A double tragedy: Colgan Air Flight 3407

The crash and the consequences

The crash of a DHC-8-400 (Q400) on approach to Buffalo, N. Y. brought on the all-time most egregious case of smoke and flames rulemaking by the FAA. It was dictated by Congress, it makes no sense, and it will have a lasting deleterious effect on air service to smaller cities and on airline flying as a profession.

The rule established new standards for first officers. Basically they now have to have, going in, an Airline Transport Pilot certificate and 1,500 hours total flying time. In some cases, primarily related to the nature of training, education, or military flying, the hour requirement can be reduced by a bit. The requirement used to be a commercial certificate and 250 hours total time.

Colgan Air crash scene

Colgan Air Flight 3407 resulted in “the all-time most egregious case of smoke and flames rulemaking.”

Everyone knows that new first officers on regional airlines make little money. It is barely a minimum wage job. Yet now the price of admission has been multiplied many times. One media report said there was not a shortage of airline pilots, there is just a shortage of pilots who are willing to spend that much time and money training for a job that pays so little. That makes sense.

Already, regional airlines are retiring airplanes and cutting service. They could pay pilots more and charge passengers more but the public doesn’t want that. The public wants dirt-cheap, 100-percent safe air transportation. Well, people who live in smaller cities will find less risk from flying simply because affordability will go one way and accessibility will go the opposite way. That will leave a lot of people without the option of air travel.

To me, the sad thing about all this is that what happened that night near Buffalo wouldn’t have been prevented by the new rule. There is much about that accident that cries out for change but it was not addressed in the rulemaking. It’s like that mouse running around in your house. You can go to Ace Hardware and buy a mouse trap and get rid of the rodent, or, you can use your double-barrel 12 gauge shotgun to dispatch the critter, peripheral damage be damned. Our government chose the shotgun approach in the aftermath of the Colgan Q400 accident. They missed the target but did a massive amount of peripheral damage.

The crew of the Q400 was an accident looking for a place to happen on that fateful night. The captain was properly certified but had a history of failed check rides and remedial training. He had over 3,000 flying hours and over 100 hours in the Q400. The first officer had 2,200 hours with almost 800 of those in the Q400. So, no shortage of flying time, as was suggested by the resulting rulemaking, just some truly lousy flying and management.

A reading of the accident report suggests that both pilots were tired and before takeoff the first officer complained of not feeling well. She had commuted from Seattle to Newark to fly the flight. The FAA did address fatigue in a rulemaking but did not address the role of commuting in putting tired pilots in the cockpit. A large number of airline pilots commute and often have been on the move for many hours before flying their flights.

The crew carried on a rather continuous conversation about things unrelated to the flight and the NTSB noted that all this conversation could have kept them from properly preparing for the Buffalo arrival.

Colgan animation

The captain pitched up 19 degrees, even while the airspeed decayed rapidly.

Both crew coordination and cockpit resource management (to use buzzwords) pretty much fell apart as the airplane moved closer to Buffalo. The captain was given the no-ice Vref for the approach where the system that operates the stick shaker and pusher was at the setting for operation in ice. This basically increases the speeds at which the shaker and pusher activate by 18 knots. The speed the captain had in mind for Vref was actually lower than the shaker/pusher speed for initial activation.

As the approach began, the airplane was originally going too fast. The power was reduced almost to idle to reduce the airspeed, which it did, rapidly. When it moved below 131 knots, the stick shaker activated and the trouble started. It was at this point that you have to acknowledge that thousands of hours of instructing and doing stalls in Piper or Cessna trainers would have done the crew less than no good on this night.

At 2,000 feet you can slow too much in a light trainer and not get near a serious problem. You can stall at 2,000 feet and even a ham-handed and slow-thinking pilot can recover. That is not true in a large turboprop as was evident in this Q400.

The gyrations they went through in the Q400 before it plunged to the ground were notable. The autopilot had disconnected when the shaker first activated and the pilot was hand-flying. I don’t think he had the remotest idea about what was going on because he actually pulled against the shaker/pusher three times and at one point manually put the airplane in a 19-degree nose-up attitude.

The first officer was of no help and did not question the captain’s actions. At one point she actually retracted the flaps without being told to do so.

While this was going on, the airplane rolled rapidly left and then rapidly right, finally doing what we used to call “spinning in.” The time elapsed, from the beginning of the loss of control, to the end, was about 25 seconds. If that doesn’t sound like much, time 25 seconds and imagine yourself at night, at low altitude, in an airplane you apparently don’t understand, and consider how much shake, rattle and roll was experienced in those 25 seconds.

When I read this report I thought back to an evaluation flight in a Fairchild (nee Swearingen) Metro commuter airliner. It is a much smaller airplane than a Q400 but similar in a lot of ways.

I did an approach to a stall, got the stick shaker, and followed the recovery procedure of reducing angle of attack and increasing power. At that time the goal was to lose no altitude in a recovery from an approach to a stall. Since that time, someone awakened and it is now okay to lose a little altitude if it results in a quicker recovery. Just don’t lose a foot more than is available.

I asked the demo pilot if the Metro had a shaker/pusher because of bad stall characteristics. Mistake! He turned the system off and invited me to stall the airplane and see for myself.

The airplane broke sharply and rolled rapidly. Clearly, this recovery would bear no similarity to the one used for recovery from an approach to a stall. I rapidly reduced angle of attack, got a windshield full of Texas, and then recovered from the dive.

I offer that as an illustration that if one of these airplanes is truly stalled at a relatively low altitude, chances of a recovery are from slim to none. That is why they have shaker/pushers.

The crew this night was at a significant disadvantage. Both were tired. The first officer had made that long Seattle-Newark trek and she had told the captain that she didn’t feel well. The captain hadn’t had normal rest, either, and the time of the accident was at about his normal bedtime.

I thought about a long commute I once made when thinking about how sharp the first officer might have been. I left London one morning on Concorde, got to New York at an earlier hour local time than I had left London, and was in my office at One Park Avenue not long after 9 a.m. To be honest, I didn’t get a lot done that day and took an early train home to New Jersey that afternoon. I would have hardly been effective in the cockpit of an airplane and my commute was a lot shorter in time than the first officer’s.

The captain of that Colgan flight had those other issues as well. He had, over time, failed four FAA check rides for a certificate or rating and had three unsatisfactory airline check events. Nobody is perfect and anybody can have a bad day and perform poorly on a check ride but seven times and you become a captain? Just think how the public would react if a person had seven drunk driving arrests and wasn’t walking.

Flight instructor

Does 1500 hours instructing in a piston single do anything to prepare pilots for airline flying?

The 1,500 hour requirement for first officers is purely ludicrous when you consider how little good light airplane time might do in a heavy airplane cockpit. I heard a USAF training person once say that light airplane training is valuable to them, up to a point. Past that point, it has little additional value. The USAF actually uses GA airplanes in a 25 hour screening/introductory program.

Consider the fact, too, that a USAF pilot does not have many more than 200 hours when winged and I am sure there are many who fly out a tour of duty without amassing that supposedly magical 1,500 hours.

If anyone had been interested in doing something constructive instead of something for show, the rulemaking would have not been so unrealistic.

It is quite logical that a type rating should be required before a pilot could act as a first officer in Part 121 airline operations. Taking away the artificial flying time requirement would free professional training institutions to do what they do best, prepare people for a specific role. Schools often pair with airlines so programs could be developed to prepare pilots for positions with those airlines, flying the airplanes that they use. This is done to some extent here now, and it is widely done in other countries.

The role of commuting and fatigue should be addressed, too. There are valid reasons for a pilot to commute but there needs to be a framework of how much in terms of time or distance is too much. To allow coast-to coast commutes without prescribed rest between that and duty is just too much.

When you look at the captain of the Q400 it becomes clear that this was not pilot error, it was system error. There is simply no system in place to weed out the weaker pilots. Aviation may be unique in that a person can try and try again as many times as it takes to pass the test. That might be okay for most pilots but not for airline captains. The public expects and deserves the best there. The FAA and the airlines need to develop a system that ensures that.

Clearly the Congress/FAA reaction to the Q400 accident was a gross error. But do you think they will ever fess up to the mistake and fix the real problems?

Finally, here is a challenge for your imagination: What kind of rulemaking furor do you think would come from our reactive Congress/FAA were a U. S. carrier to be involved in an event like that in Malaysia?

19 Comments

  1. Seth says:

    Commuting is covered by the Fitness for Duty rule, which is a shared responsibility between the individual pilot and the certificate holder (airline). So to say that commuting wasn’t addressed isn’t entirely factual. The FAA just stopped short of creating a matrix of commuting factors and how that might limit your Flight Duty Period. But I think they did that intentionally so that pilots would simply assess themselves as fatigued, or not fatigued. This Go/No-Go aspect of Fitness for Duty makes sense, as a pilot may feel pressured to take a flight if they technically meet the requirements of a commuting matrix, but are feeling fatigued. The rule also adds pressure to the certificate holder by making them equally responsible for the pilots Fitness for Duty. I don’t think they’ll want to take the shared risk and liability of a fatigued pilot and we’ll see airlines impose company rules that limit commuting distances/times to reduce their risk levels.

    • Dave Oberg says:

      Self-assessing addresses the problem? Really? What new hire on probation, that wants to keep his/her hard-won job in the right seat, is going to tell their employer that they showed up to work unfit to fly? The whole idea is ludicrous!

  2. Jeff Izo says:

    I will happily turn this argument around. It is tragic that Colgan happened, but finally the regional airlines are going to have to start paying pilots a wage that is fair relative to the responsibility and training required of them. This week alone, my regional airline upped the stating pay for an FO by seven dollars an hour, and they’re still having a problem keeping pilots. I was an Air Force pilot, and yes we were able to fly large aircraft with less flight time, but we also had superior initial training under a situation where one or two major mistakes means you were kicked out of the program. When the average university graduate is starting out flying a 300 knot jet, doing aerobatics. low level flying, instruments and formation flying, I’d buy the 250 hour argument…..but they’re not. Most of my fellow regional pilots who came up the civilian route said a 250 hour pilot in the right seat of a regional aircraft was dangerous because they do not have the experience. the ab initio schools, and the “250 hours to the right seat” idea was created so that the airlines would have cheap labor. It was done so they could up the pilot supply and then work those pilots like slaves for ten years for horrible pay, and create a “Due paying period” before they could move up to the majors. These pilots were paid next to nothing while having a huge amount of debt for the “opportunity” to move up. And let’s talk about what happened with that whole situation with “moving up”. All the senior regional Captains I fly with now, paid 100,000 bucks or more for their education, and spent 10 years or more at slave wages while the Union and the majors sold them down the river by changing the age 60 rule to 65 to give the wealthiest pilots another five years at the six figure range. Now, to even get hired by a Major you need close to 6,000 flight hours. The other thing is that middle time pilots who had 2000 to 3000 hours couldn’t get hired by the majors or the regionals because the were too qualified for the regionals and not qualified enough for the majors from about 1998 to the present. I was one of those pilots. An Air Force pilot who, after seven years in the military only had 2000 flight hours. I couldn’t find aviation work for three years as a result of how badly managed and manipulated the aviation industry has become. The rule change to 1500 hours was created to for the industry to have to pay their pilots a fair wage. It seems to finally be working.

    • John Zimmerman says:

      What a truly awful way to negotiate a pay raise. I don’t dispute that regional airline FO’s are poorly paid, but what kind of precedent does it set if we turn to Congress to fix that? Should auto workers or police officers try this route?

      If that’s the only purpose of this law (not sure the families of 3407 victims would agree), then it’s sad.

  3. Jeff Izo says:

    Correction…2008 till the present…not 1998

  4. Richard says:

    as far as the 1500 hours, the FAA only did what congress mandated them to do. Public law 111-216 section 216 require an ATP and section 217 the 1500 hours to get the ATP. Here is a link to the law that congress passed, and the president signed. The FAA had no choice but to write the regulation. http://www.gpo.gov/fdsys/pkg/PLAW-111publ216/pdf/PLAW-111publ216.pdf

    Do not get me wrong, the FAA has a fine history of knee jerk / smoke and flames reactions, but this one is on congress. Go back and watch the appropriate senate and house sub committee hearings on the colgon flight and ALPA telling congress there is two levels of safety and you might get a good idea on where congress got the idea that more hours is a good idea. Pay may come with experience and Safety may come with experience, but safety should be a mine set, that is what we really need to train and build experience in.

  5. Old Bob Siegfried says:

    Good Afternoon Richard,

    If I had written something about this subject it would have come out as just what you have written. The Captain should have been washed out years before.

    And commuting should include a 24 hour rest period in base before getting onboard as part of the working crew.

    I once had a brand new copilot in a Boeing 720 who had less than 250 hours total pilot time when he was assigned as my copilot. Did a good job and ended his career as chief pilot at our San Francisco Base.

    He did have one year as a B-720 Flight engineer and I am sure that helped, but still a very low time pilot. Quality shows both ways!

    I had 850 hours when I started as a new DC-3 copilot and managed to do OK.

    Completely counter productive rule!

    Happy Skies,

    Old Bob

  6. Herb Jacobs says:

    Longevity does NOT mean proficency. There are 200 hr pilots who cannot fly and 2000 hr pilots who cannot fly.. There are 200 hr pilots far more skilled than some 2000 hr pilots. Some have “it”. Some do not. There is no easy solution.

  7. Bob Brewer says:

    Congress is notorious for getting it wrong most of the time. This one may be on them, but the FAA did little to fight this ludicrous rule when they know from experience that total hours is the WORST way to measure proficiency.

    Screening for aptitude, intensive training and the quality of the flight time are much more accurate measurements of future performance.

    The military has been proving this for years and Flight Safety did it with their Airline Transition Program back in the early 90′s.

  8. Frank V says:

    The 1500 hrs may be too much if other problems were not addressed, such as rest time, commuting time, proficiency, etc. That said, 250 hours was far too little.

  9. James Macklin says:

    Pilot training is rarely done by highly experienced CFIs with a logbook full of 135 single and two pilot charters flights. Most pilot mills operate on a very tight budget, using the newly minted 275 hour CFI who just completed the schools FAA approved “professional pilot/CFI course.”

    That means perhaps only six to eight months from 0 time to CP/CFI ASMEI and never in or near a thunderstorm, icing conditions or all the other things experience might bring.

    I’d be willing to bet 98% of all instrument instructors tell their students what to do when a hold is issued. It starts with slow down on the way to the holding fix to spend time enroute and not just in the racetrack.

    So like Florida Air’s 737 pilots who failed to turn engine anti-ice ON causing the DC crash, the typical new hire in an airliner slows down to save expensive fuel.

    But aircraft approved for flight in icing conditions have a minimum speed [angle of attack] necessary to keep any ice on the protected surfaces. I won’t say it is listed in the POH as a limitation, but the Beech Baron BE58TC/P list 140 KIAS as the minimum holding speed in icing conditions which Beech also say as +5°C and lower in visible moisture.

    Keeping the speed up, the flaps retracted and having all ice protection turned ON [with inflatable boots you have to wait until 1/2" of ice accumulates before you pop the boots.] With a turbine engine, engine anti-ice and if available, ice vanes extended, with a de-ice system wait until the correct amount of ice has formed.

    Basic flight skills seem to take a back seat to radio and restricted airspace. Stall recovery, dead reckoning navigation and use of backup aids [helps to avoid landing a 747 on a 6,000 foot single runway instead of the airport with parallel 15,000×500 foot intended airport a few miles ahead.

    The new LSA aircraft could be useful, except most are not approved for spins, so CFI training, taildragger, that old skill called pilotage, and full use of the outside world indicators are skills that are not part of the FAA approved and mandatory approved school course.

  10. John Brundage says:

    A number of years ago NASA did some research on ice induced tail stalls & produced a training video about them. They said that an ice induced tail stall was most likely to occur when the initial flap extension is made at near the maximum flap extension speed & would result in a drop in nose attitude. Their recommended recovery procedure suggested pulling back on the yoke & reducing the flap setting to the previous setting.

    The NTSB produced a computer generated video of the Colgan accident that shows the instrument panel & control inputs. It shows the stall occurs during the initial deployment of flaps. The occurrence of the stall at the time of initial flap extension, the FO’s decision to raise the flaps, & the captain’s decision to pull back on the yoke would lead me to believe that they thought that they had a tail stall. The fact that the stall occurred well below the max flap extension speed & the activation of the stick shaker should have been an indication that they didn’t have a tail stall, but a wing stall.

    Once NASA told us that tail stalls as well as wing stalls were possible, with two completely opposite recovery techniques required, simulator training should have been changed to include multiple exposures to both conditions, with the goal of correctly identifying whether one had a wing or tail stall & applying the correct procedure. This had not been done at the time of the Colgan accident, & maybe it still hasn’t been done.

    For years FAA approved training programs for large aircraft have included stall recovery training consisting of setting a low power setting & holding altitude until the 1st indication of a stall. This results in a stall at a slight nose up attitude. The recommended recovery procedure was to hold that slight nose up attitude at the onset of the stall & apply full power. When a stall entry is set up this way, this recovery technique works. Test pilots complained to the FAA for years that this recovery technique was potentially deadly for real conditions where a stall may occur in some other configuration than level flight at a low power setting or stalls that occur at high altitude. After the Colgan accident this issue was addressed in an AC.

    As mentioned in the article, turning on the wing & tail boots on the Q 400 automatically increases the stick shaker/pusher speeds by 18kt & that the captain was given the no ice Vref for the approach. For this airplane, the approach checklist probably should have an item, “Vref deice on or off.” Was such an item on the checklist? Is it now? When the airplane stalled both airspeed indicators agreed & indicated a speed high enough that the pilots would have thought it couldn’t be stalling, but a tail stall can occur well above normal wing stall speed.

    When two instruments read the same in a case like this, it can be very confusing & one may only have a few seconds to figure it out. Northwest lost a 727 years ago when the crew forgot to turn the pitot heat on. During climb both pitot tubes iced over causing both airspeed indicators to read too high & to read the same. The crew apparently thought the airspeeds must be right, since they were both reading the same, which led to a stall & loss of the airplane.

    Maybe a different crew would have handled this Colgan flight better & maybe not. I am sure the crew could have benefited from better sim training on identifying ice induced tail stalls vs. wing stalls, more realistic stall training as recommended in the new AC, & maybe from the addition of one checklist item.

    • Old Bob Siegfried says:

      Good Morning John,

      Very complete and very true.

      The time I first saw that tail icing video, I figured it was going to kill someone, some day.

      The two Colgan pilots were discussing the NASA video before the crash. Combine that with a couple of weak and tired aviators and the disaster was set.

      I am definitely not as knowledgeable as the NASA folks, but I do know that we have been training for stalls wrong for at least the last sixty years. It has always been my contention that airliner stalls should be taught out of descending turns. I flew the 727 after the tail anti icing had been rendered inoperatve. We were taught about the potential for a control reversal being needed to recover from a tail stall. It is my contention that the recovery from a tail icing incident or a true stall is better accomplished in the same manner.

      1)Get the nose down and add power.

      2) Stop the descent before you hit the ground.

      Minimum loss of altitude in a stall is good, but is not the only criteria. The Colgan accident and the Air France disaster could both have been averted had the crews had and heeded pre 1956 style stall training. The simulator training style which was initiated after we started to get simulators worked fine for passing a check ride, but is lousy for scaring us into avoiding the stall!

      Don’t get me wrong, I think simulators were, and are, a very good thing. We learned a lot about what did and did not work well, but we lost some very important things that we had been learning by actually stalling those air carrier style flying machines.

      Nuff rambling from this old guy, but all that is ancient is not bad.

      Happy Skies,

      Old Bob

  11. Dan Jenkins says:

    First, let me say that there is no question about who is at fault here. The pilots are and I don’t dispute that one bit. However, as always in an accident, we have to ask why they did what they did. There were numerous factors involved and all played a part. However, Richard, your comment about the knee jerk or shotgun approach was wrong. 250 hours to be in the cockpit of an airliner is ridiculous at best. I know you have a fantastic aviation career, but how much was spent in the cockpit and the environment of airline operations? The 1500 hour rule is long overdue, no matter how much it costs to acquire those hours. If the airlines need pilots bad enough, they will train them (I’ve seen it happen). Commuter pay and first year pay needs to be addressed, there is no question about that. So, in the case of public safety, it’s realist for the government to get involved in this “initiation” pay the companies offer. I know airline operations reasonably well. I spent 32 years in the industry with one of the best companies there is.

  12. Joe says:

    Hi All,

    What I find the strange is the required hours can be lower if you have an Aviation Degree. I was a helicopter gunner in AF Special Operations for 14 years, with 3500 hours, most of that on night vision goggles. I have flown all over the planet, with at least 90 different pilots from all walks of life, with varying college degrees and I can tell you without a shadow of a doubt the degree they had made no difference on how good a pilot they were; you either have it or you don’t.

    Some of the worst pilots (not all)held college degrees in aerospace engineering and advanced mathematics and some of the best held degrees in physical education, mineralogy and art history; there is just no way to tell from the degree they held how good of aircraft commander / copilot they were going to be.

    Having an Aviation Degree in your pocket is not going to help you with the basics like, Airspeed!

    Joe

    • Old Bob Siegfried says:

      Hi Joe,

      I agree. It is attitude and aptitude that count. I have had lots of brand new copilots with low time that did great and eventually made fine captains. ‘Seen others that should have chosen some other career.

      Military or civilian, high time or low time, college educated or hgh school dropouts did not seem to make any difference. Some folks are just cut out to be a professional pilot and some are not.

      Attitude, attitude, attitude.

      Happy Skies,

      Old Bob

  13. Dave F. Ryan says:

    Okay the comment relating to 1500 hours in a piston single is wrong. An applicant seeking an ATP in most world countries, the 1500 hours ruling is in effect. But it’s not gained in a single engined aircraft. The starting point is either one of two Commercial Pilot Licence avenues. Most schools are adding a degree in business management in aviation study’s. The training covers 250 hours and includes a twin and instrument rating. Along with that is the Airline Transport Pilot rating exams. These are directed towards those pilots desiring to fly heavy aircraft or those entering an airline. Airlines have operational policies that differ from charter companies flying the same type of aircraft. Now when the applicant as passed the exams; which usually cost $1200+ to sit. The applicant is issued with a Frozen ATP licence. Once the magic 1500 hours has been achieved, the licence is un-frozen. Along the way to the right hand front seat on a heavy. The applicant may have flown Barons, KingAirs or Citations as Pilot In Command before an airline accepts the application. So in effect, said applicant many have logged 2000 hours in command.

  14. Jerry Lawler says:

    Two factors in this unnecessary accident. Regional pilot training is outsourced to training organizations that may never step foot on the airline. In America, major airlines have in house training by fellow employees. A check pilot is not going to sign off on a weak pilot that may some day fly his family. Many foreign carriers like Asiana also have outsourced training. Number two and in my opinion more important, the crew violated the sterile cockpit rule. If the captain was not engaged in conversation with the first officer during the approach, there is a good chance he would have paid attention to his airspeed and situational awareness.

    • Dave F. Ryan says:

      Something that is coming out of this, is: ATP’s flying DC-9′s ATR- and non heave aircraft for and airline as a regional. Are being scripted to the GA image when seeing other aircraft as in B47′s 67′s and 77.s along with A320′s and so on. Having an ATP and holding behind a heavier aircraft a crew member wants to be in change of, is there a scripting tendency of sub-coniously relating to being in a lighter aircraft as instructor student? Thought energy creates a future reality. Flying an Aero that isn’t a heavy with a heave licence, does that project an attitude of this is a glorified Citation or some other type. Does the training now need to impart having a respect for the aircraft regardless of weight. The operation being a smaller carrier operation is to be seen as having the same level of professional insight as the crew of a much heavier aircraft on the same airport.