How far it is from anybody
Antarctica has 114 research facilities on it, run by 31 national programmes, and almost every map you have seen of them is a map of the dots. The interesting question is the space between: how far can you get from all of them at once? Measured on the ice sheet grid, cell by cell, the answer is 924 km — and that is in midsummer, with every seasonal camp open. In winter 59 of the 103 open facilities shut, and the furthest point on the ice goes to 1,815 km from the nearest one.
What February takes away
The Antarctic field season runs roughly October to February. What is left after it is not a thinned-out version of the same map — it is a different map. Of the 114 facilities on the COMNAP list 11 are recorded as temporarily closed, and 59 of the 103 that are open shut for the winter. The 44 that remain are unevenly spread, so the median patch of ice goes from 300 km to 515 km from the nearest of them, and the share of the mapped ice more than 1,000 km from anywhere goes from nothing at all to 12.3%. Well over half the human presence on this continent is a summer visit.
| Distance from the nearest facility | Summer · 103 open | Winter · 44 open | Difference |
|---|---|---|---|
| more than 100 km | 92.2% | 96.9% | +4.7 |
| more than 250 km | 61.6% | 84.1% | +22.5 |
| more than 500 km | 17.2% | 52.6% | +35.4 |
| more than 750 km | 2.6% | 25.5% | +22.9 |
| more than 1,000 km | 0.0% | 12.3% | +12.3 |
| more than 1,500 km | 0.0% | 1.7% | +1.7 |
Percentages are of the 12.55 million km² of grounded ice and floating shelf the velocity mosaic maps, which is the mask this measurement is made on and not a coastline product — see the method below. The headline is a distance rather than a share for exactly that reason: the remotest point sits deep in the interior, where no coastline convention can reach it.
The emptiest place, and what surrounds it
The winter answer is at 73°41'S 122°18'W, on the ice behind the Amundsen Sea coast. What makes it the emptiest point is not that it is far from one station — it is that the three nearest places that stay open are 1,815 km, 1,864 km and 1,866 km away and all three lie in the same direction, across the pole and into the Ross sector. There is no wintering facility anywhere in the sector it sits in. In summer the answer moves to 71°45'S 98°45'W and drops to 924 km, bounded by Sky Blu, WAIS Divide and Estacion Polar Cientifica Conjunta Glaciar Union.
| Season | Furthest ice is | from | Operator | Season |
|---|---|---|---|---|
| winter | 1,815 km | Amundsen-Scott South Pole | United States | year-round |
| winter | 1,864 km | Scott Base | New Zealand | year-round |
| winter | 1,866 km | McMurdo | United States | year-round |
| summer | 924 km | Sky Blu | United Kingdom | seasonal |
| summer | 930 km | WAIS Divide | United States | seasonal |
| summer | 984 km | Estacion Polar Cientifica Conjunta Glaciar Union | Chile | seasonal |
What a station is the nearest thing to
Turn the question round. Instead of asking how far the ice is from a station, ask how much ice each station is the nearest human presence to — the patch of the continent for which that facility, and no other, is the closest thing to a person. Through the winter Amundsen-Scott South Pole holds the largest such patch at 1.52 million km², and ten of the 44 wintering facilities are between them the nearest thing to 72.6% of the mapped ice. Open the summer camps and the same ten hold 18.8%.
Who is on the ice
31 national programmes run something in Antarctica, and the ranking by number of facilities is not the ranking by winter presence: Chile keeps the most facilities on the continent at 14, of which 5 stay open. Of the 114 facilities on the COMNAP list, 83 are stations, 10 are camps, 8 are refuges, 7 are airfield camps, 3 are depots and 3 are laboratorys.
| Programme | Facilities | Open all winter |
|---|---|---|
| Chile | 14 | 5 |
| Argentina | 13 | 7 |
| Russia | 11 | 5 |
| Australia | 7 | 4 |
| United States | 7 | 3 |
| China | 6 | 3 |
| Italy | 6 | 1 |
| Brazil | 5 | 1 |
| Japan | 5 | 1 |
| United Kingdom | 5 | 1 |
When they were built
97 of the 114 facilities record the year they were established, the oldest in 1904 and the newest in 2024. Two periods stand out and neither is the one people expect: 9 were established in the three years around the International Geophysical Year of 1957–58, and 17 in the four years 1988–91. Read this as a record of what is still standing rather than a history of Antarctic occupation — every base abandoned before today is missing from it, and that is a large number this dataset cannot see.
The people and the penguins are in the same place
The atlas on this site already says that four of the six penguin species here never leave one twelfth of the coastline, and it says plainly that the density of recorded sites inside that window is partly a map of where the ships go. This measurement is that caveat with a number on it. The median gentoo breeding site is 13.0 km from the nearest research facility and 90.8% of them are within 50 km. The median emperor colony is 204.2 km away and only 13.2% are within 50 km — 5 times the distance, on the same continent, from the same list of facilities.
Two explanations fit this and the data cannot separate them. One is survey effort: a colony gets onto a list when somebody walks to it, and people walk from bases. The other is that both want the same ground — the five rock-nesting penguins need bare coastal rock that clears of snow in spring, there is very little of it, and a research station wants exactly the same thing. Under perfect survey coverage they would still be neighbours.
What tilts it, without settling it, is the emperor. It is the one species here that is mapped from orbit rather than visited — it breeds on fast ice, so it is not competing for the rock, and its colonies were found by looking for guano stains in satellite imagery rather than by anyone arriving. Its map bears no relation to where the people are. Where the record does not depend on somebody being there, the geography changes completely. That is suggestive and it is not proof, because the substrate changed at the same time as the method.
Where the penguins breed, all 729 sites → · The sea ice around the 67 emperor colonies →
How this was measured
Every cell of the ice mask is unprojected to its true longitude and latitude and the great-circle distance is taken to each facility position in the set — 103 for summer, 44 for winter; the smallest is that cell's value. Planar distance in the projection would have been wrong by up to 8% — EPSG:3031 is true to scale at 71°S and runs long towards the pole, which at these ranges is a hundred kilometres. Cell areas carry the same correction: a polar stereographic cell is not a constant patch of ground, so each is weighted by 1/k² rather than counted. The inverse projection is round-tripped against the forward one before anything is measured and agrees to 2.07 km, inside one 4.9 km grid cell.
The ice is the valid-data mask of the ITS_LIVE velocity mosaic — grounded ice plus floating shelf, the same mask this site draws its continent silhouette from. It totals 12.55 million km² against the continent's usual 14.0 including shelves, so it is not a coastline product and nothing here is described as a share of Antarctica. Every percentage on this page is a share of that mask and says so. Contours are marching squares run only on cells whose four corners are all inside the mask, so a line stops at the coast rather than being drawn round a value that was invented there.
Facility positions, seasonality and status are the COMNAP Antarctic Facilities Master list as published. "Winter" here means the 44 facilities it records as year-round and open — that is a property of the facility, not a headcount, and this site does not have one: the list carries a peak population for 93 of the 114 and no winter population for any of them. The 11 the list records as temporarily closed are in neither set.