Gremline Flight Safety Report: Spatial Disorientation / Optical Illusion at Night

the online flight safety digest

Text Box: spatial disorientation / optical illusion at night

the gremline digest — spatial disorientation / optical illusion at night

Spatial Disorientation

We will never know how many accidents to GA aircraft have either been directly caused by spatial disorientation or where spatial disorientation has been a contributory factor. An understanding of the phenomenon and its causes should assist GA pilots to avoid exposure to what can be either a totally insidious occurrence or a sudden and violent clash between your perception and the real world outside the aircraft. Either situation is extremely unpleasant and can lead to difficulty in controlling the attitude of ones aircraft or to total and catastrophic loss of control.

Spatial disorientation is described in “Aviation Medicine” [edited by Ernsting and King, available from our Bookshop]. The following edited extract is relevant to this discussion.
      “Pilots have described many different types of spatial disorientation that occur in different flight conditions. Not surprisingly, the mechanism underlying the disordered perceptions is commensurately varied. It is convenient to discuss aetiology (the causation of diseases and disorders as a subject of investigation) under two main headings, even though they are not mutually exclusive: (1) when erroneous or inadequate sensory information is transmitted to the brain (an input error); and (2) when there is an erroneous or inadequate perception of correct sensory information by the brain (a central error).”


Loss of External Visual Cues
Disorientation is very uncommon when the pilot has well-defined external visual cues. It is when he attempts to fly when sight of the horizon is degraded by cloud, fog, snow, rain, smoke dust or darkness that he quickly becomes disorientated unless he transfers his attention to the aircraft instruments. The ability to maintain control of an aircraft without adequate visual cues is quite short, typically about 60 seconds when the aircraft is in straight and level flight at the time vision is lost, and shorter still if the aircraft is in a turn. In such circumstances, loss of control occurs because the non-visual receptors give either inadequate or erroneous information about the position, attitude and motion of the aircraft.
      Spatial disorientation can be insidious and creep up on one gradually, or it can occur suddenly when there is a violent clash between the information being fed to the pilot’s brain from his physical balance mechanisms and that being transmitted by his eyes. This can easily occur when transferring ones point of reference between the instrument panel and the outside world (or vice versa), particularly when the outside world is presenting an unclear picture. Perhaps the worst manifestation of spatial disorientation is when the pilot suddenly loses outside reference during a turn. He has to immediately transfer to instrument flying only to find his balance mechanisms SCREAMING that they are in straight and level flight when the instruments indicate the beginnings of a spiral dive in one direction or another. This type of spatial disorientation can, with skill and regular practice, be overcome by reference to the flight instruments and by acceptance of what the instruments are telling the pilot. But it takes total concentration and confidence on instrument flying to overcome disorientation. A slight head cold, slightly blocked sinus passages and several ‘harmless’ medications can exacerbate the onset of spatial disorientation. The old saying, “He who prescribes for himself has a fool for a physician” is particularly pertinent to aircrew members.



Misleading Visual Cues
The gentle, insidious type of spatial disorientation that creeps up slowly is far more difficult to recognise and overcome. It can be triggered by any number of slightly false visual references that are just slightly wrong and lead you gently and progressively astray until things get out of control. One example is a sloping cloudbase ahead that, in marginal visibility, can look like a horizon but is not horizontal. Approaching higher sloping ground in similar conditions can also confuse our information processing abilities. Poor visibility or a totally empty field of vision provide ideal conditions for the onset. A combination of erratic visibility and the onset of darkness should sound caution bells in the mind of any pilot who is trying to fly by visual reference. These types of gradual onset spatial disorientation can, in my experience, induce a gradually growing sense of unease in a pilot. Something is not quite right, but the pilot cannot isolate the cause of his unease. This is a good time for a careful instrument scan and also time to consider making a 180° turn and getting away from the deterioration.
     An ‘amusing’ example of getting erroneous information from ones eyes happened to me one Friday evening when bumbling up the east coast of Northern England in a Chipmunk. I was heading from Leconfield to a weekend party as Leuchars and flew into deteriorating weather with low cloud and patchy visibility in rain showers. It was forecast to improve just a few miles further north, beyond the front. I told myself that it wasn’t worth the bother of climbing into cloud and going IMC for just a few miles. I peered ahead through the rain-spattered windscreen, following a route that was familiar to me. I was surprised to see a long, thin very straight roadway ahead, paralleling the coastline I was following. I’d never seen this roadway before. I was about to check my map when I spotted something odd about the roadway. It was actually a very tall grey tapered factory chimney quite close below me. I had allowed the Chipmunk to creep down below the lowering cloudbase as I concentrated on remaining in visual contact with the surface. I was flying on the wrong side of a line feature. I was tired after a very busy week. I began to pay more attention to what I was doing, rather than thinking about the weekend ahead. What did I learn from this experience? There is no fool like an old fool!


Optical Illusion at Night

This fatal accident happened more than 20 years ago, but the lessons to be learnt are still valid today. Two very close friends of mine died on the evening of Thursday 8 December 1983. Wing Commander John Parker and his wife were passengers in a Cessna Citation 500 that crashed into the night sea off Stornoway. They almost certainly survived the impact, but did not live long enough to be rescued from the cold winter sea. The most probable cause of the accident was an optical illusion that affected the pilot performing a visual letdown over dark, featureless water while approaching scattered lights on the island ahead. Pilots need to be aware of this illusion.

At about 16.30 hrs on Thursday 8 December 1983 Cessna Citation 500 G-UESS took off from Liverpool Airport en route for Stornoway. On board were eight passengers including two infants and what is described as a pilot’s assistant. At about 17.45 hrs when at a range of about 10nm from Stornoway and while descending at night to a low altitude, the aircraft disappeared from radar and ceased to reply to radio messages.
      An intensive search was mounted that evening and continued for several days but was successful only in recovering the bodies of the pilot and six of the passengers, although the other bodies were recovered over the next few months. The main wreckage was not recovered.
      The AAIB Report concludes that the accident was probably caused by the pilot’s lack of awareness of his true altitude which resulted in his allowing the aircraft to descend until it struck the sea. Likely contributing factors were that he was distracted (from his instrument scan) by the need to establish visual contact with another aircraft and that he was misled by false cues from lights on the ground ahead of him.
      This was by no means the first time that an accident of this kind had occurred, and the probable cause had been known for many years before this accident. Similar accidents include those to British Airways Boeing 747 G-AWNC near Sebang International Airport in Malaysia on 11 May 1976 and to Piper PA31 Navajo G-BBPC at Walney Island Channel on 26 November 1976. In both those cases the aircraft were making visual approaches at night over unlit terrain without glidepath assistance. According to research by the Boeing Aerospace Company, in these circumstances a pilot may position his aircraft so that the pattern of lights on the ground subtends a constant angle at his eyes. This would result in a curved flight path in the vertical plane, flown below the straight line approach path. In other words, the pilot will fly his aircraft into the surface some distance short of his intended destination. This will be discussed and illustrated later in this article.



The Cessna Accident
First, we will look in more detail at what appears to have happened to the Cessna Citation on 8 December 1983. Earlier in the day the aircraft left Biggin Hill with only the pilot, the pilot’s assistant and two passengers on board to fly to Le Bourget, Paris where Wing Commander and Mrs Parker embarked. John was the Air Attaché at the British Embassy in Paris and they were travelling to Stornoway to join a party. The aircraft then flew to Liverpool where it was refuelled with 800 litres of Jet A1. Two more adults and two infants joined the passengers at Liverpool. The aircraft departed on an IFR flight plan at FL310 via Deans Cross and Glasgow to Stornoway.
       The pilot reported to Scottish ATCC that he was at FL280 climbing to FL310 when approximately over Deans Cross at 16.53 hrs. This radio call was heard by the pilot of a Cessna Citation II N40GS that had been leased by the owners of G-UESS and was carrying other members of the private party to Stornoway. N40GS had taken off from Biggin Hill and was en route to Stornoway via Deans Cross at FL350. When just north of Deans Cross the pilot of N40GS saw G-UESS ahead of him and established contact with the pilot on a company discrete frequency. Thereafter, they remained in intermittent radio contact.
       Both aircraft were cleared direct to Stornoway after passing Deans Cross. G-UESS was climbed to FL330 at 17.00 hrs to clear crossing traffic. The pilot of N40GS reported that he still had G-UESS in sight at 17.18 hrs and would be ready to descend in 3 minutes. His aircraft was slowly overtaking G-UESS as Scottish ATCC directed both aircraft to maintain 330° M and cleared N40GS to descend. Three minutes later, at 17.23 hrs, G-UESS was cleared to descend. N40GS was cleared progressively to FL65 and G-UESS to FL85. Scottish ATCC released both aircraft at 17.29 hrs with no traffic to affect them. Scottish ATCC advised N40GS that G-UESS was 5 miles to his right and slightly behind him. The pilot of G-UESS responded by reporting that he had the other aircraft in sight. Scottish ATCC then handed control to Stornoway.
      At 17.34 hrs Stornoway ATC passed Stornoway weather to both aircraft and asked them to report when 25 miles from Stornoway. The weather was reported as fine with a light wind, good visibility and one eight low cloud. The pilot of G-UESS acknowledged the weather but did not read back the QNH. He was then 49 miles from Stornoway descending through FL140. At 17.38 N40GS reported 25 miles from Stornoway and immediately afterwards G-UESS reported at 30 miles range. N40GS was then cleared to 2000 feet on the QNH of 1001.
      At 17.40 the pilot of G-UESS reported 25 miles from Stornoway, with N40GS in sight. He was cleared to continue his descent, with the aircraft ahead in sight. A moment later he was asked to report when he had the airfield in sight for a visual approach to Runway 01. He acknowledged this message. There was no further contact with G-UESS and, at 17.51 hrs, the Stornoway controller reported to Scottish ATCC that he had lost contact with the Citation.
      The pilot of N40GS said that they had passed through some layered stratus cloud, patches of altocumulus and cumulus on the descent from FL350. The co-pilot described a layer of lower cloud over the sea with tops between 3000 and 4000 feet, lying across their descent to Stornoway. N40GS was tracking directly to Stornoway airfield during the descent, using Omega/VLF area navigation equipment. The pilot had reported to Stornoway that he was just breaking cloud at 1400 feet during the descent, but stated later that the base of the lowest cloud was between 1100 and 1000 feet. He also stated that the visibility below cloud was very good even though the night was dark and he could not see the surface of the sea. There was no icing and no significant turbulence during the descent and approach and landing.



The AAIB Investigation
The AAIB Report points out that although a small amount of wreckage was found on the sea surface no significant wreckage was recovered from the sea bed. There was no indication of any emergency before the aircraft struck the sea. Transponder returns indicated that electrical power was available down to the corrected last return at somewhere between 175 feet and the surface. The landing gear was locked down and the impact was severe enough to disrupt the fuselage, but not severe enough to render all of the passengers unconscious, at least two passengers escaping from the aircraft after the impact.
      AAIB investigators made a series of flights in an aircraft identical to G-UESS to observe the lights of Stornoway from overhead the impact position, some 10 miles to the southeast of the airfield. Although there was bright moonlight and no cloud, the sea surface could not be seen from 500 feet. The only lights visible were those of Stornoway and the Eye peninsular that formed a near continuous line of lights subtending a horizontal angle of about 40° to the pilot’s eyes. Some of the lights north of the town and on the peninsular are on higher ground than the town itself and gave the impression of depth, as if being viewed from a higher altitude. Because of the combined effect of the width and apparent depth of the lights, all three pilots on the test aircraft underestimated their distance from Stornoway and greatly overestimated their height above the sea. Two runs were made, one at 1000 feet and one at 500 feet. The visual aspect of the lights was much the same on both runs. Even at 500 feet the lights gave no visual clues to cause concern and all three pilots experienced a false impression of being at a safe altitude.


Research by Boeing
The AAIB Report drew attention to the research programme conducted by Dr. Kraft and Dr. Elworth of the Boeing Aerospace Company to investigate the problems of night visual approaches. The results of this research were first published in the USAF ‘Interceptor’ magazine in October 1968. They demonstrated that the visual cues available to a pilot approaching a lighted area at night over unlit terrain are misleading and inadequate. During simulator trials as part of the research programme under these conditions, with no altimeter available for reference, eleven out of twelve highly experienced pilots ‘crashed’ the simulator at between eight and five miles from touchdown although they had only begun the approach at 18 miles and had been instructed to be at 5000 feet when 10 miles out and at 1240 feet 4.5 miles out.
      The most critical visual cue was found to be the angle subtended at the pilot’s eye between the nearest and the furthest lights, i.e. the apparent depth of the light pattern. The simulator flights disclosed a tendency, during the descent, for pilots to maintain this angle at a constant value. This resulted in their flying an approach path along the arc of a circle centred over the pattern of lights, with its circumference intersecting the surface short of the target airfield, i.e. to fly into the surface short of the destination airfield. There was also a common tendency for pilots to grossly overestimate their altitude in these circumstances.
      The diagram ILLUSION IN DESCENT will make this concept clearer. Consider an aircraft in level flight flying towards a static pattern of lights on the ground (Point A to Point B on Figure 1). At Point A the angle between the nearest light (L1) and the farthest light (L2) will subtend a certain angle to the pilot’s eyes. As the aircraft continues towards the lights in level flight, this angle will increase, reaching a maximum when the aircraft is centred above the lights at Point B. Now consider the aircraft back at Point A, but instead of being in level, horizontal flight towards Point B it is in a vertical descent towards Point C. The angle subtended by the two lights (L1 and L2) will decrease until it reaches zero when the aircraft arrives at the surface at Point C.

      Therefore there must be a flight path between these two extremes when the angle subtended between the two lights remains CONSTANT while approaching the lights. This flight path will be below the straight line joining Point A to the nearest light L1 and will follow the path ADX.

The Boeing research indicated that both the tendency to fly into the surface short of the airfield and to overestimate the height above the unseen surface are exaggerated if the plane of the lights is inclined slightly towards the flight path i.e. L2 is slightly higher than L1. The critical visual clue that the descent path was deviating below the ideal had to come from the motion of the light pattern relative to the pilot’s eye. But, to be perceived by the pilot, this motion must exceed a rate change of approximately one minute of vertical arc per second, otherwise the pilot will simply not register any change in the light pattern. Thus, when flying an approach over a dark area at 1000 feet and 120 kt, few pilots are capable of perceiving the relative motion of the tilt in the plane of the lights that would alert them to an excessive rate of descent until they are within some 3.5 miles of the light pattern. The Citation struck the sea some 10 miles from the airfield.


Additional Factors
Dr. Kraft and Dr. Elworth went on to point out that other factors that also can contribute greatly to the visual approach problem at night are:


An approach over a dark surface where lights to the side and below the aircraft don’t exist

A long, straight-in approach to an airfield located on the near side of a town

A runway length/width relationship that is unfamiliar to the pilot

An airfield situated at a lower elevation and on a different slope from the surrounding terrain

A navigational facility located some distance from the airfield

A sprawling city within an irregular matrix of lights spread over hillsides behind the airfield

Industrial smoke or other obscuration that decrease the brightness of the lights as they interact with the distribution of lights around the airfield


An examination of a chart of the Stornoway area shows that a number of these factors apply to the night, straight-in approach being flown by the Citation that crashed. Additional factors were that the pilot of G-UESS probably needed to get below cloud quickly so as to maintain visual contact with the aircraft ahead of him, he needed early visual contact with Stornoway as, unlike the other Citation, he did not have an accurate positional navigation aid in his aircraft and he may have been alone in the cockpit while his assistant was in the cabin preparing the passengers for their arrival at Stornoway. A combination of all the factors involved may explain why he was unaware of the continued descent into the sea.
      The transponder height readouts from both aircraft give an indication of the rates of descent flown by each pilot during the latter part of their descents to Stornoway. These radar plots can also give a rough indication of each aircraft’s groundspeed throughout their descents. N40GS had a decreasing rate of descent as it approached the airfield. His rate of descent began at 1400ft/min passing 8100 feet and gradually reduced so that in the remaining six minutes of descent his rates were 1200ft/min, 1000 ft/min, 1100 ft/min, 900 ft/min, 600 ft/min and 500 ft/min. The groundspeed estimation for N40GS shows a steady speed reduction throughout the descent. The speeds for each of the last seven minutes of the descent were 282kt, 264kt, 258kt, 228kt, 210kt and 180kt. In contrast, the rates of descent of G-UESS over the last seven minutes to impact were 1400 ft/min, 700 ft/min, 1500 ft/min, 1200 ft/min, 900 ft/min, 1300 ft/min and 1300 ft/min to impact. The groundspeeds for each of the last seven minutes of the descent of G-UESS were also less progressive at 270kt, 246kt, 198kt, 228kt, 204kt, 138kt and 138kt to impact. This unsteady descent may point towards the pilot trying to solve the confusing information he was getting from his eyes.

Fatal aircraft accidents are not just statistics to ponder, they involve the destruction of families and the loss of friends. That sense of loss remains today, twenty five years after the event.

AAIB Reports AAR 5/84 and FSB Spring 1985.



Search Now:
Amazon Logo



Text and Photographs © 2008 Gremline & Hill House Publications, unless otherwise stated.


landing page   about gremline   copyright/conditions/contact   information exchange   glossary

uk emergency diversions   uk links, chirp & gasco   global & misc links   forum

the gremline cockpit — index of articles   the gremline bookshop   top of page