Daily flight-time limits and minimum crew-rest requirements have been established to reduce the potential for aircrew fatigue. Examination of the flight crew's flight time records for the month of December 1985 determined that the flight-time limitations of FAR 121.521 had been exceeded twice. In the 24-hour period commencing 0206 GMT, 05 December 1985, the flight crew's flying time was recorded as 13 hours 22 minutes, that is, 1 hour 22 minutes in excess of the 12-hour maximum. In the 48-hour period commencing 1018 GMT, 03 December 1985, the flying time recorded was 22 hours 24 minutes. Following this, only seven hours elapsed before the crew initiated its next flight. FAR 121.521 requires that a minimum of 18 hours crew rest be given when a flight crew member has been aloft for more than 20 hours during any consecutive 48 hours.

A review of FAA special surveillance reports determined that, on other occasions, Arrow Air flight crew had exceeded the requirements of FAR 121.521 with respect to flight-time limitations and crew rest.

It was the stated intent of the flight crew to ferry the aircraft to Oakland, California on completion of the night to Fort Campbell. The Board estimates that, at the completion of this flight, the crew would have accumulated about 15 flight hours in the 24 hours commencing with departure from Cologne. The crew's duty day would have approached 20 hours. Because the flight to Oakland was to be conducted without passengers, it was not considered an FAR 121 flight. Rather, it was to be conducted under the provisions of FAR 91. FAR 91 does not include any flighttime limitations or minimum crew-rest requirements. Thus, the flight could be conducted within the provisions of applicable FARs.

By scheduling non-revenue ferry flights under the provisions of FAR 91 at the completion of a series of FAR 121 flights, flight-time limitations and crew-rest requirements designed to reduce the potential for aircrew fatigue can be circumvented. The Board can find no reason to justify the absence of such limits and requirements for flights conducted by FAR 121 certificated air carriers under FAR 91.

To a large extent, the prevention of flight crew fatigue is dependent on the scheduling practices and policies of the air carrier. In the United States, the FARs provide a framework within which the carrier must operate; however, it is incumbent on the carrier to devise workable policies that meet the operational needs.

The pilot-scheduling policy developed by Arrow Air makes no reference to flight-time limits, duty-day limits or minimum crew rest It was determined by the Board that company scheduling procedures did not address flight crew fatigue factors. No maximum duty-day limit was established.

A detailed analysis of available information pertaining to each flight crew member's vulnerability to fatigue was undertaken. Consideration was also given to identifying behavioural evidence that could be attributed to fatigue. and the causal sequence of events leading up to the accident.

It was the opinion of the medical expert who testified at the Board's public inquiry that, in the 12 days leading up to the accident, the flight crew had been consistently exposed to work patterns and fatigue-inducing factors which were highly conducive to the development of chronic fatigue. These factors included short layovers, night departures, multiple time-zone travel, and a flight-hour accumulation of almost 57 hours in the previous 10 days.

There are no accepted toxicological tests which can verify the presence of, or quantify the influence of, fatigue. However, research has empirically identified certain fatigue-induced behaviours and associated performance decrements.

An analysis of what was known of the flight crew's behaviour while in Cologne, during the flight, and while on the ground in Gander indicated no clear behavioural pattern that could be associated with fatigue. As a result, the Board could not determine whether any individual flight crew member was in fact fatigued nor establish any cause and effect relationship between probable fatigue and the accident sequence.