There is a Boeing 757 parked on the Antarctic ice at 79 degrees south. It is white, it has rows of passenger windows, and it got there under its own power on a runway of polished blue ice at Union Glacier, roughly 1,000 kilometres from the South Pole. Nobody bulldozed a strip for it. The wind did that, over centuries, by scouring the snow away until the glacier underneath was bare and hard enough to take a jet.

Which makes the most-asked question about polar flying a slightly odd one. People type why don’t planes fly over Antarctica into Google several thousand times a month, and the honest first answer is that they do. Not many, and almost never on a scheduled ticket, but they do.

The real question underneath it is better: why is the Arctic threaded with busy airline routes while the Antarctic is, in commercial terms, empty? The answer is not a conspiracy, a no-fly zone or anything to do with the shape of the Earth. It is a stack of unglamorous regulatory and logistical facts, and the most interesting one is that a runway existing in Antarctica does not make it an airport you are allowed to plan on.

Two polar areas, drawn by two completely different rules

American regulations define both ends of the planet, and they do not define them symmetrically. Under 14 CFR 121.7, the North Polar Area is the entire area north of 78° N. The South Polar Area is the entire area south of 60° S. That is an eighteen-degree difference in where the special rules begin, and it is worth understanding why, because it tells you what each set of rules is actually for.

The northern boundary is a navigation line. At 78° N you are deep inside the region where a magnetic compass has stopped meaning anything, where meridians converge so sharply that headings change faster than the aircraft does, and where a diversion airport is a long way away. It is drawn where the physics gets difficult.

The southern boundary is a political one. The Antarctic Treaty Area covers everything between 60° and 90° South — the same line, to the degree. The FAA does not publish its reasoning, but the coincidence is exact, and it means the South Polar Area starts hundreds of miles out over open ocean, well before anything physically difficult begins.

Half as Interesting on the north-versus-south asymmetry — the most-watched explainer on the question, with 4.4 million views.

The compass gives up first

A magnetic compass works by sensing the horizontal component of the Earth’s magnetic field. Near either magnetic pole that horizontal component shrinks towards nothing, because the field is diving almost straight down into the ground. The FAA puts a number on the point of failure: when horizontal field strength falls below 6,000 nanotesla, the compass can no longer be considered reliable.

In practice that draws a circle roughly 1,000 nautical miles wide around each magnetic pole. The FAA calls these Areas of Magnetic Unreliability, and an airline needs a specific authorisation — OpSpec B040 — before it may fly in one. The Northern AMU covers everything within 1,000 NM of the magnetic North Pole except Alaska and its territorial waters, plus the whole of Canadian Northern Domestic Airspace. The Southern AMU is the matching circle around the magnetic South Pole.

Inside those circles the variation between true and magnetic north does not just grow large. It becomes absurd. The FAA’s own application guide offers the cleanest illustration of the problem anyone has written down.

The fix is to stop using magnetic north altogether. Aircraft cleared for polar operations must be able to switch their heading reference from magnetic to true, either by a crew selection or automatically by latitude, and the operator has to document which. In Canadian Northern Domestic Airspace every heading and track is given in degrees true as a matter of course. Controllers there simply do not use magnetic.

There is a second, subtler trap that starts well before the compass fails. Above about 67 degrees of latitude the meridians converge tightly enough that a great-circle course changes its true heading very quickly for very little movement across the ground. The FAA’s warning is that small errors there develop into very large navigation errors over extremely short distances. This is also a moving target: the magnetic North Pole is drifting north-northwest at roughly 45 kilometres a year, so the shape of the unreliable zone changes decade by decade.

Fuel that can freeze in the tank

Jet fuel is not a single substance, and its freeze point is a specification limit rather than a property you can look up for the load in your wings. ASTM D 1655, the international jet fuel standard, sets a maximum freeze point of −40 °C for Jet A and −47 °C for Jet A-1. Those are ceilings. A given batch is usually better, and the operator is expected to know which.

Polar air at cruising altitude is routinely colder than that, and the fuel in a wing tank keeps cooling for hours. So every polar operator has to hold a fuel-freeze strategy and a procedure for monitoring fuel temperature in flight, either by applying the ASTM limits directly or through an FAA-accepted monitoring programme that gives equivalent protection.

What makes this genuinely interesting is the list of things a crew can actually do about it in the air. The FAA names three. Climb or descend into warmer air. Deviate off course towards warmer air. Or increase Mach number — which works because flying faster heats the skin of the aircraft, and the fuel behind it, through friction. A polar crew watching a fuel temperature gauge creep downwards has the option of solving it by going faster.

What it actually takes to get permission

Appendix P to Part 121 is blunt about the threshold. Except for intrastate operations inside Alaska, no certificate holder may operate an aircraft in either polar area unless the FAA has authorised it. The authorisation is not a formality. Section III lists eight things that must appear in the airline’s operations specifications before a single passenger is carried across either line.

Designated en-route diversion airports, and the standard each must meet at the moment of diversion. A recovery plan for passengers at those airports — waived only for all-cargo supplemental operations. A fuel-freeze strategy. A plan to guarantee communications. A dedicated minimum equipment list. A training plan. A plan for mitigating crew radiation exposure during solar flare activity. And, last on the list and the one that sticks in the mind, a plan for carrying at least two cold-weather anti-exposure suits aboard, to protect crew members during outside activity at a diversion airport in extreme conditions. The FAA will let an operator off that one if the season makes it unnecessary.

The passenger recovery plan is where the abstraction stops. It has to cover food, shelter and continuing safety, a method of getting everyone off the aircraft safely in bad weather, medical care, communications, and transport to extract passengers and crew — with execution and completion expected within 48 hours of the diversion. Two days is the planning assumption for how long people might be sitting on an ice runway.

None of it is granted on paperwork alone. An operator has to pass an FAA tabletop exercise and then fly a validation flight including a simulated diversion into an approved en-route diversion airport, which is assessed for communications, navigation and surveillance, coordination, facilities, the accuracy of NOTAMs and weather, space weather information and whether the ground equipment there actually works. There is a quietly telling rule attached: the validation flight may carry revenue passengers only if the passenger recovery element has already been validated separately. Rehearse abandoning passengers in Antarctica, and you may not have paying passengers aboard while you do it.

Why the Arctic is busy and the south is not

Everything above applies equally to both ends of the planet. Airlines pay that price in the north thousands of times a year and almost never in the south, and the reason is embarrassingly simple: there is nothing on the far side of Antarctica worth flying to.

Look at a globe rather than a Mercator chart. New York to Hong Kong, London to Tokyo, Chicago to Delhi — the shortest line between the northern hemisphere’s large cities runs over or near the top of the world, because that is where the great circle goes. Cathay Pacific proved it commercially on 5 July 1998, when CX889 flew New York to Hong Kong non-stop over the North Pole in 15 hours, 35 minutes and 53 seconds with 137 passengers aboard, the longest non-stop commercial service flown to that date. The route was called Polar One. The geometry has not changed since.

The southern hemisphere offers no equivalent. The cities that would need an Antarctic crossing to shorten their route — Perth and Buenos Aires, Cape Town and Auckland — are either small markets, already well served the other way round, or both. An airline would be buying the entire polar rulebook to save an hour on a route that carries one flight a day. It never pencils out. (For why that shortest line looks so strange on a flat map at all, our piece on great circle routes works through the geometry.)

The flights that do cross

Which brings us back to the aircraft that ignore all of the above. On 14 July 2023, Qantas flight QF28 from Santiago to Sydney flew to 74.17° S, out over the Antarctic coast, purely to dodge a band of headwinds on its normal track. That route regularly runs as far south as 70° S in ordinary service. Qantas’s 787s hold a 330-minute extended operations approval, and as Flightradar24 noted at the time, at that figure it is only the centre of the continent that stays out of reach.

One point of pedantry worth making, because the number gets misused: 330 minutes is a type-certification figure for the 787 and an approval granted under Australian rules. It does not appear as a threshold in the FAA’s ladder, which runs 75, 90, 120, 138, 180, 207, 240 and then “beyond 240 minutes”. What the FAA does do is list, among the areas eligible for that beyond-240 category, the oceanic areas between Australia and South America and the South Polar Area. The regulator has a box for Antarctic overflight. Airlines simply rarely tick it.

Then there are the flights that go purely to look. Since 1994 a chartered operation out of Australia has flown day trips over the continent — now on a Qantas 787-9, previously a 747 — covering 9,500 to 10,500 kilometres round trip in roughly twelve and a half hours, of which three to four hours are spent over Antarctica itself. Nobody lands. Every passenger is issued two boarding passes and swaps seats halfway, so the window seats are shared out.

A full trip report from one of the Antarctic sightseeing charters, filmed from the cabin of a Qantas 787.

The widebody that did land

On 15 November 2023 a Norse Atlantic Airways Boeing 787-9 registered LN-FNC, named Everglades, touched down at Troll Airfield in Queen Maud Land — the first 787 ever to land in Antarctica. It had staged from Oslo via Cape Town and carried roughly 12 tonnes of research equipment along with 45 scientists to the Norwegian Troll research station. There is no asphalt at Troll. The 787 landed on a 3,000-metre blue-ice runway.

Sky News coverage of the Norse Atlantic 787-9 arriving at Troll Airfield on 15 November 2023.

It is a genuinely impressive piece of flying, and it is also the exception that explains the rule. That flight was a chartered, months-planned, single-purpose operation into a runway operated by the Norwegian Polar Institute, with Norwegian permission, in the Antarctic summer. It was not a scheduled service and it was not a diversion.

Runways are not alternates

This is the part that most explanations miss, and it is the real answer to the question. Antarctica is not short of places to put an aircraft down. It has blue-ice runways, compacted-snow strips and skiways scattered across the continent. What it does not have is a single one of them that an airline dispatcher may write on a flight plan as a diversion airport.

The reference document is the Antarctic Flight Information Manual, maintained by COMNAP, the Council of Managers of National Antarctic Programs. It is not a public document — access runs through member national Antarctic programmes and carries subscription charges — and the current release is COMNAP-AFIM-25-04, dated 16 March 2026. Its own caveat is unusually direct.

Prior agreement is required before using anyone else’s facility, and the relevant legal instruments of the Antarctic Treaty parties govern access to everything between 60° and 90° South. A runway you must ask permission to use, weeks in advance, from a national research programme that may have nobody on station, is not an alternate. Appendix P asks for designated diversion airports and the standard they must meet at the time of diversion. Antarctica cannot supply that. Everything else — the compass, the fuel, the suits — is solvable engineering. The missing alternate is not.

And no, there is no forbidden zone

The question attracts a certain amount of nonsense, so it is worth closing the loop plainly. There is no secret prohibition, no restricted airspace covering the continent, and nothing being hidden at the bottom of the map. The Antarctic Treaty regulates access to the area and COMNAP coordinates who flies where, but both operate in the open and both publish their guidance.

The actual constraints are duller than any conspiracy and considerably more demanding: a compass that stops working, fuel that can freeze, no alternate you are entitled to use, a passenger recovery plan with a 48-hour clock on it, two survival suits in a locker, and a validation flight you have to pass before you go. Airlines cross the Arctic because the geometry pays for all that. In the south, it does not.

Sources: 14 CFR 121.7 and Appendix P to Part 121 (eCFR); FAA Flight Technologies and Procedures Division, B040 and B055 Application Guide v1.0, 15 August 2025; ASTM D 1655; COMNAP Air Operations and the Antarctic Flight Information Manual; Flightradar24; South China Morning Post; Norse Atlantic Airways; Antarctica Flights.