Spatial Disorientation and Visual Illusions in Instrument Flight

Spatial disorientation is not a failure of skill or of nerve. It is the predictable output of sensory organs that measure accelerations rather than attitude, working without the one input they were designed to be corrected by — a visible horizon. In cloud and at night they hand you a confident, coherent and entirely wrong picture of which way is up, and the control input it invites is usually the fatal one. And when a horizon is available it can still deceive you: judging an approach path by eye rests on learned expectations about runway shape, terrain and lighting, so a runway narrower than you are used to, a night with nothing lit between you and the threshold, or rain on the windscreen produces a confident misjudgement of exactly the same kind. This guide covers the physiology, the vestibular illusions, the visual illusions on approach and at night, and the defences.

Three orientation systems, and the one you lose in cloud

Three systems tell you which way up you are: the visual system, dominant and the only one that measures attitude directly; the vestibular system of the inner ear, which measures some accelerations; and the somatosensory system — the seat of your pants — which measures applied force with no information about its source. Enter cloud, or take off on a moonless night, and the dominant system is gone. The other two are left flying the aeroplane — which is why single-pilot instrument flight is trained as the discipline of believing an instrument over a sensation.

The semicircular canals: angular acceleration only

Each inner ear holds three semicircular canals set at approximately right angles, roughly aligned with pitch, roll and yaw. Each is filled with endolymph and swells at one end into the ampulla, where a gelatinous flap — the cupula — sits across the channel with hair cells in its base. Rotate in the plane of a canal and the endolymph lags through inertia, deflecting the cupula. That signal encodes angular acceleration: not attitude, not rate of turn.

• There is a threshold. The figure usually quoted for detection of angular acceleration is of the order of 2°/s², varying with the plane of rotation and between individuals. The principle matters more than the number: an aeroplane can be rolled gently enough that the canals report nothing at all.
• The signal decays. Hold a constant rate of turn and within roughly ten to twenty seconds the endolymph is rotating with you, the cupula returns to neutral and the sensation fades. A steady turn feels like straight flight. Stop the turn and the fluid overshoots, giving a sensation of turning the other way.

The otoliths: linear acceleration and gravity, indistinguishable

Alongside the canals sit the utricle and saccule, each carrying a sensory macula covered by a gelatinous membrane loaded with calcium carbonate crystals — the otoconia. The utricular macula lies roughly horizontally and responds to fore-and-aft and lateral acceleration and to head tilt; the saccular lies roughly vertically and responds to vertical acceleration. Being denser than their surroundings they shear under any applied force, which gives an unavoidable ambiguity: the otoliths cannot distinguish gravity from linear acceleration. They report the resultant, and your brain calls that direction "down".

Somatogyral illusions: the leans, the spiral and Coriolis

Illusions from the canals are somatogyral; those from the otoliths are somatogravic. Exam questions routinely ask you to attribute an illusion to the correct organ.

The leans

The commonest form of disorientation, and a direct consequence of the threshold.

1. The aircraft rolls slowly into a bank. Sub-threshold, so nothing is sensed: you believe the wings are level and they are not.
2. You notice the attitude indicator and roll briskly back to level. That correction is above threshold, and it is sensed.
3. Your canals have reported exactly one roll — the one back towards level — so the brain concludes you are now banked the other way.

The result is a straight-and-level aeroplane that feels badly wrong, with pilots leaning bodily in the seat towards a false vertical. A rapid rollout after a prolonged turn does the same. It can persist for minutes; fly the instruments and let it decay.

The graveyard spiral

A chain that kills because every link feels correct. In a prolonged turn the sensation decays and you feel wings level. This only kills you if the attitude indicator is out of the scan or lying — a partial-panel aircraft, a slowly failing vacuum-driven AI that your inner ear agrees with, or simple fixation on the unwinding altimeter (see the defences). Given that, the chain runs as follows. In the bank the vertical component of lift is reduced, so the nose drops and airspeed builds. The altimeter unwinds — and believing the wings level, the natural response is to pull, which in a bank simply tightens the turn, raises the load factor and steepens the descent. Roll level and the canals report a turn the opposite way, so the temptation is to roll back in. Airspeed, rate of descent and load factor climb together until the aircraft hits the ground or breaks up.

Note that a spiral is not a spin — it is unstalled and accelerating — so spin recovery actions do not apply. The spin has an illusion of its own.

The graveyard spin

The somatogyral pair to the spiral, and the reason the two are examined together. Hold a spin for the same ten to twenty seconds and the endolymph catches up with the rotation, the cupula returns to neutral and the sensation of spinning fades. Apply the recovery, stop the rotation, and the fluid overshoots: the canals now report a spin the other way. Feeling that, the pilot re-applies pro-spin control — and puts the aircraft straight back into the spin it had just recovered from, usually while pulling against the unwinding altimeter. The defence is the one that runs through this whole guide: complete the recovery against the turn indicator and the attitude indicator, and disbelieve the sensation that arrives the moment a rotation stops.

The Coriolis illusion

The most disabling of the canal illusions, and the easiest to trigger by accident. Hold a steady turn long enough for the endolymph in the canals lying in the plane of the turn to settle, then move your head out of that plane — down to a chart, across to a fuel selector. The movement swings quiet canals into the plane of rotation and takes the settled ones out of it, so several report acceleration at once in planes that cannot all be true. The brain resolves that as a violent sensation of tumbling about a different axis, and the instinctive input is large and wrong.

The somatogravic illusion: the one that kills on the go-around

Because the otoliths report the resultant of gravity and inertial force, and the brain treats that resultant as vertical, any sustained linear acceleration is read as a tilt. Accelerate forward and the inertial reaction shears the otolithic membrane rearward; a nose-up pitch attitude does precisely the same thing as the head tilts back. The two stimuli are physically identical, so the sensation is identical: forward acceleration produces a compelling illusion of pitching up. With a horizon in sight the eyes veto it instantly; remove the horizon and it becomes one of the most lethal illusions in aviation. Two settings account for most of the accidents:

• Night or IMC take-off over dark terrain or water. Full power, a clean acceleration, no external reference. The aircraft feels as though it is pitching up towards a stall, and the instinct is to push.
• The go-around. Go-around thrust at low weight, with drag disappearing as flap and gear retract, generates a large sustained acceleration — often greater than at take-off — at the moment of highest workload and least external reference. The aeroplane feels grossly nose-high, the pilot pushes, and it descends back towards the ground it has just gone around from.

It works in reverse too. Deceleration produces an illusion of pitching down. Close the thrust levers, deploy speedbrake, or extend gear and flap on an approach in cloud, and the resultant tilts forward: the nose feels as though it has dropped, and the instinct is to pull — going high and slow, with the stall a good deal nearer than the sensation suggests, and a low-energy arrival at the missed approach point.

Two more otolith illusions carry names. The elevator illusion: an abrupt updraft stimulates the otoliths as a climb would, inviting a nose-down input, and a downdraft gives the reverse. The inversion illusion: an abrupt transition from a steep climb to level flight can produce a sensation of tumbling backwards, and with it a sharp push nose-low.

Visual illusions on the approach

The vestibular illusions come from having no visual reference. The approach illusions are the opposite: you have one, and it is lying. Your judgement of an approach path is built on learned expectations, so any scene violating them produces a systematic error.

The black hole approach

A night visual approach to a lit runway with no lighting whatever between you and it — over water, desert or unlit countryside — often with a lit town beyond the far end. No horizon, no ground texture, no peripheral cue; all that remains is the runway light pattern. Deprived of everything else, pilots tend to hold the runway's visual appearance constant, which produces a curved, progressively shallowing profile that ends up dangerously low well outside the threshold. Rising terrain short of the airfield, or a lit city mistaken for a horizon, makes it worse.

Geometry illusions and the way they push you

Anything making the runway look bigger or closer than expected drives you high; smaller or further drives you low. The dangerous half is the low half.

Autokinesis

Stare at a single stationary light against a dark, featureless background for a few seconds — commonly quoted as six to twelve — and it will appear to move, because the brain has no reference to anchor it against. Pilots have manoeuvred to avoid a "converging aircraft" that was a star. Break the fixation by scanning, and check anything that appears to move against the instruments before you manoeuvre for it.

The false horizon

A separately named illusion, and the one most likely to leave you in a genuinely unusual attitude. Given no true horizon, the brain accepts the most horizon-shaped thing in view and levels the wings on it. The standard triggers are a sloping cloud deck, a line of ground lights running along rising terrain, a lit shoreline or motorway, and on a clear night a star field taken for ground lights — or ground lights taken for stars. Having accepted the false reference, the pilot banks or pitches to align with it and flies an attitude that feels level and is not. At night and in cloud, the attitude indicator is the horizon.

Night vision: what the eye can and cannot do in the dark

Every illusion above assumed you could at least see the lights that were deceiving you. On a dark night you often cannot, and for reasons that are anatomical rather than a matter of effort or experience. The eye you use in daylight is not the eye you fly with at night. The retina carries two receptor types. Cones are packed into the fovea — the small central area you point at whatever you want to examine — and give colour and fine detail, but need light. Rods occupy the periphery, are entirely absent from the fovea, work at very low light levels and detect movement well, but give no colour and poor resolution. Hence the single most important fact about night flying: at low light levels the centre of your vision stops working.

The night blind spot and off-centre viewing

Because the fovea is all cones, a central blind area of roughly 5° to 10° appears at night. Look directly at a distant navigation light and it can vanish; look slightly away and it returns.

The technique is off-centre viewing: look about 10° to one side of what you want to see, placing its image on rod-rich retina. Then keep moving, because rods bleach if an image is held stationary on them. A night scan is a series of short, deliberate movements with a pause of a second or two in each sector — not a smooth sweep, which sees almost nothing.

Dark adaptation and empty-field myopia

Cones adapt in around five to seven minutes. Rods take far longer — roughly 30 minutes for practical purposes — and that adaptation is destroyed in seconds by bright white light, after which the full period is needed again. Red cockpit lighting preserves rod adaptation but makes anything printed in red vanish, so low-intensity white light is the usual compromise. The related Purkinje shift moves peak sensitivity towards the blue-green as light levels fall, so reds darken towards black.

Two further effects are worth knowing. Empty-field myopia: with nothing to focus on, the eye relaxes to its resting focus, roughly one to two metres ahead, so a distant aircraft can go unseen — force accommodation onto a wingtip light or a star. And night vision is the first casualty of hypoxia: rods have a high oxygen demand, and degradation is generally reported from around 5 000 ft cabin altitude, well below any altitude at which you would think about oxygen. Carbon monoxide from smoking or a leaking exhaust heater does the same by occupying haemoglobin.

Put the anatomy back against the approach and the black hole stops looking like bad luck. The night blind spot is why a distant threshold or navigation light can disappear at the moment you look straight at it; empty-field myopia is why unlit water and unlit countryside return nothing to focus on at all. Between them they are the reason a runway light pattern is the only cue left over dark terrain — and the reason judging a vertical profile from that pattern is a losing proposition.

The defences: cross-check, discipline and recovery

None of this can be trained out of you — the vestibular system will give the same wrong answer for your whole career. What can be trained is what you do about it.

1. Believe the instruments, and know why

State it mechanically rather than as a slogan: the attitude indicator measures attitude; your inner ear measures accelerations, and cannot tell a tilt from an acceleration. In unaccelerated flight it is right; the moment the aircraft accelerates, its attitude answer is unreliable and there is no way for you to know which case you are in. When the two disagree, one of them is doing a job it was not built for, and it is not the instrument. The only exception is instrument failure — hence the cross-check.

2. Cross-check discipline

Build the scan around the attitude indicator as master instrument and return to it between every other instrument. Three scan failures account for most losses of control: fixation (chasing one instrument while the aircraft quietly rolls), omission (dropping an instrument — usually the one that would have shown the problem) and emphasis (over-weighting one instrument instead of the whole picture).

A slowly failing vacuum-driven attitude indicator is the classic killer: it fails gradually and plausibly, and your inner ear agrees with it. Cross-check against the turn indicator, altimeter, VSI, airspeed and heading. Your standby heading reference is the magnetic compass, with its own turning and acceleration errors.

3. Keep your head still, and use the autopilot

The head discipline set out above — head-down work wings level and unaccelerated — removes the Coriolis illusion almost entirely, which makes it the one illusion here you can design out rather than fly out. And if you are disorientated with a serviceable autopilot, engage it: it has no vestibular system. Monitor it though, because a slow autopilot excursion is exactly the sub-threshold roll that starts the leans, and it will not announce itself.

4. Recover in the right order

• Nose low, spiral: power to idle, roll the wings level, then ease out of the dive, watching airspeed and load.
• Nose high: apply power, lower the nose towards the horizon, then level the wings — here the priority is the stall, not the bank.

In the nose-low case the wings come first because pulling while banked only tightens the turn and adds load without arresting the descent, and airspeed and load factor are already rising. In the nose-high case the nose comes first because the aircraft is decelerating towards the stall, and stall speed rises with load factor and therefore with bank — see principles of flight.

5. Manage the pilot

Fatigue, dehydration, hypoxia, stress and workload all lower the threshold at which disorientation appears, as does anything affecting the inner ear: a head cold can produce alternobaric vertigo — classically during the climb, when one middle ear vents before the other, and also after a forceful Valsalva on descent — and alcohol diffuses into the endolymph to produce positional alcohol nystagmus many hours after blood alcohol has returned to zero. In a two-crew flight deck, "I am disorientated, you have control" is the fastest fix you have.

How this is examined, and where the marks are lost

Spatial disorientation sits in Human Performance — subject 040 at IR, CPL and ATPL level, and its own Human Performance paper at PPL, examined in more depth as the licence level rises. At IR level the framing is single-pilot instrument flight, so expect scenarios in cloud, on a go-around or on a night approach rather than bare definitions. Three things reliably lose marks.

Attributing an illusion to the wrong organ. Canals detect angular acceleration; otoliths detect linear acceleration and gravity. The leans, the graveyard spiral, the graveyard spin and Coriolis are somatogyral. The false pitch-up on acceleration, the false pitch-down on deceleration, the elevator illusion and the inversion illusion are somatogravic.

Getting the direction backwards. Almost every question turns on which way the illusion pushes the aircraft.

This is the one area of theoretical knowledge where the correct answer contradicts what your body insists is true. Knowing the table is not the same as having decided in advance to fly the instruments.