2:
I'm somewhat sure that the Palm Springs area has its own ground water source (thus the springs in Palm Springs and the wells in Indian Wells). But I doubt that there is really enough water there to support the water needs without being supplemented from the same sources as San Diego, LA, etc. Imperial valley didn't bother us as much since we at least know people eat what they use the water for (or at least animals do)
3: I'm beginning to wonder if Tara and I have actually made a noticeable increase in the fraction of San Diego area fen who travel outside of easy driving range for "major" conventions? By my count, there were only 8 people I can clearly identify as San Diego area fen at Westercon. And I know of at least 3 from Chicago, one from Boston and one from (I don't recall which) either Boston or New York, on top of quite a few from the greater LA area and the Bay Area. Of course, if you take into account that someone can be a $region fan while not actually living in $region (or just consider a fan to belong to a different region based on their fanac) then perhaps there were only 6 San Diego fans and 5 from Chicago.
4: Wondering about the number of people within specific travel times of some locations I did a quick search but could not find anything. Of course the major obstacle would be to get the zip codes within a specific travel time of a given point, since once zip codes are available then population is available from the US census (at least until you cross an international boundary). I suspect that Google maps could be hacked to get the list of zip codes -- by brute force at least: iterate over all zip codes, route to the set location, if travel time is less than limit, add zip code to final list. But I don't have the time or energy to work up such a hack.
What, of course, I'm really wondering about is how many people are within about 90 (or perhaps 120) minutes drive of the hotels hosting various cons just to compare that to their actual attendance to see which ones are likely drawing most or all of their area's fen or potential fen (which would, of course, presume that the fen are evenly distributed within the US population -- a false presumption I'm sure)