Purpose-built pumped storage plants, such as Helms near Fresno, tend to have much smaller reservoirs, about enough for a daily cycle, and they go from full to empty and back each day. This annoys some people. They're not recreational lakes. Efficiency is 60% - 80% round trip from electricity in to electricity out.
Actually, the variance is worse than assumed by enthusiasts. Here's today's output from wind, hour by hour, for the entire CAISO area, which is about 400 miles across and covers most of California.[1] Today was a bad day; about 10x variation between peak wind and minimum wind. 5:1 is typical; a good day is 2:1.
You can also see similar data for PJM, which covers the Northeastern US.[2] Today's wind variation is about 2:1.
There's a myth that, with enough transmission distance, fluctuations in wind will average out. The two largest power grids in the US aren't seeing that. Maybe for the whole planet, but not for any area one can reasonably interconnect.
[1] http://content.caiso.com/green/renewrpt/DailyRenewablesWatch... [2] https://dataviewer.pjm.com/dataviewer/pages/public/windpower...
Also of note you can smooth out or shift solar by shifting power east-west without involving storage.
And the actual numbers are considerably less optimistic than you report, and that paper was rather optimistic to begin with.
But I agree with the gist of it: storage worries are quite a bit overrated,and oversupply and transmission capacities are underrated. We can fix those easily enough, though
Yes, except for those who live next to the reservoir (or actually, in the would-be reservoir). That could be people, or nature.
Or in the extreme case, consider Lake Erie is next to Lake Ontario. Pumping extra water back up for a few hours a day would have minimal issues as boat bilge water has already cross contaminated them.
PS: The water feature is even down for matinee. http://www.refinery29.com/2016/02/102431/new-york-niagara-fa...
But isn't that what it does? Unless you start to limit (more than before) the waterflow to the falls themselves, which probably would meet resistance from conservationists.
(I understand the Lewiston power plant already does have capability to pump water up.)
And, many places don't have such reservoir capacity at hand. Over here any new reservoir project meets fierece resistance, and it's rather the other way round: nature conservationists want to shut down existing hydroelectric capacity.
It's complex there are few environmental effects from blocking the falls as they are already one way and tend to kill fish that make the trip. However, there is a treaty that set's minimum flow levels and more water is diverted at night to a holding area which is then used for hydro in the day.
https://en.wikipedia.org/wiki/Niagara_Falls_hydroelectric_ge...
Anyway, the area already has plenty of pumped storage it's only by importing a lot of power that you would need to expand. But, at that point you can scale really far with few environmental costs.
Pumped storage is vastly better than batteries. Also the variance between median days and calm days for wind is far less extreme than generally assumed. PV solar also provides power in cloudy weather.
Anyway the US can get to ~99.99% from ~30% oversupply and ~6 hours of grid storage which is vastly less than many people assume.
PS: Cloudy calm days over large areas are fairly rare. Heavy clouds are generally associated with weather and increased wind. EX: Hurricane incoming, solar down wind UP.