You said it yourself, the Atacama is the driest desert on earth. I imagine that environment would be the most difficult to extract water from.
Develop the tech in easier locales then, once the major details are worked out, target the outliers.
Develop the tech in easier locales then, once the major details are worked out, target the outliers.
We build our space equipment the same way... arrange for/synthesize/simulate a close analogue of an environment for initial development and then move on to the real thing once the initial quirks are solved. I certainly don't know the science behind water extraction but I imagine an extremely dry locale would not be the best place to be performing all your validation tests. Maybe start with a kind-of-dry desert, or some other biome. Like Joshua Tree instead of Death Valley...