My obsession with Antarctica has awakened from dormancy, at least for a little while. Here's some thoughts I had.
If I'm going to establish a colony on the southernmost continent, it needs to be mostly autonomous. Currently, every research station is heavily dependent on food, fuel, and building materials from the mainland. This makes sense: their purpose is science, not settlement.
An autonomous, long-term settlement is going to need three things to support its inhabitants. First, it has to provide them with enough food. Second, it needs to keep them warm. Third, it has to be stable.
Antarctica isn't exactly known for being a lush paradise. The predominant natural resource is snow. Near the coast, there's the possibility of fishing, best conducted during the summer months. Considering these fisheries are largely untapped, it shouldn't be too difficult to store up plenty for winter.
There's also the possibility of horticulture. Despite being low in the sky, the sun shines continuously for about half the year, allowing enormous crop growth. They already do this in Alaska, but they have something we won't: fertile soil. We'll have to import our own, and maintain it (through recycling our own wastes). Hopefully food plants like dead fish bits.
We can supplement our food supply by growing things under sun lights in winter time, which brings us to the next problem: energy. Right now, pretty much all power is produced by generators, burning fuel brought down by plane. Importing oil is both unclean and expensive, so we need a better alternative.
The first possibility that I think of is nuclear power, which has actually been done before. However, nuclear reactors are very expensive, not to mention the international community frowns on them, Once a large community has been established it might be viable, but until then we need an alternative.
Even if battery technology improves drastically over the next few years, the months of night make solar still a no-go. I'm not going to rule out geothermal just yet, but I think wind is our best bet.
Consider it. Antarctica has a huge wind potential--after snow, that's probably the easiest resource to exploit. You just have to find a site with relatively constant flow, set up some turbines, and enjoy cheap, regular energy for your base.
That's one constraint on where you build. The other is stability.
One of the big issues for Antarctic explorers has been ice flows. Most of the continent is covered with glaciers, which are moving. If you build on top of them, your base moves with it. Obviously, this isn't something we want for our fledgling colony. We'd like to avoid a Little America situation. There's two ways we can deal with this.
The first is to build our colony where it won't move. This may be as simple as affixing it to one of the rock formations poking through the ice, or could involve building the settlement in one of the Antarctic mountains. Of the two, the former is significantly easier, though offers less security. I anticipate large communities will develop along protected mountains than in the open.
Plains settlements have a possible solution, however. Properly constructed, bases could be moved regularly. Such bases are harder to insulate, which is why permanent Antarctic stations are built on rocky coasts or so far inland a few kilometers drift is nothing to worry about. We'll probably have to do the same thing.
Anyway, those are my thoughts on the matter as of right now. I'll let you know if anything else comes to mind.