It doesn't require specific geography. You can create entirely artificial pump systems isolated from any natural water. These are called closed loop systems. You could create an closed loop pumped hydroeletric system in the middle of the Sahara if so desired. Incidentally you could power the entire world with a solar area taking up a single digit percent of the Sahara.
These haven't been actively developed in the past because we don't have much need for massive storage and they also take something on the order of a couple of years to plan, develop, and execute. You need the demand to be there before the storage is built, but the demand won't exist until the storage is built. Fun problems. Because of this batteries are a more practical immediate solution. They can be deployed anywhere, at practically any scale, with negligible time requirements. And similarly for manufacturing. Since they aren't 'geo locked' their market flexibility is much greater.
The Sahara-global-solar-facility scale starts bumping up a lot when you factor in realities.
PV efficiency, spacing factors, panel replacement cycles, storage requirements, the fact that we're looking at total energy use and not just electricity, first-world rather than third-world per-capita use rates (presuming we're not going to freeze the entire world at its present state of energy consumption), and projected population growth.
You can still provide most or all the hypothetical demand from the Sahara, but you're well above 1% land use. I've sketched this out elsewhere previously, don't have numbers handy.
I worked out the numbers several years back and it was around 5%. Efficiency improvements since then should have improved this a fair amount. Even if you bump it up by an order of magnitude it's still rather remarkable how easily we could power the entire world on solar alone.
However, I completely agree on the real issue being one of longterm consumption. I would say this is something that's regularly ignored. The developing world starting to consume developed world electricity/capita alongside increasing world population is easily going to increase energy consumption by some orders of magnitude in the foreseeable future.
This poses unique challenges few are considering. For instance nuclear also runs into problems here with resource availability. The technology is already rather cost prohibitive and for future energy needs if it became a primary source you'd absolutely need to move to breeder reactors alongside saltwater uranium extraction which would both push the prices up significantly higher than even present. High energy demands alongside high energy prices might make the production owners/shareholders happy, but not much of anybody else.
In any case sooner or later we'll end up relying on solar simply because nothing else can compete on gross energy availability. The sun's a fusion reactor that could fit about 1.3 million Earths inside of it. That enables practically unlimited power out there just waiting to be harnessed one way or the other.
These haven't been actively developed in the past because we don't have much need for massive storage and they also take something on the order of a couple of years to plan, develop, and execute. You need the demand to be there before the storage is built, but the demand won't exist until the storage is built. Fun problems. Because of this batteries are a more practical immediate solution. They can be deployed anywhere, at practically any scale, with negligible time requirements. And similarly for manufacturing. Since they aren't 'geo locked' their market flexibility is much greater.