I too have long been concerned about the viability of an earth-mass-coupled house in colder (and especially moister) climates. The incoming solar radiation needs to enter the soil in summer considerably faster than heat conducts deeper into the earth from the mass, as it will be moving down all year as well as moving out to the winter air. The lower the average ground temperature, the more severe that constant loss will be. My family house in upstate NY has an uninsulated concrete and cinder block foundation, open on the east face to a hillside. In the summer, the cellar would always be chilly and damp, and if air was allowed to blow through it would even condense on the undersides of the floor joists as well as walls and concrete floor.
Builditsolar.com has some useful information about soil temperatures for the US:
Ground Temperatures as a Function of Location, Season, and Depth
One item is that it isn't until you get to 12' depth that the annual temperature range diminishes to ~10F min to max. At 5', you still see a range of ~22F.
Another item is that the temperature lag is not as great as one might wish, with maximum soil temps at 5' occurring about early September, and at 12' about late October. The minimum soil temps occur at 5' in March and at 12' in April. The dates may shift a bit later in dry soils, I don't know for sure.
Finally, the average annual soil temperature for a given location fairly well tracks average annual air temperature, and especially average well water temperature. So for a colder location, the earth makes it that much harder to get warm.
The upshot of all this is that real insulation is going to be needed all around the soil mass, top and bottom, and in a colder climate the deep earth temperature might overbalance any gain that could be gotten from summer warmth. I think the viability of a wofati may be limited in cold climates to areas with very dry soils, which will conduct heat slower than moist soils.