John Jeavons wrote a book called 'How to grow more food than you thought possible on less land than you could imagine'. He included estimates on yield per square foot for a wide range of crops using his french intensive biodynamic approach (compost, double dig, clean cultivation, frequent irrigation etc..). His system assumes unlimited nutrient, carbon, and water inputs. He figured 1/4 acre for vegetarian diet with grains and legumes as the core and yearround vege production in central california, (USA zone 9?). The vege's don't take up much space. You could get you vitamin needs off of weeds if you had to, but 1 year of protien and complex carbos (grain and dry beans, maybe potatos) is the hurdle.
So... in short... the footprint is inversely proportional to the amount of energy/nutrient/water subsidy you can gather from your neighborhood. Contemporary permaculture designs on small sites typically takes advantage of your neighbors waste (I currently import arborist wood chips, spoiled straw, and horse manure, and have mooched other peoples vegetable scraps and coffee grounds as well.)
But if everyone is scavenging nutrients and organic matter, then the game changes. I recently read that your urine contains enough nutrients to produce a grain crop to feed one human for a year.
Then there is energy, fiber, and construction materials...
Permculture designs should theoretically increase yield by diversifying yield... but I have not seen any rigorous accounting like that done by Jeavons.
A final thought... a household is not necessarily the appropriate scale for sustainable system design, or the most effective unit for self sufficiency.
Then there is the hidden subsidy of ecosystem services... whoops... running out of brain power...
Paul, Puget Sound, USA
Olympia, WA, USA