Perhaps, if/when the oil crash causes the collapse of transport-based economies, such farms beneath farmland-deprived cities may help keep them somewhat viable.
Nuclear, hydroelectric, wind, or solar.
Sorry, I wasn't very clear - I meant in terms of efficiency that the energy cost of running all those lightbulbs would be far greater than transporting the produce.
Would it really be sensible to have solar cells produce electricity @ 25% efficiency, use the 50% efficiency electricity grid to transport it into the cities and then use it to power ~50% efficiency lightbulbs, all to reproduce the light available 'free' on the surface? At 45 degrees lattitude, the average solar radiation is ~200W/m^2 over the growing season, which tots up to an awful lot of energy - nearly 5KwHrs per day per m^2, for many months.
They're not lightbulbs, they're LEDs, which are a lot more efficient.
Of course, the other option would be to bring produce in from farms by electric trains or lorries using fuel cells/flywheel-stored electricity. Though, AFAIK, electric air transport would be more difficult, so the days of Kenyan beans and New Zealand lamb ending up in European supermarkets could be limited.
I'm ever eager for a return to commercial sailing boats - no more horrendous house-brick shaped cargo boats booming along. I've no idea what a 21st Century designed sailing sloop would look like - but its bound to be sleek by hydro + aero dynamic demands alone...
BTW; LEDs are not that wonderfully efficient - large halogens or fluorescents with a central ballast easily beat them. 'White' LEDs are worse still, as they rely on a Blue-LED chip illuminating flourescent coatings - exactly the same process as UV Mercury Discharge tube --> White Light. However, they're practically unbreakable (no filament to snap, no glass case to crack) + last for a ridiculous length of time - hence their phasing in for traffic lights + indicator bulbs on the roads...
If the oil crashes, what are you going to power the lights with? And please don't say Solar :^)