The crash of Colgan 3407

N200WQ, the accident airplane. Photo taken on approach at YYZ on Oct. 5, 2008, courtesy of Michael Fast and airliners.net



I've seen my share of reports on aviation accidents caused by human frailties.

There was the 1972 Eastern Airlines crash in the Everglades that occurred when a crew became so distracted by a burnt-out landing gear bulb that they were unaware the auto-pilot had disengaged and their L-1011 had begun a gentle descent into the swamps.

There was the crash of a United Airlines DC-8 that occurred near Portland Ore. in 1978, when the pilot, circling because of a potential landing gear problem, ran out of fuel. His FO had brought their fuel crisis to his attention; the captain chose to ignore it.

No less than the worst accident of all time was caused by an impatient KLM 747 pilot, who, despite the objections of his co-pilot, commenced his takeoff roll in heavy fog without clearance from air traffic control and smashed into a Pan Am 747 taxiing on the same runway, killing 583 people.

You can blame these accidents on denial, arrogance, whatever. In the end, though, the pilots knew how to fly the planes.

I'm not so sure about the pilot of Colgan 3407.

I'm hesitant to rip a flight crew, especially the deceased pilot-in-command who oversaw a terrible crash. Seems like bad karma. But reading the NTSB preliminary report, there's really no way to argue around the clear-cut evidence.

Gross incompetence cost 50 people their lives near Buffalo, N.Y.

(Not ice, as was originally and continuously lamented as the culprit in knee-jerk media reports. But when it comes to aviation news, is there another kind?)

The inattentive crew of 3407 let a bad situation fester, then when the warning bells alerted them to the situation at hand, they did the exact opposite of what they should have done to correct the problem. That's no exaggeration.

It wasn't a complicated solution, either. The fixes were things my flight students could describe and accomplish after three or four lessons in a Cessna 172. Again, I'm not exaggerating.

But it never should have gotten to that point aboard Colgan 3407, a Dash-8. As I have said a few times here, it's not one thing that causes an accident, but a chain reaction of four or five different problems that, had any been resolved differently, the accident chain could have been broken.

Colgan 3407 is no different.

Let's look at the accident chain:

The fatigue factor

1. It begins with a departure from Newark that's more than five hours late, which means a flight crew that's been on duty much longer than expected by the time they take off circa 9:30 p.m.

Fatigue, according to the NTSB report, will be a topic addressed when the board convenes in May to discuss the accident. I'm glad to hear that, because crew fatigue is a topic that gets brushed aside too often by the FAA, which although it has rules about mandated rest periods and maximum work hours, does nothing to actually enforce them.

Fatigue, you may remember, was also cited as a factor in the last major U.S. plane crash, when a Comair flight took off from the wrong runway in Lexington, Ky.

Pilots need protection from draconian schedule-makers and dispatchers, and also from their own employers. You've never met a group of employers as outright hostile to their indentured servants/employees as the regional airline companies.

It's one of the reasons I chose to abort my aviation career before takeoff; I really don't want to work for an employer who has no respect for me.

The Colgan crash is another example as to why more stringent government intervention and oversight is needed when it comes to crew fatigue. The airlines aren't going to police themselves.

Sterile cockpit

2. According to NTSB officials who have reviewed the cockpit voice recorder (CVR), the crew ignored FAA rules about maintaining a "sterile cockpit" during their approach and descent into Buffalo.

A sterile cockpit, you ask? Below 10,000 feet, commercial pilots can only talk about essential flight topics. This keeps the focus on the task at hand during departure and arrival phases of the flight, when mistakes and oversights are magnified.

(Ever wonder why you have to keep your electronics off early and late in flights? It's so operations-critical communications aren't disrupted below 10,000 feet).

Although there's no sterile-cockpit requirement for private pilots, I taught my students to adopt a similar tactic.

We weren't ever really above 10,000 feet AGL, but my rule of thumb for students was that once they contacted approach and/or copied the ATIS information (a broadcast of local weather and airport conditions), then the chatter stopped and concentration commenced.

I'm sure that the crew of Colgan 3407 isn't the only one to kind of shrug its shoulders at the sterile cockpit requirement -- on the contrary, I'd guess the rule is treated with a grain of salt by thousands of crews every day.

But when you study the accident chain of 3407, it has to be considered the second step, because there's no other way to really conceive of how they got to No. 3 without it.

Low and slow

3. So here comes 3407 into the Buffalo area, and the crew is ostensibly chatting away unaware that the airspeed is slowly bleeding off ... slowly bleeding off ... until the stick-shaker activates and warns of an oncoming stall.

That was one of the more elementary rules broken that night: Don't get low and slow.

One of the key things we're all taught -- pretty much from the first time we step in the airplane -- is to nail target airspeeds at low altitudes, because there's simply a lower margin for error when you're approaching the airport. In airspeed, you're flying closer to the stall speed of the plane ... finally approach speed is generally 1.3 times stall speed; In altitude, there's less leeway for recovery from a mistake.

On their approach, Colgan was at approximately 1,500 feet MSL, a little low, according to the report, and they got slow enough to where they stalled the plane.

This is ultimately a little more complex, and maybe we'll save for another post discussions about cross-controlled stalls, the horizontal component of lift, how g-force affects stall speeds and stall-spin accidents.

For now, moral of the story is simple: don't get low and slow.

That's what they did. They got low and slow, and it's the next chain in the accident.

Stall recovery

4. Back to our troubled plane. The stick shaker activates, warning of an oncoming stall.

The proper response in this situation is to initiate a stall recovery by 1) lowering the nose 2) applying full power and 3) retracting flaps in stages. There might be some small variations to this in jets, such as holding the nose steady instead of lowering it, but the general concept is the same.

When it came to step one, the PIC of Colgan 3407 did the exact opposite. He raised the nose.

In explaining why that's important, I need to back up and explain a few things:

You shouldn't mistake a stall for the engines quitting; in aviation-speak, a stall is code for the wings are no longer producing lift.

A stall is an aerodynamic event.

Ever stick your hand out the window and let it float in the wind? It's floating on the same aerodynamic principles that allow airplanes to fly. Lift is being created.

Lift in airplanes is affected by a lot of factors: air density, surface area of the wing, air speed and angle of attack -- the angle at which the wings meet the wind.

This oversimplifies things a little, but pilots can control the amount of lift being produced largely through two factors: airspeed and angle of attack.

(OK, when you start adding flaps and spoilers and such, they can also control the surface area of the wing, but I'm trying to keep it simple).

The faster you go, the faster the wind goes over the wings. More lift is produced.

The more "tilt" there is to the wing, the more lift is produced. If you plotted this on a graph, the amount of lift being produced would grow steadily as the angle of attack increases.

Think back to your hand out the window of the car. The more you tilt your hand upward, the more your hand wants to rise through the air. Until you reach a point where you tilt your hand so far, it plops like an anvil.

In aviation speak, that point is called the critical angle of attack. That is, the angle at which there's so much tilt to the wing, that air no longer flows smoothly around the surface. It's at this point the wing stops producing lift.

Each plane has a particular angle of attack that, if exceeded, lift ceases and the plane will stall. In most planes, the critical angle of attack is between 17 and 21 degrees.

Although airspeed plays a significant factor in stalls, it's ultimately the angle of attack that plays the most important role in the lift/stall equation.

So back to 3407.

The stick-shaker activates, enough of an event on its own to merit an investigation. The recovery procedure is simple, simple enough that my students all know it by the end of their fourth lesson.

The first two steps in stall recovery account for airspeed and angle of attack, controlling those two key things that help control lift: Power increases airspeed and the nose is lowered so that the critical angle of attack is not exceeded and air begins again flowing smoothly around the wing surfaces.

But does the captain of 3407 lower the nose, (also known as relaxing the back pressure)? Or does he maintain his current pitch, which is the proper procedure in some planes?

Nope, he yanks back on the yoke applying 25 pounds of force and pulls the Dash-8 to a plus-30 degree pitch attitude, way beyond the critical angle of attack.

This significantly worsens the stall, and with only 1,500 feet of altitude to play with, there's just not enough room for a recovery.

Raising the nose is a crazy response. It's so contrary to what's ingrained in a pilot's head from one of his earliest lessons. His response is the aviation equivalent to saying "Well, I saw that light was red, officer, so I stomped on the accelerator."

It's hard to fathom that a professional flight crew with thousands of hours of flight time and many more in a simulator erred so very, very badly.

Conclusions

I hope that I'm wrong. I hope that when the final NTSB report comes out that there's some unforeseen problem that's not in the preliminary report that explains all this.

I hope that I can come back with a post that says, "You know what? I really whiffed on my Colgan analysis and here's the extenuating circumstance that no one caught the first time around."

I'd much rather hear those things, because an accident of such simplicity scares the hell out of me.

When I think of my family and friends flying around the country and the possibility of something catastrophic occurring, I think it'd be easier to accept if a fan blade cracked because of metal fatigue or if the plane rolled after a mysterious uncommanded rudder deflection.

But this business with a bungled stall recovery? The 50 people killed by something so simple deserved much better.