Some time in the next few days, at a spot on the East Antarctic plateau where the nearest other occupied building is 558 kilometres away, the sun will come back. Concordia Station has not seen it since 2 May. That is 104 days of continuous night, and for every one of them the station has been unreachable. Nothing flies over the Antarctic interior in winter. We know because we watch the flight feeds, and for months at a time the query for the entire continent comes back empty.
Concordia sits at 75.100 degrees south, 123.333 degrees east, on the rise in the ice sheet called Dome C, at an elevation of 3,233 metres. In the COMNAP register of every facility on the continent it is the fourth highest, behind China's Kunlun at 4,087 metres, Japan's Dome Fuji at 3,810 and Russia's Vostok at 3,488. It was established in 1997, it is listed with a peak population of 80, and it is one of only 44 facilities out of 114 that stay staffed all year.
Why anybody built there
Nothing about Dome C is convenient. Dumont d'Urville, the French coastal station the plateau traverse sets out from, is 1,109 kilometres away. Vostok, the only other occupied place within a thousand kilometres, is 558. There is no wildlife at Dome C. There is no rock. There is no liquid water. What there is, is the best sky on Earth.
Air at 3,233 metres over the interior of Antarctica is the driest, stillest and darkest air anywhere on the planet. Almost no water vapour, because the cold has wrung it out. Almost no turbulence, because the plateau is flat for hundreds of kilometres and the air above it barely moves. Almost no light, for a third of the year. Astronomers and astrophysicists rate Dome C and the South Pole as the two best ground sites in the world for the same reasons, and both are staffed through the winter because that is when the observing is.
Underneath the station is the second reason. The ice at Dome C is old, thick and undisturbed, laid down in annual layers that have not flowed sideways enough to scramble them. Ice cores drilled here and at Vostok and Dome Fuji carry an atmospheric record reaching back more than 800,000 years, trapped as bubbles of ancient air. The hunt for a core older still is an active project on this plateau. There is no other archive of the atmosphere like it, and it exists only where the ice is deep, cold and slow.
The third reason is the people
The European Space Agency uses Concordia as a Mars analogue. That phrase gets used loosely about deserts and volcanic fields, and usually it means the ground looks right in a photograph. Here it means something harder. A winter crew a fraction of the size of the summer's 80 stays through the dark in a sealed habitat, at an air pressure lower than 3,233 metres would give you anywhere else on Earth because the polar atmosphere is thinner over the pole, at temperatures that make going outside a procedure rather than a walk, with no possibility of evacuation and no resupply for the whole of the polar night. That is not a simulation of isolation. It is isolation, and it is the closest thing on this planet to the crew problem of a Mars transit.
What comes out of it is human biology and medicine: sleep, immune function, cognition, group dynamics, how a small team degrades and recovers over a long confinement. The station is the instrument and the crew is the sample.
| Field | Value |
|---|---|
| Position | 75.100 S, 123.333 E |
| Elevation | 3,233 m |
| Established | 1997 |
| Seasonality | Year round |
| Peak population | 80 |
| Polar night | about 104 days, 2 May to 13 August |
| Rank by elevation | 4th of 114 facilities |
How long the dark actually lasts
The 104 day figure is worth showing the working for, because polar night is quoted carelessly all the time. At latitude 75.1 south the sun fails to clear the horizon on any day when its declination is above 14.9 degrees north. Run the solar declination through the year and that condition holds from about 2 May to about 13 August. That is the geometric definition: the centre of the sun, a flat horizon, no atmosphere.
Reality is kinder at both ends. Atmospheric refraction lifts the sun's image by roughly half a degree, and standing 3,233 metres up pushes the horizon down further still, so the first and last slivers arrive a few days earlier and leave a few days later than the geometry says. Twilight is longer again. What the crew gets is not 104 days of black, but 104 days without the disc of the sun, bracketed by weeks of blue dusk.
A small defect, and what it says about open data
Concordia is one of 15 facilities that publish a webcam address directly in the COMNAP data, which is one of the reasons that dataset is the best open source on this list. Until yesterday, our link to it was broken, and had been since the station page was built.
The cause is a single missing prefix. COMNAP publishes Concordia's camera as a bare hostname with no scheme in front of it, and a bare hostname in a web link is not an address, it is a relative path. Our page was pointing at a file inside our own station directory that has never existed. It resolved cleanly, returned nothing, and looked fine until somebody clicked it. That is the whole business of running a data site: an HTTP 200 is not proof of anything, and the failures that survive longest are the ones that do not announce themselves.
The fixed version is on Concordia's page, along with its position on the polar map. All 114 facilities are in the station directory, filterable by operator and season, and the live continent is on the map.
The source
Everything factual above about the station itself comes from the COMNAP Antarctic Facilities Master list, published by the Council of Managers of National Antarctic Programs and mirrored openly by the Polar Geospatial Center. It is 29 kilobytes of CSV, it needs no key, and it carries operator, seasonality, status, year established, elevation, peak population and webcam URL for every facility on the continent. The polar night dates are computed from the station's latitude, and the method is stated above so you can check them.