Satellite Availability Prediction: What a Pre-Flight Check Actually Proves

A pilot studying a satellite availability prediction in a cockpit at dusk, with the aircraft on the ramp and a control tower in the background.

A satellite availability prediction is one of the few pre-flight checks whose output is routinely misread. It is not a forecast of whether the receiver will work. It is a statement about the geometry of a constellation over a place and a time window, computed from an almanac and a list of declared outages, under assumptions about a receiver that is probably not yours. Understanding exactly what it covers is what turns a green line on a printout into a defensible departure decision.

What the prediction is a prediction of

The tool takes four things: a position or a route, a time window, an almanac describing where each satellite will be, and a list of satellites that have been declared out of service. It then propagates the orbits, works out which satellites will be above the horizon at that place and time, computes the resulting geometry, and asks a single question — with that geometry, can the receiver still guarantee the integrity performance the navigation specification demands?

It is also worth being precise about the word availability. A prediction does not usually return a yes or a no; it returns a proportion of the requested window during which the requirement is met, together with the times of any interruption. A result of ninety-eight per cent is not a pass mark. It is a statement that there are gaps, and the operational question is where those gaps fall relative to the estimated time of arrival, not what the percentage is.

Everything in that sentence is astronomy and arithmetic. Nothing in it is about your aeroplane, your antenna, the radio environment around the aerodrome, or the state of the receiver in the rack. That boundary is the whole subject.

The prediction is a math problem about satellite positions, not a test of your plane’s equipment. It tells you if the satellites can prove your position is correct, but not if your gear will work.

Integrity is the subject, not accuracy

Crews often speak about prediction as though it were about position error. It is not. Satellite navigation is, for most of the time, considerably more accurate than the operations built on it require. The problem it has is a different one: a satellite can transmit a signal that is wrong without saying so, and a receiver that trusts it will produce a confidently wrong position.

Integrity is the ability to detect that and to warn the crew within a time short enough to matter. Every requirement expressed as an alert limit and a time to alert is a statement about that ability, and the prediction exists to answer whether the constellation will support it. An operation can be perfectly accurate and still be unavailable, because availability here means the ability to prove the position is trustworthy, not the ability to compute one.

An Airbus A320 using GPS for an RNP AR approach at 200 feet. The prediction shows 95% availability, but a 5-minute gap during descent. If that gap hits the final approach, the system can’t guarantee integrity – even if the position is technically accurate.

Why five satellites, and why the count is not the answer

The classical arithmetic is worth having in the front of the mind, because it explains why a single satellite outage sometimes matters enormously and usually does not matter at all.

The classical satellite counts for receiver autonomous integrity monitoring
Function Satellites required What it buys
Compute a three-dimensional position Four A solution, with no redundancy and no way to check it
Detect that one satellite is faulty Five A consistency check — the receiver knows something is wrong
Detect and exclude the faulty satellite Six Continuity — navigation carries on without the bad measurement
Same functions with barometric aiding One fewer in each case Pressure altitude substitutes for one measurement

Now the qualification that makes those numbers less useful than they look: the count is a floor, not the answer. Six satellites clustered in one quadrant of the sky produce a worse solution than five well spread around it, because the integrity check works by looking for inconsistency between redundant measurements, and measurements that all come from the same direction are not usefully redundant. Geometry, expressed as dilution of precision, dominates.

This is why a prediction run at one aerodrome can show a fifteen-minute hole while an aerodrome eighty miles away shows none, on the same day, with the same satellites out of service. Nothing about the constellation changed between the two. What changed is which satellites are above the local horizon and how they are arranged.

Barometric aiding, and the receiver the prediction is not modelling

Barometric aiding deserves a moment because it is where predictions and reality most often part company. If the receiver can use pressure altitude as an additional measurement, it effectively gains a satellite. Whether it does so, and under what conditions it stops doing so, is a property of that specific equipment and its installation.

Most prediction services model a generic, deliberately conservative receiver algorithm. They do not know your part number, your software standard, your antenna installation or the masking produced by the airframe at low elevation angles. The consequence runs in both directions: a prediction may show unavailability that your equipment would actually have handled, and — more rarely but more dangerously — the assumptions in the model may not be conservative for a particular installation.

The trap. The predictive integrity function inside the flight management system is not the same product as a ground prediction service. It typically uses the almanac the receiver has collected, which can be stale, and it usually computes for a single point at a single estimated time. It answers a narrower question, and it answers it with older data.

What the prediction cannot see

This is the section that matters most, because the failures that have actually disrupted operations in recent years are all in the right-hand column.

Inside and outside the scope of an availability prediction
Covered by the prediction Not covered, and invisible to it
Orbital geometry at the place and time requested Radio interference, jamming and spoofing
Satellites declared unserviceable before the run Satellites failing after the run
The number and spread of satellites above the mask angle Terrain and airframe masking at your actual attitude
The integrity requirement of the navigation specification requested The health of the receiver, antenna and wiring
The effect of barometric aiding, as modelled Whether your installation actually applies it
A modelled ionosphere under normal conditions Severe ionospheric storms and scintillation
  • Radio interference, jamming and spoofing
  • Satellites failing after the run
  • Terrain and airframe masking at your actual attitude
  • The health of the receiver, antenna and wiring
  • Whether your installation actually applies barometric aiding
  • Severe ionospheric storms and scintillation

Read the right-hand column as a list of things that require a different mitigation. None of them is improved by running the prediction again.

Interference and spoofing: a different problem entirely

Deliberate interference has moved from a curiosity to an operational planning item across several regions, and it breaks the mental model that predictions encourage. Jamming denies the signal; the receiver loses the solution, annunciates a loss of navigation capability, and the crew reverts to other means. That is unpleasant but honest.

Spoofing is worse, because the receiver keeps working. It produces a position, a velocity and a time that are all internally consistent and all wrong. The downstream effects reported by crews and by the European agency’s safety information bulletins on the subject include inertial systems being corrupted by a false position or a false time, spurious terrain warnings, loss of the ability to re-align, and surveillance reporting a position the aeroplane is not at. A prediction run two hours earlier has nothing to say about any of it.

There is a second-order effect that pre-flight preparation habitually misses. Interference does not only affect the aeroplane inside it. Ground stations, augmentation reference stations and surrounding traffic can all be affected at once, so the fallback the crew assumed would be available — a nearby aerodrome using the same augmentation, a diversion within the same region — may be degraded by the same source. A reversion plan that stays inside the affected area is not a reversion plan.

The mitigations are procedural rather than predictive: read the notices for the airspace concerned, brief the conventional alternative in advance, know which of your systems degrade and in what order when the position input is false, and know how to isolate the receiver if the equipment allows it.

When the check is required, and by whom

The requirement is not universal, and this is where generalisations do the most damage. It attaches to specific operations, and its detail comes from the navigation specification, the operator’s approval and the requirements of the State in which the aerodrome lies.

Broadly, an availability check is expected where the operation depends on satellite integrity monitoring computed on board with no external augmentation. Approach operations flown to lateral navigation minima on a receiver using only autonomous monitoring are the usual case in continental operations. Long-range operations over areas without conventional coverage are the other, and there the check concerns fault detection and exclusion rather than detection alone, with a maximum tolerable outage duration set by the specification being flown.

Where an augmentation system is providing the integrity, the question changes shape rather than disappearing. The relevant items become the published status and service area of the augmentation, and the notices affecting it — not the number of core satellites in view.

When augmentation changes the question

Space-based augmentation broadcasts corrections and, more importantly, integrity information from a ground network, so the receiver no longer has to derive integrity from redundancy alone. This is what supports approach minima whose vertical guidance comes from the satellite signal itself rather than from barometric altitude.

Two consequences follow for pre-flight preparation. First, the augmentation has a defined service area, and its performance degrades towards the edges — an aerodrome near the boundary is not in the same position as one at the centre, whatever the core constellation is doing. Second, the augmentation itself has published status information and its own notices, and those are the documents to read. Outside the service area, or when the augmentation is unavailable, the receiver falls back to autonomous monitoring and the original question returns.

Turning a red window into a departure decision

A prediction that shows a gap is information, not a prohibition. What it does is force the operation to be re-planned around a stated window, and there are more responses available than most briefings consider.

Responses to a predicted unavailability window
Response When it fits What it costs
Shift the departure or arrival time The gap is short and the schedule has slack Delay, and the geometry may not improve much
Plan a conventional approach at destination The aerodrome has ground-based aids and the aeroplane is equipped Higher minima, possibly a different runway
Plan an augmented approach instead The augmentation is available and the approval covers it None, if the procedure is published
Nominate an alternate with conventional coverage Destination options are thin Additional fuel, possibly a payload penalty
Accept the window with a briefed reversion The gap does not overlap the critical phase Requires a genuine plan, not an intention

The right reflex. Record what you ran, for what position, for what window, and against which navigation specification. A prediction without those four parameters attached to it is not evidence of anything, and it is the first thing an investigator or an auditor will ask for.

The habit worth building

The professional posture is to treat the prediction as one input among three. The prediction tells you about geometry. The notices tell you about outages, interference and augmentation status. The equipment list and the aeroplane’s technical log tell you what is actually fitted and serviceable. A departure decision is defensible when all three have been consulted and the reversion has been briefed out loud; it is not defensible because a tool returned a green result for a generic receiver at a nominal position.

A pilot reviews a checklist in an aircraft cockpit at dusk.
Pre-flight checks ensure satellite availability for critical navigation and communication.

Crews consolidating their performance-based navigation knowledge will find structured material on the subject at the FL330 training centre.

Commentaires

Leave a Reply

Découvrez l’écosystème FL330

FL330, c’est toute l’aviation professionnelle — explorez nos autres espaces.