Ask the pilot

Chances are you've come across one or more recent stories about the marked rise in so-called runway incursions at airports across America. That's a euphemism for when a plane or other vehicle erroneously enters or crosses a runway without permission from air traffic control, setting up a potential collision hazard. In FAA-speak, it is "any occurrence at an aerodrome involving the incorrect presence of an aircraft, vehicle, or person on the protected area of a surface designated for the landing and take off of aircraft."

Not every incursion is the proverbial "near miss." The vast majority are in fact harmless. But the numbers are rising and a handful of incidents were indeed close calls. In some respects, the uptick is no real surprise. After all, there are twice as many airplanes flying today as there were 25 years ago. But the numbers don't match -- incursions have been rising well out of proportion with traffic growth. (In 2007, commercial air carriers were involved in eight separate incursions deemed by the FAA to be "serious.") The problem isn't the volume of planes, per se, but the congested environments in which many of them operate. LaGuardia, Reagan National, Boston and JFK are among airports that were laid out decades ago for a fraction of today's capacity. Their crisscrossing runways and lacework taxiways are inherently more hazardous than the parallel and staggered layouts seen at newer airports. That does not imply these locations are unsafe, but they are more challenging for both crews and air traffic controllers, particularly during spells of bad weather and low visibility.

Essentially one of two things causes an incursion: pilot error or controller error. The relationship between pilot and controller is, and has to be, one of mutual trust. Indeed this is one of the few aspects of flight safety in which the pilot relinquishes direct control. Flying along through dense clouds, or cleared to land on a fogged-in runway at a busy airport, he or she assumes that unseen others have not made some terrible mistake. Maneuvering through the skies and along taxiways, pilots listen not only for their own instructions but for those of other pilots as well. By creating a mental picture of what other aircraft are doing, they can orient themselves in the vast choreography of a crowded sky or tarmac. Should anybody offer an incorrect read-back, acknowledge the wrong clearance or otherwise screw up, other pilots often detect the mistake. Errors are rare -- dangerous ones even more so -- but the potential is always there, and there's no guarantee they'll be caught.

The last serious runway collision in this country occurred in 1991 in Los Angeles, when a USAir Boeing 737 landed atop a SkyWest commuter plane that a controller had cleared onto the runway and forgotten about. There have been numerous near misses since then. Last year in Fort Lauderdale, Fla., jets belonging to Delta and United came within a hundred feet of hitting each other. In a 2005 near collision in Boston, an Aer Lingus A330 and a US Airways 737 missed each other by approximately 170 feet at Boston's Logan International. Even more heart-stopping was the 1999 incident at Chicago's O'Hare, in which an Air China 747 wandered onto an active runway and was nearly struck by a Korean Air 747.

In 2001 at Milan's Linate airport, 118 people were killed when an SAS (Scandinavian Airlines) MD-87 collided with a corporate jet. A ground controller, an air traffic control director and three airport executives were later given prison sentences of up to eight years. In Mexico City a few years ago, an Air France 777 crossed in front of a departing DHL cargo plane. The DHL captain averted disaster by swerving into the grass, and the entire crew was later given an award.

(As a former DHL employee I knew all three pilots aboard that freighter in Mexico. Although I have personally never been involved in a runway incursion, longtime readers might recall that I was once privy to a near miss of another kind.)

The FAA has been working fast and furiously on developing new programs and technologies that will reduce the number of mistakes, and/or mitigate the consequences when they occur. These include an upgrade of tarmac markings at 75 large and medium-sized commercial airports and mandatory anti-incursion training programs for pilots and controllers. Now under testing are improved runway and taxiway lighting systems and an emerging, satellite-based technology known as ADS-B (automatic dependent surveillance-broadcast). An enhancement of ADS-B, called CDTI (cockpit display of traffic information), has been installed in over a hundred aircraft flown by cargo giant UPS. It provides pilots with a detailed view of surrounding traffic, both when aloft and during ground operations.

These are all good ideas, but the FAA has a habit of overengineering complicated fixes to simple problems. There will be no magic technological bullet. At heart this is a human factors issue.

As it stands today, runway safety relies on simple trust and acknowledgment. There are no railway-style gates, stoplights or other mechanical means to stop a crew from errantly crossing a runway or taxiway. There are, however, standardized tarmac markings, signs and very specific protocols that all commercial pilots are familiar with.

Crews acknowledge "hold short" instructions that allow them to taxi up to, but not onto, an active runway. When cleared "into position and hold" (or, if overseas, "line up and wait") they may proceed onto the runway but are not yet authorized to take off. Finally comes "cleared for takeoff," meaning just what it says.

On arrival pilots are "cleared to land," meaning the plane is allowed to continue its approach and touch down without further communication. Implicitly, whether or not it's visible to the crew, the runway will be clear of other aircraft. (This generally happens around the time you hear the landing gear clunking into place. The "we've now been cleared to land" announcement commonly made by flight attendants is bogus; they haven't the slightest idea when the pilots are cleared to land.) During taxi in, pilots may be instructed to "hold short" of other runways. Each of these directives must be acknowledged using specific language -- basically, the crew reads back the exact instruction:

Control tower: "United 502, taxi into position and hold, runway one-six." Pilot: "Position and hold, runway one-six, United 502."

A truncated "Roger, 502" or other muddled reply will get the controller snapping back at you, asking for clarification.

Communication is sometimes hampered by radio congestion. Foreign crews with marginal English skills and the lack of standardized worldwide phraseology are also negative factors. But the premise is simple and extremely effective: You do not land on, depart from, taxi onto or otherwise cross a runway without explicit permission to do so. If you're at all unsure, don't do it.

Of course, this is only half of the picture. A pilot can do everything perfectly and still be victimized by a controller's error. To that end, advanced ground surveillance radar, which allows ATC to closely monitor airplanes -- and certain vehicles -- as they maneuver around the tarmac, is becoming increasingly common. Thus far, however, the FAA has not come up with a solution for the chronic understaffing found at many ATC facilities. Call me old-fashioned, but if the agency believes that high-tech radars and fancy lighting schemes can compensate for reduced manpower, it is probably, and perhaps dangerously, mistaken.

The trouble with, for instance, flashing lights or a highway-style traffic signal is that we might come to rely on them too thoroughly. As any number of motorists who've been broadsided at intersections know, putting your trust in a stoplight isn't enough. As a driver, I feel more comfortable at a four-way stop than at a crossing controlled by lights. Technology, for all its usefulness, isn't the solution. In the end, staying out of harm's way relies on basic vigilance and situational awareness.

A little-discussed aspect of this issue is the influx of thousands of extremely low-time pilots into the system. Entry-level qualifications at regional carriers have fallen sharply, putting new hires with fewer than 500 flight hours at the controls of $30 million R.J.s, flying into some of the most congested airports in the country. An acquaintance of mine was recently hired by one regional affiliate with 280 total hours. When I had that level of experience, I could not have fathomed the idea of working for an airline. He had never landed at a busy commercial airport in his life; his first day on the job began with a rush-hour trip into LaGuardia. As discussed in this column in December, even the lowest-time first officer is highly trained, and is always accompanied by a more experienced captain. Experience, per se, is not a good indicator of a pilot's potential for screwing up. But on some level this is one more risk factor.

Which is not to inspire undue worry or panic. Remember, close to 20,000 commercial flights operate daily in the United States, virtually all of them safely. I don't feel the current situation represents a crisis, or reveals any large-scale breakdown of the system. But even the strongest chain has its weakest link, and there's no harm in pointing that out -- or trying to make it stronger.

Incursions are still very rare, and I'll remind passengers that certain experiences shouldn't be construed as close calls. For example, you'll occasionally be on approach, close to the ground, when suddenly your plane accelerates and begins to climb away. The captain comes on and explains that traffic on the runway hadn't cleared yet. This is not the same thing as an incursion, and does not, except in highly extraordinary circumstances, propose that you were close to hitting another aircraft. Go-arounds are not near misses. That's what the spacing parameters are there for -- to prevent a near miss from happening.

Meanwhile, the FAA's most valuable contribution to the problem might be something it has already done: stirred up awareness. When it comes right down to it, the best way of preventing collisions is for pilots and controllers to always be conscious of their possibility.

Had those planes in Chicago not missed each other that night in 1999, it wouldn't have been the first time a pair of 747s met in disaster. Not to close on a morbid note, but I'll remind you that aviation's worst-ever catastrophe involved two 747s that never left the ground. It happened at Tenerife, in Spain's Canary Islands, in March 1977. An unlikely chain of events, capped by an overanxious captain who took off without permission, led to a foggy collision that killed 583 people. Two summers ago I spent an afternoon with Bob Bragg, the surviving copilot of the Pan American plane involved in the crash. Hearing Bragg's riveting firsthand account, described here, I reckon that the smartest thing the FAA could do is hire Bragg to travel around the country, sharing his story with every pilot and controller.