A Swing and a Bunch of Near Misses

It has been 14 years since the last fatal airline crash in the US. For the white-knuckle flyers, that may seem hard to believe but the the crash of Colgan Flight 3407 on February 12, 2009, which killed 50 people including one person on the ground, was the last time an American airliner went down. The safety streak is amazing considering that major crashes happened every year or so prior to 2009 and the stringent safety practices in the industry paired with good fortune deserve much of the credit. Now, many wonder if that streak may be coming to an end after the FAA reports that near misses are becoming more common.

Among the stories of recent near misses are a close call between a Southwest airliner and a FedEx transport in Austin in which the two planes came within 100 feet of colliding and another incident in Boston in which a private jet took off without a clearance causing a conflict with a JetBlue airliner on approach. A New York Times investigation this week revealed that such close calls are far more common than most people are aware, with at least 46 near misses in July.

The Times analyzed NASA’s Aviation Safety Reporting System (ASRS) and found that reports of near misses had doubled in the past decade. The Times admits, however, that “it is unclear whether that reflects worsening safety conditions or simply increased reporting.”

Let’s talk about the ASRS. NASA also deals with aeronautics, for those who don’t remember the first “a” in the acronym. The ASRS is a long-running self-reporting system that encourages pilots, controllers, and other aviation workers to report errors or unsafe conditions anonymously. In exchange, the reporters are offered immunity from punitive action (withcertain limitations). In the aviation community, the ASRS form is known colloquially as the “get-out-of-jail-free” card for inadvertent mistakes or rule violations (with an emphasis on “inadvertent”).

Having said that, the Times is right to be suspect about the data reported to ASRS. I’d say that there are probably more near misses than are in the database because pilots are only likely to report problems that they think the FAA already knows about. “Don’t ask, don’t tell” could describe the attitude of most crews to violations that aren’t likely to bring on an FAA enforcement action.

The flip side is that the number probably isn’t a lot higher than the Times estimate, at least not with respect to airliners. The nature of jet and airline flying is that most of it occurs in controlled airspace and at airports with control towers. The majority of the time, airline and other jet flights are in radar contact with ATC, which is monitoring their distance from other aircraft.

In its response to the Times article, the FAA said that it welcomed the scrutiny and pointed out that near-miss incident reports are on its website, just not compiled into the form that the Times used for its article.

“The FAA maintains extremely conservative standards for keeping aircraft safely separated. Safety experts follow up on all events — even those in which no collision was imminent or even possible — and evaluate them for safety risks,” the FAA added in its statement. “The agency publishes this information on our website, updating it as new information becomes available.”

Air traffic control (ATC), which was itself enhanced after the 1956 collision of two airliners over the Grand Canyon, typically takes most of the responsibility for separating aircraft that are operating under instrument flight rules (IFR). This would include the vast majority of jets and airliners of all types. Minimum separation varies by phase of flight with cruise flight requiring larger separation than the congested areas around airports, but minimum separation is often several miles horizontally and/or 1,000 feet vertically. It’s no surprise that most near misses occur near airports because that’s where the airplanes are.

In recent decades, technology has come a long way in providing an extra margin of safety against near misses. The Traffic Alert and Collision Avoidance System (TCAS) displays the location and altitude of nearby aircraft to pilots in the cockpit. If the TCAS units, which actually communicate between aircraft, sense that a collision is possible, they will even coordinate and issue resolution advisories (RAs) to their respective pilots.

Since TCAS works primarily in the air, there have also been enhancements to airports to help prevent runway incursions, simply described as going onto an active runway without an ATC clearance. A variety of lights and signs are available at many airports to alert pilots to the proximity of runways and even whether the runway is safe to use. Some airports even have surface monitoring equipment that allows controllers to track the location of taxiing aircraft and vehicles.

The weak link in these systems is the human component. Aviation systems have become so reliable that accidents are almost always traced back to some sort of human error. The incident in Austin was apparently the result of a controller who issued a takeoff clearance to Southwest while the FedEx freighter was on a three-mile final. In normal circumstances, this would not have been a problem, but the Southwest crew took an inordinate amount of time (about 30 seconds) on the runway.

The Boston incident seems to have been the result of a “brain fart” in which the pilot of a charter Learjet took off without permission. A common procedure is for ATC to clear aircraft to “line up and wait” on the runway to facilitate a quicker takeoff. There is still a requirement to wait for an actual takeoff clearance, but the pilot of the Lear mistakenly (he says) thought he heard the controller clear him to take off rather than hold in position. This incident bears a striking resemblance to the most costly crash (in terms of lives) in aviation history when two Boeing 747s collided on the runway at Tenerife in the Canary Islands in 1977.

Ironically, the Lear crew may be able to submit the incident via the ASRS and escape severe punishment for their mistake. The ASRS submissions often allow crews to be sanctioned with remedial training for honest mistakes rather than losing their licenses. Many companies have similar amnesty programs for internal errors as well.

There are several reasons that near misses might be on the rise, although they don’t necessarily apply to the two headline incidents. Chief among these are labor shortages at both the FAA and the airlines, which means that new hires may lack experience and old hands may be overworked. While there are stringent rules aimed at preventing fatigue, a legacy of the Colgan crash, even legal work schedules can be fatiguing under certain circumstances. Even the major airlines, which for years have had a steady supply of experienced regional airline and military pilots to draw upon, are increasingly having a hard time recruiting enough new pilots to replace pilots reaching the mandatory retirement age.

The pilot shortage is at least partly a result of the Colgan crash as well. In the wake of that tragedy, Congress increased the minimum experience for airline pilots to 1,500 hours. That experience can take years to amass and, combined with a dearth of airline hiring in the decades since the September 11 attacks which dampened pilot training, the result is a severe shortage of qualified pilots.

To underscore the severity of the pilot shortage, I’ll note that I frequently get emails and cards inviting me to apply for airline jobs. As I was writing this, I got one such invitation from a regional airline offering a direct-entry captain position. Normally, airlines operate on seniority and a new hire would have to wait years before upgrading to captain.

A few years ago, such solicitations were unimaginable. The airlines had more pilot applications than open slots and it was almost impossible to get an interview. Today, some airlines hire more pilots in a month than they did over a period of years in doldrum decades.

Forty-six close calls in a month is a tiny number given the hundreds of thousands of flight hours generated by airlines and private aircraft in the US every month. The FAA website estimates that there are 45,000 flights handled by the ATC system every day. If only one or two of those come into close contact, the overwhelming majority are still very safe and clear of conflicts.

But that’s still a massive problem. While midair collisions are rare events, they also tend to be catastrophic when they do occur. For that reason, the FAA is going to be driven to do something.

One obvious solution is to educate pilots and controllers about the problem. In the past, such educational pushes have been used to reduce the number of runway incursions and altitude deviations (when an aircraft is not at the assigned altitude). Cockpit Resource Management (CRM) is also an educational concept that was born out of a number of crashes in which crewmembers didn’t speak up to contradict the captain. Sometimes just being aware of the increased risk is enough to reduce occurrences.

Regulation is another possibility. There is a saying in the aviation industry that the Federal Aviation Regulations are written in blood. Many of our rules are based on behaviors or accidents that have killed people in the past.

An obvious path here is to increase separation requirements for aircraft, especially in busy terminal environments. The downside to this approach is that more separation can mean fewer operations and more delays at the busiest airports.

Ultimately, the responsibility for safely operating the airplane is on the pilot’s shoulders. When humans err, it is often up to other humans to identify the problem and take action to avoid a catastrophe. Despite all the gee-whiz electronics in the cockpit and around the airports, it was the quick instincts of the FedEx and JetBlue crews who averted disaster.

Those instincts are honed by years of experience. That’s what passengers expect when they see an airline pilot and that level of skill and judgment is what airlines are paying for when they negotiate lucrative salaries for their pilots.

But the problem goes back to the shortage of qualified pilots in the labor pool. It’s easy to train a pilot to manipulate the controls of a jet. Consider that the military puts neophyte pilots at the controls of some of the most high-performance and sophisticated aircraft on the planet with only a few hundred hours of training.

Experience, judgment, and good situational awareness, especially in high-workload environments, are harder to teach. And that’s where both pilots and controllers can find themselves in trouble.