> If one tries this approach I'd recommend pumping into a holding pond, ideally up hill from the land to be irrigated. Tanks would be absolutely massive for a crop on any meaningful scale, especially if its big enough to be putting a dent in the aquifer.
Sure, but almost all the countries TFA mentions are generally farmed by small landholders, so irrigation via stored water is not infeasible. The mammoth style broadacre tiny-return-per-hectare, heavily mechanised, fertiliser- and insecticide- reliant style agriculture popular in the USA isn't so common in these places.
The 'dent in the aquifer' you mention (and TFA dwells upon) is because there's thousands of people doing the same thing at the same time to the same aquifer for years on end.
Regrettably TFA doesn't talk volumes, other than three mentions of 'cubic miles'.
Apart from an exasperated 'why do they use anything but metric?' sigh, it's also obviously a heavily aggregated figure across large geo regions.
These small holdings with their panels + pump per field, pulling from an increasingly deep water table, are presumably extracting probably in the range of 10-20 kilolitres of water per day - the size of an average plastic water tank that has a one-off cost about the same as a submersible DC pump + panels. (Lined steel, concrete, etc, become more attractively priced at larger volumes.)
So, yes, absolutely, if your geography allows, then put in a megalitre+ sized hole in the ground up-hill, cover it to reduce evaporation & algae, fence it to prevent animal damage, etc. Happy days.