It seems that, even in this desert, perhaps even especially here.the
temperature of a building should be a matter of how much coupling it
chooses to have to the radiant objects in its environment --- within
the
limits of their temperatures, of course. If the outdoors ever reaches
15 degrees (or 10, or 5), you can increase the coupling of your cool
reservoir to the 15-degree objects until the cool reservoir is
arbitrarily close to that temperature ...
By coupling to a clear night sky, those could be degrees K, not C:
http://www.mathcad.com/library/LibraryContent/MathML/clearnightfrost.htm
(the ancient Egyptians are said to have made ice in this manner)
Citrus growers used to use smudge pots to decouple their crops from the
sky during frost danger nights; now that filling the air with fine
particulates is no longer considered cool, they attack the other end of
the equation and use wind machines to increase the convective flux at
the expense of radiative.
It's even possible to arrange for negative coupling, albeit with a
great deal of damping, by using massive construction. One of Fourier's
first applications of his eponymous series was in heat flow problems,
and his solutions indicated that the daily heating term should be
completely out of phase with a couple of decimeters of soil, and the
yearly with a couple of meters. Fourier remarked that the yearly term
was a convenient depth for cellars; I don't know if he said anything
about the daily term, but it's consistent with the thick walls of
alpine stone buildings, and with my memories of wall depth for Spanish
adobes in California.
I wonder if it's possible to get any work out of desert temperature
differentials? Frost heave and pingos imply that the arctic annual
term can do some work, but at nanoHz frequencies, it'd be even more
difficult to do something interesting with than the usual examples of
low-rpm torques.
-Dave
