Ask the pilot

How can an airplane, loaded with every high-tech gizmo imaginable, accidentally land at the wrong airport?

Published July 2, 2004 7:30PM (EDT)

On Saturday, June 19, Northwest Airlines flight 1152, an Airbus A319 en route from Minneapolis to Rapid City, S.D., carrying 117 passengers and five crew members, touched down smoothly just after noon. Trouble was, they weren't in Rapid City. They'd arrived at Ellsworth Air Force Base, a military airfield six miles from the Rapid City Regional Airport.

Remarkable as it might sound, the unannounced landing -- which, as my mailbox burstingly attests, did not go unnoticed by news-savvy readers of Salon -- was not unprecedented. Years ago a Western Airlines 737 pulled a similar maneuver in Wyoming, and in 1995 a DC-10 touched down in Brussels instead of Frankfurt. A regional turboprop once confused the northern Maine outposts of Caribou and Presque Isle, while more recently, in January of this year, a US Airways Express flight bound for State College, Penn., ended up in Philipsburg, Penn.

If you're like me, you wouldn't know the difference between State College and Philipsburg if you were sitting in the lobby of City Hall, but the idea of highly trained aircrews with troves of technology at their behest landing astray sounds, I'll agree, amusing, quaint and even patently ridiculous. The letters I've received in the past several days are equally aghast, incredulous and cackling.

How does it happen? Anticipated by some is an answer of tangible malfunction: improperly keyed coordinates or a navigational computer gone crazy. Others suspect a more visceral, seat-of-the-pants explanation: a tired crew mixing up a pair of similar-looking runways.

After the Ellsworth incident, an Air Force spokeswoman reminded the dumbstruck news media that the crew's intended runway at Rapid City airport is "just over the hill" from Ellsworth. Implication: Pilots find their airport the way a lost motorist finds his way home -- second hill on the left, slight bank over the lake, watch for the farmhouse. Comments like these entice you to miss the big picture, in this case the vast context of the airspace system. I'm betting you don't want a dissertation on the innards of instrument approaches or a glossary of air traffic control vernacular, but the complexities of even the most ordinary good-weather landings are a bit more involved than lazy pilots craning their necks and saying, "Yeah, that looks like the right place, let's give it a try."

In addition to whatever human errors catalyze such events, weather and air traffic control, to name two other factors, can lend a hand in getting from point A to, as it were, point C. In the case of that DC-10 finding its way to Belgium instead of Germany, a confusing string of foul-ups caused controllers essentially to lead the plane astray. By the time the crew realized it was being vectored to the wrong country, they decided to land at Brussels rather than have to recalculate fuel reserves and orchestrate a last-second re-routing. In the end, it was safer to land in Brussels than Frankfurt.

At the same time, yes, pilots cleared for what we call a "visual approach," a procedure used with fair routine when sky conditions permit, will merely eyeball an airport through the windshield and report "field in sight," to air traffic control.

Descending toward Rapid City, the Northwest crew was flying a so-called VOR (very high frequency omnidirectional range) approach, a fair weather, non-precision-instrument procedure designed to guide aircraft toward the general vicinity of the runway, if not to the very threshold as with more accurate ILS (instrument landing system) approaches. Sixteen miles out, they'd intercepted the inbound course and began tracking it toward the airport. Emerging below a layer of clouds, they identified what was believed to be the correct runway and configured their jet for landing. What they'd seen, of course, was Ellsworth, its runway sharing an almost identical alignment with the intended strip at Rapid City, still about six miles away.

The carrier refuses to comment on who or what may have been responsible for the Ellsworth touchdown, and is keeping mum on everything from the pilots' names to their levels of experience. This is neither surprising nor unfair, and it's any investigator's policy to refrain from passing judgment when events are fresh and particulars still unknown. This is a business where the determined causes of accidents and incidents rarely conform to the earliest suppositions and theories.

Disclaimer out of the way, I'll tell you what I think: The crew, it appears, committed the simple bungle of assuming one airport was another. "Pilot error," a buzzword we're all acquainted with. Here we find it unusually, even shamefully, obvious.

In the winter of 1990 I was part of a crew cleared for a nighttime visual approach to New Haven, Conn. Suffice it to say that the lights and orientation of Bridgeport appear strikingly similar to those of New Haven, and after a minute or two we realized the error and corrected course. All of this happened far from the runway, and we were flying a 15-seat Beech-99, a vintage '68 relic with as many electronic accoutrements as my Trek mountain bike.

What makes the Northwest example so baffling is the seeming nullification of all the safeguards on hand. At the pilots' disposal were more than enough cockpit gizmos and published charts reminding them of Ellsworth's vicinity to Rapid City. Since when do we entrust our lives and fancy planes entirely to the humble predilections of judgment, no matter the complications or contributing factors? People make mistakes, big and small, and technology is intended to catch them. (And, understandably, vice versa). Flight 1152 was a small but relatively state-of-the-art Airbus A319 (a sub-model of the popular and somewhat indistinguishable A320). The captain and first officer sat before a multiscreen panel of high-tech computers and "moving map" guidance systems.

(Still to be determined is why air traffic control was unable to relay a warning to the wayward Airbus. Both Ellsworth and the Rapid City personnel could have tracked the diversion, though it's possible the plane had dropped beneath radar coverage.)

"I find it startling," says a Northwest first officer assigned to the A320, "that the crew managed to do what they did, especially with such an advanced airplane. The proper flight path -- right to the very runway -- is displayed graphically on the cockpit nav displays, regardless of what instrument procedure is in use. The mistake would have been very apparent."

Asked if any type of technical miscue may have led to the confusion, the pilot, who asks not to be identified, claims it a virtual impossibility. "The nav displays would not let this happen. I suspect the pilots were simply locked in, visually, to what they saw outside."

Last year, when an Air Canada jet came close to landing at Vernon, B.C., instead of Kelowna, about 40 miles away, we heard similar reasoning. Due to forest fires, the plane had been assigned an unorthodox routing, and the crew became fixated on topographical landmarks. "My guess," said a pilot, "is the guys just had their heads out the window and weren't aware of where they were."

The act of landing a plane is at once inordinately simple (subject to no more fancier gizmos than a pilot's eyes and ears) yet simultaneously beholden to the guidance of technology (high-tech instrument and procedural rigors). Thus, this becomes one of the most challenging phenomena for a pilot -- even one who moonlights as a columnist -- to dissect or analogize for the layperson. Call it a cop-out, but a part of me wants to shout back at all those e-mails asking "How the hell does it happen?" with a simple and equally frustrated, "I DON'T KNOW!"

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

A normal San Diego takeoff is toward the west, up over the ocean, but occasionally we take off toward the east, heading straight at the steep mountain incline just off the airport! Isn't that dangerous?

Takeoff procedures are a lot more complicated than just gunning it and hoping you clear the hills. Climb gradient profiles, based on worst-case scenarios (for example, engine failures during liftoff) are computed for every departure, specific not only to the airport, but to each individual runway, taking all obstructions into account, whether hills, mountains, skyscrapers or broadcast towers. A plane's weight shall not exceed what's required to meet these performance prerequisites.

Each night the local UPS hub sends several departures toward Anchorage, from where they make the hop into Asia. Anchorage sits to the north and yet, almost without fail, the vast majority of these planes take off heading south. Is there a reason why a plane would take off in the opposite direction of its destination?

Planes rarely take off in the direction of their destination unless by chance. The runway a plane uses depends on any number of things -- the ceiling and visibility, the direction and speed of the wind, the length of pavement, off-airport obstructions, and whether any noise restrictions apply. The latter is important in your case, since freighters tend to be back-of-the-clock operators.

After liftoff, the published departure routes may take a plane in a circuitous pattern before finally delivering it in the general direction of its destination. For example, here at Boston, a plane bound for California may climb east over the harbor, then swing to the north, up over southern New Hampshire before proceeding westward. Just last evening I caught sight of a Virgin Atlantic jet, flight planned to London, lumbering due west for several miles before executing a lazy arc toward the northeast.

Arriving aircraft will be subject to similar course reversals and spirals before landing.

In your book you talk about how parallel runways are given "left" or "right" designations in addition to their magnetic numbering. OK, so what about pairs of parallel runways, as you'll find in Atlanta or Los Angeles? With four strips all pointing the same direction, you can't have two lefts and two rights.

That's an excellent question, if a bit more in-depth than I'd expect from a layperson. Either you're an airplane nut or I've really inspired some observation. (And somebody is reading the book, apparently, which is at once flattering and intimidating.)

To answer: airports will usually tweak one of the pairings to the next cardinal. Rather than have four runway 27s, for example, you'll have a pair of 27s and a pair of 26s (with a Left and Right for each). Or a pair of 27s and a pair of 28s. A strip's precise orientation is usually some fraction of 10 degrees; runway 27 is likely to be pointing 274 degrees, or 267 degrees, rather than a clean 270. Thus, rounding up -- or down -- isn't so egregious.

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

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