[Schroedingersat]

You're double counting that 20% the 100W is thermal output of a resting human, not work done. As evidenced by not needing 5000 Calories a day just to survive

It's still about 50-150kW average of sunlight for the absolute minimum to sustain a person with a mostly closed ecosystem (this will also feed numerous insects/bacteria/fungi/etc). A very optimized hydro setup might do it with 2~kW based on C4 photosynthesis being about 4%.

But yes, excluding sunlight of a reasonable amount of agricultural and urban land seems sensible.

[netjiro]

> ... double counting ...

Thanks for catching that. Sedentary lifestyle is around 2000kcal/day, ca 100W.

I assumed the ca 20% efficiency was for efficiency of chemical energy extraction from the food (on average, varies a lot between carbohydrate, protein, fats) to storage as glucose/fat before we run it to ... > AMP > ... > ATP & NADH.

But that should probably already be taken into account by the output calculations from the agricultural production. Right? (so my original comment was wrong).

Embarrassing that I don't know this off the top of my head. I used to be good at it. Need to dust off the books. Or probably buy new ones since the field has progressed since last I dug around in it.

[Schroedingersat]

I was thinking more in terms of inputs to known systems.

There are small scale farming methods that feed one person on about 400m^2 with some experience and almost no inputs other than water, or up to 6 people on a quarter acre/1000m^2 block with minimal inputs.

160m^2 in a very good climate is very roughly 50kW average including day/night and some weather.

So a hundred or so kW seems like a pretty reasonable total budget to aim for, with 10% used by things that aren't plants (which would come to ~2kW of work at current efficiencies).

Re. the idea that I'd believe 2kW was possible (if beyond current methods):

These methods generally dedicate around 30% of the land to energy limited crops like potato, and of that land maybe half is covered by a leaf.

You can beat sunlight->photosysnthesis in efficiency by only producing photons that plants use well (currently the best is blue+phosphor white LEDs), and I'm willing to believe without evidence that precisely scheduling and dosing your light could double output again. As could switching to some king of genetically modified corn or sorghum from potatoes. Recycling nutrients on a budget of a few hundred watts plus whatever is in discarded leaf matter seems somewhat scifi but not impossible.