Safety Publications/Articles


The surprise turn-back

Simulation is just not the same

Turning back to the departure runway following a sudden and total loss of engine power is consistently a hot topic of aviation-related discussion. Of particular interest is the determination of a maximum performance procedure and a minimum, or "target," altitude from which a turn back to the departure runway can be successfully executed. In reality, the edge of this performance envelope is fuzzy and based on individual differences influenced by human factors.

Theory versus practice and the human factor

The turn-back theory sounds perfectly logical, but something is missing. It is significant that discussion of the topic is often from the standpoint of demonstrating success rather than optimizing safety. There is a difference between a maneuver being proven possible--even easy--and that same maneuver being safe. Will there be a difference between the recovery from an intentional spin and the recovery from an accidental spin? Of course!

In determining an optimal target altitude it's not enough to analyze just the technical factors such as aircraft performance, weight and balance, and wind. You need to consider the effect of human factors and individual differences. The effect of individual differences (the ability or skill of one pilot compared to another) is the fuzzy part. Only an individual can determine his or her ability to perform at the level of precision required to fly a maneuver during a true emergency. This is why a target altitude should be personal and cannot be tabulated in a manual.

So, if an individual gets into a specific airplane, simulates the take-off conditions, and demonstrates that the turn can be made with certain loss of altitude, this should account for the human factors, right? Probably not. The demonstration omits the very important element of surprise.

Physics versus psychology

The airplane--the machine itself--has associated with it a specific performance envelope. This is physics. The airplane with a pilot on board has a theoretical performance envelope that can be demonstrated. This is physics combined with psychology. Having an individual involved will separate demonstration from reality. To be more specific, a human's reaction to the same event will be markedly different depending on who or what initiates the event.

Surprise is a factor that is most desirable and almost impossible to achieve in training and simulation. However, the effect of surprise can be easily demonstrated.

Try this. Drop a dollar bill held lengthwise with your left hand. Try to catch it from the mid-point with your right hand. Easy, right? Now have somebody else drop the bill for you. Your task is the same; try to catch it. Youll find that your reaction time is significantly different if you're not prepared for the bill to drop. Now try some variation of warning such as a count to three before the bill is released.

Reaction time

Any reaction time or performance difference experienced during the preview exercise is caused by an "event initiation variable." There are three basic alternatives:

  • The event can be self-initiated
  • There may be a prior warning, but the event is not self-controlled
  • The event may be completely random (surprise!)

The exercise illustrates a physical barrier of human performance and perception. Perception-to-motor processing takes time, especially if the event is unexpected. The process creates a chain that looks something like this:

  • Event
  • Perception
  • Interpretation
  • Reaction
  • Correction

Each link of the chain requires some finite period of time.

When a loss of power is simulated by the pilot pulling the throttle, the reaction is like the reaction to a self-initiated event. It is a two-link processing chain--event and reaction--such as the one that occurs when you drop the dollar bill yourself.

If an instructor pulls the power for the pilot, the pilot will most likely see the instructor move the throttle. Thus, an expectation is produced. This is similar to being warned that the bill is about to be dropped ("one, two, three, drop"). Expectation will have an impact on reaction performance. The interpretation or diagnoses link is removed from the chain. Reaction is slower than self-initiation but can be improved with practice. If a pilot has the self-discipline to maintain a constant high level of expectation, performance in the real world may approximate the demonstrated values--if there's warning.

Being caught by surprise is the worst-case scenario. The surprise event will produce a processing chain with the largest number of links and take the longest time to work through. If the event in question is a sudden engine failure after takeoff, any additional time taken will result in a corresponding loss of energy, altitude, and options.

If the target altitude is determined by self-initiation or with warning, the real-world variable of being caught by surprise is not being considered. The impact of this variable is an unknown, but we do know that it will degrade performance.

What to do?

You should add a healthy buffer to your target turn-back altitude to account for performance degradation caused by the element of surprise. But how much buffer should you add? The exact answer depends on both individual differences (how well will you react in a real emergency?) and the nature of the unexpected loss of power (was there any warning?). Keep in mind that the chance of surviving a straight-ahead, controlled descent into terrain, even if an obstacle is struck, is much better than surviving a stall/crash (with or without the spin). Bending the airplane should not be of significant concern in this situation.

Eric E. Geiselman is a line pilot and cockpit resource management instructor at Comair. He has been an Air Force and Navy researcher, a general aviation flight instructor, and an adjunct professor at Embry-Riddle Aeronautical University.

By Eric E. Geiselman

Back to the Index of Instructor Reports