After doing energy modeling for my own earthship, here are some of my conclusions on the best ways to keep the house warm without requiring any artificial heat in winter in the coldest city in the world here.
First, thermal buffer zones. In winter at night heat transfer occurs from the inside to the colder environment outside. Practically what happens is heat in the inner air circulates around the living space and makes contact on a window or the ground where a heat exchange occurs. Air can freely waft through the room to a surface which interfaces with the external environment. Still air is an insulator. Relatively still air has about an R-1 value per inch. Completely still air has about an R-3 value per inch. A lot of insulation works on the principle of trapping still air to take advantage of these facts. A packaged earthship or vol. 1 earthship had the living space and the external environment. People inside walk around and hot air travels to the glass and there is heat loss. If we have the living space in an envelope with almost no direct contact with the outside environment, then thermal buffer zones are created. Generally people don't walk through the thermal buffer zones at night. The global model has buffer zones in the 2nd greenhouse and optionally in the 3rd greenhouse. However, if we can encase the entire living space in thermal buffer zones (to the south, north, east, west, above and below), the results are improved. This is one of the best ways to deal with cold sub-arctic climates.
Second, maximize solar gain. An earthship or passive solar house is heated by solar gain. Therefore, if we're in a cold climate, we have to maximize solar gain--period. In Mongolia I was in an unheated room in early January. The room only had 2 single pane windows. In the room I had a bike with a thermometer. Around noon and until about 1:30 - 2:00 it was in direct sunlight. The high of the day was about -20C. In the sun it reached over +20C/70F. In a cold climate it's absolutely necessary to maximize the solar gain during the day. If it's too much, we can always draw shades. If it's day time in winter, what can obstruct solar gain? There are various things. One pane of standard width (6mm) glass reduces the amount of solar gain by 10-15% depending on construction assuming there is no tinting or covering. We went into some local glass stores and they were even selling quadruple glazing (4 panes with air between them). If you had quadruple glazing over the living space, the 2nd greenhouse and the outside of the 3rd greenhouse, that's 12 panes of glass light has to travel through. If we say the glass is completely clear and we optimistically assumed 90% of the solar gain passed through each layer, the best case scenario would be that 28.2% of the sunlight reaches the living space. If we assume 85%, then 14.2% reaches the living space. If condensation, ice, snow or dirt is present on the glass, even less. If we had used a total of 3 panes of glass, 72.9% of the solar gain would enter the living space, which is over 2.5 times as much solar gain, which becomes heat at night time.
Third, minimize heat loss at night. Earthships in middle latitudes, like New Mexico, are probably warm enough at night, even on the coldest nights. Using insulated glazing is OK in such environments. Double insulated glazing might be OK is many places where some heat loss at night is acceptable. It may not acceptable in winter in the coldest city in the world or sub-arctic climates. If we use single pane glass, we get 72.9% of the solar gain. Once the sun goes down we cover one (or more) of the windows with thermal curtains, which can have over R-10 value. These curtains must seal around the edges. A good strategy is also to trap about 25cm of air between the glass and the curtains. This way we maximize solar gain, but trap that extra heat inside the house. The best way is to put the curtains on automated control, so they open in the morning, close around sunset or dark periods (whenever heat loss occurs) and open at sunrise automatically. Quadruple glazing may have R-7. If may have more with certain gases (which leak after years of use). It costs at least 4 times as much as single glazing, which could be tens of thousands of dollars more. R-7 is OK, but not warm enough in sub-arctic environments. Also, once condensation infiltrates that kind of glass, it generally remains until maintenance is done.
Fourth, thermal mass. Thermal mass can store heat over time. Heat absorbed during the day can radiate into the living space at night. It also stabilizes the temperature in the room. Without thermal mass artificial heat in cold climate is pretty much necessary. Even in earthships if we have a tire wall in the air locks to the east or west, the airlocks can stay warm due exposure to the sun and the thermal mass in the whole tire wall.
Fifth, heat storage. Some locations closer to the poles don't have enough sunlight, so they can store solar gain in the form of heat in a hot water tank, then spread it into the living space when needed. This makes it possible to build earthships in Alaska, Scandanavia and Europe. I assumed I would need it in Mongolia, but though it's very cold in January, the day is 8.5 hours long on the shortest day and solar gain is fairly constant. My energy model concluded that it's not necessary at all.
Sixth, insulation. How warm is it under your floor? Many places in middle latitudes it's well above freezing down under the frost line not too far below the surface. In middle and northern Mongolia there's permafrost down there, maybe 15 meters deep or more. It is about +0C down there. If I didn't insulate the floor, then heat would be constantly sucked out of the living space into the ground. Therefore, floor insulation is necessary. Without some floor insulation artificial heat is necessary (or heat storage with a radiator system), and floor insulation is probably less money and trouble than constantly heating the place. If the floor is mostly below the frost line, then if you put some vertical insulation around the perimeter, that will keep things above freezing all the time. Even in southern Canada it might be OK. The living space might not be +20C, though. I modeled our situation. 15cm of floor insulation is pretty much necessary if the roof is done well at R-75. Just increasing the thermal wrap wall doesn't negate the necessity of floor insulation. Materials high in thermal mass typically have high conductivity, so without enough insulation heat will quickly be drawn out of the living space. Heat rises so the roof is where the most insulation needs to be. Also wooden doors have very little insulation r-value, so they are like thermal holes. Having a wall of windows with no covering has an inadequate r-value. There are thermal shutters. The problem with earthships is they have to go somewhere when open, so sliding shutters (pieces of insulation) reduce the window area and incoming solar gain during the day (unless one found a way to get them off the front face during the day).
