AF447 VANITY FAIR REPORT.pdf

8/10/2019 AF447 VANITY FAIR REPORT.pdf

1/23

Twitter

Facebook

Instagram

Tumblr

Pinterest

YouTube

SHOP VFSTORE.COM

POLITICS BUSINESS CULTURE

HOLLYWOOD STYLE

SOCIETY BLOGS

PHOTOS VIDEO

INSIDE THE ISSUE SEARCH

VF AGENDA

business

October 2014

The Human Factor

Airline pilots were once the heroes of the skies. Today, in the quest for safety, airplanes are meant tolargely fly themselves. Which is why the 2009 crash of Air rance light !!", which killed 22# people,remains so perple$ing and significant. William %angewiesche e$plores how a series of small errorsturned a state&of&the&art cockpit into a death trap.

'(William %angewiesche)*+T+ %%-T/AT+ '( ean 1ca3e

EXPLOREAN INTERACTIVE LIST OF THE 2014 NEW ESTABLISHMENT


2/23

PHOTO ILLUSTRATION BY SEAN MCCABE

TROUBLE AHEADInside the automated cockpit of an Airbus A330like the one belonging to Air France that crashed into the

equatorial Atlantic in 2009.

I. Into the Night

On the last day of 1ay in 2009, as night enveloped the airport in /io de 4aneiro, the 256passengers waiting to 3oard a flight to )aris could not have suspected that they would never seedaylight again, or that many would sit strapped to their seats for another two years 3efore 3eing founddead in the darkness, 57,000 feet 3elow the surface of the Atlantic +cean. 'ut that is what happened.Air rance light !!" carried a crew of nine flight attendants and three pilots8their num3ersaugmented 3ecause of duty&time limitations on a ,"00&mile trip that was e$pected to last nearly 55

hours. These were highly trained people, flying an immaculate wide&3odied Air3us A770 for one ofthe premier airlines of the world, an iconic company of which all of rance is proud. :ven today8withthe flight recorders recovered from the sea floor, rench technical reports in hand, and e$haustiveinquests under way in rench courts8it remains almost unimagina3le that the airplane crashed. Asmall glitch took light !!" down, a 3rief loss of airspeed indications8the merest 3lip of aninformation pro3lem during steady straight&and&level flight. t seems a3surd, 3ut the pilots wereoverwhelmed.

To the question of why, the facile answer8that they happened to 3e three unusually incompetentmen8has 3een widely dismissed. +ther answers are more speculative, 3ecause the pilots can no


3/23

longer e$plain themselves and had slid into a state of frantic incoherence 3efore they died. 'ut theirincoherence tells us a lot. t seems to have 3een rooted in the very advances in piloting and aircraftdesign that have improved airline safety over the past !0 years. To put it 3riefly, automation has madeit more and more unlikely that ordinary airline pilots will ever have to face a raw crisis in flight83utalso more and more unlikely that they will 3e a3le to cope with such a crisis if one arises. 1oreover, it

is not clear that there is a way to resolve this parado$. That is why, to many o3servers, the loss of Airrance !!" stands out as the most perple$ing and significant airline accident of modern times.

The crew arrived in /io three days 3efore the accident and stayed at the ofitel hotel on opaca3ana'each. At Air rance, the layover there was considered to 3e especially desira3le. The ;unior co&pilot,)ierre&s introduction, in 599!,not a single A770 in line service had crashed. -p in the cockpit, =u3ois occupied the left seat, thestandard captain>s position. Though he was the )ilot in ommand, and ultimately responsi3le for theflight, he was serving on this run as the )ilot ot lying, handling communications, checklists, and3ackup duties. +ccupying the right seat was the ;unior co&pilot, 'onin, whose turn it was to 3e the)ilot lying8making the takeoff and landing, and managing the automation in cruising flight. 'oninwas a type known as a ompany 'a3y@ he had 3een trained nearly from scratch 3y Air rance andplaced directly into Air3uses at a time when he had only a few hundred flight hours under his 3elt. 'ynow he had accumulated 2,976 hours, 3ut they were of low quality, and his e$perience was minimal,3ecause almost all of his flight time was in fly&3y&wire Air3uses running on autopilot.

'onin switched on the autopilot four minutes after lifting off from /io. This was standard procedure,as is the practice of flying 3y autopilot until ;ust 3efore touchdown. The route of the flight had 3eendecided 3y company dispatchers in rance and entered into the airplane>s flight&managementcomputer at the gate@ it was a direct course up the coast of 'rail, over the city of atal, thennortheast across the Atlantic. The initial cruising altitude was to 3e 7,000 feet. The only weathercomplication was a line of thunderstorms associated with the ntertropical onvergence Bone,spanning the Atlantic ;ust north of the equator. atellite pictures suggested a developing patternperhaps stronger than normal, and with storm clusters too high to top, 3ut with gaps that could 3enegotiated laterally.

or now the night was smooth and clear. Thirty&one minutes after takeoff, the autopilot leveled theairplane at 7,000 feet, nearly as high as the Air3us could fly, given the outside air temperature andthe airplane>s weight? the automatic throttles set the thrust to achieve the selected 0.#2 1ach, whichin thin air translated into an aerodynamic speed of 2#0 knots, and, with the tailwind factored in,delivered a ground speed of !0 miles an hour. 1ore than a thousand parameters were registeredstart to finish, for the entire duration of the trip, 3y the airplane>s data recorder. The cockpit voicerecorder, 3y contrast, was a self&erasing loop, a 3it more than two hours long, restricted 3ecause of

long&standing privacy concerns 3y pilots. As a result, the voice recording opened on the scene twohours and five minutes 3efore the end, or one hour and forty minutes into the flight.


4/23

It was 9@09 ).1. /io time. aptain =u3ois and the young 'onin had settled in for the ride, and thecockpit was mostly quiet. omeone shuffled papers. omeone ad;usted a seat. At 9@2!, =u3oismentioned that they might have to wait a 3it longer for dinner, and 'onin replied affa3ly that he toowas getting hungry. Though they had not previously 3een acquaintances, the two men addressed eachother using the informal Ctu,D a mannerism that has 3ecome de rigueur among Air rance pilots. 'utas su3sequent e$changes would demonstrate, 'onin was almost too deferential, and perhaps tooaware of rank.

A flight attendant entered the cockpit to deliver the meal. he said, CAll is wellED

'onin answered 3rightly, CTutti va 3eneFD

=u3ois said nothing. Apparently he was wearing headphones and listening to opera on a porta3ledevice. Addressing him, the flight attendant said, CAnd you tooE All is wellED

=u3ois said, C*uhED

CAll is wellE o coffee, no teaED

CAll is well,D he said.

=u3ois handed his porta3le device to 'onin, urging him to listen to the opera piece. 'onin did notsay, CThank you, no, we>re on autopilot, 3ut >m supposed to 3e the )ilot lying,D or CThank you, no,

>m not interested in your girlfriend>s music.D *e put on the headset, listened for a few minutes, andsaid, CAll that>s missing is the whiskeyFD

That was the end of the opera. =u3ois indicated a line on an electronic map and said, Ct>s theequator.D

C+.G.D

C(ou understood, suppose.D

'onin did not say, C%ook, aptain =u3ois, >ve already flown five rotations to outh America.D *esaid, C figured . . . D

=u3ois said, C like to feel where we>re going.D

'onin agreed. *e said, C(eah.D

A weather te$t came in from the dispatchers in )aris, accompanied 3y a depiction of the developingline of thunderstorms ahead. either pilot made mention of it, 3ut later comments hint that 'oninwas growing nervous. =u3ois then sowed confusion 3y answering an air&traffic controller>s call toanother Air rance flight and insisting on it despite 'onin>s weak suggestions that he had gotten thecall sign wrong. After a few minutes the controller gracefully sorted out the tangle and gave light !!"a frequency change. imilar confusions arose over required reporting points and frequencies ahead,


5/23

3ut 'onin did not intervene. onversation in the cockpit was desultory, generally a3out flightplanning, sometimes not. The airplane sailed over the port city of atal and headed out to sea.

=u3ois said, CWe were not hassled 3y thunderstorms, huhED This might have 3een an opportunity for'onin to e$press his uncertainty a3out the weather ahead, 3ut at that moment the cockpit door

opened and a flight attendant walked in, asking that the temperature in the 3aggage hold 3e lowered3ecause she was carrying some meat in her suitcase. 'onin lowered the temperature. ifteen minuteslater a flight attendant called the cockpit on the intercom to report that passengers in the 3ack werecold. 'onin mentioned the meat in the 3aggage hold.

'y 50@70 ).1., the airplane had moved well offshore, and 3eyond view of air&traffic&control radar.=u3ois checked in with 'railian oceanic control, known as Atlantico. *e gave a position report andthe time estimates for two waypoints to come. The controller thanked him and instructed him tomaintain 7,000 feet. 'onin said, C:h, well, there you are.D =u3ois radioed, CWilco.D The controlleranswered, CThank you.D t was the flight>s last ver3al e$change with land.

'onin was an$ious to cross the ntertropical onvergence Bone at a higher altitude in order to stay insmooth air 3y remaining a3ove the clouds if possi3le. *e was distur3ed 3y =u3ois>s acceptance of thealtitude assigned. *e said, CWe won>t delay asking to clim3 nonetheless.D =u3ois answered, C(eah,D3ut did not make the request. As he saw it, there was nothing unusual a3out the onvergence Bonethat night@ they might encounter some tur3ulence during the crossing, 3ut the heavy stuff could 3eavoided 3y using the airplane>s weather radar in a normal manner to igag loosely around the largeststorms. urthermore, there was no reason to 3elieve that 3y flying a 3it higher they would encountersignificantly different weather. inally there was this@ the ne$t&highest standard altitude for theirdirection of flight was 7",000 feet, which was shown on a screen as the current Crecommendedma$imum,D or /: 1AH. This was an altitude where, under current conditions, the performancemargins would 3e tight, 3ecause the airplane would 3e flying at a relatively low airspeed and close to

an aerodynamic stall. tandard procedure at Air rance was to maintain greater margins 3y avoidingflight as high as /: 1AH. 'oth pilots understood this. +ne of the enduring mysteries of Air rance!!" is why 'onin kept wanting to clim3.

All was 3lack outside. 'onin saw the first storm on the radar, perhaps 200 miles in front. *e said, Cowe have a thing straight ahead.D =u3ois 3arely answered. *e said, C(eah, saw that,D and dropped thesu3;ect. A minute later, he commented on the outside air temperature, which was frigid at thataltitude 3ut 52 degrees elsius warmer than standard. 'onin said, C(eah, yeah, still, otherwise we>dhave, we>d have a lot higher cruising altitude.D =u3ois said, CAh yeah . . . D *e was reading a magaine.*e steered the conversation to an article a3out ta$ havens. 'onin tried to match his nonchalance. At50@! he said, CWe>re crossing the equator. =id you feel the 3umpED

C*uhED

C=id you feel the 3umpED

C+h shit, no.D

CWell, there you are.D

There were no 3umps? the night remained smooth as the airplane gradually approached the weather.=u3ois said, CBon,we>ll ;ust take whatever measures are required.D t was the closest he came to

advising 'onin of a plan. 'onin lowered the cockpit lighting and switched on the landing lights toilluminate the outside. They entered a cloud layer. =u3ois answered an intercom call from a flight


6/23

attendant, who told him she was taking the night duty in case he needed anything. *e answered witha rench endearment, C(es, my flea,D and ended the call. Although thunderstorms lay ahead and wereshowing on the radar, no lightning was visi3le. They were in mild tur3ulence, without any need yet todeviate from the straight&line course. 'onin said, Ct would have 3een good to clim3, huhED =u3oissaid, Cf there>s tur3ulence.D *e meant significant tur3ulence, which the record later showed they

never encountered. /eferring to rules associated with distance from potential diversionary airports,=u3ois said, CWe>re entering the :T+) one, the death one,D and 'onin answered, C(eah, e$actly.DThe airplane was 3uilding up a static charge, causing some popping on the radios. 'onin got theimpression that they were flying close to the top of the cloud layer. +nce again he suggested a clim3.CWe try to ask for 7I6 J76,000 feetK nonstandardE We>re really at the limits Jof the layerK. :ven 7I6would 3e good.D =u3ois for once was unam3iguous. *e said, CWe>re going to wait a 3it, see if thispasses.D The ghostly lights of aint :lmo>s fire danced across the windscreen.

With most of the weather still lying ahead and an an$ious ;unior pilot at the controls,

=u3ois decided it was time to get some sleep. The chief rench investigator, Alain 'ouillard, later saidto me, Cf the captain had stayed in position through the ntertropical onvergence Bone, it wouldhave delayed his sleep 3y no more than 5 minutes, and 3ecause of his e$perience, may3e the storywould have ended differently. 'ut do not 3elieve it was fatigue that caused him to leave. t was morelike customary 3ehavior, part of the piloting culture within Air rance. And his leaving was notagainst the rules. till, it is surprising. f you are responsi3le for the outcome, you do not go onvacation during the main event.D

4ust 3efore 55 ).1. /io time, =u3ois 3rightened the cockpit lighting, limiting the view outside, and herang the flight&rest compartment, a small ca3in containing two 3erths ;ust 3ehind the cockpit. Asecond co&pilot had 3een doing there, and he knocked on the wall in response. *e was =avid /o3ert,7", another ompany 'a3y who, however, had more than twice the flight e$perience of 'onin and wasthe senior of the two. /o3ert had graduated from :A, one of the elite Lrandes Mcoles, and hadrecently migrated into the airline>s e$ecutive ranks, where he now had a management ;o3 at theoperations center. *e had opted for this trip in order to maintain his currency as a pilot, and hadflown the out3ound leg from )aris, and had made the landing in /io, his first in three months. Afterhis summons to the cockpit, he took two minutes to arrive.

II. Cockpit Resource Management

In the short history of airline safety, the great turning point occurred in the 590s with theintroduction of ;et airplanes, which were far more relia3le and easy to fly than the comple$ piston&engine 3ehemoths that preceded them. +ver the ne$t two decades, as the glo3al ;et fleet grew, wholecategories of accidents related to mechanical failures and weather were largely engineered away. Thesafety improvement was dramatic. t opened the way to airline travel as we know it today.

'ut 3y the 59"0s, a new reality had come into view. Though the accident rate had 3een reduced, theaccidents that continued to occur were 3eing caused almost entirely 3y pilots8the very people, manyof them still at the controls, who had earned a nearly heroic reputation for having stood in the way ofthe mechanical or weather&related failures of the past. )ilot error had long 3een a recogniedpro3lem, 3ut after the advent of ;ets it was as if an onion had 3een peeled to reveal an une$pectedlyimperfect core. The pro3lem was glo3al. n :urope and the -nited tates, a small num3er of


7/23

specialists 3egan to focus on the question. They were researchers, regulators, accident investigators,test pilots, and engineers. The timing was unfortunate for line pilots, who had 3egun to fight a futilerear&guard action, ongoing today, against an ine$ora3le roll3ack in salaries and status. The roll3ackwas a consequence of the very improvements in technology that had made the airlines safer. implyput, for airline pilots the glory days were num3ered, and however unfortunate that was for them, for

passengers it has turned out to 3e a good thing.n the late 59"0s, a small team of researchers at a AA facility in 1ountain Niew, alifornia, 3egan asystematic assessment of airline&pilot performance. +ne of them was a young research psychologistand private pilot named 4ohn %au3er, who later served for 50 years as a mem3er of the ationalTransportation afety 'oard and went on to run the safety division at Air3us in rance. As part of theAA effort, %au3er spent several years riding in airline cockpits, o3serving the operations and takingnotes. This was at a time when most crews still included a flight engineer, who sat 3ehind the pilotsand operated the airplane>s electrical and mechanical systems. What %au3er found was a culturedominated 3y authoritarian captains, many of them crusty old reactionaries who 3rooked nointerference from their su3ordinates. n those cockpits, co&pilots were lucky if occasionally they were

allowed to fly. %au3er told me a3out one occasion, when he entered a 'oeing "2" cockpit at a gate3efore the captain arrived, and the flight engineer said, C suppose you>ve 3een in a cockpit 3efore.D

CWell, yes.D

C'ut you may not 3e aware that >m the captain>s se$ual adviser.D

CWell, no, didn>t know that.D

C(eah, 3ecause whenever speak up, he says, Of want your fucking advice, >ll ask for it.> D

At )an American World Airways, once the de facto -.. flag carrier, such captains were knownas lipper kippers, a reference to the flying 3oats of the 5970s. AA talked the airline into lendingit a full&motion simulator at the an rancisco airport with which to run an e$periment on 20volunteer 'oeing "!" crews. The scenario involved a routine departure from ew (ork>s GennedyAirport on a transatlantic flight, during which various difficulties would arise, forcing a return. t wasdevised 3y a self&effacing 'ritish physician and pilot named *ugh )atrick /uffell mith, who died afew years later and is revered today for having reformed glo3al airline operations, saving innumera3lelives. 4ohn %au3er was closely involved. The simulator runs were intended to 3e as realistic aspossi3le, including 3ad coffee and interruptions 3y flight attendants.

%au3er told me that at )an Am some of the operations managers 3elieved the scenario was too easy.CThey said, O%ook, these guys have 3een trained. (ou>re not going to see much of interest.> Well, wesaw a lot that was of interest. And it had not so much to do with the pilots> physical a3ility to fly8theirstick&and&rudder skills8or their mastery of emergency procedures. nstead, it had everything to dowith their management of the workload and internal communication. 1aking sure that the flightengineer was doing what a flight engineer needs to 3e doing, that the co&pilot was handling the radios,that the captain was freeing himself to make the right decisions.D

t all depended on the captains. A few were natural team leaders8and their crews acquittedthemselves well. 1ost, however, were lipper kippers, whose crews fell into disarray under pressure


8/23

and made dangerous mistakes. /uffell mith pu3lished the results in 4anuary 59"9, in a seminalpaper, CAA Technical 1emorandum "#!#2.D The gist of it was that teamwork matters far morethan individual piloting skill. This ran counter to long tradition in aviation 3ut corresponded closelywith the findings of another AA group, which made a careful study of recent accidents andconcluded that in almost all cases poor communication in the cockpit was to 3lame.

The airlines proved receptive to the research. n 59"9, AA held a workshop on the su3;ect in anrancisco, attended 3y the heads of training departments from around the world. To descri3e the newapproach, %au3er coined a term that caught on. *e called it ockpit /esource 1anagement, or./.1., an a33reviation since widened to stand for rew /esource 1anagement. The idea was tonurture a less authoritarian cockpit culture8one that included a command hierarchy 3ut encourageda colla3orative approach to flying, in which co&pilots Pnow Cfirst officersDQ routinely handled theairplanes and were e$pected to e$press their opinions and question their captains if they saw mistakes3eing made. or their part, the captains were e$pected to admit to falli3ility, seek advice, delegateroles, and fully communicate their plans and thoughts. )art of the package was a new approach to theuse of simulators, with less effort spent in honing piloting skills and more emphasis placed on

teamwork. This was known as line&oriented flight training. As might 3e e$pected, the new ideas metwith resistance from senior pilots, many of whom dismissed the AA findings as psycho3a33le andderided the early seminars as charm schools. As in the old days, they insisted that their skill andauthority were all that stood in the way of death for the pu3lic. Lradually, however, many of thosepilots retired or were forced to change, and 3y the 5990s 3oth ./.1. and line&oriented flight traininghad 3ecome the glo3al standard, al3eit imperfectly applied.

Though the effect on safety is difficult to quantify, 3ecause these innovations lie insepara3ly amongothers that have helped to improve the record, ./.1. is seen to have 3een so successful that it hasmigrated into other realms, including surgery, where doctors, like pilots, are no longer the little godsthey were 3efore. n aviation, the change has 3een profound. Training has changed, co&pilots have3een empowered, and the importance of airplane&handling skills 3y individual pilots has implicitly3een de&valued. 'ut the most important point as it applies to Air rance !!" is that the very design ofthe Air3us cockpit, like that of every recent 'oeing, is 3ased upon the e$pectation of clearcommunication and good teamwork, and if these are lacking, a crisis can quickly turn catastrophic.

The tenets of ./.1., which emerged from the -nited tates, fit naturally into the cultures of Anglo&a$on countries. Acceptance has 3een more difficult in certain Asian countries, where ./.1. goesagainst the traditions of hierarchy and respect for elders. A notorious case was the 599" crash of aGorean Air 'oeing "!" that hit a hillside on a 3lack night, while on approach to Luam, after avenerated captain descended prematurely and neither the co&pilot nor the flight engineeremphatically raised concerns, though 3oth men knew the captain was getting things wrong. n theimpact 22# people died. imilar social dynamics have 3een implicated in other Asian accidents.

And Air ranceE As ;udged from the cockpit management on display in light !!" 3efore it wentdown, AA>s egalitarian discipline has devolved within the airline into a self&indulgent style of flyingin which co&pilots address the captain using the informal CtuD 3ut some captains feel entitled to dowhatever they like. The sense of entitlement does not occur in a void. t can 3e placed in the conte$t ofa proud country that has 3ecome increasingly insecure. A senior e$ecutive at Air3us mentioned to methat in 'ritain and the -nited tates the elites do not 3ecome airline pilots, whereas in rance, as inless developed countries, they still do. This makes them difficult to manage. 'ernard Biegler, thevisionary rench test pilot and engineer 3ehind the Air3us design, once said to me, Cirst you have tounderstand the mentality.D

said, C=o you really think they are so arrogantED


9/23

*e said, Come, yes. And they have the flaw of 3eing too well paid.D

Co there must 3e no pro3lem in the -nited tates.D

'ut Biegler was serious. *e said, Cecond, the union>s position is that pilots are always perfect.

Working pilots are perfect, and dead pilots are, too.D

n the case of Air rance !!" the union has gone so far as to suggest that it is immoral to 3lame thepilots 3ecause they cannot defend themselves. At the e$treme, a !!" victims> family group has eventaken their side. t is an old pattern, deeply rooted. n 597, when an Air rance crew flew a perfectlygood onstellation into a mountain during a routine descent into ice, Biegler>s father, who was theairline>s managing director, went with the chief pilot to report to the rench prime minister. Theprime minister opened 3y saying, CWhat did your pilot do wrongE,D and the chief pilot answered,C1onsieur, the pilot is never wrong.D

Biegler smiled ironically. *e is so 3lunt that for a while he required police protection. *e was 3uilding

airplanes so docile, he once declared, that even his concierge could fly them. We spoke soon after Airrance !!" had crashed, and 3efore the recorders had 3een recovered. rance is a great aviationnation. And Biegler is a patriot. 'ut he is also a modernist. *e has designed the most advancedairliners ever 3uilt. *is point was that at Air rance the piloting culture has not changed with thetimes.

III. Loss of Control

On the night of 1ay 75, 2009, the pilots of light !!" certainly did not serve their passengers

well. After aptain =u3ois left the cockpit to get some sleep, /o3ert, the senior co&pilot, sat on theleft, serving as the )ilot ot lying. 'onin, on the right, continued to handle the 3asic flying chores.The airplane was on autopilot doing .#2 1ach, progressing toward )aris at 7,000 feet, mushingslightly with its nose two degrees up and its wings meeting the oncoming air at a positive angle ofa3out three degrees8the all&important, lift&producing angle of attack.

As the angle of attack increases, so does lift efficiency83ut only up to the point where the angle3ecomes too steep and the oncoming air can no longer flow smoothly over the tops of the wings. Atthat point, the airplane stalls. The phenomenon is characteristic of all airplanes and has nothing to dowith the engines. When an airplane stalls, it loses lift and its wings 3egin to plow through the sky with

enormous drag, far greater than engine thrust can overcome. The airplane enters a deep, mushing,nose&high descent, often accompanied 3y difficulties in roll control. The only solution is to reduce theangle of attack 3y lowering the nose and diving. This is counter&intuitive 3ut 3asic to flight. Therecovery requires altitude, 3ut in cruise there is plenty of altitude to spare.

As usual with airliners at high altitude, Air rance !!" was flying ;ust shy of a pro3lematicangle of attack. Three degrees higher, at degrees, a warning would have sounded in the cockpit, and degrees higher still, at an angle of attack of a3out 50 degrees, theoretically the airplane would havestalled. The last is theoretical 3ecause in the A770, under an all&encompassing automation regime


10/23

known as ormal %aw, the flight&control system intervenes to protect against the stall@ it lowers thenose and advances the power in a manner that cannot 3e overridden 3y the pilots. uch interventionsare e$tremely rare. )ilots spend their entire careers without e$periencing them8unless somethinggoes really wrong with their ;udgment.

omething went really wrong here, 3ut for now nothing was out of the ordinary. n front of each pilot,

'onin and /o3ert, were two independently sourced flat&screen displays. The easiest for casualo3servers to understand were the navigational displays8moving maps showing heading, course,waypoints, and ground speed, with weather radar superimposed. 'ut the more important were theprimary flight displays, each 3uilt around a sym3olic representation of the airplane in relation to ahorion line8showing pitch Pnose up or downQ and 3ank Pwings level or notQ, along with heading,altitude, airspeed, and clim3 or descent rates. A third, stand3y display showed much the same, thoughin smaller form. t is on the 3asis of such marvels of informational presentation that pilots maintaincontrol while flying 3y hand at night or in clouds, when the actual horion cannot 3e seen.

After =u3ois turned up the cockpit lights, the view outside was 3lack. The airplane entered anothercloud layer and was ;ostled 3y light tur3ulence. n the passenger ca3in the seat&3elt sign was on.

'onin rang the forward flight&attendant station and said, C(es, 1aryline, it>s )ierre up front. %isten, ina3out two minutes we ought to 3e in an area where it will start moving around a 3it more than now.D*e advised the ca3in crew to take their seats and rang off with C>ll call you when we>re out of it.D As ithappened, he never did.

The tur3ulence increased slightly. 'onin kept lamenting the ina3ility to clim3. *e mentioned againthe unusually warm temperature outside@ Ctandard plus 57.D Then he said, CPutain la vache. Putain!DNery roughly this translates into Cucking hell. uckFD There was no particular reason for his out3urst.*e was an$ious. *e said, CWe>re really at the very top of the cloud deck. t>s too 3ad. >m sure thatwith a nonstandard 7I6&0 J76,000 feetK, if we did that, it would 3e good . . . D

/o3ert did not respond. *e was looking at his navigational display, which showed a thunderstormdead ahead. *e said, C(ou want to go a little to the leftED The suggestion was posed as a question.'onin said, C:$cuse meED /o3ert said, C(ou can eventually go a 3it to the left.D This was closer to acommand. 'onin selected a heading 20 degrees to the left, and the airplane dutifully turned. Thee$change was the first step in a confusing shift 3y which 'onin 3egan to acquiesce to /o3ert>sauthority without acceding to it completely.

They entered an area of heavier weather, and the cockpit filled with the muted roar of ice crystalshitting the windscreen. 'onin dialed 3ack the airplane>s speed 3y selecting .#0 1ach. /o3ertshrugged ver3ally. *e said, Ct costs nothing.D The automatic throttles responded 3y reducing thethrust. The angle of attack slightly increased. The tur3ulence was light to occasionally moderate. Thenoise of the ice crystals continued.

-n3eknownst to the pilots, the ice crystals 3egan to accumulate inside the airplane>s three air&pressure pro3es, known as pitot tu3es, which were mounted on the underside of the nose. Theclogging of that particular pro3e design was a known issue on certain Air3us models, and though itoccurred only under rare high&altitude conditions and had never led to an accident, it was consideredto 3e serious enough that Air rance had decided to replace the pro3es with ones of an improveddesign and had sent out an advisory to warn pilots of the pro3lem. The first of the replacement pro3eshad ;ust arrived in )aris and were waiting in a storeroom to 3e installed.

or light !!", it was too late@ the pro3es were quickly clogged. 4ust after 55@50 ).1., as a result of the3lockage, all three of the cockpit>s airspeed indications failed, dropping to impossi3ly low values. Alsoas a result of the 3lockage, the indications of altitude 3lipped down 3y an unimportant 760 feet.either pilot had time to notice these readings 3efore the autopilot, reacting to the loss of valid


11/23

airspeed data, disengaged from the control system and sounded the first of many alarms8anelectronic Ccavalry charge.D or similar reasons, the automatic throttles shifted modes, locking ontothe current thrust, and the fly&3y&wire control system, which needs airspeed data to function at fullcapacity, reconfigured itself from ormal %aw into a reduced regime called Alternate %aw, whicheliminated stall protection and changed the nature of roll control so that in this one sense the A770

now handled like a conventional airplane. All of this was necessary, minimal, and a logical response3y the machine.

o here is the picture at that moment@ the airplane was in steady&state cruise, pointing straight aheadwithout pitching up or down, and with the power set perfectly to deliver a tranquil .#0 1ach. Thetur3ulence was so light that one could have walked the aisles8though perhaps a 3it unsteadily. Asidefrom a minor 3lip in altitude indication, the only significant failure was the indication of airspeed83ut the airspeed itself was unaffected. o crisis e$isted. The episode should have 3een a non&event,and one that would not last long. The airplane was in the control of the pilots, and if they had donenothing, they would have done all they needed to do.

Naturally the pilots were surprised. At first they understood only that the autopilot haddisengaged. %ight tur3ulence tilted the airplane into a gentle 3ank. 'onin reached for the side&stick tohis right, a device similar in appearance to a gaming stick. *e said, C>ve got the controlsF,D and /o3ertanswered, C+.G.D A &chord alert sounded 3ecause the indications of altitude had deviated from theselected 7,000 feet. t is likely that 'onin was gripping his control stick much too hard@ the datarecorder, which measures stick movements, later showed that he was flailing from the start, trying tolevel the wings 3ut using high&amplitude inputs like a panicked driver over&controlling a car. t caused

the airplane to rock left and right. This was possi3ly the result of 'onin>s unfamiliarity with handlingthe Air3us in Alternate %aw, particularly at high altitude, where conventional roll characteristicschange. *ad he 3een more seasoned, he might have loosened his grip83acked off to his fingertips8and settled things down. The record shows that he never did.

'ut worse8far worse8was what 'onin did in the vertical sense@ he pulled the stick 3ack. nitially thismay have 3een a startle response to the false indication of a minor altitude loss. 'ut 'onin didn>t ;ustease the stick 3ack8he hauled it 3ack, three&fourths of the way to the stop, and then he kept onpulling. Alain 'ouillard, the rench investigator, equated the reaction to curling instinctively into afetal position. The airplane responded 3y pitching up into an unsustaina3le clim3, causing its speed toslow and its angle of attack to increase.

i$ seconds after 'onin assumed control, with the &chord altitude alert chiming in the cockpit, a3rief stall warning sounded. t was a loud synthetic male voice. t said TA%% one time. The &chordalert resumed. /o3ert said, CWhat was thatED The airplane answered, TA%% TA%%, and again the &chord sounded. either pilot grasped the message. The angle of attack had increased to a3out degrees, and the wings were still flying well, 3ut it was time to do something a3out the warning. 'oninsaid, CWe don>t have a good indication of . . . speedF,D and /o3ert concurred, saying, CWe>ve lost thespeedsFD

With that realiation8that the airspeed indications had dropped out8the pro3lem should have 3eensolved. Though 'onin had reacted wildly on the controls, the crew had assessed the failure correctlywithin 55 seconds of the onset, a3out as quickly as could 3e e$pected. The nose was 55 degrees up,which was e$cessive at high altitude 3ut not in itself e$treme. The solution was simple, andfundamental to flying. All 'onin had to do was to lower the nose to a normal cruising pitch8a3out to


12/23

the horion8and leave the thrust alone. The airplane would have returned to cruising flight at thesame speed as 3efore, even if that speed could not for the moment 3e known.

'ut 'onin continued to pull 3ack on the stick, ;erkily pitching the nose higher. Was he yearning forthe clear sky he 3elieved was ;ust a3oveE Was he remem3ering an Cunrelia3le airspeedD procedure

that is meant for low altitude, where power is ample and the 3iggest concern is to clim3 away from thegroundE =id he think that the airplane was going too fastE :vidence emerged later that he may have,3ut if so, whyE :ven if he did not hear the stall warning, the nose was up, the availa3le thrust was low,and with or without valid indications, high&speed flight in those conditions was physically impossi3le.A renowned cockpit designer at 'oeing8himself a transport pilot8once said to me, CWe don>t 3elievethere are any 3ad pilots. We 3elieve there are average pilots who have 3ad days.D *e called this aprinciple that underlies 'oeing>s cockpit designs. 'ut if 'onin was an average pilot, what does that saya3out the averageE

At least one answer takes the form of the man on his left. After /o3ert concurred that theairspeed indications had 3een lost, he looked away from the main flight displays, there3y a3andoninghis primary role as the )ilot ot lying, which according to the tenets of ./.1. should have 3een tomonitor 'onin>s actions. nstead he started reading aloud from a message screen that ranks anddisplays certain system conditions, and in some cases provides a33reviated advice on procedures. nthis case the advice was irrelevant to the situation, 3ut it led 'onin to switch off the thrust lock, whichcaused the engines to spool up automatically to full thrust. t was the first of a series of seesaw powerchanges that complicated the picture for the pilots and must have caught the attention of somepassengers.

/o3ert kept reading from the message screen. *e said, CAlternate %aw. )rotections %ost.D This at leastwas relevant. t meant that the wings could stall, and that the warnings had to 3e heeded. t is notclear, however, that /o3ert had processed his own words or that 'onin had heard them.

/o3ert said, CWait, we>re losing . . . D *e stopped. Twenty seconds had passed since the loss ofairspeed indications. They were soaring upward through the thin air at 76,000 feet and 3leeding offspeed. The nose was 52 degrees up.

/o3ert returned to the primary flight displays. *e said, C)ay attention to your speedF )ay attention toyour speedFD 'y this he must have meant the airplane>s pitch, since the airspeed indications remained

o3viously invalid. 'onin may have understood the same, 3ecause he said, C+.G., >m going 3ackdownFD *e lowered the nose, 3ut 3y only half a degree. The airplane continued to clim3.

/o3ert said, C(ou sta3ilieFD

'onin said, C(eahFD

C(ou go 3ack downFD /o3ert pointed to a measure of clim3 rate or altitude. CWe>re clim3ing,according to thisF According to all three, you>re clim3ingF o you go 3ack downFD

C+.G.FD

C(ou>re at . . . Lo 3ack downFD


13/23

This is not the time for a dissertation on the Air3us flight&control system, which is criticied 3y'oeing, 3ut to the e$tent that it em3odies a mistake in design, it is that the pilot>s and co&pilot>s side&sticks are not linked and do not move in unison. This means that when the )ilot lying deflects hisstick, the other stick remains stationary, in the neutral position. f 3oth pilots deflect their sticks at thesame time, a =-A% )-T warning sounds, and the airplane responds 3y splitting the difference. To

keep this from causing a pro3lem in the case of a side&stick ;am, each stick has a priority 3utton thatcuts out the other one and allows for full control. The arrangement relies on clear communication andgood teamwork to function as intended. ndeed, it represents an e$treme case of empowering the co&pilot and accepting ./.1. into a design. 1ore immediately, the lack of linkage did not allow /o3ertto feel 'onin>s flailing.

'onin pushed the stick forward, and the nose pitched down, 3ut a little too quickly for /o3ert>s taste,lightening the load to 0." L>s, a third of the way to weightlessness. /o3ert said, CLentlyFD Apparentlyhe realied only now that the engines had spooled up. *e said, CWhat is thatED

'onin said, CWe>re in climb!D t seems that one of the pilots now pulled the throttles 3ack to idle, and

si$ seconds later the other advanced them again. t is not clear who did what, 3ut it seems likely that'onin opted for idle and /o3ert for thrust. 'onin 3y then had gotten the nose down to a si$&degreepitch, and the clim3 had tapered. Though they remained in an untena3le position, all he had to do waslower the nose another few degrees and they would have 3een 3ack where they started. 'ut 'onin forsome reason did not do it, and /o3ert seemed to run out of ideas. *e kept trying to rouse the captain,=u3ois, 3y repeatedly pushing the call 3utton to the flight&rest compartment, 3ehind the cockpit. *esaid, Cuck, where is heED

Bonin 3egan to pull 3ack on the stick again, raising the nose 57 degrees a3ove the horion. The

angle of attack increased, and three seconds later the airplane 3egan to shake with the onset of a stall.The shaking is known as a 3uffet. t occurs as the flow of air 3oils across the wings. As the stalldevelops more fully, it 3ecomes rough enough in the cockpit to make the instruments hard to read.

arried 3y inertia, the airplane continued to clim3. A flight attendant called onto the intercom,apparently in response to /o3ert, who may unintentionally have rung her while trying to rouse thecaptain. he said, C*elloED As if the 3uffet weren>t enough of an indication, the stall warning eruptedagain, alternating 3etween TA%% TA%% TA%% and a chirping sound. The warnings soundedcontinuously for the ne$t ! seconds.

The flight attendant said, C(esED/o3ert ignored her. *e may have realied that they had stalled, 3ut he did not say, CWe>ve stalled.D To'onin he said, C:specially try to touch the lateral controls as little as possi3le.D This is a minor part ofstall recovery, and nothing compared with lowering the nose.

The flight attendant said, C*elloED

truggling with the controls, and with increasing difficulty keeping the wings level, 'onin said, C>m atT+LA, huhED T+LA is an acronym for ma$imum thrust. t is another minor part of stall recovery,especially at high altitude, near an airplane>s propulsive ceiling, where ma$imum thrust means very

little thrust at all. 'onin kept raising the nose, pulling it as high as 5# degrees.


14/23

/o3ert said, Cuck, is he coming or notED

The flight attendant said, Ct doesn>t answer,D and hung up with a click.

'y then the pitot tu3es had unfroen, and the airspeed indicators were working normally again8

though this would not have 3een o3vious to 'onin or /o3ert, in part 3ecause they had no idea of thespeed that the indications at this point should have shown, and apparently did not have the presenceof mind to e$trapolate from the L.)..&derived ground speed, which had 3een displayed on thenavigational screen all along. or the ne$t 52 seconds, neither pilot spoke. Amid repeated stall alarms,the airplane ran out of the inertial a3ility to clim3, topped a para3olic arc at 7#,000 feet, and starteddown on the far side with its nose up and, out at the wings, an angle of attack as steep as 27 degrees.+ne minute and 5" seconds had passed since the trou3le had started, and that is a very long time. Thedescent rate rapidly grew to 7,900 feet per minute, and as a result, the angle of attack furtherincreased. The 3uffeting grew heavy.

=u3ois finally knocked on the cockpit wall, signaling that he was coming. /o3ert kept urgently ringing

the call 3utton anyway. *e said, C'ut we>ve got the enginesF What the hell is happeningED TA%%.TA%%. TA%%. *e said, C=o you understand what>s happening, or notED

'onin said, Cuck, don>t have control of the airplane anymoreF don>t have control of the airplane atallFD 'ecause the right wing was stalled more deeply than the left, the airplane was rolling in thatdirection.

/o3ert said, Controls to the leftFD -sing the priority 3utton on his side&stick, he assumed control ofthe airplane. *e had it for only a second 3efore 'onin, using his own priority 3utton, and withoutsaying a word, took control 3ack. This left /o3ert with a sense that his side&stick had failed. *e said,Cuck, what>s going onED

'onin said, C have the impression we>re going craily fast.D With the nose up and little thrustavaila3leE *ow could he have 3een so confusedE We do not know.

The cockpit door opened, and =u3ois entered. All was commotion. /ather calmly he asked, CWhat>shappeningED TA%%. TA%%. TA%%. The cockpit was shaking heavily.

/o3ert did not say, CWe lost airspeed indications, and this guy pulled up. We>re in Alternate %aw. Weclim3ed to 7#,000 feet, and now we>re going down.D *e said, C don>t know what>s happeningFD

'onin said, CWe>re losing control of the airplaneFD

The Air3us was passing through the original altitude of 7,000 feet? the nose was 5 degrees up? thedescent rate was 50,000 feet per minute and increasing? the angle of attack, though not indicated inthe cockpit, was an incredi3le !5 degrees? the right wing was down unstoppa3ly 3y 72 degrees? andthe airplane was arcing off course through the 3lackness over the mid&Atlantic.

/o3ert said to =u3ois, CWe completely lost control of the airplane, and we don>t understand anythingFWe tried everythingFD

IV. Flying Robots


15/23

Ro3ert>s confusion was later reflected in the frustration of engineers and air&safety specialistsworldwide. The A770 is a masterpiece of design, and one of the most foolproof airplanes ever 3uilt.*ow could a 3rief airspeed&indication failure in an uncritical phase of the flight have caused these Airrance pilots to get so tangled upE And how could they not have understood that the airplane hadstalledE The roots of the pro3lem seem to lie parado$ically in the very same cockpit designs that havehelped to make the last few generations of airliners e$traordinarily safe and easy to fly.

This is as true for 'oeing as for Air3us, 3ecause, whatever their rivalries and differences, 3othmanufacturers have come to similar cockpit solutions. The first was the elimination of the flight&engineer position, despite loud o3;ections 3y the pilots> unions, which claimed that safety would 3ecompromised. This occurred in the late 59"0s, at the same time that 4ohn %au3er and the AAresearchers were pursuing their systematic studies of flight&crew performance and were coming up

with the idea of rew /esource 1anagement. 'y then the individual aircraft systems8engines, fuel,electronics, pressuriation, hydraulics, and so on8had 3ecome sufficiently self&regulating that therewas no longer a need for a third crew mem3er to control them manually. Air3us was the underdog,hemorrhaging pu3lic funds and making airplanes that did not sell. t decided on a no&compromisegam3le to produce the most technologically advanced airliners that could 3e designed. gnoring theunion clamor, it started 3y imposing a two&person cockpit on its models, kicking off an argumenta3out the value of pilots that still comes into view every time an Air3us crashes. 'oeing, which wasdeveloping the "" and "6" concurrently, took a more polite position, 3ut the writing was on the wall.The 'oeing "7" and =ouglas =&9 had already 3een certified to operate with two&pilot crews, withouta flight engineer a3oard. After a presidential task force in the -nited tates studied the matter andconcluded that a third crew mem3er in the cockpit constituted, if anything, a distraction, the unionsaccepted defeat.

The question was how to design cockpits for the two&pilot crews, particularly in light of advances inmicro&computing power, digital sensing, 3right&screen displays, and new navigational possi3ilitiesthat invited the use of electronic moving maps. The manufacturers scrapped the crowded electro&mechanical panels of the past and, using proof&of&concept work done 3y AA, equipped their newairplanes with CglassD cockpits 3uilt around flat&panel displays. The new displays offered manyadvantages, including the a3ility to de&clutter the cockpit 3y consolidating 3asic flight informationonto a few screens, using improved sym3ols, and 3urying much of the rest83ut in readily availa3leform. %ike ./.1., it was all a3out getting 3etter, more consistent performance from pilots8and it hasdone that.

Automation is an integral part of the package. Autopilots have 3een around since nearly the start ofaviation, and component systems have 3een automated since the 5960s, 3ut in glass&cockpit designs,the automation is centralied and allows the systems to communicate with one another, to act as partsof an integrated whole, and even to decide which information should 3e presented to the pilots, andwhen. At the core are flight&management computers8with keypads mounted on central pedestals8which are largely pre&programmed on the ground according to optimiations decided upon 3y airlinedispatchers, and which guide the airplane>s autopilots through the full comple$ity of each flight. 'ythe mid&59#0s, many such airplanes, 3oth Air3uses and 'oeings, had entered the glo3al fleet, for themost part leaving their pilots to simply o3serve the functioning of the systems. n 59#", Air3us tookthe ne$t step 3y introducing the first fly&3y&wire airliner, the smallish A720, in which computersinterpret the pilots> stick inputs 3efore moving the control surfaces on the wings and tail. :veryAir3us since has 3een the same, and 'oeing has followed suit in its own way.


16/23

These are generally known as Cfourth generationD airplanes? they now constitute nearly half theglo3al fleet. ince their introduction, the accident rate has plummeted to such a degree that someinvestigators at the ational Transportation afety 'oard have recently retired early for lack ofactivity in the field. There is simply no arguing with the success of the automation. The designers3ehind it are among the greatest unheralded heroes of our time. till, accidents continue to happen,and many of them are now caused 3y confusion in the interface 3etween the pilot and a semi&ro3oticmachine. pecialists have sounded the warnings a3out this for years@ automation comple$ity comeswith side effects that are often unintended. +ne of the cautionary voices was that of a 3elovedengineer named :arl Wiener, recently deceased, who taught at the -niversity of 1iami. Wiener isknown for CWiener>s %aws,D a short list that he wrote in the 59#0s. Among them@

Every evice creates its o!n opportunity for human error.

E"otic evices create e"otic problems.

#igital evices tune out small errors !hile creating opportunities for largeerrors.

Invention is the mother of necessity.

$ome problems have no solution.

It takes an airplane to bring out the !orst in a pilot.

%henever you solve a problem& you usually create one. 'ou can only hope thatthe one you create is less critical than the one you eliminate.

'ou can never be too rich or too thin (#uchess of %insor) or too careful about!hat you put into a igital flight*guiance system (%iener).

Wiener pointed out that the effect of automation is to reduce the cockpit workload when the workloadis low and to increase it when the workload is high. adine arter, an industrial engineer at the-niversity of 1ichigan, and one of the pre&eminent researchers in the field, made the same point tome in a different way@ C%ook, as automation level goes up, the help provided goes up, workload is

lowered, and all the e$pected 3enefits are achieved. 'ut then if the automation in some way fails,there is a significant price to pay. We need to think a3out whether there is a level where you getconsidera3le 3enefits from the automation 3ut if something goes wrong the pilot can still handle it.D

arter has 3een questioning this for years and recently participated in a ma;or .A.A. study ofautomation usage, released in the fall of 2057, that came to similar conclusions. The pro3lem is that3eneath the surface simplicity of glass cockpits, and the ease of fly&3y&wire control, the designs are infact 3ewilderingly 3aroque8all the more so 3ecause most functions lie 3eyond view. )ilots can getconfused to an e$tent they never would have in more 3asic airplanes. When mentioned the inherentcomple$ity to =elmar adden, a former chief of cockpit technology at 'oeing, he emphatically deniedthat it posed a pro3lem, as did the engineers spoke to at Air3us. Airplane manufacturers cannotadmit to serious issues with their machines, 3ecause of the lia3ility involved, 3ut did not dou3t theirsincerity. adden did say that once capa3ilities are added to an aircraft system, particularly to theflight&management computer, 3ecause of certification requirements they 3ecome impossi3ly


17/23

e$pensive to remove. And yes, if neither removed nor used, they lurk in the depths unseen. 'ut thatwas as far as he would go.

arter has written e$tensively a3out Cautomation surprises,D often related to control modes that thepilot does not fully understand or that the airplane may have switched into autonomously, perhaps

with an annunciation 3ut without the pilot>s awareness. uch surprises certainly added to theconfusion a3oard Air rance !!". +ne of the more common questions asked in cockpits today isCWhat>s it doing nowED /o3ert>s CWe don>t understand anythingFD was an e$treme version of thesame. arter said, CWe now have this systemic pro3lem with comple$ity, and it does not involve ;ustone manufacturer. could easily list 50 or more incidents from either manufacturer where thepro3lem was related to automation and confusion. omple$ity means you have a large num3er ofsu3components and they interact in sometimes une$pected ways. )ilots don>t know, 3ecause theyhaven>t e$perienced the fringe conditions that are 3uilt into the system. was once in a room with fiveengineers who had 3een involved in 3uilding a particular airplane, and started asking, OWell, howdoes this or that workE> And they could not agree on the answers. o was thinking, f these fiveengineers cannot agree, the poor pilot, if he ever encounters that particular situation . . . well, good

luck.Dn the straight&on automation incidents that concern arter, the pilots overestimate their knowledgeof the aircraft systems, then do something e$pecting a certain result, only to find that the airplanereacts differently and seems to have assumed command. This is far more common than the recordindicates, 3ecause rarely do such surprises lead to accidents, and only in the most serious cases ofaltitude 3usting or in&flight upsets are they necessarily reported. Air rance !!" had an additionalcomponent. The 3lockage of the pitot tu3es led to an old&fashioned indication failure, and theresulting disconnection of the autopilot was an old&fashioned response@ trust the pilots to sort thingsout. There were definitely automation complications in what followed, and to that mi$ one can addthe design decision not to link the two control sticks. 'ut on Air rance !!", the automation pro3lemran still deeper. 'onin and /o3ert were flying a fourth&generation glass&cockpit airplane, and unlikethe pilots who think they know more than they do, these two seemed to fear its comple$ities. TheAir3us was reacting in a conventional manner, 3ut once they ventured 3eyond the routine of normalcruise they did not trust the nature of the machine. t is hard to imagine that this would havehappened under the old lipper kippers, the stick&and&rudder 3oys. 'ut 'onin and /o3ertE t was asif progress had pulled the rug out from 3eneath elementary aeronautical understanding.

V. +he Final #escent

Captain =u3ois entered the cockpit 5 minute and 7# seconds after the pitot tu3es

malfunctioned. t is not known whether he knelt or stood 3ehind 'onin and /o3ert, or sat in the ;umpseat. %ikewise, the conditions in the passenger ca3in are not known. Though the unusual motionsmust have 3een noticed 3y some, and the passengers seated in front may have heard the cockpitalarms, there is no evidence that panic 3roke out, and no screams were recorded.

n the cockpit, the situation was off the scale of test flights. After =u3ois arrived, the stall warningtemporarily stopped, essentially 3ecause the angle of attack was so e$treme that the system re;ectedthe data as invalid. This led to a perverse reversal that lasted nearly to the impact@ each time 'oninhappened to lower the nose, rendering the angle of attack marginally less severe, the stall warningsounded again8a negative reinforcement that may have locked him into his pattern of pitching up,assuming he was hearing the stall warning at all.


18/23

=u3ois pointed to an indication on a flight display. *e said, Co, here, take that, take that.D

/o3ert repeated the order more urgently. CTake that, take thatF 'ut try to take thatFD

The stall warning erupted again. 'onin said, C have a pro3lem8it>s that don>t have a vertical&speed

indication anymoreFD =u3ois merely grunted in response. 'onin said, C have no more displaysFD Thiswas not correct. *e had displays 3ut didn>t 3elieve them. The descent rate was now 5,000 feet perminute.

/o3ert was suffering from the same dis3elief. *e said, CWe don>t have a single valid displayFD

'onin said, C have the impression we>re going craily fastF oE What do you thinkED *e reached forthe speed&3rake lever and pulled it.

/o3ert said, Co. oF A3ove all don>t e$tend the 3rakesFD

CoE +.G.FD The speed 3rakes retracted.

At times 3oth of them were on their side&sticks, countermanding each other on the controls. 'oninsaid, Co, we>re still going downFD

/o3ert said, C%et>s pullFD

or 27 seconds aptain =u3ois had said nothing. /o3ert finally roused him. *e said, CWhat do youthinkE What do you thinkE What do you seeED

=u3ois said, C don>t know. t>s descending.D

t is said in his defense that he faced an indeciphera3le scene, having arrived after the loss of control,3ut his o3server status was actually an advantage. *e knew nothing of the original airspeed&indication failure. ow he had a functional panel, showing low airspeeds, a low ground speed, a nose&high attitude, and a 3ig descent under way. Add to that the repeated stall warnings, the telltale3uffeting, and the difficulty in controlling roll. t might have 3een helpful to have an angle&of&attackdisplay8one capa3le of indicating such e$tremes83ut what else could this 3e 3ut a stallE

'onin had managed to come out of the sustained right 3ank. *e said, CThere you areF There8it>sgood. We>ve come 3ack to wings level8no, it won>t . . . D The airplane was rocking 3etween left& andright&3ank angles up to 5" degrees.

=u3ois said, C%evel the wings. The horion, the stand3y horion.D

Then things got even more confused. /o3ert said, C(our speedF (ou>re clim3ingFD *e pro3a3ly meantthat 'onin was raising the nose, 3ecause the airplane was emphatically not clim3ing. *e said,C=escendF =escend, descend, descendF,D again apparently referring to pitch.

'onin said, C am descendingFD

=u3ois picked up the language. *e said, Co, you>re clim3ing.D

'onin may have realied that the reference was to pitch. *e said, C>m clim3ingE +.G., so we>re goingdown.D


19/23

ommunication in the cockpit was withering. /o3ert said, C+.G., we>re at T+LA.D

'onin asked, CWhat are we nowE n altitude, what do we haveED Apparently he was too 3usy to see forhimself.

=u3ois said, Cuck, it>s not possi3le.D

Cn altitude what do we haveED

/o3ert said, CWhat do you mean Oin altitude>ED

C(eah, yeah, >m descending, noED

C(ou>re descending, yes.D

'onin never got his answer, 3ut the airplane was dropping through 20,000 feet. t rolled into a steep,

!5&degree 3ank to the right. =u3ois said, C*ey, you, you>re in . . . )ut, put the wings levelFD

/o3ert repeated, C)ut the wings levelFD

CThat>s what >m trying to doFD

=u3ois was not happy. *e said, C)ut the wings levelFD

C>m at full left stickFD

/o3ert moved his own side&stick. A synthetic voice said, =-A% )-T.

=u3ois said, CThe rudder.D This did the trick, and the airplane righted. =u3ois said, CWings level. Logently, gentlyFD

n confusion, /o3ert said, CWe>ve lost everything on the left wingF have nothing left thereFD

=u3ois answered, CWhat do you haveE,D then Co, waitFD

Though precise modeling was never pursued, the investigators later estimated that this was the lastmoment, as the airplane dropped through 57,000 feet, when a recovery would theoretically have 3een

possi3le. The maneuver would have required a perfect pilot to lower the nose at least 70 degrees3elow the horion and dive into the descent, accepting a huge altitude loss in order to accelerate to aflying angle of attack, and then rounding out of the dive ;ust a3ove the waves, pulling up withsufficient vigor to keep from e$ceeding the airplane>s speed limit, yet not so violently as to cause astructural failure. There are perhaps a handful of pilots in the world who might have succeeded, 3utthis Air rance crew was not among them. There is an old truth in aviation that the reasons you getinto trou3le 3ecome the reasons you don>t get out of it.


20/23

Bonin said, CWe>re, we>re there, we>re getting to level 500FD %evel 500 is 50,000 feet. t>s astandard call in normal operations. t used to 3e said that 3elow 50,000 you were in Cndian country.Dow it>s said that the cockpit should 3e sterile, meaning there must 3e no distractions./o3ert said, CWaitF 1e, have the, have the controls, meFD *e did not push his priority 3utton, and'onin did not relinquish his stick. The synthetic voice said, =-A% )-T. The airplane>s angle ofattack remained at !5 degrees.

'onin said, CWhat is itE *ow is it that we>re continuing to descend so deeplyED

/o3ert directed aptain =u3ois to the overhead switching panel. *e said, CTry to see what you can dowith your controls up thereF The primaries, etc.D

=u3ois said, Ct won>t do anything.D

'onin said, CWe>re getting to level 500FD our seconds later he said, Cine thousand feetFD *e wasstruggling to keep the wings level.

=u3ois said, C:asy on the rudder.D

/o3ert said, Clim3, clim3, clim3, clim3FD *e meant, )itch upF

'onin said, C'ut >ve 3een at full&3ack stick for a whileFD =-A% )-T.

=u3ois said, Co, no, noF =o not clim3FD *e meant, =o not pitch upF

/o3ert said, Co go downFD =-A% )-T.

'onin said, CLo ahead8you have the controls. We are still in T+LA, eh.D omeone said, CLentlemen .. . D +therwise, for the ne$t 57 seconds none of them spoke. ount it on a clock. /o3ert was doing theflying. The cockpit was lousy with automated warnings.

=u3ois said, CWatch out8you>re pitching up there.D

/o3ert said, C>m pitching upED

C(ou>re pitching up.D

'onin said, CWell, we need toF We are at !,000 feetFD 'ut pitching up is what had gotten them intotrou3le to start with. The ground&pro$imity warning system sounded. A synthetic voice said, G/AT:. )-%% -).

=u3ois said, CLo on, pull.D With that, it seems, he had resigned himself to death.

'onin was younger. *e had a wife in the 3ack and two little children at home. *e assumed control,

saying, C%et>s goF )ull up, pull up, pull upFD


21/23

/o3ert said, Cuck, we>re going to crashF t>s not trueF 'ut what>s happeningED

n sequence the alarms were sounding )-%% -), &chord, TA%%, &chord, )-%% -), )/+/T(/L*T. At the same time either /o3ert or 'onin said, Cuck, we>re dead.D

=u3ois calmly said, CTen degrees pitch.D

Thousand one, thousand two. light !!" then pancaked into the equatorial Atlantic. The time in /iowas 55@5! ).1., ! hours and 5 minutes into the flight, and ! minutes and 20 seconds into the upset.Two years later, when the flight&data recorder was retrieved, it showed that 3y the last moment theairplane had turned 22 degrees off course and was flying due west with its nose 56 degrees up and itswings nearly level? thoroughly stalled, it was progressing at merely 50" knots, 3ut with a descent rate,despite full thrust, of 55,000 feet per minute. The impact was shattering. :veryone a3oard diedinstantly, and the wreckage sank in deep water. n the small de3ris field soon found floating on thesurface lay 0 3odies, including that of aptain 1arc =u3ois.

VI. ,rave Ne! %orl

For commercial&;et designers, there are some immuta3le facts of life. t is crucial that yourairplanes 3e flown safely and as cheaply as possi3le within the constraints of wind and weather. +ncethe questions of aircraft performance and relia3ility have 3een resolved, you are left to face the mostdifficult thing, which is the actions of pilots. There are more than 700,000 commercial&airline pilotsin the world, of every culture. They work for hundreds of airlines in the privacy of cockpits, wheretheir 3ehavior is difficult to monitor. ome of the pilots are super3, 3ut most are average, and a feware simply 3ad. To make matters worse, with the e$ception of the 3est, all of them think they are3etter than they are. Air3us has made e$tensive studies that show this to 3e true. The pro3lem in thereal world is that the pilots who crash your airplanes or simply 3urn too much fuel are difficult to spotin the crowd. A 'oeing engineer gave me his perspective on this. *e said, C%ook, pilots are like otherpeople. ome are heroic under pressure, and some duck and run. :ither way, it>s hard to tell inadvance. (ou almost need a war to find out.D 'ut of course you can>t have a war to find out. nstead,what you do is try to insert your thinking into the cockpit.

irst, you put the lipper kipper out to pasture, 3ecause he has the unilateral power to screw thingsup. (ou replace him with a teamwork concept8call it rew /esource 1anagement8that encourageschecks and 3alances and requires pilots to take turns at flying. ow it takes two to screw things up.

e$t you automate the component systems so they require minimal human intervention, and youintegrate them into a self&monitoring ro3otic whole. (ou throw in 3uckets of redundancy. (ou addflightmanagement computers into which flight paths can 3e programmed on the ground, and you linkthem to autopilots capa3le of handling the airplane from the takeoff through the rollout after landing.(ou design deeply considered minimalistic cockpits that encourage teamwork 3y their very nature,offer e$cellent ergonomics, and are 3uilt around displays that avoid showing e$traneous information3ut provide alerts and status reports when the systems sense they are necessary. inally, you add fly&3y&wire control. At that point, after years of work and 3illions of dollars in development costs, youhave arrived in the present time. As intended, the autonomy of pilots has 3een severely restricted, 3utthe new airplanes deliver smoother, more accurate, and more efficient rides8and safer ones too.

t is natural that some pilots o3;ect. This appears to 3e primarily a cultural and generational matter.n hina, for instance, the crews don>t care. n fact, they like their automation and rely on it willingly.'y contrast, an Air3us man told me a3out an encounter 3etween a 'ritish pilot and his superior at a


22/23

1iddle :astern airline, in which the pilot complained that automation had taken the fun out of life,and the superior answered, to paraphrase, C*ey asshole, if you want to have fun, go sail a 3oat. (ou flywith automation or find some other ;o3.D

*e kept his ;o3. n professional flying, a historic shift has occurred. n the privacy of the cockpit and

3eyond pu3lic view, pilots have 3een relegated to mundane roles as system managers, e$pected tomonitor the computers and sometimes to enter data via key3oards, 3ut to keep their hands off thecontrols, and to intervene only in the rare event of a failure. As a result, the routine performance ofinadequate pilots has 3een elevated to that of average pilots, and average pilots don>t count for much.f you are 3uilding an airliner and selling it glo3ally, this turns out to 3e a good thing. ince the 59#0s,when the shift 3egan, the safety record has improved fivefold, to the current one fatal accident forevery five million departures. o one can rationally advocate a return to the glamour of the past.

Nonetheless there are worries even among the people who invented the future. 'oeing>s

=elmar adden e$plained, CWe say, OWell, >m going to cover the 9# percent of situations canpredict, and the pilots will have to cover the 2 percent can>t predict.> This poses a significantpro3lem. >m going to have them do something only 2 percent of the time. %ook at the 3urden thatplaces on them. irst they have to recognie that it>s time to intervene, when 9# percent of the timethey>re not intervening. Then they>re e$pected to handle the 2 percent we couldn>t predict. What>s thedataE *ow are we going to provide the trainingE *ow are we going to provide the supplementaryinformation that will help them make the decisionsE There is no easy answer. rom the design pointof view, we really worry a3out the tasks we ask them to do ;ust occasionally.D

said, C%ike fly the airplaneED

(es, that too. +nce you put pilots on automation, their manual a3ilities degrade and their flight&pathawareness is dulled@ flying 3ecomes a monitoring task, an a3straction on a screen, a mind&num3ingwait for the ne$t hotel. adine arter said that the process is known as de&skilling. t is particularlyacute among long&haul pilots with high seniority, especially those swapping flying duties inaugmented crews. +n Air rance !!", for instance, aptain =u3ois had logged a respecta3le 7!6hours over the previous si$ months 3ut had made merely 5 takeoffs and 5# landings. Allowing agenerous four minutes at the controls for each takeoff and landing, that meant that =u3ois wasdirectly manipulating the side&stick for at most only a3out four hours a year. The num3ers for 'oninwere close to the same, and for /o3ert they were smaller. or all three of them, most of their

e$perience had consisted of sitting in a cockpit seat and watching the machine work.The solution might seem o3vious. 4ohn %au3er told me that with the advent of ./.1. and integratedautomation, in the 59#0s, :arl Wiener went around preaching a3out Cturn&it&off training.D %au3ersaid, C:very few flights, disconnect all that stuff. *and&fly it. ly it like an airplane.D

CWhat happened to that ideaED

C:very3ody said, O(eah. (eah. We gotta do that.> And think for a while may3e they did.D

arter, however, is continuing with variations on the theme. he is trying to come up with improved

interfaces 3etween pilot and machine. n the meantime, she says, at the very least revert to lowerlevels of automation Por ignore itQ when it surprises you.


23/23

n other words, in a crisis, don>t ;ust start reading the automated alerts. The 3est pilots discard theautomation naturally when it 3ecomes unhelpful, and again there appear to 3e some cultural traitsinvolved. imulator studies have shown that rish pilots, for instance, will gleefully throw away theircrutches, while Asian pilots will hang on tightly. t>s o3vious that the rish are right, 3ut in the realworld arter>s advice is hard to sell. The automation is simply too compelling. The operational

3enefits outweigh the costs. The trend is toward more of it, not less. And after throwing away theircrutches, many pilots today would lack the wherewithal to walk.

This is another unintended consequence of designing airplanes that anyone can fly@ anyone cantake you up on the offer. 'eyond the degradation of 3asic skills of people who may once have 3eencompetent pilots, the fourth&generation ;ets have ena3led people who pro3a3ly never had the skills to3egin with and should not have 3een in the cockpit. As a result, the mental makeup of airline pilots

has changed. +n this there is nearly universal agreement8at 'oeing and Air3us, and among accidentinvestigators, regulators, flight&operations managers, instructors, and academics. A different crowd isflying now, and though e$cellent pilots still work the ;o3, on average the knowledge 3ase has 3ecomevery thin.

t seems that we are locked into a spiral in which poor human performance 3egets automation, whichworsens human performance, which 3egets increasing automation. The pattern is common to ourtime 3ut is acute in aviation. Air rance !!" was a case in point. n the aftermath of the accident, thepitot tu3es were replaced on several Air3us models? Air rance commissioned an independent safetyreview that highlighted the arrogance of some of the company>s pilots and suggested reforms? anum3er of e$perts called for angle&of&attack indicators in airliners, while others urged a new

emphasis on high&altitude&stall training, upset recoveries, unusual attitudes, flying in Alternate %aw,and 3asic aeronautical common sense. All of this was fine, 3ut none of it will make much difference.At a time when accidents are e$tremely rare, each one 3ecomes a one&off event, unlikely to 3erepeated in detail. e$t time it will 3e some other airline, some other culture, and some other failure83ut it will almost certainly involve automation and will perple$ us when it occurs. +ver time theautomation will e$pand to handle in&flight failures and emergencies, and as the safety recordimproves, pilots will gradually 3e squeeed from the cockpit altogether. The dynamic has 3ecomeinevita3le. There will still 3e accidents, 3ut at some point we will have only the machines to 3lame.

REGISTRATION ON OR USE OF THIS SITE CONSTITUTES ACCEPTANCE OF OUR USER AGREEMENT(EFFECTVE JANUARY 2, 2014) AND PRIVACY POLICY(EFFECTVE JANUARY 2, 2014).VANITY FAIR COND NAST DIGITAL. YOUR CALIFORNIA PRIVACY RIGHTS. THE MATERIAL ON THIS SITE MAY NOT BE REPRODUCED, DISTRIBUTED, TRANSMITTED, CACHED OR

OTHERWISE USED,EXCEPT WITH THE PRIOR WRITTEN PERMISSION OF COND NAST DIGITAL.

AD CHOICES

Documents

AF447 VANITY FAIR REPORT.pdf