The Gimli Glider

How general aviation knowledge averted a near disaster

July 1, 2000

When a giant Boeing 767 runs out of fuel at 41,000 feet, hearts beat faster and knuckles turn white. It happened to Air Canada Flight 143, carrying 61 passengers and a crew of eight, at 8:15 p.m. on July 23, 1983. En route from Montreal to Edmonton with an intermediate stop in Ottawa, the flight was piloted by Capt. Robert Pearson and First Officer Maurice Quintal.

Never in the history of commercial aviation have events involved with a scheduled flight gone so wrong in the beginning—and so right in the end.

Bearing the tail number 604, the aircraft designated as Flight 143 was one of the first of a dozen 767s that Air Canada purchased from Boeing. These were—and still are—among the largest and most sophisticated airliners in service. Two turbofan engines provide 96,000 pounds of thrust, a cruising speed of Mach 0.80 or about 480 knots (552 mph), and a 5,000-mile range. Seating seven passengers abreast with two aisles, the fuselage is four and a half feet wider than those of single-aisle jetliners. But it's the computerization that really sets a 767 apart; virtually every instrument and function runs by computer, and all critical systems have backups.

Pilots are a bit double-minded about flying these behemoths: They marvel at the state-of-the-art "gee whiz" computerization but, at least subconsciously, they resent relinquishing so much control to black boxes. Perhaps for this reason, many airline pilots who learned yesteryear's stick-and-rudder flying skills maintain this kind of proficiency by flying gliders or aircraft dwarfed by the jetliners that they fly professionally. In a totally unpredictable, even unimaginable way, Pearson's extensive glider experience would force the world's aviation community to reconsider the value of these "out-of-date" skills.

Most new aircraft have snags, usually minor, but they need attention. The 767 was no exception, and one of its most significant problems was the fuel processor. It had malfunctioned on several 767s newly delivered to various carriers, including Air Canada. The failure of this assembly, and its backup, were central to the problems of Flight 143.

There were human errors too, but a major contributing factor was that Air Canada yielded to insistent government pressure for metrification, and beginning with the new 767s, measurements, including fuel, were part metric and part Imperial (nonmetric). These two factors led to a miscalculation of the fuel on board that certainly would have resulted in a major tragedy if not for the nearly unbelievable airmanship of the flight crew.

Pearson had a chance meeting in the parking lot with Capt. John Weir, who had just flown 604 to Montreal from Edmonton, and they briefly discussed problems relating to the fuel system. There was one fuel gauge for the left tank, one for the right, and a third showed total fuel aboard. Pearson understood Weir to say that he had flown the aircraft with all three gauges inoperative after he had the fuel level in both tanks checked manually. When Pearson entered the cockpit all three fuel gauges were blank, and like Weir, he ordered the fuel levels to be checked manually. He obtained verbal assurance this had been done when he pursued the matter with two mechanics.

The minimum equipment list (MEL), developed cooperatively by the U.S. Federal Aviation Administration (FAA), Air Canada, and Transport Canada, lists circumstances under which an aircraft can be dispatched even when some equipment is inoperable. It clearly states that two of the three fuel gauges must be operable. However, Air Canada's maintenance system had just been reorganized, and maintenance control was given considerable say about the airworthiness of airplanes. For a multitude of reasons, Pearson thought that maintenance control approved the dispatch of 604 despite its inoperative fuel gauges. This assumption, coupled with twice-repeated manual measurements of the fuel in each tank and considerable extra attention given to the matter by qualified mechanics and First Officer Quintal, convinced Pearson to accept and fly the airplane.

The flight to Ottawa and the en route weather on this midsummer evening were near perfect. It was a welcome respite from the confusion involved with the refueling episode in Montreal.

Troubles started shortly after the aircraft departed Ottawa and climbed to a cruising altitude of 41,000 feet. Rapid beeps and near-simultaneous amber warning lights were followed in seconds by displays on computer-like screens in front of both pilots, indicating that one of two fuel pumps in the left wing tank was failing.

Pearson decided to divert to Winnipeg, the closest major airport, 120 miles to the southwest. Quintal explained the emergency to Winnipeg Air Traffic Control (ATC) and was immediately cleared to Winnipeg Airport. When Pearson began a gradual descending turn, more buzzers, lights, and displays told the crew that fuel pressure was failing in the right main tank. In scant minutes, the left engine failed, followed by failure of the right engine. Incredibly, Flight 143 was totally out of fuel. Miscalculations of pounds to kilograms of fuel by mechanics, fuel servicing personnel, and copilot Quintal; the broken fuel processor and inoperable fuel gauges; Pearson's assumption that Air Canada's maintenance control had authorized an exception to the MEL; and, above all, miscommunications were all reasons why. But fixing blame would have to wait.

The lights went out and engine-driven generators stopped delivering power; every electronic device throughout the aircraft simply stopped working, leaving only a 24-volt nicad battery to operate the radio and standby instruments.

The flight management computer (FMC) no longer provided vital information to the pilots, and only four standby instruments remained: a magnetic compass, an artificial horizon, an airspeed indicator, and an altimeter. Worse, the compass was virtually impossible to read because it was so high above the windscreen.

One of the largest and most sophisticated aircraft in the world was gliding at 35,000 feet with fewer instruments than Charles A. Lindbergh's Spirit of St. Louis. The 767 was 65 miles from Winnipeg and 45 miles from a largely abandoned Royal Canadian Air Force training field at Gimli, Manitoba.

No one dreamed that a modern airliner would run completely out of fuel, so gliding performances had not been computed. Pearson knew that the 767's rate of descent with both engines idling was 1,500 feet per minute, but how much should he allow for the parasite drag of two dead engines? What airspeed would give him the greatest range? He settled on just over 200 kt.

ATC in the Western Hemisphere is able to track virtually all aircraft with radar, provided they have transponders. But Flight 143's transponder was not operating because it lacked the necessary power. Fortunately, Winnipeg ATC had a primary radar that could track aircraft without transponders. Without a readable compass, however, Pearson had to estimate heading changes from ATC by reference to the top of the clouds.

Most passenger aircraft manufactured between the mid-1930s and the late 1970s had hydraulically boosted controls, usually powered by electric motors; if the power went out and the hydraulic system failed, it was possible to fly the aircraft manually. But as airplanes and their control surfaces grew larger, this was no longer possible. Boeing and other wide-body airframe manufacturers designed ram air turbines (RATs) that, in case of major power failures, could be extended from the aircraft to provide supplementary power from the flow of the slipstream over a small impeller. Quintal wasted no time in extending 604's RAT from a compartment near the right wheel well.

Power from the RAT was just enough to operate the primary control surfaces—the ailerons, elevator, and rudder—but not enough to operate the flaps or speed brakes. Nor, as time would tell, was there sufficient power to completely lower the landing gear.

To avoid under- or overshooting the runway, it was essential for the crew to know their altitude, rate of descent, and distance from the point of intended landing. The FMC would ordinarily provide this information as a descent profile, but it wasn't functioning and, worse, the vertical speed indicator, which displays the rate of descent, was also not working. By remembering the altitudes and constantly getting distances from ATC, Quintal improvised a makeshift descent profile.

It was Quintal who voiced what both pilots knew deep down: They would never make Winnipeg. Gimli was the only alternative. Gimli's eastern runway was still used by small aircraft, but unknown to either pilot, part of the parallel western runway they selected was being used as a dragstrip by a sports-car racing club.

ATC gave the pilots radar vectors to Gimli, cleared all traffic in the vicinity, and called all available emergency equipment to the scene.

Lowering the landing gear prematurely would have slowed the aircraft and shortened its range even further, so Quintal put off trying to do so until about 90 seconds before touchdown. His first attempt failed because there was insufficient electrical and hydraulic power to extend and lock the gear into position. The emergency and abnormal procedures section of the 767's operating manual on Quintal's lap made no reference to landing gear free-fall, so on his own initiative he pulled an alternate gear-extension switch. The heavy main wheel assembly dropped into the down-and-locked position, but the lighter nose gear only partially deployed.

With flaps and speed brakes extended, the approach speed in a 767 ranges from 125 to 150 kt (depending on weight), and there is usually good braking and reverse thrust to slow the landing roll. But not one of these systems was available to Flight 143. This presented Pearson with a terrible dilemma. To avoid a stall and disastrous undershoot as he approached the beginning of the runway, his airspeed had to be at least 180 kt, but with this much speed and so few options for slowing the landing roll, it was distinctly possible he would overshoot the runway.

Then the final—and by far most challenging—problem occurred: With only about 35 seconds of gliding time left, two miles short of the "runway" threshold, Pearson saw that he was too high—not high enough to circle and land, but still much too high. He instantly executed a sideslip, a maneuver few other pilots might have thought possible in one of the world's largest jetliners. The nose of the aircraft pointed to the right, and its left wing pointed down toward the runway. This maneuver compensated for the lack of flaps and speed brakes; the aircraft lost altitude without gaining unwanted airspeed.

Teenager Art Zuke, riding his bicycle down the dragstrip, saw Flight 143 silently bearing down on him. Along with two of his friends, he barely managed to get out of the way. His comment after the narrow escape was telling: "I saw this thing flying sort of sideways and cockeyed...." Zuke's observation was accurate; he saw the sideslip.

Pearson maintained the slip to within a few feet of the ground, straightened out just before touchdown, and landed 800 feet from the threshold. Tires on the right main gear blew out, and the aircraft skidded down the runway 4,000 feet on its nose and the right engine nacelle, stopping just 500 feet from amazed sports-car racers and their families.

There were no deaths or serious injuries, and within a few days the airplane was flown back to its base for repairs. Known as the "Gimli Glider," it is still in service.

Air Canada at first laid the blame for the accident on Pearson, Quintal, and two mechanics, but public opinion pressured the Canadian government into launching a yearlong inquiry into all the circumstances. As stated in the final report concerning the accident, "There would have been no accident on July 23, 1983...had the fuel processor on aircraft 604 not malfunctioned. The human errors would never have occurred had the fuel processor that operates the fuel gauges been functioning properly. Nor would the shortcomings of Air Canada, Transport Canada, and the fueling companies have been exposed." No punitive action resulted, and blame for the accident was distributed among numerous individuals and organizations.

The Diploma for Outstanding Airmanship, first awarded in 1985 by the world-acclaimed Fédération Aéronautique Internationale, is conferred on only two pilots in the world each year. Capt. Robert Pearson and First Officer Maurice Quintal were the first recipients.


Bruce Frazer, AOPA 3502125 , of Arlington, Virginia, is a former Army aviator and demonstration pilot for Bell Helicopter who is currently writing a book about outstanding airmanship. The "Gimli Glider" story will appear in his book.