The Classroom of the Skies
Airline travel is so safe because the aviation system is so relentless about learning from failures. New lessons from an ex-astronaut at age 90, two 90-year-old aircraft, and two 737 MAX planes.
This post is a return to online action after a cold, rainy, but sublime week-plus immersion in Amsterdam with young Jack Fallows, as described here. Now here is an overview of some recent, publicized aviation accidents and close calls. My purpose is to say something about them in their own right and to suggest what air-safety systems may try to learn from them. Let’s start with the best known:
1) William Anders and the predicament of ‘older’ pilots.
Two weeks ago William “Bill” Anders, who as an astronaut on Apollo 8 took the world-famous “Earthrise” photo in 1968, died in a small plane crash at age 90. Anders was flying the plane and was the only person aboard when it plunged into the waters off the San Juan Islands north of Seattle.
It’s natural to ask: What was a 90-year-old doing in command of a plane generally used as a military trainer? Even if this particular 90-year-old had been a lifelong aviator and was an Air Force fighter pilot before his time as an astronaut? Shouldn’t the rules be tightened up?
Despite the tragedy of this episode for Anders and his family, I think it should be seen as a sad outlier rather than as a warning sign about larger systemic problems. Overall the aviation world is more sophisticated about issues of aging-and-competence than are most other parts of modern life. Consider:
In principle a pilot “certificate”—the term of art, rather than “license”—never expires. Once a Marine, always a Marine. Same for being a pilot.
The catch is that to “exercise the privileges” of your piloting credentials—that is, to be legal when flying a plane —you have to keep meeting a multi-layered, never-ending, increasingly stringent series of tests and standards.
—Pilots have to pass and keep passing medical exams and hearing and vision tests. For airline pilots there are regular EKGs. Pilots must report every problem a doctor has discovered during a checkup and every new medication they have been prescribed. As a “private pilot” you have to do this every few years, to legally fly yourself or your family around. As a “commercial” or “air transport” pilot carrying a plane full of passengers, you have to do it every few months—for as long as you want to fly.1
—Apart from those ongoing measures of your physical health, you have to keep showing that you know what you’re doing at the controls. All pilots are supposed to complete a “Flight Review” with an instructor at least every two years, ideally more often. (I’ve been doing it once a year.) To be legal when flying in “instrument conditions”—clouds and other bad weather—you need to remain “instrument current,” by logging real-world instrument approaches or passing an “Instrument Proficiency Check” every six months. To land after dark carrying even one other person in an airplane, you need to have done at least three night landings (in the same kind of plane) in the preceding 90 days. And these are just the rules for private pilots! They’re much tighter for the airlines.
In short, there are few arbitrary age limits in the flying world. (Airline pilots must retire at age 65, but they can still fly charters, corporate jets, freight planes, and so on.) Instead the system is based on performance. You may fly the plane, as long as you continue to show that you can fly the plane.2 It’s not how old you are, but how sharp. It would never be practical to apply the same re-testing standards to everyone who drives a car. But imagine how much safer the roads might be that way.
Still, Bill Anders managed to fly himself to his death, as about 200 US pilots do each year. (For comparison, about 120 American drivers kill themselves or others in car crashes each day.)
What went wrong? Most news coverage simply reported that Anders had “crashed” or that the plane had “plummeted” or “dived into the water.” In fact, videos of the crash (like this one) strongly suggest that he was attempting a demanding airshow-type aerobatic stunt and did not pull it off.3
Presumably Anders had flown daring “loops” of this type earlier in his career. Was his minor-but-fatal misjudgment of altitude and safety margins on this final attempt related to his age? Was his go/no-go decision even to try the maneuver also distorted by age? Or was this one more case of “Fate is the Hunter”? Some times you’re lucky, some times you’re not. Many pilots half his age, or younger, have died during similar maneuvers.
We don’t know exactly what happened, or why. And again, while it’s small comfort to the Anders family, the larger point is that the aviation system has thought thoroughly and creatively about the realities of age.
2) Southwest and the ‘Dutch Roll.’ And what is a Dutch Roll, anyway?
The Netherlands are wonderful. After our latest sojourn there, Deb and I have only good things to say.
Which makes the connotations of Dutch, used as an adjective in English, so strikingly one-sided. Dutch treat is not a treat. A Dutch uncle is a scold. Dutch courage is alcohol-fueled bluster. Let’s not even get into Dutch elm disease.
But the Dutch roll phenomenon that apparently led a Southwest 737 MAX into well-publicized trouble recently? At least as a matter of word-history it is more positive. As best I can tell (ie, “the internet”) it appears to be a flattering reference to the graceful sideways-and-forward swaying movement of Dutch skaters as they speed along the ice. As shown here (and note that in most of the video you are seeing a team of two male skaters, so closely coordinated that they appear to be one):
In the sky, a “Dutch roll” looks the opposite of graceful and can feel very bad. You can find some technical descriptions in this footnote4. In essence a plane is in Dutch roll when it is swaying (with its nose “yawing,” or turning right and left, which pivots the tail left and right in the opposite direction) at the same time it is rolling, with wings moving up and down. Managing a controlled, intentional version of this weird motion is something you learn in basic pilot training. But in an airliner it’s almost guaranteed to make people airsick. You’ll get a vivid idea from this short video, starting about 8 seconds in. Imagine sitting on this plane, especially in the back:
Why haven’t even nervous flyers spent much time worrying about Dutch roll? Mainly because in most modern planes the problem goes away before most passengers notice. Planes have elaborate “yaw damper” and “rudder damper” systems to offset such oscillations. Most airplanes are designed for “positive dynamic stability”—that is, to naturally return themselves to stable flight after being disrupted. You’ve probably been on flights whose pilots offset incipient Dutch roll or let it cure itself.
What was unusual about a recent Dutch roll episode, on a Southwest flight in May from Phoenix to Oakland, is that it appears to have damaged the airplane’s control systems. An airplane’s rudder is the key to managing Dutch roll; in many airliners, devices called the PCU, or Power Control Unit, are keys to controlling the rudder; and this week the NTSB tersely reported that the “standby PCU” for the Southwest plane had been damaged in the episode.
This could be another outlier. Maybe the PCU already had something wrong with it? Maybe something the pilots did (by stepping on the rudder pedals) affected it? Maybe some random bad-luck failure occurred? Those are the kinds of possibilities the NTSB will explore.
But as always the question will be: Are we seeing some sign of a deeper problem? The episode naturally drew the aviation-world’s attention because it involved control systems in the Boeing 737 MAX, which has lost the benefit of the doubt about those control systems. Aviation safety experts will also be fully aware that twice in the 1990s, earlier-model 737s actually crashed because of rudder-related problems. One was a United flight in Colorado, in 1991; the other, a USAir flight in Louisiana Pennsylvania [sorry] in 1994.
Probably those old problems have no bearing on the recent Southwest Dutch roll. They involved different (older) models of the airplane, different control systems, different apparent causes from anything suspected now. But the strength of the air-safety system lies in saying: Something unusual happened. And we won’t write it off as coincidence until we’ve explored every other possibility. Which they’re doing now.
ALSO: The cockpit audio from this Dutch roll flight is “not available,” because of the stupid two-hour recording limit on airline cockpit systems. (After two hours, the tape loops around and records over itself.) This is nuts and needs to be changed right away.