Antarctica has 729 known penguin breeding sites south of 60 degrees south. Ask where penguins breed in Antarctica and the honest answer is that almost none of them breed in most of Antarctica. Cut the continent into 36 slices of ten degrees of longitude and three adjacent slices hold 510 of the 729. The rest of the coastline, eleven twelfths of the way round the bottom of the world, holds 219.
The window runs from 70 degrees west to 40 degrees west. It contains the Antarctic Peninsula, the South Shetland Islands and the South Orkney Islands, and that is more or less the entire penguin geography of the continent. One slice inside it, 60 to 50 degrees west, holds 210 sites on its own, which is very nearly as many as all 33 slices outside the window hold between them.
The grid was drawn before anything was counted
A window that holds 70 per cent of the sites is only interesting if somebody did not draw it around the answer. So the grid here is fixed: ten degrees, starting at the date line, 36 slices, no adjustable edges. The window is simply the three fullest adjacent slices that the grid produces.
There is a second check, and it comes from a body with no interest in this argument. CCAMLR, the treaty commission that manages Southern Ocean fisheries, divides the ocean into numbered subareas for regulatory reasons that have nothing to do with penguins. Subareas 48.1 and 48.2 cover roughly the same patch of sea. They hold 504 of the 729 sites, six from the sector count, on a boundary nobody here chose.
The concentration is a species list, not a crowd
The natural reading of that number is that penguins pile into the Peninsula because it is the warmest, most ice free, most productive corner of the continent. That is part of it. The larger part is that four of the six Antarctic penguins simply do not live anywhere else.
| Species | Sites | Inside window | Sectors of 36 |
|---|---|---|---|
| Chinstrap | 369 | 99.5% | 4 |
| Adelie | 287 | 43.2% | 23 |
| Gentoo | 142 | 100% | 3 |
| Emperor | 68 | 11.8% | 34 |
| Macaroni | 44 | 100% | 3 |
| King | 8 | 100% | 2 |
Gentoo, king and macaroni are inside that window at every single recorded site. Chinstrap misses at two sites out of 369. Those four species are not Antarctic in the way the word suggests. They are birds of the Scotia arc and the Peninsula, which happen to fall south of the line.
Two species behave completely differently. The emperor is recorded in 34 of the 36 sectors and the Adelie in 23. They go right round. And these are exactly the two whose habitat this site can measure from orbit, which is not a coincidence: a circumpolar bird on sea ice or on open coastal rock is a bird a satellite can find, count and follow. Every site is drawn on the atlas, six species, six frames, one per species.
The emperor keeps to itself, and the reason is rock
Run the co-occurrence: how often does a species share its breeding site with another species? For five of the six the answer runs between 28 and 100 per cent. For the emperor it is 7 sites out of 68.
All seven are shared with the Adelie. Not one is shared with a gentoo, a chinstrap, a macaroni or a king without an Adelie also present.
The explanation is underfoot. Five of the six species need bare rock that clears of snow early enough in spring to build a nest on, and on a continent buried under an ice sheet the supply of such headlands is tiny. The same few capes turn up in list after list and species stack on top of each other. The emperor does not want rock. It breeds on fast ice, the sea ice welded to the coast, which is why it is the only penguin that incubates an egg through the polar winter and the only one not queueing for a rocky beach.
That habit is also why the emperor is the species this site watches hardest. Fast ice is a floor that has to arrive in autumn and stay until a chick can swim, around mid December. We track when the sea ice around each emperor colony arrives and when it goes, at 61 sites, back to 1979. Some of that record is unsettling and quite a lot of it is a flat "nothing measurable has happened here", which is a finding people rarely publish. Both are on the break-out map.
Two claims here, and only one is clean
That four species are confined to the Peninsula and the Scotia arc is a statement about range, and it is safe. Range is a map of where a bird can live. It does not care how many ships went where.
The density of sites inside the window is a different claim and we will not make it. 144 recorded sites in the South Shetlands against 31 in Wilkes Land is partly a map of where the tourist ships and the research vessels go. The Peninsula is a day from Ushuaia and Wilkes Land is not. Anyone quoting the site count as a population map is reading survey effort and calling it biology. The range claim survives that objection because the two circumpolar species have been censused from orbit, over the whole coastline at once, with no ship involved.
And there are no counts here at all. This site publishes breeding site positions and species lists, never a number of birds. We built a detector to read colony sizes off guano staining in satellite imagery and it does not work: hand tuned thresholds correlated at minus 0.16 against the field counts, which is worse than useless. Until that is fixed with labelled training data, a colony count on this site would be a fabrication. The positions come from the MAPPPD open data archive, the Oceanites penguin database, published as a Darwin Core archive on the Antarctic Biodiversity Portal with no key and no account.