The Crash Detectives

by Christine Negroni

After reading about the flight of Qantas 32—and I promise, I’ll get back to that story shortly—I called Denny Fitch, who told me that all that combined experience would have been an enormous asset for de Crespigny. He should know; he was a hero pilot himself.

  In 1989, Fitch was a passenger aboard United Airlines Flight 232 from Denver to Chicago. One hour into the flight, the engine mounted on the tail of the DC-10 came apart at cruise altitude, and a piece of it sliced through a section at the back of the plane, where three separate hydraulic lines came together. Severing the lines caused the fluid to drain, leaving the pilots with no way to turn, slow, or brake the airplane.

  Al Haynes was in command of the flight, with First Officer William Records and Second Officer Dudley Dvorak. “Somebody has set a bomb off” was Haynes’s first thought when he heard the noise. He was so startled he dropped his coffee.

  The pilots were still trying to figure out what happened when they were interrupted by another crisis. The plane began a descending turn to the right. The plane’s right wing angled sideways at thirty-eight degrees, far steeper than commercial airline passengers are accustomed to. The DC-10 was on its way to rolling over. Haynes closed the throttle to the left engine and slammed open the lever controlling fuel to the right one. The uneven engine power brought the right wing back up. It was an act of instinct and creativity, gleaned from Haynes’s early days flying. He was relying on his knowledge of basic aerodynamics. “You reduce thrust, and that reduces lift,” he explained.

  Right side up again, the plane began to nose up and down in a near-constant cycle of ascents and descents called phugoids, which would last throughout the flight. Still, the successful righting of the airplane allowed Haynes to reframe his thinking about what just moments earlier had seemed an impossible situation. The crew could continue to maneuver the airplane using the only control mechanism available: fuel to the engines. Into this scene of spontaneous piloting walked Denny Fitch.

  Fitch was a United DC-10 training captain, and he’d gone to the cockpit to see if he could help. He found the men focused on the technique Haynes had just thought up. The added complication was that they could not keep the thrust the same on both engines because that made the plane want to roll over.

  “Take one throttle lever in each hand,” Haynes told Fitch. “You can do it much smoother than we can.” So, positioned between Haynes and Record, Fitch did as instructed. “The throttles became my assignment,” Fitch said. None of the four experienced airmen on United 232 had ever tried to fly an airplane this way. No one had ever imagined an airliner losing all flight controls.

  Haynes was the commander of the flight, but in the many talks he has given on this event since 1989, he has acknowledged that the skill, talent, and knowledge of all four combined worked to avert complete disaster. “Why would I know more about getting that airplane on the ground under those conditions than the other three?” he said.

  When the plane slammed down onto the runway at Iowa’s Sioux Gateway Airport three-quarters of an hour later, 185 of the 296 people aboard, including all four of the pilots, survived the crash landing and subsequent fire. One hundred eleven people died, so at best it was a mitigated calamity. It was also a demonstration of the Wright metaphor: a plane otherwise not flyable was wrestled through the air and down onto the runway because control was in the hands of the pilots, and not just any pilots, but a coordinated team whose knowledge, maturity, and experience had a synergetic effect.

  Fitch died in 2012, but when I spoke to him about Qantas 32 in the fall of 2010, he reminded me that, as with United 232, those four men on the flight deck that day represented an abundance of hours at the controls of an A380, so it wasn’t coincidence that things turned out so well. “You cannot have all the experience in your life to equal seventy-six thousand hours,” he said when I told him the combined flight hours of the Qantas crew. The combined flight hours of the pilots on United Flight 232 was even higher: eighty-eight thousand hours. Machines will break, Fitch said, so “at the end of the day it is the human factor that counts.”

  Decision Making

  When the number two engine on de Crespigny’s A380 flew apart, the pressure turbine disk fractured into three crescent-shaped pieces, each roughly six feet long and a foot wide. They flew out of the engine like supersize medieval chakrams, taking the back end of the engine cowling with them. Other shrapnel peppered the fuselage and tore holes in the plane’s left wing, puncturing the fuel tank and severing a number of wire bundles.

  There was no mystery that the problem was with the number two engine, but everything else was uncertain, including why two of the three remaining Rolls-Royce Trent 900 engines were not performing properly. The pilots could not jettison or transfer fuel, and the pumps were not working. Fearful that the last engine might fail, de Crespigny made a request to ATC to climb to ten thousand feet. “I wanted enough altitude so we could glide back to Changi,” he reasoned.

  Nine months earlier, Captain Sullenberger found himself in a similar situation with even less altitude. He was at three thousand feet following takeoff from New York’s LaGuardia Airport when geese flew into the engines, knocking them out. The A320 began a one-thousand-feet-per-minute descent. In his book Highest Duty, Sullenberger said he and Skiles knew in less than a minute that they were not going to get to any of the nearby airports. “We were too low, too slow, too far away and pointed in the wrong direction,” he wrote. The Hudson River was “long enough, wide enough and on that day, smooth enough to land a jetliner.” So he did.

  Worrying about whether the Qantas A380 might also turn into a 550-ton glider, de Crespigny calculated just how much altitude he would need to get back to Changi Airport. He wasn’t thinking about Sullenberger, he was thinking about the astronaut Neil Armstrong, remembering that when Armstrong was a test pilot flying the X-15 at NASA in the 1960s, he helped develop a technique for gliding the rocket-powered plane back to earth once its fuel was spent.

  Armstrong reached altitudes as high as two hundred thousand feet, then glided back to Edwards Air Force Base in California, harnessing gravity to descend in an ever-diminishing spiral. This bit of pioneering aviation was something de Crespigny thought he might need to emulate.

  “I was going to do a slow climb to ten thousand feet, to be in gliding zone using the calculation that I could get thirty miles” at that altitude, he explained. His unilateral decision alarmed the other pilots. They wanted to get the plane lower, not higher. For all his charm, de Crespigny is no pussycat. He is opinionated and sometimes stubborn, and he was perturbed that the other airmen did not agree with him. Still, de Crespigny yielded, realizing, as had Al Haynes, that when flying a plane with such grave damage, no one was an expert and everyone was.

  “The total number of flight hours accumulated by pilots does not predict the quality of their decisions,” Robert Mauro, a professor of psychology at the University of Oregon, wrote in a paper on pilot decision making. “It is experience within a situation that confers expertise.” When everyone is a novice, communicating during decision making becomes even more critical.

  James Reason describes this as “a willingness on the part of subordinates to speak up and a corresponding willingness on the part of the leader to listen.” So concerned was de Crespigny that the three senior pilots on the flight deck not smother the input of the two younger men that he asked the most junior officer, Mark Johnson, to offer his opinions first, followed by his copilot, Matt Hicks.

  Using past experience to guide a decision is called associative decision making. And while it can be a fast and effective method, the danger, according to Mauro, is that past experience may not be helpful in “unstable environments or ambiguous situations.” Worse still is applying a reflective by-the-book response when creativity or innovation is needed.

  As the crew of Qantas 32 flew in circles a mile and a half above the sea, they were consumed by the checklists that were constantly being generated by a computerized airplane trying to diagnose itself and guid
e the pilots through possible remedies. Fuel was draining overboard from the hole in the left wing tank, and this created several fuel imbalances. When the checklist for wing imbalance appeared, it called for the pilots to open the valves to send fuel from the good tank to the one that had been breached.

  “Should we be transferring fuel out of the good right wing into the leaking left wing?” de Crespigny asked his crew. “No,” they replied. Many airlines expect pilots to follow standard procedures strictly. Determining when to follow and when to ignore them requires knowledge, experience, logic, mindfulness, communication, and strength, but decision-making strategies are still evolving.

  In The Pilot’s Handbook of Aeronautical Knowledge, the FAA uses the mnemonic 3P, for perceive, process, and perform, to help pilots remember what steps ought to precede a decision. At Lufthansa, the cadets learn a different acronym, FORDEC, for facts, options, risks, decisions, execute, and check. That final C could also stand for circle back because the big lesson for pilots is that a decision isn’t made and done; it’s an ongoing cycle.

  Captain Norhisham opted to keep the autothrottles engaged while maneuvering MH-124 back to Perth, but he had to revise that plan because of the constant revving and powering back of the engines. That’s one example of reviewing a decision. On Air Canada Flight 143, Pearson and Quintal wanted to land at Winnipeg Airport because emergency equipment would be available and big-city hospitals were nearby. But it was too far away. They considered ditching the plane in Lake Winnipeg, but as Quintal continued updating his distance calculations, he realized they could glide to Gimli. The constant revising of the plan continued, leading to the innovative piloting that has made Air Canada Flight 143 one of aviation’s most talked about recoveries.

  Airmanship

  The day James Reason gave his presentation on heroic recoveries to the attendees of the conference on risk in 2009, he shared the stage with forty-five-year-old British Airways captain Peter Burkill. Burkill’s crash landing in London the year before was the darkest swan among the flock because the 3Ps, FORDEC, CRM—all those intended-to-be-helpful alphabet formulas—were irrelevant. Burkill and First Officer John Coward were faced with a failure so sudden they had only seconds to react.

  The pilots and the relief first officer, Conor Magenis, were at the tail end of the ten-hour flight of British Airways 38 from Beijing. As the plane flew over the outskirts of London, Burkill had no worry bigger than whether the gate would be available when they arrived at the airport. The captain could see where they would touch down, on runway 27L, off to the west, on the other side of the borough of Hounslow. Less than a minute before landing, Coward, who was flying the leg, said suddenly, “I can’t get any power.”

  “I remember looking at his hands on the throttles, and I could see the demand: the autothrottle was fully forward,” Burkill told me. He was still processing what was going on when Magenis chimed in from behind them, saying it looked like a double engine failure.

  “I remember every second of that event. It seemed like the event was three minutes long,” Burkill said. In truth, the time that passed from Coward’s stunning discovery to the plane hitting the ground was just thirty seconds.

  After realizing they would have to make a landing without power, Burkill first decided to leave Coward flying the plane while he concentrated on their options. Ahead loomed frightening obstacles: factories, multistory residences, the Hatton Cross tube station, and a gas station, all of which they would have to fly over to reach the airport. The airfield was surrounded by a high perimeter fence, and on the other side of that, a nine-foot-tall lineup of antennas and airport lights would block a too-low approach.

  Burkill eyed the gauges showing ten tons of fuel still in the tanks and worried about fire. Again he checked the throttles, but there was no improvement. Then the first warning horn sounded in the cockpit. The autopilot had kept the flight on the approach path, but the lack of fuel to the engines was causing the plane to slow. The yoke started to vibrate in Coward’s hands, and an audible airspeed warning sounded. The rate of descent increased until it was more than double the normal seven hundred feet per minute.

  “I knew what it was supposed to feel like, and it was not this,” Burkill said. The only way to keep the plane from slamming into the football-size two-story building dead ahead was to reduce the airplane’s drag. The landing gear, which had been lowered before the crisis began, was slowing them down, but there wasn’t enough time to retract it. Burkill thought it might also help absorb the impact of the inevitable crash landing. He kept thinking.

  Before the trouble began, in preparation for the approach, Coward had asked Burkill to set the flaps to thirty degrees. This makes the wing more comma shaped so the plane can fly at slower speeds. Burkill now considered whether slightly flattening the curved surfaces might be enough to keep the plane flying. He reached over to the flaps lever and, after a moment’s hesitation, moved it from thirty to twenty-five degrees. He did not consult Coward; there just wasn’t time. The effect was immediate: the descent slowed.

  Gus Macmillan, a musician from Melbourne, was in a window seat just behind the wing on the right. “I remember thinking, ‘We only just cleared that fence’ as the grass of the runway unfolded beneath us,” he told the Australian newspaper The Age. The plane hit the grass field 890 feet short of the runway, at 124 miles per hour.

  Burkill’s decision had given the plane an additional 164 feet in the air and enabled Flight 38 to fly over the ominous white building, pass the gas station, clear the highway, and cross safely over the imposing electrified metal barrier of antennas and runway lights. All 152 people on board survived.

  “I wish I had time to actually communicate” with the others, Burkill told me later. A believer in the benefits of CRM, he says that not only did they face a situation for which they were never trained, but also there was no time to use any of their crisis management tools. It was nothing like the sessions in the simulator with all that time to talk to ATC and the cabin crew, glorious minutes to consider options with others on the flight deck. In a real-life emergency, he had to make everything up as he went along.

  “I’m in this gray area,” he said, “this gray area that no pilot wants to be in, with no checklists for my situation and nothing written down.” In the modern jet age, the loss of all engines is so exceedingly rare that it is not a scenario practiced by pilots in their simulator sessions.

  Without diminishing the horrifying experience, one that leads many hero pilots to wrestle with posttraumatic stress disorder long after the public accolades subside, this lack of guidance is where humans excel.

  Uncertainty and Surprise

  If you are wondering if I’ve forgotten you over the Singapore Strait in a noisy and unstable jumbo jet, flying a horse track holding pattern with the 469 frightened travelers on Qantas Flight 32, I have not. I left you to experience just a portion of the hour and forty-five minutes during which the pilots and their passengers flew on, uncertain of their fate.

  One of the mighty plane’s four engines was out, and two others were degraded. Imagine Second Officer Mark Johnson walking down the aisle of the airplane, straining to assess the damage by looking out the passengers’ windows.

  Perhaps, like me, you wonder why the pilots didn’t just put that airplane back on the ground ASAP? This was a subject of discussion in the cockpit.

  De Crespigny said, “We reconsidered this option every fifteen minutes in the air. This was not the time to panic and make irrational decisions. We were on a fact-finding journey; we had to understand how much of the A380 we had left before we could hope to land.”

  Yes, flying was a hazard. Still, before committing to landing a plane that was still too heavy because of all the fuel loaded for the flight to Sydney and only 65 percent of its braking power, de Crespigny had some questions he wanted answered. Could any of their many problems be fixed? How would those they could not solve affect their ability to land? They could not know the answers. Too much was w
rong with the airplane.

  Nearly an hour into the flight, Dave Evans and Harry Wubben set about calculating how much pavement the plane would need to stop. The longest runway at Changi Airport was 13,000 feet. Evans and Wubben calculated that the plane would need 12,700 feet of it, more than twice that for a normal landing. It was achievable, albeit with a very tight margin. That was good news.

  When the wheels of the giant A380 touched the ground, de Crespigny jammed both feet on the brakes, and miraculously the plane slowed, stopping within the distance calculated by Evans and Wubben. Yet the drama was still a long way from over.

  Fuel continued to dump out of the wing, but now it was pooling within range of the brakes that had just done all that stopping and were heated to sixteen hundred degrees Fahrenheit. The fire trucks could not approach the aircraft until the engines were powered down, but when the crew switched them off, the plane was thrown into darkness. Nine of the ten cockpit display screens had failed. Six of the seven radios were dead, and the number one engine kept on spinning.

  For nearly an hour, everyone sat in the dark and sweltering airplane, while the crew worried about the ignition of the fuel and the threat posed by an engine that still whined as if it were in the air. Dave Evans told the Royal Aeronautical Society, “We’ve got a situation where there is fuel, hot brakes, and an engine that we can’t shut down. And really the safest place was on board the aircraft.”

  It took three hours of dousing the engine with water and foam to finally get it to stop turning.

  As with the other dramatic events described in this chapter, the pilots knew the uncertainty would come to an end—in hours for Qantas Flight 32, forty-five minutes for United Flight 232, half an hour for Malaysia Flight 124, a quarter of an hour for Air Canada Flight 143, and less than a minute for British Airways Flight 38. They did not know what the end would look like. It’s easy to forget that.

  All the pilots I’ve written about here were deliberate in their decisions, and they struggled to improve the odds weighing heavily against them—and they all had to contend with an eleventh-hour disruption. It was an engine that would not shut down and the ongoing risk of fire on Qantas 32; and it was a wind shear alarm that rattled Norhisham and Foong as the Malaysia Airlines 777 approached Perth. With United Flight 232, as the DC-10 approached Sioux Gateway Airport, another of the phugoids sent the jet plummeting just three hundred feet above the runway.