https://en.wikipedia.org/wiki/List_of_pumped-storage_hydroel...
EX: A single location is storing 4,200 gigawatt-hours per year:
https://en.wikipedia.org/wiki/Huizhou_Pumped_Storage_Power_S...
Siting and water flows for such installations are somewhat lacking.
http://physics.ucsd.edu/do-the-math/2011/11/pump-up-the-stor...
There math is also wrong on energy density, you get ~30% efficiency with gas > electricity. So one gallon of gas takes less water on balance. ~4 cubic meters of water dropping one mile ~= 1 gallon. A 10 meter drop in water level over 1 km^2 = ~2,500,000 gallons of gas per day and 0.9 billion gallons of gas per year.
Sure, you would need two ~1 km^2 square body's of water within ~3 miles of each other with one ~1 mile higher than another. But, building two 10-20m deep pools is cheap, so now you just need a large flat hill or an unusually deep valley.
PS: People talk about how variable wind is, but if you ignore the peak and compare median power to minimal power it's actually surprisingly steady. Yes, we will waste a few days at peak power, but that's not actually a lot of wasted energy.
If you want to see seasonal trends, the Fraunhoffer Institute in Germany has excellent actual production plots for Germany. Solar does well in summer, not so much in winter. Some of the slack is made up for with wind.
There's additional other sources, including wastestream processing, though that itself is just derivative of biofuels. Germany has surprisingly little geothermal potential.
And though Germany is a high-latitude country, it's fairly representative of about 450 million people living in modern high-energy lifestyles, throughout Europe.
Solar hot water heating is much more efficient ~90% and easy to store locally. Payback is ~3-5 years in most of the US even if you only need heat for a few months out of the year. Granted, it does not work in Alaska, but it does work in Mane.
Thus, Wind for base load 24/7/365. Solar for peak power usage. Toss in a 30% buffer + ~20% extra generating capacity get's you to 5 9's.
This isn't guaranteed to be true. Before electrical household appliances were common, we used far more energy at night because indoor lighting was the primary use case for energy. And we are likely to revert to this pattern again if we ever have a mass move toward electric cars which will be charging at night.
Presently with inflexible base-load power generation (coal and nuclear), that's at night.
Given nondispatchable opportunistic generation, that's likely to be during daylight hours.
Storage is the electrical demand of last resort, along with other very-high-energy uses (e.g., thermal banking, aluminium smelting, etc.).
Using Tesla numbers, 230 miles on 60kwh. Average person drives 13,000 miles per year which averages out to 9.3kwh per day. Split that over 12 hours and your looking at ~780watts.
Further, if it cost less we would charge cars during the day.
PS: Average household uses 31kwh per day. But, offices are really daytime biased.
As to geography you can also move electricity ~1,000 miles with minimal losses, So there is some dependence on geography but outside of some islands you can use pumped storage just about anywhere in the world as all you really need is a large hill to start with.
That seems like a lot, but for comparison purposes, China uses 5,400,000 GWh/yr. I don't know how many sites China has available to build storage plants of this capacity; the terrain there looks kinda special --- Wikipedia claims the lower dam is over 400m high...
The scale of solar and wind build-out required, and of energy storage, are both absolutely staggering.
Largest built infrastructure, possibly short of highway and road systems (which are significantly lower tech, and have far fewer moving parts).
Though highway maintenance costs might give you pause as well. Lifetime for solar and wind plant is roughly 20 years, due to various degredation factors.