Ask the pilot

Winding up the safest seven years in modern aviation history. Plus: A federal task force weighs in on tarmac delays.

By Patrick Smith

Published November 21, 2008 11:26AM (EST)

Before getting on to the bigger stuff, let's talk about doors for a minute -- cockpit doors. "One point that never gets made in discussions about post-9/11 air safety," posted a reader in response to last week's column about Sept. 11 and security, "is that of secure cockpit doors."

As most fliers are aware, cockpit entry doors, once flimsy and relatively easy to breach, were redesigned in the wake of the 2001 terrorist attacks. I’m not at liberty to discuss the specs of this redesign, but everybody knows they are heavily armored. "Even if a terrorist somehow manages to bring a sharp object on board," the reader continued, "or fashions one in flight, he will not be able to gain access to the cockpit."

That's one of the reasons, yes. A plane full of angry passengers, most of whom are averse to dying, is the other, even bigger reason. And partly because of that, I have mixed feelings about the armoring of cockpits, as do many safety experts. It is true the new doors would be very effective at keeping unwanted guests out of the flight deck. But think about it. What happens if, difficult as it would be, one or more attackers did make it up there and managed to incapacitate the crew? The culprits are now barricaded in the flight deck, and there is no easy way to get in after them. In a way, especially with the hijack paradigm shift after 9/11 (passenger awareness and expectations during a hijacking), we would be safer with the old doors. Not the exact old doors, but perhaps a redesign not as impervious as what we have today. True, passengers were unable to breach even the old-style door on United 93, but apparently they came close. Why give a hijacker more time? I didn't always think this way, but I've come around on it.

Not that it matters much. Doors or no doors, the 9/11 blueprint is defunct. Replicating the attacks would present such high probability of failure that I can’t imagine anyone being stupid enough to try it.

Curiously, there was a push to reinforce cockpit doors long before Sept. 11. We go back to 1987, when a recently fired ground worker, David Burke, used his credentials, which his employer had failed to recover, to carry a concealed handgun onto Pacific Southwest Airlines Flight 1771 from Los Angeles to San Francisco. En route he broke into the cockpit, shot both pilots, and nosed the airplane into the ground near Harmony, Calif., killing all 44 on board.

Unbelievable as it might sound, the government’s response to the incident was not to implement security screening for ground personnel, but for pilots and flight attendants instead! It was a public relations gimmick -- passengers now saw crews having to wait in the same annoying security lines, making them feel safer -- completely useless at preventing another David Burke.

The airlines meanwhile resisted the idea of upgrading the doors, on the grounds that it was expensive and unnecessary. How right or wrong they were I can't be sure. Would it have prevented the attacks of 2001? Maybe, maybe not.

Don't tell me, though: another good reason to hate the airlines.

Speaking of which, passenger rights groups were disappointed last week after a federal task force failed to recommended stricter guidelines in cases of extended tarmac delays. Passenger advocates were hoping for a formal rule limiting how long an airplane is allowed to await takeoff before returning to the gate. The ongoing push for a so-called Passenger Bill of Rights, including a time limit for delays, got going two winters ago after a Valentine's Day storm at Kennedy airport resulted in thousands of passengers being stranded in icebound airplanes for up to 10 hours.

The task force did make several smaller, common-sense recommendations for airlines to follow during prolonged delays. These include providing refreshments, keeping lavatories usable and clean, and making available a secure lounge for those passengers diverted from overseas. None of this is binding, and I need to point out that the 36-member board consisted primarily of airline industry representatives.

To be clear, all airlines should have contingency plans for storms, mechanical breakdowns and other situations that cause delays, and there is no excuse for keeping hundreds of people confined on a jetliner against their will. I am not excusing any carrier for doing so. But, as I pointed out in 2007, because delays are so complex and fickle, enforcement of an arbitrary time limit is bound to cause more problems than it solves.

Returning a plane to the gate is not as simple as pulling a bus over to the curb and dropping a few people off. There are security issues to deal with, fuel, flight plans, air traffic control slot times (which are prone to change without warning), and a host of logistical challenges (gate availability, gate staffing, etc).

When we discussed this issue back in 2006, I painted a possible scenario to illustrate why this was a bad idea. It went something like this: Imagine you are on a delayed airplane going from New York to Chicago. You have been parked on the taxiway for the past two hours. Your assigned wheels-up time is only a half-hour away, but because the new, federally mandated time limit has just elapsed, the crew is required to ask if anybody wishes to get off. Imagine now that one person raises his hand.

There are 150 people on the plane, but only one of them wants to leave? Well, the rule is the rule, and so the pilots ask for clearance back to the terminal. Using portable stairs won't work because the guy has two suitcases somewhere below, and security regulations require they have to come off.

There are no vacant gates at the moment, and waiting for one to open takes half an hour. Then, the ground staff needs to locate the correct luggage bin in order to pull the guy's suitcases. Twenty minutes there.

And because taxiing to and from the terminal will burn 2,000 pounds of kerosene, the plane must also be refueled. (The weather in Chicago is foggy; en route and alternate airport minimums have put things right at the legal limit, so there's no way around this.)

Together this will entail a new weight-and-balance manifest, and possibly, because the flight will miss its wheels-up time, a whole new flight plan. And, oh, missing that time means you're assigned a new one, and lo and behold it's another hour away.

What initially was less than a three-hour delay has now become a five-hour delay. Throw in the need to de-ice, or the possibility of the crew's running up against duty time limits, and it’s easily worse.

A Passenger Bill of Rights? In some respects it's a good idea, but be careful what you wish for.

Ironically, the board came out with its recommendations on, of all days, Wednesday, Nov. 12. That happened to be a very important day for commercial aviation, yet as best as I can tell, nobody was paying attention. Heck, even I had forgotten about it, being busy that week preparing an unrelated column.


Nov. 12 was the seventh anniversary of the crash of American Airlines Flight 587.

On Nov. 12, 2001, American 587, an Airbus A300 bound for Santo Domingo, Dominican Republic, went down shortly after departure from Kennedy airport in New York, killing everybody on board.

Seconds after takeoff, the Airbus had encountered a virulent burst of wake turbulence spun from a Japan Airlines 747 that had departed just ahead of it. The wake itself was nothing deadly, but the first officer of Flight 587, who was at the controls, overreacted, repeatedly moving the widebody jet’s rudder back and forth to maximum deflection. (The rudder is a large hinged surface attached to the tail, used to aid turns and help maintain lateral stability.)

Planes can take a surprising amount of punishment, but airworthiness standards are not based on such rapid applications of extreme force. In addition, the A300's rudder controls were designed to be unusually sensitive, meaning that pilot inputs, even at low speeds, could be more severe than intended. In other words, first officer Sten Molin didn't realize the level of stress he was putting on the tail. The ferocity of the back-and-forth movements caused the entire tail to fracture and fall off.

Quickly out of control, the plane plunged into the Belle Harbor neighborhood of Queens, a skinny section of Rockaway only a few blocks wide, with ocean on both sides. All 260 passengers and crew were killed, as were five people on the ground. It remains the second most deadly aviation accident ever on U.S. soil, behind only that of American Flight 191 at Chicago, in 1979.

What makes this seventh anniversary so important has nothing to do with the disaster itself, but with the remarkable streak that has followed it. Eighty-four months have passed since then. That’s 2,555 days and counting. In that span, our carriers have transported approximately 5 billion people and made more than 51 million takeoffs and landings. Yet the crash of Flight 587 was the last large-scale mishap involving a major U.S. carrier -- the longest such streak since the advent of the jetliner five decades ago.

Don’t get me wrong: There have indeed been fatal incidents on U.S. soil. The worst of them was the wrong-runway takeoff crash of Comair Flight 5191 in Lexington, Ky., in 2006. Forty-nine people perished in that one. Others include the 2003 crash of a US Airways Express turboprop in Charlotte, N.C., in which 21 people died, and the October 2005 crash of Corporate Airlines Flight 5966 near Kirksville, Mo., which killed 13. Less than a week before the Kirksville incident, a Pinnacle Airlines regional jet went down near Jefferson City, Mo., during a repositioning flight, killing two crewmen. In December 2006, the crash of a vintage seaplane killed 20 in Florida. And at Chicago’s Midway Airport in December 2005, a Southwest Airlines 737 overran a snowy runway and collided with a car. Although none of the 103 occupants of the aircraft were seriously hurt, a 6-year-old boy, a rider in the crushed car, was killed.

Duly noted. (Each of the above received due attention in this column.) But when it comes to the streak, I emphasize both "large-scale" and "major carrier." For those are the ingredients that constitute air disasters as we tend to think of them, and as history records them.

If you insist otherwise, well that’s fine too. Parse it however you like, but the past seven years have been astonishingly safe by any measure. (Not just domestically, but globally as well. The raw number of crashes is up, but the accident rate, the number of fatalities per miles flown, has steadily declined.)

With all the anti-airline vitriol out there, rarely if ever is this streak acknowledged. And neither, I am afraid, will it be acknowledged when the inevitable accident comes. It will pass unmentioned by newscasters and reporters, caught up in what will surely be an orgy of hype and fear.

In the meantime, perhaps you’re wondering what, exactly, is responsible for such a run of good fortune. Better training and technology, mostly. We have engineered away the most common causes of past accidents. But another thing has played a role, much as I hate to admit it: a long stretch of very good luck.

So you can argue that we're due, and I suppose we are. Can I say that, without triggering panic among those of you skittish about flying? Truly, I don't mean to suggest a catastrophe is imminent, or that you should forgo that vacation. Flying will always be safe, even on the very day of our next terrible tragedy.

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

Do you have questions for Salon's aviation expert? Contact Patrick Smith through his Web site and look for answers in a future column.

Patrick Smith

Patrick Smith is an airline pilot.

MORE FROM Patrick SmithLIKE Patrick Smith

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

Air Travel Ask The Pilot Business