Regarding water supply, the best solution is to live near the water. Building a huge city in the desert is asking for problems, and the long pipes and canals and treatment will certainly be expensive. But the best solution isn't always practical, and can have problems of its own, such as supply limits (the creek is only so big; the aquifer recharges only slowly), and water quality (see: Flint MI, well water in many parts of the central and western San Joaquin Valley, well water in much of western WY). Water quality can often, but not always, be handled with treatment, but that can be expensive; sometimes, it's more expensive than building a pipe or canal to someplace that has good water already.

Then there are water management decisions, which are important in places like California where it doesn't rain all year (was on a trip to DC and the taxi driver was complaining about the horrible drought and dry lawns (they were a bit patchy) - it hadn't rained in 3 weeks). This year, for instance, the big reservoirs in the northern Valley (Folsom, Oroville, and Shasta) all topped out at 80-90% full; other places to the south, though, didn't do as well. So as usual there will be a lot of water released to keep the Delta water quality reasonably good - besides the smelt, LA drinks that water. The the managers like to use Folsom when flushing is needed because releases can hit the Delta within a day or so (Shasta releases take more than a week to arrive, and more of Shasta belongs to the farmers than at Folsom), but it's a much smaller lake than the other 2. So despite starting at nearly 90% full (rather than just over 50% last year) I expect it to be emergency-conservation-level empty again by late summer. Think of it as base-load (Shasta, nuke power plant) vs. peaking (Folsom, local gas turbines).

What does it have to do with trains? When you look at the quantities of water involved, even unit trains are the literal drop in the bucket. For a city of 100K, even operating under strict conservation, a 100-car train isn't likely to provide supply for more than a few days to perhaps, at most, a week. So trains might be useful for short-term emergency supplies to conveniently-located places (on or very near the track, at a good place to park a train), but not as a large-scale drought alternative.

Then there's water rights in a world where potable water is becoming permanently scarce in many places, for both population and overdevelopment reasons. Good luck with that.

Yes, it's been discussed before. And yes, India is different - they don't have the kind of high individual water use the U.S. does so a trainload will last a bit longer for a given population served.