I will start this article with a short foreword. Although it may sound as though I am being critical of airline pilots, that is not the case. I believe there is a consequential gap in the way airline pilots are currently being trained. Yes, there are exceptions to this rule and there are some great airline pilots out there. My good friend Stan Humphrey in Galveston, Texas is one of them. When he is not flying a Boeing 737, he is out doing loops and rolls in his personal Starduster bi-plane. There are plenty of airline pilots out there who I would (and do) trust my life to. That doesn't mean that there is not room for improvement in our training programs.
There was a time when the first flight lesson for a pilot-to-be started by tying off the tail of a Stearman and hand propping its old 220 horsepower radial engine to life. The instructor would climb into the front hole, and the student in back. Most of the flight was briefed on the ground, because there were no headsets or intercoms in these days, which made communication once airborne nearly impossible. The student would often be flying 'solo' within just 8 hours of dual instruction. When the U.S. Military was cranking out pilots to keep up with the war effort, pilots with just 200 hours or less were being put into the P51 Mustang fighter and sent off to combat. These were the days when pilots were .... well .... pilots. They figured a lot of things out for themselves. It took a certain type of personality and mindset to become a pilot. In today's 'technologically advanced aircraft,' as they are so affectionately referred to as, pilots are no longer pilots. Instead, they have become computer operators. We can train almost anyone to become a pilot. Our stick and rudder skills have been long since forgotten, and pilots either don't know how or are afraid to 'think' through a situation. They have been trained to refer to a book when anomalies are encountered, and work through anything with the presumption that the computers are always right. They are taught that if the procedure isn't clearly delineated in a manual, it cannot be authorized and should therefore not be considered. I will cite the AirFrance 447 and Asiana 214 crashes as examples of what happens when we allow computer operators instead of pilots to fly airplanes.
Let's take AF447. The summarization is: Airbus flies into an intense high altitude storm which iced over the pitot/AOA tubes, causing air data discrepancies and an autopilot disconnect. The 'pilots,' who had a combined 20,000 hours of experience put the plane into a deep stall and held it there for over 40,000 ft until they impacted the water. So, how could such seasoned pilots make such a 'rookie' mistake just because of a little conflicting air-data information? How could they fail to understand one of the simplest concepts in aerodynamics? This concept which is drilled into student pilots by the third lesson. Let's analyze 20,000 hours of airline experience. In a typical 10 hour long haul flight in an Airbus, the autopilot is engaged just seconds after takeoff. The autopilot remains engaged for the entire flight, usually through-out a coupled approach, with the pilot disconnecting only a few hundred feet above the runway for touchdown. So, how much of that 20,000 hours was hand flown? Perhaps 1 or 2 percent? Most of that time is spent sitting straight and level cruise, with the pilots just working the radio, changing frequencies and checking in with air traffic controllers. The truth is that our current training mantra is 'USE THE AUTOPILOT,' and 20,000 hours of twisting knobs and pushing buttons does not make you a pilot. Another excellent example is the recent Asiana 214 accident in San Francisco. These 'pilots,' and it nauseates me to call them that, were actually 'scared' to fly an approach by visual reference. Why? Trading airspeed for altitude is lesson number 2 in your initial pilot training. I know student pilots with 4 hours in their logbooks who can hand fly a better visual approach than a 10,000 hour airline pilot. The airline pilots might be able to rattle off every speed and button sequence from memory, but often times they do not understand the 'why' behind the answer to a question.
So, what is the solution? I believe we need to get back to the basics. I am type rated in a small jet, which has the same training requirements and checkride as the airliners. When speaking with other pilots rated in the same aircraft, I am shocked to learn how they were trained to engage the autopilot as soon as possible and disconnect it as late as possible. I was once flying the right seat with another Eclipse rated pilot who had an ADC (Air data computer) failure which disconnected the autopilot while getting vectored for an instrument approach in IMC. He nearly put the plane inverted while trying to set the approach up. I stepped in, and got him back on track, but even once on the ILS, he nearly went full deflection several times. It was apparent that he had been using the autopilot as a crutch and was not actually capable of hand flying the aircraft.
The airlines need to take a lesson from the military training program. With just 100 hours in a Naval Aviator's logbook, they are doing actual aircraft carrier landings, both day and night time. Airlines need to put more emphasis on ACTUALLY flying the airplane and reinforce the abandoned concept that pilots are to be decision makers and problem solvers. Perhaps a 1 hour aerobatic flight in a Pitts to keep stick and rudder skills sharp should be required as part of our annual 61.58 checkrides.
A monkey can monitor systems. I want a pilot flying my airplane.