Reading this article (highly entertaining, as always, Maciej; thanks for brightening my day with your cynicism) led me to reflect on architecture. It seems pointlessly wasteful to have built a bunch of Quonset huts and trailers that people have to walk between, and also to have built the buildings above ground in the first place. I suppose digging in permafrost is difficult, and bringing in machinery so that you can have economies of scale is difficult, even if we no longer have the problems of cold steel embrittlement and tin pest that bedeviled the old polar explorers.

A little quick calculation, since calculation is always what I end up doing when confronting stories of human folly and suffering.

If you need to house, say, 1024 people, with 128 m² of floor area for each one (home plus office plus bar, etc.), with a mean ceiling height of 4 m, that’s 524 288 m³. If you want to enclose that volume inside a hemispherical dome, the radius (and thus the height) of the dome is about 64 m, or 16 floors. The skin of the dome — the part that insulates the people from the cold wind — is about 26000 square meters, 25 square meters per person. The cross-sectional area that the dim sun illuminates during the summer is about 6400 m², so if we assume a bit less than 1000 W/m², you receive about 6 megawatts of solar energy during the summer, and about 3 megawatts year-round (3 kW per person).

How much insulation do you need? If the inside-outside temperature difference is 40° C, and you need to maintain that on 1500 W per person (maybe your thermal solar collection is only 50% efficient and you don’t have significant other sources of heat) then you need insulation with an average R-value of about 3.8. 1-inch polyisocyanurate foam panels have an R-value of about 6, and they only cost US$19 for a 4’×8’ panel (US$6.40/m²; this is the Home Depot retail price and includes aluminum facers) which works out to about US$160 000 to cover the whole dome. Unfortunately they aren’t transparent, so you can’t get solar radiation through them; you kind of need some kind of non-imaging optics heliostat if you want to gather the solar heat to illuminate and keep warm with. As far as I know, these don’t exist yet.

At this point, and certainly when McMurdo was built, it would make more sense to use heavier insulation, and do your climate control by dissipating energy that you generate by some other means, such as with the nuclear reactor or by burning fuel oil. If you’re dissipating 500 watts per person, you need three times the R-value (11.4, a bit under two inches of foam insulation, or US$320 000 of insulation). You need to use countercurrent heat exchangers to keep the air from going stale.

16 stories is small enough that people can avoid using elevators most of the time, at least if the common areas they usually travel to are intelligently located; 128 meters diameter is small enough that you can walk anywhere (in about two minutes), but large enough that bicycles or skateboards would occasionally be convenient.

If each floor is concrete 250mm thick (suitable for essentially any purpose that doesn’t involve armored vehicles), we need 32768m³ of concrete, or about 65536 tonnes, if we use somewhat lightweight concrete. Concrete typically costs about US$120/m³, so that’s about US$4 million of concrete, plus probably a similar cost in rebar. This cost isn’t particularly sensitive to whether you build a bunch of separated buildings or a single giant arcology like I’m suggesting above, as long as the buildings are more than two or three stories tall. It is sensitive to things like whether each person gets 128m² or 32m² and to the flooring material.

If you could somehow get by with more inexpensive floor materials like expanded steel sheet with drywall under it, you could reduce the cost dramatically — the 250mm reinforced concrete I suggested above costs about US$60/m², while 9-gauge expanded steel sheet might cost US$24/m² (according to MetalsDepot.com). I feel like that kind of thing might be acceptable for a lot of floors that don’t separate unrelated strangers.

This is important not so much for the cost of the materials (although that is kind of important) but more because you don't have a cement plant or even a quarry onsite there in McMurdo; all your manufactured materials have to be shipped in, as if you were in Alaska or something. A heavy-tested TEU only holds 28 tonnes; the amount of concrete suggested above is 2340 TEUs’ worth of mostly sand and rocks. That’s because the concrete weighs a ton per square meter, while the expanded sheet metal weighs 8.8 kg per square meter. That way, you might only need 25 or 30 TEUs instead of 2300 of them.

(The above conveniently omits the sheetrock...)

What about trash? Is it really necessary to haul it away from Antarctica? Let’s make some pessimistic assumptions: suppose we need to plan to store 64 years’ worth of garbage — ideally, frozen — and that the McMurdo Base residents and visitors produce the same amount of garbage per capita as New Yorkers, who are twice as productive of waste as any other major metropolis, at 7.8 million tons per year (220 kg/s) out of 8.6 million people (26 mg/s per person). Over 64 years and 1024 people, this is 54 000 tonnes of garbage, or about 54 000 m³.

This is about an order of magnitude smaller than the size of the people dome. If we just build a garbage dome near the people dome and put garbage in it every day and let it freeze, the garbage dome will only be five stories tall if it’s built for 64 years’ worth of garbage.

Presumably it will take less than 64 years for it to become economic to mine the rich deposits of refined mineral resources (indium, gallium, gold, copper, maybe even aluminum if energy prices don’t fall dramatically) in the garbage pile.

Hopefully, the McMurdo residents will demolish less buildings, buy less new clothes, junk less taxis, and compost more of their food than New Yorkers do, so hopefully their garbage volume will be even smaller.

So you could probably build a new, much better McMurdo Base for under ten million dollars.