Ask the pilot

Don't blame the pilot for the crash of Flight 587. The truth is much more complicated.

Published November 5, 2004 7:30PM (EST)

I never liked the name Airbus, and who would? It's a moniker whose negative connotations are plainly obvious and offensive to the sensibilities of any good air-travel romantic. We're used to it by now, but when the maker's first model, the A300, debuted 30 years ago, we could hardly get over it. Air Bus? Then again, why not? There we were on the cusp of deregulation, civil aviation's crowning thrust into all things dreadful and ordinary. It was tempting to see the name, as the writer James Kaplan puts it, as a "capitulation to the inevitable."

Airbus Industrie, to use its full, letterhead name, was born in 1970, a consortium of aerospace expertise from France, Germany, the U.K. and Spain. Five baseline models have emerged, with a sixth on the way. What it lacks in graceful branding, the line has made up for in technological trailblazing. The Airbus pedigree is marked by high levels of cockpit automation, fly-by-wire control systems, and a reliance on high-strength, lightweight composites. Today all commercial plane makers take advantage of these concepts, but Airbus, more than anybody else, pioneered them.

Development of these advances was sometimes controversial, most memorably after a spate of crashes (in aviation context that means two or three) involving what was, at the time, the world's most high-tech airliner, the Airbus A320. All told, however, the Airbus record is no more or less exemplary than Boeing's, its products no "better" or "worse," in any statistically meaningful way.

In the mid-1980s, the original A300 -- a widebody twin-jet intended for short- to medium-range operations -- underwent significant redesign and was given a new designation, the A300-600. Changes included a two-place digital flight deck, numerous internal systems upgrades, and an all-composite tail. The -600 met with decent success, and among its buyers was American Airlines, which introduced the plane in 1988. It became a staple on the carrier's Latin and South American runs.

Maybe you already know this, because American and its A300s are back in the news again. Last week, the National Transportation Safety Board (NTSB) passed down its rulings on the crash of American Flight 587, the A300 that went down near John F. Kennedy airport three years ago, killing 265 people. It attributes the disaster to a combination of pilot error and faulty design.

The doomed plane was a 13-year-old ship, registered N14053, bound for Santo Domingo on the morning of Nov. 12, 2001. Flight 587 was well known among New York City's Dominican community. In 1996, long before the accident, a merengue star named Kinito Mendez paid a sadly foreboding tribute with his song "El Avion." "How joyful it could be to go on Flight 587," sang Mendez, immortalizing the popular daily nonstop.

It was airborne for less than three minutes before plummeting into the Rockaway section of Queens. Moments after takeoff, the flight encountered the wingtip vortices of a preceding departure, a Japan Airlines 747. As it roiled in the 747's wake -- an event neither uncommon or particularly hazardous for an A300 -- something happened. To put it crudely and accurately, the tail fell off.

Out of control, the Airbus began shedding parts -- contrary to what you might see on the latest network mockery, ABC's "Lost," nobody emerges alive from a disintegrating jet with no tail -- before impacting a row of houses. All 260 passengers and crew were killed, along with five people on the ground. The death toll makes it the second-worst airline disaster ever on U.S. soil, eclipsed only by the 1979 crash of a DC-10 at Chicago-O'Hare.

Investigators quickly located the detached tail -- or, as you might hear it described in some reports, vertical stabilizer -- which had fallen into the water of Jamaica Bay, more or less intact. Initially they wondered if the wake encounter alone may have somehow been strong enough to shear the structure away, perhaps with help from a heretofore undetected crack in its composite architecture.

Eventually, attention turned from the turbulence itself to the way the crew reacted to it. Specifically, the veracity of rudder inputs made by first officer Sten Molin, who was flying the aircraft at the time. The rudder is the large movable surface hinged to the aft edge of the vertical stabilizer, used to control the side-to-side "yawing" motion of an airplane. It is controlled either through the autoflight systems, or, if need be, manually through a pair of foot pedals at either pilot's station. Here is a view of the first officer's station of an A300-600. The pedals are found on either side of the control yoke stanchion.

The voice and data recorders show the first officer having commanded a full deflection of the rudder. Fully deflecting a plane's rudder is somewhat akin to yanking a car's steering wheel 90 degrees, so most larger planes, including this one, are equipped with segmented rudders and automatically adjusted limiters, reducing the available rudder travel as speed increases. The faster you're flying, the less available movement. Keep in mind that full deflection of any control surface, rudder included, is rare. In 587's case, the crew was reacting to a strong burst of turbulence from the 747.

Even had the limiting systems somehow failed, Flight 587 was, at the moment of its doom, sufficiently below the speed at which any max deflection, intentional or otherwise, should damage the structure. Pilots call this "maneuvering speed." So, barring any structural anomaly, it seems there was no reason for Flight 587's vertical stabilizer to fail.

Except for two things. First, the first officer applied maximum pressure rapidly and in both directions, repeatedly swinging the rudder to the left and to the right. A plane's airworthiness certification standards are not based on such unusual, alternating applications of extreme force. Secondly, and perhaps more importantly, Airbus is alleged to have known the A300's rudder controls are unusually sensitive, and that resultant movements, even at low speeds, may be more severe than a pilot intends. In other words, the first officer didn't realize the level of stress he was putting on the tail.

Airbus claims to have passed this knowledge on to American Airlines numerous times, presuming the carrier's crews would be warned and trained accordingly. American says it was not informed of the rudder's potentially dangerous idiosyncrasy until after the accident. This is where things get murky, contentious, and, dare I propose, litigious.

In the meantime, and for the most part unfairly, Sten Molin is socked with most of the blame. The NTSB, for all its good work, sticks us yet again with, as I've called it before, that nastily ambiguous buzzword: pilot error. It lays primary fault to Molin, whose overcontrolling stressed the vertical stabilizer to the point of separation.

The full NTSB critique is quite long and in-depth, as you'd hope and expect it to be, hitting each of the critical nuances of the accident. But it's that single vague catchphrase that sticks in the craw of the flying public. The report tells you, clearly and succinctly, that the pilot "did it," leaving you to sift and digest the complex matters of what, precisely, he did, and for what reasons.

The NTSB describes Molin's actions as "unnecessary and excessive." Two of his colleagues cited past examples in which Molin overreacted to wake turbulence. The board points out that use of the rudder "was not an appropriate response to the turbulence, which in itself provided no danger to the stability of the aircraft."

That's a valuable point: In all but the most severe cases, wake encounters, although often uncomfortable, are brief and harmless, particularly when the aircraft involved are of similar size and weight (the vortices of a Boeing 747 are certainly strong, but the A300 is itself a large and powerful aircraft). The trouble is, Molin didn't know his inputs were unnecessary and excessive. Why he didn't know is the crux of this accident.

The NTSB does its best to answer that question through its 18 itemized conclusions. Culpability is spread roughly evenly between American's training regimen and the jet's rudder design. American's simulator techniques may have caused the first officer to have an unrealistic and exaggerated view of the effects of wake turbulence, erroneously assuming the need for "aggressive roll upset recovery techniques." Meanwhile the A300-600's rudder control system was described as having "the lightest pedal forces of all the transport-category aircraft evaluated."

Exonerated, by the way, are the tail's composite materials, an early suspect in the crash. The tail had been subject to more than twice its certified stress limits. Also laid to rest, thankfully, are assorted conspiracy theories. The incident occurred only a month after Sept. 11, and a bevy of alleged plots were quickly peddled. According to the most popular one, a terrorist bomb had caused the catastrophe, and the government, afraid of mass panic and further emotional meltdown, was covering it up. Says the NTSB, numerous witnesses who spoke of Flight 587 being on fire prior to impact were likely seeing compressor surges from the engines, and/or ignited fuel vapors once the engines, under stress while the Airbus flailed, broke from the wing.

So where does this all leave you, the passenger?

First and foremost, don't be misled by the broad brush of pilot error. The layperson is led to envision a negligent first officer kicking at the rudder controls for no good reason, like a lousy driver hitting the accelerator instead of the brake. The truth is far more complicated. The crash of Flight 587, like many accidents, was not the result of any single, overt act. It was the culmination of an unlikely chain of events: a wake encounter; the overuse of control force; and an unusual design quirk -- compounded by possible mistakes in how knowledge of that quirk was communicated and interpreted. Was the accident unfortunate, tragic and ultimately avoidable? Yes, though you could say exactly the same thing about almost every air crash. Such conclusions do not imply blatant negligence, incompetence or commission of a crime on the part of the airline, manufacturer or crew.

Asks an e-mailer: "Did you see the NTSB decision? Why would the pilot be so stupid as to break off the tail?

He wasn't.

From another: "I am flabbergasted that Airbus makes a tail that can be broken off when a pilot simply uses the controls."

They don't.

From a third: "Doesn't the FAA oversee training at the airlines! Why did American encourage the use of such dangerous practices?"

They didn't.

The investigation itself has not scapegoated Molin. The NTSB conclusions are comprehensive and fair. The board's Oct. 26 press release begins like this: "NTSB Says Pilot's Excessive Rudder Pedal Inputs Led to Crash of American Flight 587; Airbus Rudder System Design and Elements of Airline's Pilot Training Program Contributed."

Unfortunately, if not predictably, it's the first part of that sentence that is being seized upon by media outlets and a nervous public. The lessons here -- the point of this article -- are not to highlight danger or sensationalize the awful ramifications of error. The intent is to emphasize the complexity of accidents and their official summations.

For what it's worth, excluding cargo variants sold to UPS and Federal Express, American Airlines remains the A300-600's only U.S. customer. Domestically there are few options that will put you aboard one, but nonetheless I'll answer the obvious question: Is the Airbus A300 safe to fly on?

Absolutely. It was safe to begin with -- which I'll agree is no consolation for 265 victims that November morning, and is safer now. So goes the evolution of air safety. Bad things occasionally happen. Flaws come to light and we act to remedy them. All parties are now knowledgeable about the A300's rudder peculiarities, and crews have been retrained. The NTSB has issued 14 recommendations to the FAA covering all mechanical and human-factor aspects of the accident.

But, you're thinking, even as no single act caused the 587 tragedy, could we not argue that multiple, independently innocuous breakdowns combined to create an environment of impending danger? If so, what else must loom out there?

Yes, a strange brew of coincidence can result in calamity. Simultaneously, however, we see the statistical anomaly that was this, and every other, plane crash. All those black stars in perfect alignment.

How often does it happen? Worldwide, the past three years have been the safest in aviation history. In the United States, close to 2 billion passengers have arrived safely since the morning of Nov. 12, 2001. It's a jolt of irony, on one hand disturbing, but at the same time instructive, that our last serious accident was that of American Airlines Flight 587.

Personal note: I was flying from Amsterdam to New York on the day of the 587 mishap. About an hour from landing the captain emerged from the flight deck, stood at the front of the cabin and informed us we'd be diverting to Hartford, Conn., because of an accident at JFK. Everyone on the plane thought exactly the same thing. The young Kuwaiti businessman sitting next to me looked especially mortified by the prospect of another terror attack.

Eventually we broke off our approach to Hartford and continued to JFK. Our 767 was the first plane to touch down after the airport had reopened. Smoke from the crash was still visible -- a gray gossamer haze rising over the long, thin stretch of Rockaway. I'd never seen a crash site before, but now, albeit from a distance and a wee bit figuratively, I could see two of them: a smoldering swath from the downed Airbus, and the empty piece of sky where the World Trade Center used to stand.

- - - - - - - - - - - -

Do you have questions for Salon's aviation expert? Send them to AskThePilot and look for answers in a future column.

By Patrick Smith

Patrick Smith is an airline pilot.

MORE FROM Patrick Smith

Related Topics ------------------------------------------

Air Travel Ask The Pilot Business