Gimli Glider: The 22,300 kg That Arrived as 22,300 lb

July 23, 1983. Air Canada Flight 143, a Boeing 767 four months old, climbs out of Ottawa bound for Edmonton. At 41,000 feet over the Manitoba lakes, both engines quit within minutes of each other — not from mechanical failure, but from fuel starvation. The aircraft was carrying exactly what the paperwork said: 22,300. The paperwork said kilograms. The tanks held pounds. Sixty-nine people were now riding a 132-ton airliner that had become a glider, with seventeen minutes to find somewhere to put it down. The place they found was a decommissioned air base that had been turned into a drag-racing strip, packed with families on a summer afternoon.

~15 min read  ·  Published September 18, 2026

By EnginStack Editorial Team | Verified against the Lockwood Board of Inquiry report and FAA accident archive About →

Most unit-conversion disasters are easy to summarize. Mars Climate Orbiter: pounds-force where newtons were expected. Korean Air 6316: meters read as feet. The Gimli Glider is different, and it is different in a way that makes it the purest unit error of them all. Nobody mistranslated anything. The arithmetic was checked twice and came out clean. The aircraft left the ground with a number on the fuel slip that matched the flight plan exactly — 22,300. The disaster was that the plan wanted 22,300 kilograms, and the slip meant 22,300 pounds. The number was identical. Only the unit was wrong, and a unit is not a small thing. It is the whole meaning of the number.

This is the accident that turns the abstract phrase "metric conversion" into a physical event. A brand-new airliner, the first metric aircraft in an imperial fleet, ran dry at 41,000 feet because a single fuel-density value — 1.77 lb/L where 0.803 kg/L was required — was taken from the wrong side of the metric-imperial divide. To understand what happened you have to understand three things: the coincidence that hid the error, the density number that carried it, and the country-sized transition that made the whole fleet unsafe. None of them is complicated. Each one, on its own, was survivable. Together they dropped 69 people onto a drag strip.

Quick answer

Air Canada Flight 143 was required to carry 22,300 kg of fuel, but a fuel-density value of 1.77 lb/L was mistakenly treated as kilograms per litre. The aircraft therefore departed with about 22,300 lb (10,100 kg) of fuel — less than half the required mass. When it later ran out of fuel at 41,000 feet, the crew glided it to the former Gimli air base.

The 22,300 Coincidence That Hid the Error

Here is the detail that turns this from a routine fuel miscalculation into something almost elegant, and it is confirmed in the official inquiry. The flight plan for the Montreal–Ottawa–Edmonton run required 22,300 kilograms of fuel. Because of the conversion error, the aircraft was loaded with 22,300 pounds. When the captain and first officer cross-checked the arithmetic on the ground, the figure on their scratchpad matched the target perfectly. The number was right. The unit was wrong. And because the number looked right, no alarm went off.

That coincidence is why the error survived two independent checks by the flight crew, and why it survived a second fuel measurement at the Ottawa stopover. A number that agrees with the plan is the most dangerous kind of error there is — it does not announce itself. If the aircraft had been loaded with 10,100 kg instead of the 22,300 kg it believed it had, someone would have noticed. Instead the error produced a number that was supposed to be there, and the unit difference — invisible on a handwritten slip — did the rest of the damage silently.

The scale of that damage snaps into focus when you convert it. One kilogram is 2.2046226218 pounds, so 22,300 kg is about 49,160 lb. But the tanks actually held 22,300 lb — which, converted the other way, is 10,100 kg (the factor runs through 0.45359237 kg per pound). The aircraft departed carrying less than half the fuel required to reach Edmonton. The difference between "half full" and "full enough" was not a math error that any calculator would flag. It was the quiet gap between two systems of measurement, and the aircraft was pointed straight down it.

A Metric Aircraft in an Imperial Fleet

The real cause of the Gimli Glider was not one careless person. It was a country in the middle of switching measurement systems, and an airline that had put a metric aircraft into a fleet that was still thinking in pounds. In the early 1980s Canada was several years into a politically fraught conversion to metric: weather had gone Celsius in 1975, road signs to kilometres in 1977, petrol pumps to litres in 1979. Aviation was caught in the gap. Fuel was sold by suppliers in metric litres. But nearly every aircraft Air Canada flew — the DC-9s, the 727s, the 747s, the L-1011s — measured fuel weight in imperial pounds.

The new Boeing 767s were the exception. They were the first aircraft in the fleet calibrated entirely for metric: the computers, the gauges, the load sheets, and the manuals all required fuel to be calculated in kilograms. Air Canada had introduced a metric machine into a non-metric operation without retraining the people who fed it numbers. The Lockwood Board of Inquiry would later identify this as the systemic deficiency at the root of the accident. The error was not that someone forgot how to convert. It was that nobody in the chain had been told the conversion was now their job.

Compounding it was a structural change in the cockpit. The 767 was one of the first widebody airliners certified for a two-pilot crew, eliminating the flight engineer. Historically, the flight engineer was the person who monitored systems, managed the fuel panel, and ran the fuel-burn numbers. When the role disappeared, Air Canada never clearly reassigned those fuel-checking duties to the two remaining pilots. The 767's two-pilot configuration removed a role that had historically handled those aircraft-system and fuel-management duties, but the responsibilities were not fully and clearly reassigned within Air Canada's new operating structure — so the fuel check no longer had a single named owner.

0.803 and 1.77: The Same Fuel in Two Numbers

The error lived inside a single fuel-density value: 1.77 lb/L was used where 0.803 kg/L was required. To work out how much more fuel to load, the crew had to convert the volume already in the tanks into a weight, subtract that from the required 22,300 kg, and convert the remainder back into a volume. That three-step calculation is the entire accident, and it hinged on one number.

The tanks were measured with dripsticks — mechanical sticks that drop from under the wing until a float bobs on the fuel surface — and read 7,682 litres (about 2,029 US gallons). Jet fuel does not weigh a round number per litre. Its density varies with temperature; on that afternoon the correct figure was 0.803 kilograms per litre. The correct calculation was therefore:

7,682 L × 0.803 kg/L = 6,169 kg already on board
22,300 kg − 6,169 kg = 16,131 kg of fuel still to add
16,131 kg ÷ 0.803 kg/L = 20,088 litres to load

But the ground crew used a different number, one that was perfectly correct for a different unit: 1.77 pounds per litre. That is the imperial specific weight of jet fuel — the figure every other aircraft in the fleet used. Applied to the same calculation, it produced:

7,682 L × 1.77 = 13,597 "kg" (really pounds)
22,300 kg − 13,597 kg = 8,703 kg to add
8,703 kg ÷ 1.77 = 4,917 litres to load

The same 22,300 kg — two different fuels FLIGHT PLAN: load 22,300 kg of fuel one requirement, two possible densities CORRECT — 0.803 kg/L metric density 20,088 L to load 22,300 kg on board ≈ 49,160 lb — enough fuel WRONG — 1.77 lb/L imperial, read as kg/L 4,917 L to load 22,300 lb on board ≈ 10,100 kg — half the fuel ENGINES FLAME OUT at 41,000 ft over Red Lake L × kg/L = kg. L × lb/L = lb — the unit, not the number, tells you which fuel you have.

Why the Unit Would Have Caught It

Dimensional analysis makes the error visible before you touch a calculator. In the correct calculation the units cancel the way they should: 7,682 L × 0.803 kg/L = 6,169 kg — litres cancel and kilograms remain, because L × kg/L = kg. In the mistaken calculation they point to the wrong quantity: 7,682 L × 1.77 lb/L = 13,597 lb, and L × lb/L = lb, which can never be kilograms. The unit on the answer does not depend on what you wanted; it depends on what you multiplied in. That is the entire lesson in one line — the unit, not the number, is what tells you whether the tanks hold 22,300 kg of fuel or 22,300 lb of it.

Instead of taking on 20,088 litres, the aircraft took on 4,917. The density number 1.77 was not wrong — it was the weight of a litre of jet fuel in pounds, to two significant figures. The error was treating pounds as if they were kilograms. And because 0.803 kg/L and 1.77 lb/L are describing the same physical fuel (the difference is just 1 kg = 2.2046 lb), the error carried a factor of 2.2 through every line of the arithmetic. Half the fuel, hidden inside a perfectly ordinary-looking density. If you want the SI view of that density, 0.803 kg/L is 803 kg/m³ — the same number a chemical engineer would use in g/cm³ (0.803) or lb/ft³.

Why the Gauges Were Blank: The FQIS Failure

The failed FQIS removed one of the safeguards that could have exposed the incorrect fuel quantity. The 767's fuel-quantity indication system, or FQIS, was a dual-channel processor: two independent channels each computed the fuel load and cross-checked against the other, so that if one faulted the other would take over. On C-GAUN, the aircraft in question, the FQIS had been misbehaving for weeks, with processor faults logged in Edmonton, San Francisco, and again in Edmonton in early July.

The failure chain was a study in how redundancy dies quietly. A technician in Edmonton found a workaround: pulling the circuit breaker on the faulty Channel 2 let Channel 1 take over and restored the gauges. He tagged the breaker "inoperative" and logged it. Later, in Montreal, a second technician entered the cockpit, saw the log entry, and — not understanding the workaround — pushed the breaker back in to run a self-test. At that moment the fuel truck arrived, he left the cockpit to deal with it, and never pulled the breaker out again. With the faulty channel re-engaged, the FQIS failed entirely. Both fuel gauges went blank.

The aircraft was dispatched under a Minimum Equipment List provision that allowed one fuel-gauge channel to be inoperative — provided the fuel load was physically confirmed with measuring sticks before every departure. That confirmation, done by hand, is exactly where the 0.803-versus-1.77 error entered. The one instrument designed to tell the pilots "you have half the fuel you think you have" was dark, and the manual procedure that replaced it was the very calculation that went wrong. At Ottawa the crew re-measured with dripsticks, got 11,430 litres, applied the same wrong density, and entered 20,400 kg into the flight computer — believing it. In reality the tanks held about 9,250 kg. The error had now been made twice, independently, and confirmed.

41,000 Feet of Gliding: The Dead-Stick Landing

Shortly after 8:00 p.m., cruising at 41,000 feet (about 12,500 metres) over Red Lake, Ontario, the cockpit warned of low fuel pressure in the left pump. Then the right. Within minutes the left engine failed, then the right. Without engines there was no generator power, and the glass cockpit went dark — all that remained were standby instruments: a magnetic compass, an artificial horizon, an airspeed indicator, and an altimeter. A 132-ton airliner had become one of the largest aircraft ever flown as a glider.

A small ram-air turbine deployed from the belly, using airflow to spin a generator and supply just enough hydraulic and electrical power to keep the basic controls alive. The crew now faced the purest physics problem in aviation: convert altitude into distance, spend it precisely, and land with nothing left to spare. A clean 767 has a glide ratio of roughly 12:1 — it travels about twelve feet forward for every foot it descends. From 41,000 feet that bought roughly a hundred miles of reach. It was not enough for Winnipeg, the nearest major airport. First Officer Maurice Quintal, a former Canadian Forces pilot, remembered an abandoned air base at Gimli, Manitoba. He gave the captain the heading.

What neither of them knew was that Gimli was no longer an airfield. The former RCAF station had been converted into a drag-racing strip, and that Saturday afternoon it was hosting a community motorsport event — families, cars, and campers lining the tarmac. Coming in high and fast, Captain Robert Pearson — an experienced glider pilot — used a technique borrowed from sailplanes called a forward slip: crossing the controls to drop a wing and apply opposite rudder, deliberately flying sideways through the air to steepen the descent and bleed off speed without gaining any. No airliner had been slipped like that before. The 767 crossed the strip, touched down hard, and the nose gear collapsed. The aircraft slid to a stop. Sixty-one passengers and eight crew walked away. Ten had minor injuries. No one died.

The powerless descent lasted roughly 17 minutes and covered about 50 kilometres, setting a record for the longest glide by a commercial airliner — a record the 767 held until 2001, when another Canadian aircraft, Air Transat Flight 236, glided even farther into the Azores. Investigators later put other crews in a simulator with the same scenario, to see whether the outcome was luck. Every simulated attempt ended in a crash. The landing was not the accident avoided by chance; it was the accident survived by skill, and by a glider pilot's instinct for energy management.

What the Gimli Glider Teaches Engineers

The temptation is to file this under "aviation" and move on. That would miss the point, because the failure here is a unit-conversion failure, and it belongs to every engineer who has ever multiplied a number by a factor and trusted the result. Three lessons survive the accident reports.

The first is that a number and its unit are a single object, and you cannot check one without the other. The flight crew checked the arithmetic twice and it was correct — correct arithmetic applied to the wrong unit is still wrong, and no amount of re-checking the multiplication catches a mistake that lives in the label. This is the sharpest version of the rule every engineer learns the hard way: never hand a number to anyone without its unit attached. A bare "22,300" is not a quantity. It is a quantity waiting to be misread.

The second is that density is a bridge, and a bridge has two ends. The 1.77 that caused the crash was a perfectly good number — for pounds. The 0.803 that would have prevented it was the same fuel, expressed for kilograms. Both describe the same physical jet fuel, and the difference between them is exactly the 2.2046 conversion from pounds to kilograms. The error was not an inaccurate density; it was a density whose unit had been silently assumed. When you convert volume to mass, the factor you multiply by is a density with a unit, and that unit determines whether your answer lands in pounds or kilograms. Skipping the unit is how you load half the fuel.

The third is that system transitions are where these errors breed, and the cure is not better arithmetic — it is redundancy that survives the transition. Canada's switch to metric did not cause the accident; the decision to run metric and imperial aircraft side by side, with a missing flight engineer and no reassigned responsibility for the fuel math, is what caused it. Every layer of protection — the working FQIS, the flight engineer, a properly unit-labelled density — had been allowed to fail or was never assigned, one hole lining up behind another. Engineers call this the Swiss-cheese model. The Gimli Glider is the model's textbook illustration: no single hole was fatal, but enough of them aligned that a brand-new aircraft departed with half its fuel, and 69 people glided to a drag strip because one number was the right number in the wrong unit.

Related Converters

The kg-vs-lb error at the heart of this accident is a mass conversion, but it ran through a density and a volume. These are the tools for working it out precisely:

Related Guides

Frequently Asked Questions

Was the Gimli Glider caused by a unit conversion error?

Yes — it is one of the purest unit errors in aviation history. The aircraft needed 22,300 kg of fuel but was loaded with 22,300 lb (about 10,100 kg), because the crew converted litres to mass using 1.77 lb/L instead of the correct 0.803 kg/L. The number on the paperwork matched the flight plan, so the unit error went unnoticed until both engines quit at 41,000 feet.

How did 22,300 kg become 22,300 lb?

The flight plan required 22,300 kilograms. The ground crew used the imperial fuel density of 1.77 pounds per litre instead of the metric 0.803 kg/L, so the volume-to-mass conversion produced pounds where kilograms were expected. The result, 22,300 lb, matched the required 22,300 kg digit-for-digit, which is why the error survived two checks and a second fuel measurement at Ottawa.

What is the correct density of jet fuel, and what was used instead?

Jet fuel density varies with temperature; on the day it was about 0.803 kg/L (803 kg/m³). The refueller supplied 1.77, which is the weight of a litre of jet fuel in pounds — correct for the airline's imperial fleet, wrong for the metric 767. The two describe the same fuel; the gap between them is the 2.2046 kg-per-pound conversion.

How far did the Gimli Glider glide without engines?

From flameout at 41,000 feet (about 12,500 m) the 767 glided roughly 50 kilometres over about 17 minutes — a then-record for a commercial airliner — using a 12:1 glide ratio and a forward-slip manoeuvre from Captain Pearson's glider experience. It landed at Gimli, a decommissioned air base turned drag strip, with 10 minor injuries and no deaths.

Sources

Primary sources

Secondary references