones.

    Malaysia Airlines 777, at Kuala Lumpur airport ( Wikipedia photo )

    I have been offline most of these past few days and thus not weighing in on daily developments. But let me mention three items whose similarity concerns cast of mind.

    1) Adam Gopnik on Crimea. This is several days old in The New Yorker but very much worth reading if you have missed it. For instance:

    With Ukraine and Crimea suddenly looming as potential [WW I-style] Sarajevos, the usual rhetoric of credibility and the horrors of appeasement comes blaring from the usual quarters. People who, a week ago, could not have told you if Crimea belonged to Ukraine—who maybe thought, based on a vague memory of reading Chekhov, that it was Russian all along—are now acting as though the integrity of a Ukrainian Crimea is an old and obvious American interest. What they find worse than our credibility actually being at stake is that we might not act as though it always is.  

    As the years go by, I am more and more convinced that the immediate, fast-twitch talk-show responses on what we "have" to do about some development are almost always wrong, and the calm, day- or week-after reflections about proportion, response, and national interest are almost always wiser. If I could, I would put all cable-TV discussion of breaking-news crises on a 24-hour delay. Maybe there has been a case in which immediate reflex-response to big news has seemed wise in the long run. Right now I can't think of any.

    Naturally this reminds me of an adage from the piloting world: In most emergencies, the crucial first thing to do is ... nothing. Take a deep breath, calm down, steady your nerves, count to 10, and then "fly the airplane" as you begin applying knowledge rather than panicked instincts to the options at hand. Which brings us to:

    2) Patrick Smith on Malaysia Airlines. At Ask The Pilot, airline pilot and aviation writer Patrick Smith makes the frustrating but unavoidable point about the still-missing Malaysia Airlines flight: We have no idea what happened, and it may be a long time (if ever) before we do.

    Here are the tactical points involved in this argument:

    • Commercial airline flight is now statistically so safe that when something does go wrong, the causes are often mysterious by definition. That is because the non-mysterious risks for airlines have been buffed away. The most famous recent exception was the Asiana crash at SFO last year. It looked from the start like a simple case of pilot error, and that is where all subsequent evidence points. But many other tragedies have taken months or years to sleuth out. 
    • The first reports after a crash should be viewed with great suspicion, because experience shows they're probably wrong. What the NYT says in its current headline about Malaysia Airlines applies to most disaster coverage:

      For this reason it  would be great to have a 24-hour tape-delay on most disaster coverage as well. 

      This goes in spades for any coverage on the lines of, "This latest tragedy proves that [theory X] is true." Most instant-analyses of this sort I can think of were grossly wrong; when they're right, that's often due to luck rather than insight. This principle applies not only to air crashes but also to mass shootings, bombings, episodes of suspected terrorism, and similar tragedies for which people crave an explanation.
    • Might the Malaysian plane have broken up in flight? Yes. Might it have been hijacked? Perhaps. Might both pilots have conked out? Maybe. Could there have been an on-board bomb? Perhaps. Does this show a problem with the Boeing 777? Likely not. Does it have anything to do with the Asiana 777 crash in San Francisco? Hard to imagine how it could. Did the stolen passports matter? Conceivably. Might the plane have been hit by a meteor? Or undone by pilot suicide? I suppose anything is possible. But these are all in the realm of "would King Kong beat Godzilla?" until there is more evidence, which can take a long time.

    The strategic point is: We do crave explanations, especially for bad news. Pilots are more prone to this tendency than anyone else. If you pick up an aviation magazine, you'll see that half the stories concern disasters, usually with the theme: Here is why bad things happened, and how to keep them from happening to you. But sometimes bad things happen for reasons no one can explain. Let's hope there is at least an instructive explanation, eventually, for this one. 

    Update: I am sorry to see that the usually excellent Foreign Policy has gone in for speculation-ahead-of-facts in a big way, e.g. here and, with the caveat that it is reporting on speculation, here.

    3) Jim Sleeper on the New Cold War. In an item about Leon Wieseltier for The Washington Monthly, Jim Sleeper gives another instance of what I'm suggesting is a larger point: that rushing, quickly, to larger self-confident, self-righteous stands is usually a source of error. He reminds us of what a group of "strategists" told the public a few days after the 9/11 attacks:

    [E]ven if evidence does not link Iraq directly to the attack, any strategy aiming at the eradication of terrorism and its sponsors must include a determined effort to remove Saddam Hussein from power in Iraq. Failure to undertake such an effort will constitute an early and perhaps decisive surrender in the war on international terrorism.

    People who react this way have the right temperament for cable talk shows but the wrong one for decisions about the national interest. Cable pundits are in business to say, "The evidence is not yet in, but we know this means [xxx]." Give us leaders (and accident investigators) willing to say, Calm down. Breathe. Let's wait a minute, and think.

    Previous post                                                                    Next post

  • The Malaysian Airlines Flight from Kuala Lumpur to Beijing

    No firm news, but the indications are grim

    As I write (at 9:40pm EST in the US, 0240 March 8 GMT), things look bad for Malaysian Airlines flight 370 from Kuala Lumpur to Beijing, but nothing is known for sure.

    The most illuminating information I have seen so far is this log from Flight Aware. It shows that the airplane had leveled off at 35,000 feet -- and then suddenly was not transmitting any more information about its location, speed, altitude, or rate of climb or descent.

    Flight log, from Flight Aware

    [Update: Flight Aware is imperfect, as I've written here many times. Also, the Lat/Long of its last report for this flight is over the Malay peninsula and is less than one hour after takeoff, versus the 2 hours that we've heard in many news reports. This goes into the category of "All early reports about air incidents are contradictory, confusing, and often wrong." But still the indications are not good.] 

    It is hard to imagine a systemwide failure of the transponder and other reporting equipment that could have made the plane stop transmitting any information and yet still be flying along safely. And in that case, it would presumably have tried to land somewhere along the Malay peninsula or in Indochina.

    The main insight I have to add for now is: Whatever happened, it is unlikely to reflect chronic shoddiness with Malaysian Airlines, which in my experience is a good, competent, and modern airline (I have flown MAS many times, including during the two years I lived in Malaysia in the 1980s, and more recently along this route), nor with the Boeing 777, a good and well-experienced airplane. Beyond that, we await further news with best wishes for all involved.

    12-hours-later update: Flightradar24 has more detailed reporting on what appear to be the plane's last known positions, over the South China Sea.

  • A Longtime Pilot on the Aspen Crash

    Modern jet planes are so safe that many things have to go wrong to bring them to grief. What some of the factors might have been in this latest case.

    Aspen Airport, c/o Aspen airport authority. Planes usually land from right to left, in this view.

    J. Mac McClellan is known through the flying world as the long-time editor of Flying magazine.  Now he has a regular column for the EAA, the Experimental Aircraft Association. He sent a message about this weekend's fatal crash at Aspen, when a private jet landed with strong and gusty tailwinds (as mentioned here).

    Bombardier Challenger 600-series, similar 
    to the jet that crashed this past weekend.

    This is more highly detailed than some readers may care to know or even be able to follow. But my experience is that after aviation mishaps of any kind, even people not usually interested in aviation are grateful for additional, detailed information.

    Part of the reason may be a search for reassurance that the airplane didn't just fall out of the sky, which virtually never happens but is a widespread if unspoken fear. Part of the reason may be the complex chain of bad luck, circumstance, and (often) miscalculation that leads to a crash -- today's airliners and private-jet planes generally being so reliable and redundantly fail-safe-equipped that it usually takes many things going wrong at once to bring one to grief. Part of the reason is the unavoidable Bridge of San Luis Rey-style human fascination with the chronicles of misfortune.

    In any case, here is the report from Mac McClellan. 

    As you know from landing there, the approach to Aspen is visually confusing because the runway slopes uphill. [JF note: In the photo at the top of the page, mountains are just out of view to the left, so planes approach from the right. The runway slopes up more than 150 feet from the right-hand side to the left.] Also, there is a depression on the ground short of the landing threshold which accentuates the perception of upslope of the runway.

    The Challenger series is one of only two jets I have ever flown that is nose-down during final approach instead of level to nose up as other jets are on final. This looks really odd from the cockpit in the 601 Challenger business jet like the one that crashed, but is really strange in the longer CRJ series. That’s why many regional airline pilots call those things “lawn darts.”

    The other bizjet that approaches nose down is the Hawker 4000, first called the Hawker Horizon. Originally the design called for leading edge slats which were abandoned to save weight and money. Because the slats were not included in the actual airplane the behavior of the wing is different so it flies nose down on approach. At least that’s what the experimental test pilot guys at Beech believe explains the unusual attitude.

    The first time I flew the Hawker 4000 after it received some sort of provisional certification we decided to go to Aspen so they could show off its hot and high performance. I’m on final to a runway sloping uphill in a jet flying nose down to make my very first landing in the airplane and it was very strange. When the test pilot in the right seat started to gasp I hauled back hard on the wheel and got the nose up in time, but barely.

    Throw in a tailwind, wind shear, the visual illusion of the runway slope, and a nose down attitude on very short final and that Challenger crew had its hands full. That is not an excuse, but my Hawker arrival at Aspen came to mind when I first heard about the accident....

    I don’t want to speculate on what caused this particular accident, but the conditions of the runway slope, terrain, wind, unique approach angle of the Challenger and so on are simply noting the many challenges this crew faced in trying to land. Lots of factors all came together here that made the approach unusually difficult and that’s what we know and can say for now.

    And, from another reader with experience in this field: 

    Like most pilots, I was tracking the reports out of Aspen closely.

    One aspect that hasn't been mentioned yet is the maximum tailwind component permitted in jets. In the CL[Challenger]300 and 900XP I've flown, it's 10 knots. I'm not sure about the CL600 family but I suspect it is very similar.... [JF note: As mentioned previously, tailwinds at the time were in the 16-26 knot range.]

    This doesn't answer why the crash occurred. This appears to be a loss of directional control as opposed to running off the end of the runway. I would bet you lunch (at Hardee's) the tailwind will be a contributing factor...


  • Three Crashes: Aspen CO, Buckhannon WV, Melbourne FL

    What we know, and don't, about today's fatality in Colorado.

    1) Colorado. This afternoon a private jet crashed, with at least one fatality, at the Aspen airport. Here is one of several online reports from people at the airport or in other planes:

    What is knowable, in the short run, about this sad event is that apparently there were strong and gusty tail winds in Aspen while the plane was trying to land. One of many factors that make the Aspen airport, like many other mountain airports, very demanding is that for all practical purposes you can only land in one direction. There are mountains close to the southeastern side of the airport, so virtually all flights land from the north, in the direction shown by the red arrow. (The arrow is something I've photo-shopped onto a FAA Sectional chart.)

    In aviation terms, any given runway has two different names, depending on which direction the planes are going. In Aspen, planes virtually always take off using "Runway 33" -- starting at the southeastern end and going toward the northwest, over the valley, in the direction of compass heading 330 degrees. And they virtually always land on "Runway 15," coming in from the northwest over the valley, in the direction of the red arrow, and landing toward the southeast with heading 150 degrees. (I have flown a propeller plane into and out of Aspen several times. But it is so demanding and weather-dependent, and I am so aware of not being experienced enough in mountain flying, that I choose not to do it any more.)

    The problem with today's weather is that the wind was blowing strongly from the northwest, in exactly the same direction as the final landing path. When you land into the wind, as pilots always prefer to do, your ground speed relative to the runway is your airspeed minus the windspeed. Thus a plane with approach speed of, say, 100 knots, and a 20 knot headwind, would touch down at 80 knots relative to the runway.

    In this case, the plane had a 15 to 20 knot tailwind, which meant that its speed when meeting the runway would be airspeed plus windspeed -- 120 knots, rather than 80, in the hypothetical example. The problem with going so fast is that you can use up all the runway in a big hurry. You can't compare professional jet operations with amateur piston-plane flying, but just as a benchmark: the greatest tailwind I've ever had to land with was 5 knots, and I was impressed at how quickly the runway went by. Again, at non-mountain airports, you almost always have the choice of landing in the direction that gives you a head-rather-than-tail wind.

    An archived version of today's Air Traffic Control broadcasts from the Aspen tower is available here. The accident plane's call sign is 115WF, said "one one five whiskey foxtrot." On my first-pass listening, it appears that this happened:

    • The plane "went around" -- that is, aborted its first attempted landing -- because the tailwind was more than 30 knots.
    • The pilot set up for another approach, and was informed that the tailwinds over the preceding minute had averaged 16 knots gusting to 25.
    • The last transmission with this plane is at time 20:25, when it is cleared for landing and told of the tail winds.
    • The last ten minutes or so of this archive show the tower and ground controllers deploying all the other aircraft waiting to take off and land on what has become a closed runway, after the crash.

    First reports about crashes are often misleading. We know that it was a very high tailwind; what else might have been involved, we'll learn. For now, condolences to all affected. UPDATE Please see this informative item from Minnesota Public Radio, emphasizing the problem of strong, gusty tail winds.

    2) West Virginia. A happier outcome from a difficult situation: A Cirrus SR-22, the same kind of airplane my wife and I have been flying around the country, had engine trouble yesterday afternoon near Buckhannon, West Virginia. The pilot pulled the handle to deploy the "ballistic parachute" that is a feature of Cirrus airplanes. It came safely to the ground, on top of a truck, and the occupants walked away. Details here; West Virginia news shot below.

    3) Florida. For completeness, here is an animation, from the Air Safety Institute of the Aircraft Owners and Pilots Association, of a Cirrus crash in Melbourne, Florida, that had a more tragic ending. 

    Safe travels to all.
  • Asiana 214: Airplane as Hero, and Other Analyses

    The role of engineers, pilots, coaches, and God in the path toward disaster -- or safety.

    Three weeks after the crash, I hear from several travelers that debris from Asiana 214 is still visible at SFO, apparently as investigators keep working through the clues. I am entering my last day-plus in my current Internet-impaired environment, so a few text-only updates.

    1) The landing gear succeeded; they failed. From a reader in the Seattle area:
    One factor not mentioned in your posting was the "failure" of the landing gear. That is, in an impact beyond the strength of the LG, they are supposed to detach from the wing without breaking the wing off or ripping open the fuel tanks.
    They worked!  (before the frisbee pirouette )
    Likewise the engines.  Unfortunately, one of them came to rest snuggled up to the fuselage, and was the ignition source for the post-evacuation fire....

    I work for a certain aeronautical enterprise, and actually sent a congratulatory e-mail to the 777 designers...
    2) Credit to the airplane as a whole. From another reader in the industry:
    I agree that fatigue & a little bit of culture are the broken links in this chain of events. 

    I have worked as an aircraft mechanic for United Airlines for [more than 25 years] at [a major US hub], and most of us at work believe the 777 is one of Boeing's finest achievements. The talk around the hanger has always been  the 777 "is so smart it's a very difficult aircraft to have an accident in," and unfortunately without that technology engaged on the aircraft that  is exactly what happened to flight 214. 

    3) Other airplanes are strong too. A reader says the crash is a good occasion to remember a previous unusual landing with a happier overall outcome. This was the flight of the "Gimli Glider," whose 30th anniversary occurred last week:
    It's a shame that more people in the US aren't aware of that remarkable feat. I live outside of Detroit and was able to watch a short documentary about it on CBC the other night.  Admittedly, I had never heard about this until watching it. The stories of the pilot, crew, and passengers on that flight was far more interesting than the perceived impending crash of Noah Gallagher Shannon's flight.

    I'm including a link to the piece, it's about 20 minutes. You might need to use a proxy server in Canada to actually watch it.
    Short version of this saga: because of various fuel-management miscalculations, an Air Canada Boeing 767 ran entirely out of gas over Ontario. The crew and passengers all escaped alive only because the crew glided the plane down safely, with no engine power at all, from an altitude of 35,000 feet. Eg:
    Captain [Robert] Pearson was an experienced glider pilot, which gave him familiarity with flying techniques almost never used by commercial pilots. To have the maximum range and therefore the largest choice of possible landing sites, he needed to fly the 767 at the "best glide speed". Making his best guess as to this speed for the 767, he flew the aircraft at 220 knots (410 km/h; 250 mph). First Officer Maurice Quintal began to calculate whether they could reach Winnipeg. He used the altitude from one of the mechanical backup instruments, while the distance traveled was supplied by the air traffic controllers in Winnipeg, measuring the distance the aircraft's echo moved on their radar screens.
    Lots more in the Wikipedia account and on the video.

    4) Previous Asiana landing problems. A report on the SF Gate web site contends that even before the crash Asiana flights had a much higher-than-normal rate of "go-arounds," or aborted landing attempts, at San Francisco Airport. The sources for this claim aren't named, and FAA officials decline comment, although the Airport Director for SFO does go on-record as saying he had been concerned about Asiana's performance. 

    In itself, a decision to "go around" on any given landing can be a sign of a competent and safety-conscious flight crew rather than the reverse. When learning to land a plane, you're taught to be ready to go around at any moment before touchdown, rather than trying to save a landing that is shaping up the wrong way. But a pattern of frequently needing to go around can obviously be a bad sign.

    5) More on Confucius in the Cockpit. A Western reader who has worked for years in China previously sent in an account of a Dutch soccer coach who greatly improved the Korean national team by (in this reader's view) shaking up some Confucian concepts of hierarchy and group effort. Another reader disagrees:
    I'd like to send a brief note in response to your post - specifically, the anecdote about Hiddink [Guus Hiddink, the Dutch coach.] It is usually the case, in a situation in which a new authority changes the fortunes of a team (whether in sports or in business or wherever these inspirational movie theses appear), that some nugget of aphoristic truth can be gleaned from the turnaround, like so much dropping pitch. So it is with this one, in which Hiddink reverses the team's entire trajectory via an elemental ceremony that just so happens to represent the insertion of Western values into a team ruled by Eastern culture.

    What instead turns out often to be the case is that this turnaround was actually managed through an intensive process of redesigning the team's (or department, or business, or what have you) strategy and tactical objectives, followed by an even-more intensive process of working with the team to relearn this new method of approaching competitors and the world at large. I'm sure one of the things you realize in these situations is that they rarely make for good cinema until they're condensed into the crystalline pitch-wisdom that we see in your anecdote.

    Apologies for the soccer pun.

    While I am loathe to claim that a practice of giving up shots on goal didn't doom the Korean squad pre-Hiddink (as a player myself once upon a time, I know how precious SOG are), I regrettably cannot bring myself to believe that a) this practice was the only problem the squad had, or that b) the problem was solved in a single ceremonial display of Hiddink's authority.
    6) The 'Asoh Defense.' Back in 1968, a Japan Airlines plane bound for landing at SFO had a problem somewhat like Asiana 214's. The crew guided it through a properly stabilized approach --  but "landed" two miles short of the runway, right in San Francisco Bay. The circumstances were worse than in the Asiana case -- bad weather, and a ceiling of only 300 feet (versus clear skies three weeks ago). The outcome was better, in that no one was killed. The episode is famous in aviation lore for the "Asoh defense," the explanation offered by captain Kohei Asoh: "As you Americans say, I fucked up."

    7) It's a "cockpit management" problem, not a national-culture problem. Another reader writes: 
    On whether the crew's failure was a product of Korean "culture" or simply poor crew culture: Take a look at the NTSB report from the United Airlines DC-8 crash in Portland [Oregon] in 1978.  One of the safety recommendations stemming from that accident was to direct all air carriers to indoctrinate flight crews in principles of flight deck resource management.  It's my understanding that United took this recommendation to heart, and pilot Al Haynes credits that change with saving so many lives in the United DC-10 crash at Sioux City.

    As Haynes later said:
    "As for the crew, there was no training procedure for hydraulic failure. Complete hydraulic failure. We've all been through one failure or double failures, but never a complete hydraulic failure. 

    "But the preparation that paid off for the crew was something that United started in 1980 called Cockpit Resource Management, or Command Leadership Resource Training ...All the other airlines are now using it. 

    "Up until 1980, we kind of worked on the concept that the captain was THE authority on the aircraft. What he said, goes. And we lost a few airplanes because of that. Sometimes the captain isn't as smart as we thought he was. And we would listen to him, and do what he said, and we wouldn't know what he's talking about. And we had 103 years of flying experience there in the cockpit, trying to get that airplane on the ground, not one minute of which we had actually practiced, any one of us. So why would I know more about getting that airplane on the ground under those conditions than the other three. So if I hadn't used CLR, if we had not let everybody put their input in, it's a cinch we wouldn't have made it. 

    "I don't know if any of you remember the old movie Marty, I kind of refer to that, it was Ernest Borgnine, and a group of his cronies, trying to find something to do on a Saturday night, and they said, what do you want to do Marty, and he said, i don't know, what do you want to do Joe, and that's kind of the way we flew the airplane now."
    8) A final thought from Nevil Shute. A reader who is an active pilot says the episode reminds him of this line from Nevil Shute's autobiography, Slide Rule:
    "Aircraft do not crash of themselves. They come to grief because men are foolish, or vain, or lazy, or irresolute or reckless. One crash in a thousand may be unavoidable because God wills it so - not more than that."
  • Today in Asiana 214 News

    It's all about Confucian culture. Or circadian rhythms. Or personality. Or maybe something else.

    A roundup from readers around the world.

    1) Korean pilots doth protest. According to AVweb, the union representing Asiana pilots has filed a protest against the NTSB because of "NTSB's press conferences which only give prominence to the possibility of a pilot error." 

    Hmmm. The NTSB has pointed out that under ideal weather conditions, with no indication of mechanical failure of any kind, the plane's approach path toward the runway was never "stabilized" in altitude or air speed. Also, that about a second before impact, the cockpit voice recorder showed that the crew attempted (too late) a "go-around" to climb away and set up for another approach. Most accidents occur because of (a) a mechanical failure that redundant safety systems somehow can't cope with, (b) extreme weather of some sort, or (c) an error in judgment, execution, or decision-making, for whatever reason, by the plane's crew. The NTSB has said that so far there is no sign of causes (a) or (b).

    2) Airframe improvements. On a brighter-side aspect of the crash, reader TH notes a dramatic yet under-appreciated implication of the event:
    I haven't seen it remarked upon elsewhere, but one of the most incredible aspects of the accident, to me, is the fact that the plane remained virtually intact except for the tail section that took the initial impact. The crash video shows the enormous forces the airframe withstood -- the fuselage, wings still attached, whips around almost 360 degrees horizontally and perhaps 45 vertically -- and yet neither the fuselage nor the wings seem to suffer much damage, let alone shred apart. I can't help but note the extraordinary physical strength, and of course extraordinary engineering and manufacturing, involved. 

    Leaving aside all the advances in control, reliability, sensors, etc., surely that basic physical toughness represents a giant leap forward in passenger safety compared to 50 or even 25 years ago. Similar, perhaps, to the way automobiles have gotten safer over that period? 

    Yes. And a similar change has been notable even in the small-aircraft world. The kind of airplane that I fly, whose design and origin I described in Free Flight and that is now the best-selling small plane of its type in the world, has a fiberglass cockpit and airframe that have proven amazingly robust. The best known safety feature of these Cirrus SR-22 airplanes is the parachute for the entire airplane that can be deployed to avert a crash. But even when the plane has been banged up by trees or towers while descending under the parachute, the tough fiberglass cabin structure has stayed intact.

    3) The role of fatigue. Dr. Daniel Johnson of western Wisconsin, who has long experience in both  aviation and in medicine (and has been an FAA-designated senior medical examiner since the mid-1980s), writes about another factor: 
    As a pilot, and as a physician interested in mistakes of perception and circadian biology, I think nothing could be more obvious than, whatever else happened mechanically or procedurally, this Asiana 777 crash was likely related to jet lag, sleep deprivation, and its consequences on judgment, perception, vigilance, and reaction time. The flying-pilot's limited hours in the 777 is a red herring, as he has 12,000 hours of experience.

    1: All the pilots in this aircraft were finishing a long eastbound flight from Korea.
    2: We don't know that they slept well when off duty during the flight.
    3: We don't know that their duty schedule, sleep schedule, and light-dark schedule were ideal during the prior week or two.
    4: We don' know that they had managed their circadian rhythms wisely.
    We have clear evidence of degraded judgment and reactions:
    1: A decision to perform a hand-flown visual approach to the runway...
    2: A very low visual approach to the runway implies inattention, which implies degraded awareness.
    It is much harder to judge altitude when over water than when over land, as any pilot of seaplanes knows. Nevertheless, there would have been adequate visual clues, particularly the perspective view of the runway and the PAPIs. [JF note: these PAPIs are the red and white lights to help you judge if you are too high, too low, or just right on the descent.] I have much experience with low approaches; there is no doubt about one's glide angle unless perception is clouded.
    3: Slow reactions and obvious misperception, with degraded awareness, imply fatigue.
    A: failing to correct for the low approach angle
    B: failing to notice the slow airspeed (137 kt normal, about 105 knots actual) - Normally the pilot continually checks the airspeed. Failing to glance repeatedly at this indicates degraded attention, vigilance, and perception. There was not a cockpit call for more speed until 7 seconds before hitting the seawall. [JF note: if it turns out that the actual approach speed was 32 knots lower than the normal range, that is an enormous difference. Normally you try to maintain approach speed within 1 or 2 knots of the target speed.] 
    C: Failing to respond instantly to the stick-shaker (which began 4 seconds before impact; the pilot should have initiated instantly application of more power - jet engines don't respond instantly, but the correction should typically occur in about a half-second, even with a surprised pilot. Power was applied 2.5 seconds later, just 1.5 seconds before hitting the seawall.
    D: failing to apply power before the nose was raised. This is a failure of basic pilot skills. Power first, then attitude.
    E: Raising the nose as a reaction to the stick-shaker. This is wrong! The stick-shaker means that a stall is beginning; raising the nose guarantees a stall (and caused the tail to strike the ground). This a very common error, even among professional pilots, a Delta / Air Force Reserve pilot instructor has informed me. [JF note: Go back to all the discussion of the Colgan/Buffalo and Air France/mid-Atlantic crashes for more on this theme.]
    Ironically, an airplane's effective stall speed and rate of descent are both significantly lower - the induced drag is perhaps halved - when the airplane is within a wingspan of the ground/water ("ground effect") If this pilot had not panicked and pulled the nose up, this airplane could have been landed - awkwardly but without breaking the tail off on the seawall - without trying to speed up or to climb - simply by continuing to fly in ground effect to the runway. [JF note: Whether continuing in "ground effect" could indeed have saved the flight is one of the things the NTSB will presumably figure out.]
    So we see a collection of events that likely occurred due to degraded attention, vigilance, judgment, and reaction time. This severe impairment occurs with fatigue. This is most commonly due, in professional pilots, to jet lag and sleep deprivation...

    As you know better than I, fatigue is not, as is often portrayed, "not getting enough sleep." Fatigue is a complex symptom of diverse causes. Inadequate sleep duration is only one.
    Dr. Johnson also sent a link to this very interesting medical-journal article about the surprising manifestations of fatigue, and the ways people and cannot try to overcome circadian rhythms.

    4) Maybe it's about Confucian culture. A Westerner who has long held a senior position as a manager in China sends this view. For US readers: this Hiddink he refers to is a soccer coach whose international reputation would be comparable in American terms to Phil Jackson's or Vince Lombardi's:
    Hiddink's magic transformed South Korea co-hosting the 2002 World Cup by somehow improving its record from five first round knock-outs in previous appearances to fourth place. How? 

    He very quickly changes the attitude of a country, its team and players. When Guus went to Korea there was shyness - the younger players would not talk to the older ones, 

    Hiddink realized while watching former games of the Korean team, that very often, young players were well positioned to score, but passed on to an older player, even though that blew most chances. 

    He made the older players step forward in the circle of players and bow (!!) and ask the younger and more junior players to score themselves, that it was ok not to pass on to a more experienced and senior member of the squad. 

    Hiddink had to break the pattern of Korean seniority, to get them to the semis (where they lost against Germany). He was fully aware of stories circulating at that time, that a Korean Airliner went down in Guam, as the younger pilot did not dare to correct the elder and more senior pilot 
    5) But maybe it's not. A reader sends in this reference:
    About the "cockpit culture" discussion: I'm more interested in individuals, and it seems that a captain, for good or ill, can have an enormous impact on the rest of the crew's performance during operation of the aircraft. As shown in this story.

    There is a lot of commentary on the web about this particular captain's habits and reputation. This article touches on that, as well as the inadvisability of "slam-dunk" approaches.

    As you'll see, the story is about a crash in Minnesota nearly 20 years ago in which blame was fixed firmly on the captain for disregarding his junior officers' warnings that the plane was headed for trouble. With, again, no sign that the captain had ever heard of Confucius. That's it for now, thanks to these and other readers.

  • Back to Asiana 214

    On a light-wind, clear-skies day, a team of professional pilots flew a normally functioning airplane smack into the runway. What have we learned about the cause?

    Let me try to work through a few of the leads, responses, red herrings, and insights that readers have sent in since the crash two weeks ago.

    1) An illuminating video recreation. Here is a useful animation of the difference between a normal "stabilized" approach to San Francisco's runway 28L, and what can be deduced about flight 214's path. 

    A professional pilot added this commentary about the clip:
    This is the best video I've seen so far demonstrating the cost of a flat approach. The only misleading element is the airspeed differential, probably to keep the two images in close proximity. The dramatically slower approach speed of the aircraft would have better highlighted the flaw, not just in altitude above the ground but the equally dangerous decreasing airspeed. The properly positioned "ghost" aircraft would have left the doomed Asiana far behind - running out of altitude, airspeed, and options.

    If your own flight instructor never told you - get on the proper descent profile as soon as possible. Being high or low is a formula for a more challenging landing. Waiting to correct for being high, low, slow, fast, or off center means "going to school" much too late - no rodeos required. With this international route, a good crew has been efficient for the last 12 to 14 hours and nine time zones of circadian shift, so a stabilized approach is the primary consideration during the final minutes of flight.

    2) What's wrong with "Confucius in the cockpit" / "this is how Asians fly" hypotheses. If you've been following this topic, you've seen countless circulated emails from Western pilots alarmed at what they have seen at Asian, and especially Korean and Chinese, flight schools and airlines. I don't have time to fish these all out at the moment. I will say that if you'd like a bracing retort, the place to start is with the "Ask a Korean" site, notably this post (which goes right at Malcolm Gladwell-ism) and then this omnibus followup, including a reply from Gladwell. 

    When I can I will try to give a Solomonic pronouncement on winners and losers in this dispute. Two-sentence preview: Of course Asian-style education and culture can lead to distinctive dynamics in the workplace or in an airline cockpit. But I'm skeptical of moving directly to  civilizational interpretations of events with more modest potential explanations. (OK, a third sentence: At face value it appears that for some reason there was a failure of basic flying skills here, but just the same appeared to be true in the 2009 Colgan crash in Buffalo, whose two home-grown American pilots had no known connection to Confucianism.)

    3) The NASA view of culture in the cockpit. One of the seminal papers in this discipline, by a NASA official back in 2000, is available in a grainy but legible scanned version of the original printed pages, here. Worth reading in light of the Asiana discussions.

    4) 'Children of Magenta.' In the piloting world, this crash has revived a perennial debate on whether pilots are becoming so dependent on automation that they have lost basic "stick and rudder" flying skills. This is the aviation analogy of the old debates about whether 1960s-era calculators were destroying people's ability to do math, or whether today's GPS systems are destroying a sense of direction. The difference is that pilots in most cases still actually have to land the plane. 

    A classic discourse on perils of automation is enjoying new popularity now. It's a 25-minute lecture called "Children of Magenta" (with autopilot and GPS systems, the plane is often programmed to follow a magenta-colored line). Here you go:

    There's more in the queue, but that is what I have time for now. 

    Essay-question topic for bonus credit: precisely because commercial airline accidents have become so rare, with the "normal" causes of accidents being eliminated one by one, the accidents that do occur almost always involve improbable, complex, surprising, or puzzling combinations of circumstances. In much of life, the medical-diagnostic nostrum that "when you hear hoofbeats, think of horses, not zebras" makes sense. Accidents on major airlines these days are nearly all zebras.

  • The Perilous Shoals of Memory: Back to that Dicey NYT Mag Story

    Memory is unreliable. But there's a difference between misremembering and making things up.


    [TL;DR Executive summary of what's below: Patrick Smith, of Ask the Pilot, has seen maintenance records from the flight described in a controversial NYT Magazine piece. These records show no evidence of the "landing gear failure" on which the entire story was based.]

    Let's step away from NSA and Edward Snowden, and even from Pooh and Tigger, for a moment. Earlier I said I would not re-prosecute the case against the recent, fantasized NYT Magazine "Lives" story on what a writer felt when an airliner was (supposedly) about to plunge to its doom. For a refresher on the story, see this -- plus later details from Patrick "Ask the Pilot" Smith, from Clive Irving, from the Economist's Gulliver site, and from John Warner at Inside Higher Ed.

    I have two new reasons to go back into this article. One makes me think more gently of the story and the author. The other, much more harshly.

    1. I'll start with Mr. Nice Guy. The controversy over the story involved whether what was recounted by the author, Noah Gallagher Shannon, could have happened as described. To choose one notorious example:
    The captain came out of the cockpit and stood in the aisle. His cap dangled in one hand. "All electricity will remain off," he said. Something about an open current and preventing a cabin fire. Confused noises spread through the cabin, but no one said a word. "I'll yell the rest of my commands from the cockpit." I could see sweat stains under his arms. "Not going to sugarcoat it," he said. "We're just going to try to land it."
    I was not on that plane, but I can tell you: This. Did. Not. Occur. The dangling cap-in-hand; the sweat stains; the captain coming out of the cockpit and saying he would "yell" his commands; the "not going to sugarcoat it" and "just going to try to land it." No. But the explanation from the magazine's editor, Hugo Lindgren, was that the story aspired to describe what one writer "felt and heard," what he remembered, rather than what anyone else might recall. ("Naturally, not every detail matches everybody else's experience. Surely even people on that plane would remember it differently.")

    At first, I thought this explanation was mainly artful in rendering the story un-falsifiable. I thought about it differently and more sympathetically after hearing, this weekend, a remarkable episode of the the TED Radio Hour, hosted by my former All Things Considered comrade Guy Raz.

    The subject of the show was the unreliability of even the "strongest" and clearest memories. The first segment, with a legal authority named Scott Fraser, went into the very great likelihood of error, false certainty, and wishful filling-in of facts that accompanied most eyewitness recollections. The second segment, with the sociologist and economist  psychologist (who won a Nobel prize in economics) Daniel Kahneman, was even more directly applicable to the NYT Mag case. It dealt with extremely "vivid" memories that might never have happened, or with details very different from what the rememberer became "sure" had occurred. Kahneman described one of the most powerful episodes from his childhood as a Jewish boy in Nazi-occupied France. And then he explained why he couldn't be sure the event had ever happened the way he "recalled." (The final segment was the writer Joshua Foer describing how he won the national trained-memory championship.)

    By chance, immediately after that show I listened to a recent episode of Radiolab, which in one chilling segment reinforced the Memento / Inception / Matrix- like point that we can very rarely be sure what is "really" going on. So maybe that was the story with the New York Times. Something happened on an airplane; a passenger on that plane imagined and then remembered something else; and that something-else is what he, in all sincerity, described to us.

    And then I saw an update on Patrick Smith's site.

    BookWithMap1.jpg2. Smith, an airline pilot, has for many years written the celebrated Ask the Pilot chronicles, along with his excellent new book Cockpit Confidential. (Seriously: If there is anything you don't like, or do like, about air travel, you'll find this book fascinating.) Smith has ample reason to be believed on this topic, and no reason to risk his reputation by inventing facts. And today he says that he has come across maintenance records for that "doomed" flight that made an emergency landing in Philadelphia. What he has found is not convenient for the story Noah Gallagher Shannon told.

    You can get all the details from from Smith, but here's the heart of it. Smith says that the flight crew's logbooks from the flight show this crucial notation:
               "ECAM HYD Y RSVR LO LVL

    Let's break this down. ECAM is the monitoring service for system-functions on an Airbus. HYD stands for the hydraulic systems for controlling the plane; Y stands for "Yellow," to identify one of the three color-coded hydraulic systems. RSVR LO LVL means "reservoir at low level" [see, you too can be a pilot], a warning signal about potential problems with that hydraulic system.

    This was the "emergency." It was not a "problem with the landing gear," which was the entire premise of the NYT Magazine article. (The yellow hydraulic system does not control the plane's landing gear, as the pilots would beyond any question have known -- you're always drilled and tested on which systems control which functions, and what happens if they fail.) There is nothing in this circumstance that would have made a professional flight crew panic or try to prepare passengers for the worst. Moreover, after the plane landed in Philadelphia, a maintenance crew found (according to Smith) that this was purely a faulty-indicator problem. The hydraulic level had been fine all along. Let Smith spell it out:
    An A320 captain I spoke to says that a shut-down of the yellow system would have meant, at worst, a slightly longer-than-normal landing roll (due to loss of the right engine thrust reverser and some of the wing spoiler panels), and, in newer A320s, loss of the nose-gear steering system, requiring a tow to the gate.

    There were enough red flags [in Shannon's story] to begin with, but this puts it over the top, tilting the entire account from one of eye-rolling embellishment toward one of outright fabrication.
    There was never a problem with the landing gear. There was never a reason for the pilots to come out, sweat-drenched, and say brave words to the possibly doomed souls aboard. Based on what Patrick Smith has learned, there was never a reason to shut off all the lights and electricity in the plane. Memory is unreliable. But the Times Magazine story appears to be something more than that, and worth another look by the paper.


Where Time Comes From

The clocks that coordinate your cellphone, GPS, and more


Computer Vision Syndrome and You

Save your eyes. Take breaks.


What Happens in 60 Seconds

Quantifying human activity around the world



From This Author