Air Asia: AF447 In A Different Hemisphere


Ho-hum, another month, another accident in which the aviation industry can get its pants snagged on the notion that cockpit automation is turning us all into numbskull idiots with hands of stone. But what if the pilot is already well on his way to being a numbskull? Does the automation push him over the edge or draw him back from the precipice?

I was contemplating this when slogging through the accident report (PDF) on Air Asia QZ8501, which we published earlier this week. And here, reader discretion advised: Being a pilot and all, I try not to be too harsh on accident pilots because you know the prose having to do with God’s grace. (If there is a God, but that’s another subject.) To extract the core sample from a lot of muck, here’s what happened: Last December 28, Air Asia QZ8501, an Airbus 320, was en route from Surabaya to Singapore when it summarily plunged into the Java Sea. No survivors.

As you can read in the accident report, this was—as usual—caused by a chain of events. But it may be one of the shortest chains in aviation history. The airplane had a minor technical fault related to a malfunctioning rudder travel limiter, a device that in no way immediately affected airworthiness. Bottom line: The First Officer, assisted by really bad CRM, allowed a perfectly good and utterly recoverable airplane to descend into the water. Readers of major accident reports—I am one—will have seen this movie before. It’s all but an instant replay of Air France 447, an Airbus 330, which plunged into the Atlantic in 2009. A protracted and determined accident investigation shocked the industry: A flight deck cheek-by-jowl with trained pilots flew a perfectly airworthy airplane into the ocean with the same outcome as Air Asia 8501 would suffer six years later.

Understandably, that ignited months of hairy human factors discussion about how automation has become so complicated that pilots just can’t cope with it under duress. While I think there’s merit to this argument, I’m also wondering what the automation has to do with simply flying the airplane, interpreting an instrument—mainly an ADI—and keeping it from plunging uncontrollably into what lies below. Isn’t this pretty basic stuff? It has nothing to do with troubleshooting ECAM messages or paging through the emergency checklist. It’s just flying and at least one pilot is always supposed to be doing that.

Following AF447, widely labeled as an industry wake-up call, wouldn’t pilots be getting training specific to this particular problem? I asked this question of a friend of mine, who’s done plenty of tin-kicking in major accidents. He pondered it for 10 seconds and said, “Nah, they probably assume no one could ever be that dumb again and don’t train it.”

This was more or less confirmed when one of the Air Asia executives was asked if the First Officer on 8501 had been given upset recovery training. He said he had not. This makes a perverse sense to me because it comes from the not-unreasonable assumption that a pilot on the flight deck of a transport category airplane would know enough not to turn a perfectly normal flight regime into a departure all on his own and from whence he would then be expected to recover. According to the accident report, he did have help. The Captain demonstrated how not to perform CRM by not assuming control, which he was supposed to do, and telling the FO to “pull down” when he meant “push forward.” Still, even at that, wouldn’t a pilot with reasonable training have gasket-blowing alarm bells going off in his head as he pulled an airplane through 40 degrees of pitch up? (It eventually reached 48 degrees of pitch and 104 degrees of left roll.)

Evidently not. This makes me wonder how the First Officer developed his stall sense. Was it from the visceral feel of watching the nose break in a little airplane or hearing the alarms in a simulator? Maybe more the latter than the former. The accident report noted that the FO had 2247 hours, with most of his experience in the A320. For that amount of time, he was older that you’d expect: 46. Where I’m going with this is wondering if someone who came out of light or at least small aircraft experience has a better feel for when the secondary cues—instruments—need to be consulted for stall awareness, rather than relying entirely on instruments that give but an abstract of what the wing is doing.

Does that experience convey more of the why you’re doing something rather than just doing something because the electronic script calls for it? That’s idle speculation, I’ll concede, but one thing is certain: Having entered the stall regime, neither pilot seemed equipped to sort it out, neither the FO flying, nor the Captain advising him.

Not that this kind of accident is anything new, either for the airline industry and certainly not for general aviation. If you’re a real connoisseur of accident history, you’ll recall Northwest 6231 on December 1, 1974. The 727’s pitot tubes iced over because the crew forgot to turn on the heat, turning the then-steam-gauge ASIs into altimeters. The pilots pulled the airplane into a stall trying to reverse the rising airspeed and spun it into the ground. If a pitot tube heat switch is automation, then maybe the automation contributed. But … not really.

The pilots of Air Asia 8501 had functioning instruments and, ultimately, so did AF447. The trick is to interpret them and act accordingly.

Let the flogging of the automation horse begin.

The Other Side of Airmanship

Thanks to Amazon’s Machiavellian marketing machine, my Kindle app somehow gets loaded with books I don’t even remember ordering. One I just finished is James Holland’s Dam Busters. This is not a new book, but it’s the latest (2012) written about the legendary RAF 617 Squadron, which breached two major dams in Germany’s Ruhr region in May of 1943.

I’m sure you know the story well, but the book is rich with additional technical detail that makes it well worth the read. One aspect of the story I didn’t grasp is that 617 dispatched 18 Lancasters to attack three dams. To get to the target, they flew the entire trip at extreme low level—like 100 feet or less, at night and purely by pilotage. For the North Sea crossing, they could climb briefly to use the RAF’s GEE radio navigation system to grab quick fixes, but it wouldn’t work over Europe at low level.

Reading about this gave me sweaty palms. Here’s the context. Take off at night from Indianapolis and fly, at 100 feet or less, to the Fontana dam in the Tennessee Valley system, 350 miles away. It’s about the same size as the Mohne, which 617 breached first. Those Lanc pilots and flight crews had far less total time than the average AVweb reader; like several hundred hours at most. Do you think you could find the dam with a chart, a plotter and a watch from 100 feet at night? I’m gonna go with no, I couldn’t. In a Skyhawk, much less a 30-ton Lancaster.

Of the 18 crews, only one got lost. But two crashed with the loss of all hands after hitting power lines and one returned to base after inadvertently scooping up several tons of the North Sea in the open bomb bay. Six were shot down, making it one of the costliest raids of the war, but also one of the most enduring.

AVweb Insider is a blog and a statement of opinions and commentary on current issues in general aviation. We welcome alternate points of view and guest blogs.