by Palmer, Bill
In 2005 he was assigned to Air Calédonie Internationale airline to carry out flights on the A330 between Tokyo and Nouméa, New Caledonia (a French owned island 850 miles east of Australia). While in Nouméa, he also flew the small single engine TB10.
In May 2008, he was appointed as a member of a core group at the Technical Flight Crew Division as representative of the Flight Crew hub.
Before the outbound Paris to Rio flight, his last A330 landing was March 9, 2009, nearly three months earlier. He flew the outbound flight to meet the recency of experience requirements to keep his dual A330/A340 rating up to date.
David Robert had flown 39 trips to South America since arriving in the A330/A340 division seven years earlier.
32-year-old Pierre-Cédric Bonin, was the least experienced pilot of the three. He had slightly less than 3,000 hours total time and 800 hours on the A330.
He received his private pilot license and passed his airline transport theory test in 2000, and his professional pilot’s license, multi-engine instrument type rating, and glider pilot’s license in 2001 at age 22.
He was selected by Air France in October 2003 and started training at the Amaury de la Grange flying school in Merville, France.
He received his A320 type rating in 2004, and his airline transport license in August 2007.
He finished his A340 type rating June 2008, and his A330 type rating followed in December 2008.
Pierre-Cédric Bonin had performed five trips to South America since arriving in the A330/A340 division one year earlier.
Nine flight attendants and 216 passengers were also on board, including First Officer Bonin’s wife and an off-duty flight attendant allegedly accompanying Captain Dubois.3
The Layover
After the flight from Paris, the crew had a three day layover in Rio De Janerio. While that may seem like it would be plenty of time to get adequate rest for the return flight, fighting fatigue on international flights that span multiple time zones is a challenge.
Could fatigue have been a factor in the crews inability to figure out what the problem was and solve it? Possibly. A 1997 study concluded that 17 hours awake is equivalent to a .05% blood-alcohol level.4
The flight pushed back for a twelve hour flight at 8pm local time, 11pm in Paris. No matter how you look at it, it was a long flight on the back side of the clock. It is not easy to flip your body clock over by 12 hours to simply sleep all day, in order to work all night. The presence of three pilots on the two-pilot airplane and an on-board rest facility are a testament to that, though not a complete solution.
The French investigative TV program Pièces à Conviction (French for “incriminating evidence"), in an analysis of Air France 447, claimed that Captain Dubois and a female friend and First Officer Bonin and his wife spent their three day layover in Rio together.5 The captain's fatigue state was described by a guide, who accompanied them in a helicopter they chartered a few hours before the fatal flight, as being so tired that he was probably not fit to fly. The program also said that remarks were made that they would be able to rest on the plane.
The French news magazine Le Point released a story on March 15, 2013 based on a previously undisclosed judicial report that included a voice recorder conversation not included in the official accident report. The magazine revealed a comment made by Captain Dubois at 01:04, "I didn’t sleep enough last night. One hour – it’s not enough.”6 It also claimed that the two first officers were also dangerously tired.
In his book Air Crashes and Miracle Landings, Christopher Bartlett states “it has been said that Air France pilots regarded the long flight back to France as a good chance to have a good rest and recuperate after having fun in Rio de Janeiro.”
The above allegations fall short of what I would consider as factual evidence. One must question a tour guide’s qualifications to determine the captain’s fitness to fly. The BEA reports did not comment on the activities of the layover, or the duration of it, concluding that the private lives of the pilots were out of the scope of the investigation. But these publications do raise the legitimate question if fatigue could have been a contributing factor to the crew’s inability to properly diagnose and maintain control of the situation.
About an hour before the autopilot disconnected, the captain offered First Officer Bonin an opportunity to take a nap. One must assume that he had some reason to offer this, such as looking or behaving tired. The captain said, “Try maybe to sleep twenty minutes when he comes back or before if you want.” Bonin turned down the offer. “Oh … that’s kind” he said. “For the moment I don’t feel like it, but if I do feel like it, yeah.”
One might tend to conclude that Bonin clearly said he was not tired. After all, they had only pushed back three hours earlier. But the captain follows up with, “it’ll be a lot for you.” Thus apparently knowing and trying to convince Bonin that he had already had a long day and was probably not well rested.
As the flight progressed toward a line of thunderstorms in the intertropical convergence zone, clear thinking, their skills as pilots, and their understanding of the A330 would soon be required.
Rest Breaks
The question is often asked, why was the captain taking a break then?
For long flights, additional crew members are provided so that no pilot is on flight duty for more than 8 hours. This was the case with AF447, where the flight time was scheduled at 12 hours 45 min. There were two first officers, each fully qualified to act as the pilot in command in the captain’s absence. Obviously, the captain did not suspect that the first officers would become overwhelmed and crash the plane within minutes of his departure.
While qualifications and experience among pilots vary, it is a false assumption to assume that the captain is always the most qualified or experienced. All pilots go through the same certification course on the airplane, and captains and first officers are trained together.
It is not uncommon for a first officer to have more experience on a particular plane than the captain. In fact, that was the case with AF447. First Officer David Robert, who would replace the captain in the left seat when the captain went on break, had 4479 hours in the A330, over 2.5 times as much as the captain, though slightly more than half as much total fight time.
Pierre-Cédric Bonin, the pilot flying in the right seat, at 32 years old was 5 years younger than Robert, and 26 years younger than Captain Dubois. He had 2936 total flight hours and 807 hours on the A330 with a type rating issued six months before the accident. All three pilots had flown the A320 prior to the A330 & A340 for several years, so Airbus flight controls and handling were nothing new to any of them. In fact, their A330 & A340 type ratings courses were differences training from the A320 type rating they each held prior to flying the A330 & A340.
It is often asked, “why was the junior pilot flying the plane?” The fact is that all pilots need to keep their qualifications current (takeoffs and landings) and the actual flying is shared pretty much equally. No pilot would have any business being there if he was not qualified to fly the airplane at any time.
Prior to the captain going on his rest break, he discussed that First Officer Bonin would be doing the landing, and thus was to be the pilot flying at that point. According to Air France procedures at the time, because he was the pilot flying in the right seat, he would assume command in the captain’s absence.
At a press conference on July 29, 2011, the BEA presented the 3rd interim report. It was asked, “You said that the captain had not clearly defined the task-sharing when leaving the cockpit. Did he hand over command of the airplane to the pilot in the right-hand seat? Also, do you understand why the more experienced co-pilot in the left-hand seat did not take back control of the airplane?”
Alain Bouillard, Investigator in Charge of the BEA, responded that on his departure, the captain implicitly designated the pilot flying who was on the right as his relief. “It is always the co-pilot flying who relieves the captain.” 7
First Officer Ro
bert was to take the captain’s place in the left seat. No particular briefing was conducted by the captain to designate who would serve as pilot in command (PIC) as this was apparently implied by the normal procedures. (Air France later revised their policy so that the left seat pilot would be the designated PIC.)
Typically, the pilots on a long international flight with an additional pilot will divide the flight time to achieve equal break times. Often times the third pilot (occupying the cockpit jumpseat for takeoff) will take the first break, followed by the captain, and then the other first officer for the third break. This schedule is not set in stone, but is common, and it appeared to be the case for AF447. The cockpit voice recording only covers the last 2 hours and 5 minutes so we only know for sure who was in what seat at the first break changeover. It is possible that First Officer Robert was in the right seat for takeoff before he took the first break. Even if that were true, it is insignificant.
According to Air France policy, the augmented crew members are present in the cockpit and actively monitor the flight from the departure briefing to FL200 and from the arrival briefing to the gate. Outside of these flight phases, each member of the flight crew must be able to rest for at least an hour and a half continuously during the flight duty time. The captain sets the procedures for each member of the crew taking their rest.
The pilot rest area is a small room located right behind the cockpit. Its forward wall adjoins the cockpit’s aft wall. The rest area has two bunks to allow the resting pilot to sleep (or two resting pilots for long flights with a double crew). A dedicated button on the left side of the cockpit overhead panel rings a call chime in the rest area. (There are other call buttons for other areas of the aircraft as well).
Many A330s have an inter-phone handset in the rest area that allows for voice communication between the cockpit and the rest area. A conversation with the captain is only seconds away at any time. But the inter-phone in the bunk area is optional. Air France did not take this option. The best the pilots can do to communicate is ring the call button from the flight deck, and bang on the wall from the bunk. On the transcript, First Officer Robert was called in the rest area at 01:56. Four seconds later he responded with a knock on the wall (as if to say “I’m up”). But without the inter-phone, there is no way to distinguish between a call to say “your break is up in 10 minutes” and “we have an emergency and need you ASAP.”
It was not just communication between the cockpit and the crew rest area that was difficult that night. Voice and digital communications with air traffic control also suffered.
Chapter 3: Communication Breakdown
As the flight departed the North coast of Brazil there were communications issues that night that contributed to the lengthy delay in both locating any floating wreckage (5 days) and locating the sunken wreckage (2 years).
Two communication technologies are currently in use: HF radio, and modern satellite-delivered digital communications and tracking.
Oceanic flights are often outside the range of radar and line-of-sight VHF radio communications. To provide for traffic separation and communications in these areas, pilots and controllers use HF (shortwave) radio, and a system of position reporting at designated points along their assigned route. Pilots report their time and altitude over each reporting point, the estimated arrival time for the next reporting point, and some additional data. This system had its genesis in the late 1930’s supporting flying boats (large seaplane airliners) across the Atlantic using Morse code. The HF communications now are all by voice but the medium still has limitations as a result of the physics of radio wave propagation.
HF radio uses a frequency range that can bounce off the ionosphere to provide very long range communications. However, with that capability come drawbacks. HF radio can be greatly affected by the sun, upper atmospheric conditions, sunspots, and other “space weather.” Communication can often be difficult, and sometimes nearly impossible. Call quality can range from good to non-existent. Pilots are assigned two frequencies in case the first one is not workable in the aircraft’s location. Because of the difficulty, pilots often talk to a radio operator whose job it is to relay messages between the aircraft and controllers, so that the controllers can spend their time concentrating on managing the air traffic. The process often involves slow clear talking, repeated transmissions, and an occasional message relay from other aircraft to complete the transfer of information.
The lack of real-time position information (like radar) limits the capacity of the ATC system in remote regions. Unlike a radar environment, where airplanes can be accurately controlled within a few miles of each other, separation in oceanic areas relying on position reports is often 80 miles in-trail and 50 or more miles laterally.
The same radio operator can cover huge areas of oceanic airspace with many aircraft, and the same frequency may be in use by pilots and radio operators on both sides of the ocean. Radio frequency congestion is often such that it can take several minutes to get a word in to make a report or request.
The transmissions on HF are often noisy and the frequencies congested; therefore, pilots do not normally monitor it full time. Instead they use a Selective Call system (SELCAL) that allows the radio operator to transmit an aircraft’s unique SELCAL code. When the radio receiver recognizes its SELCAL signal it sounds a call signal in the cockpit. The call notifies the pilots to contact the radio operator for a message.
It just so happened that the night of May 31, 2009 HF communications were particularly poor. When AF447 did not check in with Dakar control, the controller attributed it to the poor state of communications that night and nobody realized the flight was in trouble until hours later.
The most modern airliners, including Air France’s A330s, also use two forms of digital communications for tracking and message relay between the flight and controllers: ADS and CPDLC. Each of these technologies use cheaper and faster VHF radios when within line of site of a ground based station, but switch to satellite communications to relay the messages when out of VHF range. The satellite link takes slightly longer as the connection must be established each time, but it is often quicker than the HF system, and the messages are not subject to poor reception issues.
ADS (Automatic Dependent Surveillance) is a position reporting system. There are several varieties of this system, but the mode in question is known as ADS-C or ADS Contract. This mode requires a logon from the flight deck with the ATC facility. The controllers on the ground set parameters that tell the logged-on system on board the airplane when to make automatic position reports. The reports include the airplane’s position, time over the last reporting point, estimate for the next reporting point, altitude, speed, and a handful of other data automatically gathered from the flight management system. These transmissions are completely transparent to the crew. They have no indications of what is sent or when. Their only interaction with the ADS system is to log on and log off, though the log-off is automatic.
Typically, the system is set to make position reports passing each waypoint along the route, a report about every 15 minutes, as well as warning messages if an airplane’s altitude changes by 200 feet, route by 10 miles, or the absence of an expected position report.
The crew also has an emergency setting which will force the sending of position reports every minute. This would be appropriate if a flight was diverting to an alternate airport due to an on-board emergency as they might not pass any of the planned waypoints along the way. Setting the Emergency ADS to ON, requires making selections in the communications-system menu. It is not a one-button item and not at the top of the list of things to do when control of the airplane is in question. It would be something a crew would get to once their diversion was initiated.
CPDLC (Controller Pilot Data Link Communications) is a text messaging system. With it, pilots and controllers exchange formatted text messages to exchange requests, clearances, and messages. The messages are formatted such that key content can be interfaced with the fli
ght management system so that it can, for example, load a transmitted clearance directly into the system, or remind the crew of instructions that take effect at a given time. Pilots can also send position reports, but if ADS is in use, it is rarely necessary. A typical use of CPDLC en route would be for the crew to request a different altitude, or a lateral offset from their planned route for a weather deviation.
CPDLC also requires a logon, which is done concurrent with the ADS logon. Like ADS, an agreement between the airline and the ATC facility has to be in place before the facility will accept the logon and provide service via this method. After all, the airline is billed for this service, so the system does not accept a logon from just anyone.
Currently, the North Atlantic air traffic control agencies are initiating a program of reduced lateral and in-trail separation on routes dedicated to aircraft with ADS and CPDLC (also referred to collectively as FANS 1/A). This initiative is expected to generate greater efficiency and flexibility for both individual aircraft and the system as a whole.
AF447 was not able to log on with CPDLC or ADS that night. An error in entering the flights data into the ATC computer prevented the system from matching up the identification data transmitted from the airplane with what was erroneously typed in to the computer. So when the crew was approaching the oceanic portion of the flight and tried to log on, the logon was unsuccessful, and the crew reverted to using only HF voice communications.
The night of the accident, AF447 was in radar and VHF radio contact with the Recife controller as they left the coast of Brazil. The controller instructed the crew to contact Atlantico over the INTOL waypoint (the boundary between Recife and Atlantico), and Dakar over TASIL waypoint (the boundary between Atlantico and Dakar) and issued the frequencies for each.