How flight radar actually works, and why your trail vanishes when you land

Open any live flight map and you get several thousand aircraft moving across the world in something close to real time, for free, with no account.

It looks like the output of a government system. It mostly isn't. It is a by-product of a safety technology, collected by hobbyists, and understanding how it works explains something useful about your own flying: why the trail is there while you're in the air, and gone the moment you land.

It isn't radar

Radar works by shouting and listening for the echo. A ground station emits a pulse, the pulse bounces off an aircraft, and the time it takes to come back gives a range and a bearing. It is expensive, it needs enormous rotating hardware, and it doesn't reach very far out to sea.

What the live maps run on is almost the opposite, and it's called ADS-B — Automatic Dependent Surveillance–Broadcast. Instead of being interrogated, the aircraft simply announces itself. It reads its own position from GNSS satellites and broadcasts that position, its altitude, its heading and its identity, roughly twice a second, on 1090 MHz.

Three properties of that broadcast explain everything else:

• It is continuous. Nobody has to ask.
• It is unencrypted. There is no key, no authentication, no subscription.
• It is omnidirectional. Anyone within line of sight can hear it.

The signal is a safety measure — it lets aircraft see each other, and lets controllers track traffic without radar coverage. The fact that anyone with an aerial can also hear it was not the point, but it wasn't prevented either.

The receivers are volunteers

Because the broadcast is public, receiving it needs remarkably little: a cheap software-defined radio dongle, an antenna with a clear view of the sky, and a computer the size of a credit card. The whole setup costs about the price of a checked bag.

The tracking sites are, in effect, aggregation layers over tens of thousands of these receivers, run by enthusiasts who feed data in exchange for a free premium account. OpenSky Network, which is run as a research consortium and publishes its data openly, is the clearest illustration of the model.

This is why coverage looks the way it does. It is not distributed by importance, it is distributed by where people live. Western Europe is saturated. So are the eastern and western United States. Large parts of central Africa, Siberia and the Southern Ocean are thin or empty, not because those aircraft aren't broadcasting, but because there is nobody underneath them listening.

Why the middle of the ocean used to go dark

1090 MHz is line of sight. From an aircraft at cruising altitude the horizon is a few hundred kilometres away, which sets the practical range of any ground receiver.

A few hundred kilometres is nothing against the Atlantic, and much less than nothing against the Pacific. For most of the history of these sites, aircraft crossing an ocean simply disappeared from the map and reappeared some hours later on the far side, with a straight line drawn optimistically between the two points.

Two things filled the gap:

• Multilateration (MLAT). Older aircraft without ADS-B still reply to interrogation with a basic transponder signal. If four or more receivers hear the same reply, the tiny differences in arrival time can be solved into a position. It needs a dense receiver network, so it works over land and not over water.
• Satellite ADS-B. Putting the receivers in orbit instead of on the ground removes the line-of-sight problem entirely. Aireon hosts ADS-B receivers on the Iridium NEXT satellite constellation, which is what gave the oceans genuine live coverage for the first time.

The map you see today is a composite of all three: terrestrial ADS-B where there are volunteers, MLAT where the aircraft are old, and satellites over the water.

Why none of it remembers you

Here is the part that matters if you care about your own flying.

Everything in that broadcast is about the aircraft. The ICAO 24-bit address, the callsign, the position, the altitude. There is no passenger manifest in the signal, because a manifest has no business being broadcast unencrypted to anyone with an aerial.

So the tracking networks know, in enormous detail, that an Airbus A320 flew from Milan to Copenhagen on a Tuesday. They have no way of knowing, and no interest in knowing, that you were in row 19.

Add the practical limitations on top:

• Historical playback is generally the paywalled part of these services, because storage is the expensive bit.
• The archive is indexed by aircraft registration and flight number — the two things you are least likely to remember about a trip from six years ago.
• Nothing ties any of it to you. There is no query that returns "every flight this person has been on", because that connection was never recorded anywhere in the system.

The live trail is genuinely impressive, and it is genuinely temporary. It exists while the aircraft is in the air, and its purpose ends when the wheels touch down. Your having been aboard was never part of the data.

Which leaves the record to you

That's not a flaw in ADS-B. It is doing exactly what it was designed to do, and doing it well. It just means the thing many people assume is being kept somewhere — a complete history of where they personally have flown — is not being kept anywhere at all.

It is recoverable, though, and mostly from your own inbox. Airline confirmations, boarding pass emails and photo timestamps between them will rebuild the large majority of a flying history in about an evening.

Once you have the routes, the numbers follow from geometry. Two airport codes and a date give you the great circle distance, the flying time, the countries and the continents — which is what Orbit computes, offline, from a database of 8,801 airports, and draws as arcs on a globe.

The aircraft's trail belongs to the aircraft, and it disappears. Yours is worth writing down.

Related reading: flight tracker or flight log — which one do you actually want? and why your flight to Tokyo goes over the Arctic.