Ask the pilot

So many crashes, so little time. First stop: Toronto, where Air France 358 wasn't the first plane to go barreling off a runway, and might not be the last.

By Patrick Smith

Aug 19, 2005 | Doubtless those clicking in are curious about last Sunday's mysterious 737 crash in Greece, and the one in Venezuela on Aug. 16. You can expect a full report next week. There's plenty to talk about, but for now details are sketchy and conjecture rife. As these things go, it's prudent to hold off until more is known. Until then, it's best to season your intake of news with as much salt as you can possibly tolerate. The usual suspects, from the Associated Press to the network anchormen, have truly outdone themselves over the past several days, propagating the usual myths about cabin pressurization and engine failure. Dispatches from Greece in particular have been depressingly outrageous.

First things first, and back to Toronto we go for Part 2 on the final seconds of Air France flight 358. If lightning, wind shear and flaming engines weren't responsible, then who or what was the culprit? The full official inquiry will, of course, take many months, and if history shows us anything it's the dullness of Occam's proverbial razor when it comes to air crashes. The seemingly obvious explanation is not, often enough, the right one. Disclaimer out of the way, let's go ahead and speculate:

Central to the ongoing investigation is a report that the Airbus A340 overshot the runway's designated touchdown zone by as much as a thousand feet. The touchdown zone encompasses the first 3,000 feet of pavement from the threshold, and Pearson International's runway 24L is 9,000 feet long. (For those of you reading outside the United States, there are 3.28 feet to the meter.) If the claims are accurate, that means the jet did not strike terra firma until it was almost halfway down the strip.

Landings are more scientific than most people realize. On departure, planes are obliged to meet so-called balanced field criteria; a runway must always be long enough to ensure adequate stopping distance should the roll be discontinued at or below takeoff speed. Arrivals work much the same way; a pilot does not eyeball the runway and conclude, "that looks about right," then hit the brakes and hope for the best. Taking weight, wind and weather into account -- including penalties for a surface slickened by ice, snow or rain -- data must show that a plane is able to stop within a maximum of 85 percent of the total available distance. (I'm citing an FAA regulation, but Canadian and European authorities mandate similar, if not identical standards.) At Toronto on Aug. 2, that left a worst-case margin of 1,350 feet. Tracking runway 24L's instrument landing guidance beam, the Airbus should have floated over the threshold no higher than 50 feet. Its tires would meet pavement a few seconds later, well within the touchdown zone, guaranteeing safe stoppage even under wet and windy conditions.

That's textbook. Miss the touchdown zone, and the buffer is out the window.

During severe weather things can get squirrelly. Gusty crosswinds, wind shear, turbulence and air traffic control (ATC) requests will sometimes combine to throw an approach off kilter. The best way -- indeed the right way -- of dealing with an unstable approach is to discontinue it. All frequent fliers have experienced an aborted landing at one time or another, known to crews as a "go-around" when a plane suddenly powers up and climbs away, circling back for a second attempt. In the majority of cases this is an ATC separation issue, but once in a while it's related to weather. A crew finds itself slightly high, fast or otherwise not exactly where it should be. The decision is made to go around.

Mind you, should a plane miss the target and elect to land anyway, that alone doesn't mean it will wind up in a gully in a heap of flames. Even on a relatively short strip like Toronto's 24L, a liberal application of braking and reverse thrust is liable to get the job done. However, handicap a landing with max-limit crosswinds and heavy rain, both of which decrease braking efficiency and increase the likelihood of hydroplaning, and the margin for error is slim. Landing long, at that point, really isn't an option.

So, the unanswerable question at this point is why the Air France captain did not reject the landing and climb away. Possibly, according to one account, a go-around was initiated, albeit after the plane had already touched down. Not an unprecedented maneuver, but with 50 percent of the runway already eaten there might not have been room for stopping or re-accelerating to climb speed. Realizing this, the pilots would have had little choice but to deploy maximum braking and spoilers, throw on full reverse power, and begin praying that it's all a bad dream.

It is impossible to say what the investigation may reveal -- an ongoing emergency of some kind, a mechanical malfunction -- but the available evidence points to an error in judgment. Typical in civil aviation accidents, the official ruling is likely to speak of a central mistake (failure to execute a go-around) compounded by a serial of contributing factors (lousy weather, fatigue), but out of the fine print will emerge two simple and familiar words: pilot error.

Whether they occur during landing or takeoff, runway excursions make up a high percentage of airliner mishaps. The following list is by no means exhaustive:

1982: World Airways flight 30 slides into Boston Harbor after landing long and fast along an icy runway at Logan International. The nose section of the DC-10 splits from the fuselage, and two passengers disappear into the ocean. Their bodies are never found.

1986: After an unstable approach in rainy weather, a Piedmont Airlines 737 rolls off the end of a runway at Charlotte, N.C. Several occupants are seriously injured.

1989: A USAir 737 aborts takeoff at La Guardia and flounders into the East River. The aircraft strikes a wooden pier and breaks apart, killing two.

1999: Eleven people perish when American Airlines flight 1420 speeds off a runway in Little Rock, Ark., after arriving in stormy conditions not unlike those in Toronto two weeks ago.

1999: On landing, a DC-10 skids into a neighborhood along the perimeter of Guatemala City's La Aurora International Airport. Twenty-six are killed.

1992: TWA flight 843, a Lockheed L-1011, rejects takeoff at JFK and bursts into flames off the end of the runway. All 292 passengers and crew survive.

As you can glean from the sample, although some overrun accidents are serious, they're usually of the non-catastrophic variety. With crews doing all they can to decelerate, aircraft tend to be traveling at comparatively low speeds by the time they run out of pavement, resulting in limited damage and reduced likelihood of fire.

Next Page: Pilot error? Or runway failure?

• View blog reactions