Solar powered refrigerated root cellar?

I did a quick search before posting this, so let me know if this already exists... And I need some maths education!

Been having thoughts about either adapting a big rv system, or even brainstorming a baseline system design for a solar powered root cellar.

Basic idea is use a parabolic reflector to heat the refrigerant and have the condenser located where it will directly cool a root cellar-like space, or use an appropriate fluid(Epsom salt saturated water?) to pump the heat out thru a radiator...

It feels to me like it could be possible to keep a properly designed system below 40defF... But I don't know the proper math to figure it out so educate me if you can!!

Years ago I attended a lecture where an Ontario Farmer was using some sort of refrigerating system to keep his root cellar at an optimum temperature, but I don't think he tried to make it solar at that time.

It would help if you would give us some idea of your climate - or at least the expected temperatures during the part of the year you feel you need the root cellar cooled.

Also, some idea of the design of the cellar. The Ontario fellow had built his into a slope, so it wasn't as deep as a true cellar.

This isn't for anything existing, but I'm thinking of something that could work in the pacific northwest,  Ohler style build.

My basic idea was some kind of natural refrigerant like salt water, maybe with some ethanol, in a large pex tube system and heat exchangers.

I'm looking for a system that doesn't use electronics, just solar energy and thermal dynamics if possible.
The ultimate lowtech luxury of off grid refrigerator temperatures in a root cellar (or at least part of one.)

It also seems like an interesting tool when it comes to moving air and water around an environment.

Being able to actively cool a thing or space with a passive generation setup and some kind of "cold battery"  just sounds like a good time.  

What is the temperature of your subsoil? Can you add lots of insulation above and naturally cool from below? That's as low-tech and old-school as it gets.

Up here in Portland area I believe it's 55 on the high side.

The goal would be beyond natural cooling, possibly a freezer box in a root cellar even.

No exact location, this is a design exercise really. Thinking along the lines of an earth sheltered Ohler home or walipini.

Trying to figure out moving the heat around, circulation requirements...

I know the larger ammonia cooling systems can get cold enough to make ice or keep ice cream for days. But I'm trying to think of a way to build a passive system that doesn't use electronics, and preferably no mechanical pumps

Ah! Not cold enough to go passive then.

It's an interesting challenge in that climate. Usually some significant temperature differentials are required.

Don't play with anhydrous ammonia though. It's dangerous as hell.

Do you have enough seasonal variation to create a large, cold mass you could then draw on for cooling in the warmer months?