Spatial disorientation, in-cockpit distractions, and icing are all possible causes of the Dec. 20, 2001, crash of Eagle Air Ltd. flight EAB 220. The Cessna CE 560 Citation V took off from Zurich Kloten airport on a night ferry flight to Berne-Belp. Fog patches had formed at the snow-covered airport. The air temperature was -9 degreesC (16 degreesF). Immediately after takeoff, the aircraft lost altitude, crashed near the runway, caught fire and skidded across the frozen ground to a nearby runway. Both pilots were killed. The German Aircraft Accident Investigation Bureau (BFU), which has just released its report, reached its conclusions before the recent revelations about the criticality of business jet wing leading edges and frosted icing (see ASW, Dec. 6, 2004). The BFU attributed the accident to yet another human factors phenomenon.
The flight took off at 21:06 UTC and the aerodrome controller immediately requested the pilot to change to the approach control west (APW) frequency. Very shortly afterwards, the controller noticed that the airplane’s radar echo had disappeared from the display. The digital flight data recorder (DFDR) confirmed the initiation of a recovery maneuver. The investigation also showed that at initial impact the aircraft had an attitude of 1-5 degrees nose-up (ANU) and approximately 5 degrees right wing down.
The failure of the cockpit voice recorder (CVR) hampered the investigators. The tape’s last recording had been made on a date some 20 months before the accident. The drive belt from the motor to the tape transport roller had skipped off the drive roller. Thus, no new recordings were possible.
The visibility prevailing at takeoff was at or below the minima specified for a runway with edge and centerline lighting (i.e., 150m, or 500 ft.). Due to ATC oversight the crew was not advised by radio or Airport Traffic Information Service (ATIS) that Low Visibility Procedures (LVP) were in effect, or of the very low Runway Visual Range (RVR). Weather conditions were conducive to icing on taxi, and more ice probably formed on taxi as a result of the freezing fog. On line-up, meteorological visibility was 100m (325 ft.) in partial (i.e., not homogenous) fog. The copilot did not have any LVP training or prior experience in night instrument takeoffs into weather.
The anti-icing protocols from the company flight operations manual included a caution that, “As a guiding rule, takeoff with even the smallest amount of ice or frost on any part of the aircraft must not be attempted.” The accident report noted that: “Apart from the impact of supercooled fog droplets which would freeze relatively rapidly on the surface of the aircraft, under these conditions (an almost clear sky, rapidly falling temperature) the formation of hoar frost due to sublimation is also possible, especially in the case of aircraft whose surface temperature is already significantly below zero degrees C.”
After engine start, the commander was observed using a scraper and spray aerosol to remove ice deposits from the left wing. This was a cost-saving standard within the company.
Flight Instrumentation
Instrumentation corresponded to the basic equipment for operating the aircraft under instrument flight rules. Notably, however, this was configured predominantly for the pilot in the left-hand seat:
- EFIS equipment was on the commander’s side only.
- Parameters of the navigation management system (NMS) were displayed mainly on the commander’s flight instruments.
The most significant difference in instrumentation between the two sides of the cockpit concerns the most important instrument for IFR operations. The commander’s and copilot’s artificial horizons (attitude indicators) were of different types and from different manufacturers.
The commander had a Honeywell ED-600 electronic attitude director indicator (EADI) on his side to represent the aircraft’s attitude. An AIM electro-mechanical attitude indicator was installed on the copilot’s side. The commander’s EADI measured 103 x 93 mm, while the copilot’s attitude indicator measured 79 x 63 mm. (For the considerable difference in visual intelligibility and scanning prominence between the two horizons, see above graphic. Also see ASW, Apr 5, 2004, “Getting Your Attitude Together”). The electro-mechanical standby horizon was installed immediately adjacent to the copilot’s attitude indicator. This independent device was used as a reference attitude display in the event of any perplexing deviation between the two pilot’s artificial horizons.
Flight Details
The crew was under enormous time pressure to reach the destination airport before its closing time, which was rapidly approaching. The delays prior to takeoff due to unfavorable meteorological conditions on the ground, plus the grouping together of departures because of the nighttime restrictions on use of the Zurich runway system, further increased the time pressure and stress upon them. Their eventual departure reflected a 40-minute delay. It was known that the company CEO had made a number of cell-phone calls to the crew. In addition, he had directly negotiated with ATC a departure priority and a curfew extension at their destination. Note that frequent “hands on” direct management of his crews via cell-phone contact in flight was the CEO’s trademark.
As the aircraft was not operating under JAR/OPS 1 certification, no ground proximity warning system (GPWS) was required or installed in the aircraft involved in the accident. GPWS mode 3 “altitude loss after takeoff” would have warned the flight crew of the loss of altitude.
After receiving takeoff clearance at 21:05:54 UTC, the copilot as pilot flying (PF) initiated a rolling takeoff. The two engines had very dissimilar idle RPM settings (49% and 41%) and this affected their relative acceleration times. Since the left-hand engine was run up within six seconds to 102 percent of takeoff power and the right-hand engine to only 58 percent, for a few seconds during the acceleration phase the aircraft veered on the runway to such an extent that its heading changed 10 degrees to the right. The pilots were only able to bring the aircraft back into alignment with the runway by making a major nose-wheel steering and throttle correction. This gyration, with only 100m visibility would have been unnerving to the PF. It is one reason why his alma mater (Flight Safety International) specifies a standing start in LVP conditions, with a high matching power preset. Shortly after takeoff, the commander acknowledged the request to change frequency to departure control. This task would have required him to interrupt his flight instrument-monitoring scan of the inexperienced PF at a crucial time.
An analysis of the recording from comparable flights and the accident flight revealed no substantial differences concerning control deflections, increase in speed, vertical acceleration and attitude during the initial climb – and no steep noise abatement climb was attempted. On the basis of the DFDR recordings, it was possible to exclude an icing stall during the accident flight.
The flaps were retracted 14 seconds after takeoff at 524 ft. above ground level (AGL). At the same time the elevator trim tabs began to move in the nose-down direction. The trim tab movement lasted for 8.5 seconds. In 42 previous takeoffs, the average duration of trim operation by the crew, when retracting the flaps from 15 degrees to 0 degrees, was only 2.55 seconds. It was established that this was a pilot-initiated trimming. After retracting the flap, the PNF (in the LH seat) did not register the subsequent marked reduction in pitch attitude over the next 13 seconds, presumably because he was busy with other activities. He would have had to retune the VHF radio, as he’d been instructed to establish radio contact with a different ATC unit. The report says that he may also have been distracted by another event such as a signal inside the cockpit – the ringing of his mobile telephone.
Inadvertent Descent
Visual references to the runway lighting were lost as the aircraft’s nose was raised in the darkness and fog. Consequently, it was not possible for the pilot to determine the attitude of the aircraft by orientation using the natural horizon. The report assays that the minimum specification artificial horizon installed on the copilot’s side might not have allowed the copilot, with comparatively little experience in similar flying conditions, to orientate himself in sufficient time as to the current attitude, because of its diminutive size and low reading accuracy. On takeoff the copilot had rotated the aircraft to an attitude of 7-10 degrees nose-up and maintained it for about 13 seconds. Logically he must have assumed this attitude with the aid of the artificial horizon. But the nose of the aircraft then began to drop until shortly before impact the aircraft had attained an attitude of about 12 degrees nose-down. On the basis of the DFDR data, this change happened simultaneous with the deflection of the elevator (i.e., it was a selected outcome). Seconds before impact, the elevator control was pulled back, initiating a recovery maneuver.
Immediately after takeoff, the copilot had maintained a constant heading of 334-335 degrees for approximately 10 seconds. It can be assumed that during this phase he was maintaining the aircraft in a neutral bank attitude. The aircraft then began to change course to the right. On the basis of the recorded aileron deflections, this turn was neither initiated by the copilot nor was it corrected. It is possible that the commander inadvertently applied some right rudder while twisting in his seat to reach down right and access a ringing cell-phone from his bag. The nose of the aircraft began to drop and at the same time the aircraft began to bank and turn to the right. At this point only a GPWS could have saved them. The question becomes: “Is there an explanation for the PF’s action of continually trimming forward and lowering the nose as the flap was retracting and the aircraft accelerating?” Was it the insidious result of inattention or distraction — or was it a form of spatial disorientation (SD)? The crew had not switched off the anti- collision lights. The glare of the lights in the fog may have contributed to, or at least fostered, disorientation or vertigo in the pilot. The real clue is in the prolonged trimming.
To obtain the BFU Accident Report (101 pages), visit: http://www.bfu.admin.ch/common/pdf/u1829_e.pdf
Three Types of Spatial Disorientation
- Type I (Unrecognized): The pilot is oblivious to his or her disorientation, and controls the aircraft completely in accord with and in response to false perceptions.
- Type II (Recognized): The pilots may experience a conflict between what they feel the aircraft is doing and what the flight instruments show that it is doing.
- Type III (Incapacitating): The pilot experiences an overwhelming — i.e., incapacitating – physiologic response to physical or emotional stimuli associated with the disorientation event.
Source: http://www.spatiald.wpafb.af.mil