Hi,
This is a tempting concept.
A while back I tried to estimate how well such a system would work for my house.
Even with me driving the backhoe, I could not get the numbers to work out. The heating season energy for my house is pretty large, and the cost of an insulated hot water storage excavation plus was high. Controlling heat loss from the excavation requires a lot of care and good insulation. You need a way to keep the insulation dry -- this has been a big problem on some existing commercial installations.
But, I do live in cold climate (MT), and there is only so much I can do to get the heat loss down on my existing house, so the scheme might work better for you with a new house from the ground up.
There is some good information on building the underground heat storage "tanks" -- if you search on "seasonal heat storage" or the like you should find them. There have been a number of them built in Europe, and there are papers that analyze performance (often not up to expectations) on ones that have been in service a while. These are typically larger than needed for one home, but (I think) the issues are the same.
The Drake Landing community uses underground seasonal storage -- their storage scheme is different, but interesting -- search for "Drake" on this page: http://www.builditsolar.com/Experimental/experimental.htm
I think that it is a concept that definitely needs some careful design and analysis. You could start with a heat loss on your projected house:
http://www.builditsolar.com/References/Calculators/HeatLoss/HeatLoss.htm
This will give you and idea of the full winter heat demand.
You can get a very rough idea of the "tank" size by taking the winter heat demand from above and dividing it by the temperature range you can use for the tank -- maybe 150F down to 90F, or a 60F range. So, if you need 30 million BTU for the winter and can run your tank from 150F down to 90F, you need (30000000/60F) = 500000 lbs of water or 60K gallons. This ignores the heat you will collect in the winter, which would reduce the storage, but also ignores tank losses which would increase the tank size.
To get a very rough idea of the collector size you might need to charge the tank, you might assume that you can store about 1000 BTU/sqft of collector per day. So, if the collection season is 150 days(?), you might get 150000 BTU per sqft of collector per season. To store up the 30 million BTU, you might need about 200 sqft of collector -- not so bad.
Some people have looked at this and concluded that it makes more sense to work on getting the house envelope well insulated and tight, incorporate passive solar, and perhaps some active solar but with storage for only a couple cloudy days.
Gary