Efficient food production

[james_davis_nicoll]

I was noodling around on soc.history.what-if and made a calculation I'd never bothered with before: if a human needs enough food to produce N Watts, how many square meters are required to intercept that much sunlight? OF course I was too lazy to actually look up insolation for various latitudes but the BOTEC I committed seemed to show that it should be a few square meters.

Even Fairbanks, Alaska, gets from 90 to 350 watts/m^2. Say your mark 1 human needs at least 100 watts worth of food to keep functioning [1]: They'd need about one square meter dedicated to collecting solar powers, asssuming no losses. The entire population of North America should require a few hundred to a thousand square kilometers of converters to power themselves. Even a factor of ten losses should mean that we'd need about 300 square kilometers to feed all of Canada, assuming the lowest insolation in Alaska is what we have to work with, and about 3000 square kilometers to feed all of the USA. That's a square less than 20 kilometers on an edge for Canada and a bit over 50 kilometers on an edge for the USA. Feeding the entire planet should require about 60,000 square kilometers or a square about 250 km on an edge (or less, if we pick someplace sunnier than Fairbanks to grow food).

Clearly modern methods of coverting solar (and fossil) energy into human energy are criminally inefficient.


1: Googling says "at least 2500 kilocalories" per day so call it 4000 to be safe. That works out to about 50 watts, which I will double just because.

Comments

[liveavatar.livejournal.com]

At first I misread this as "how many square meters [of human] are required to intercept that much sunlight," and yes, I know that's not what you really meant.

But until I figured that out I was having a grand time imagining a lot of people who went around fat and naked as a statement of low energy consumption (or as a statement that they could only afford their Guaranteed Solar Income).

Yrs in the name of stfnal life through incorrect assumptions,
Liveavatar

Not Fat

[james-nicoll.livejournal.com]

Fat = volume. You want to maximize area to volume, so you'd end up extended in two dimensions and reduced in another. Like a leaf, say, or those ribbon-like organisms in the ediacaran.

Re: Not Fat

[liveavatar.livejournal.com]

Hmn, I guess it depends on how efficient the energy absorption mechanism is, and on the individual's metabolism.

Heck, if we're engineering people enough that they can absorb their calories via sunlight in the first place, then we can engineer them with photosynthetic ribbon-hair or dreadlocks. Or photosynthetic tails. (See H. Allen Smith's The Age of the Tail.) Or webbing between the fingers that can take advantage of hydropower (read: swimming)

In the case of energy-dreads, baldness would become more than an aesthetic problem. But if we've gone that far, I doubt baldness would require more than an afternoon in a bioplastic surgeons' outpatient chair.

[sunshaker.livejournal.com]

Are you suggesting that we Gene-Mod humans so that they are green and can eat dirt to make food? How much energy would be produced that way? Would it be worth it?

[autopope.livejournal.com]

1. IIRC, an adult human metabolism dissipates on the order of 400 watts at rest -- the increase when working isn't that great. I'm not sure there isn't something missing here in the Kcal calculation ...

2. Again IIRC (ancient biology/biochem lessons surfacing here) photosynthesis is bloody inefficient -- less than 1% of the incident photoelectric energy landing on a chlorophyl molecule ends up being used for ADP->ATP synthesis. When you add the not-unreasonable requirement that plants spend the majority of their metabolic energy on processes not directly contributory to human nutrition (e.g. synthesis of indigestible lignified cell walls, required to stop them flopping around in the dirt), the available energy drops further.

So as long as we're using a plant-based biochemistry, I figure you need to add three orders of magnitude to your estimate. Advanced nanotech might shave two orders of magnitude off of this, but I'm not sure how.

Plants are the problem

[james-nicoll.livejournal.com]

We need some sort of industrial process, I think. Something a lot more efficient at making stuff to power humans than what we are using.

[(Anonymous)]

Oddly enough, this is relevant to a story idea I've been working on. Aliens muck up the sun so its visible-light output drops by a factor of one thousand. The energy goes into near-infrared, above the water absorption band so Earth doesn't get any colder. It doesn't look any darker - human eyes adjust to the dim sun just like they do to artificial light. The problem is that photosynthesis shuts down. The only way to keep people fed is to grow food under artificial lights. With current technology and assuming a strictly vegan diet, how much energy would this require?
Gareth Wilson

the environment is the problem

[twoeleven.livejournal.com]

i've seen crop yields increased by large factors in greenhouses. according to the plant breeder i pestered, plant growth is limited by (roughly in order): water, nitrogen, phosphorous, micronutrients, and suitable temperature. too much or too little sunlight is also a problem in some climates. predation by insects, competition by weeds, and infection by viruses and fungi also knock down yields.

under sterile, controlled conditions, well, the plants are much happier. i imagine if we were to breed or genengineer plants that didn't even bother to try to fight disease or worry about weeds, we might be able to eek out further yield increases.

That Kcal calculation

[james-nicoll.livejournal.com]

OK, this is what I did.

1 cal ~ 4 J (roughly) so 4000 kilocalories ~ 16,000,000 J
There are about 86,400 seconds in a day so that is about 185 watts--

Well, crap. That's _not_ the number I got. I bet I forgot to convert to Joules in step one.

Re: That Kcal calculation

[james-nicoll.livejournal.com]

I was going to look up the surface area of humans and use body temperature to calculate the watts we use. Turns out "what's the surface area of a human?" isn't as straight forward as I thought.

It occurs to me that knowing the energy that must be radiated and the temperature we do it at should let us calculate the area humans must be, except I bet with convection and conduction to keep track of, that's not so straight forward either.

Maybe if I put a person in an evacuated bell-jar and do measurements from outside....

[mishalak.livejournal.com]

So.... If some bright boy came up with a more efficient molecule at doing the ADP->ATP synthesis there is huge potential to have plants produce more grain, more board feet of lumber, more fruits, more everything. I see great potential here for a story.

[oldsma.livejournal.com]

Because it will go horribly wrong.

MAO

[mishalak.livejournal.com]

Well... yes. Imagine if you will soybeans that use a mere 4% of sunlight instead of 1%. Why that means they grow faster and such like, right? They might just out compete weeds if they get out into the wilds...

[james-nicoll.livejournal.com]

Which sounds fine, if you aren't one of those "mass extinctions are bad" extremists. The blackleaf takes over and less favoured races become extinct. It's no worse than kudzu in the South and probably won't have effects any more dramatic than the whole flower/insect thing 125 million years ago.

Of course, because the stuff is meant to be et by humans, a crapload of other animals can also eat blackleaf plants. At a guess, about four times as many as can be fed by greeleaf plants. Those animals aren't going to just sit around, either. They are going to use that energy to do stuff and not all of it is going to please humans.

Oh, and if blackleaf is doing stuff like releasing O2 at four times the rate normal plants do, there could be short term side effects from that as well.

Blackleaf

[mishalak.livejournal.com]

Well it won't grow everywhere. Being a crop plant it will only disrupt life in areas where it gets enough water and such like and it isn't too cold. So we're not looking at a new ice age (thank our lucky stars we didn't do this to phytoplankton or we would). But it will definitely be a weed control problem because it grows so darn fast. Also thank our lucky stars it wasn't mustard seed for biodiesel. Those darn seeds are small...

Re: Blackleaf

[james-nicoll.livejournal.com]

Ice age? Oh, from the reduced CO2?

I was thinking black leaves = lower albedo for the Earth = warmer in the immediate area of the plants.

Re: Blackleaf

[james-nicoll.livejournal.com]

This has consequences even if the [whatever] is containable and never leaves the food-factory. Cheaper food isn't a good thing for people at the production end of the food business, because demand is inelastic. It's great for consumers, though, and most people are consumers.

Now, if you happen to be in a large nation currently run by neocons, check out how much of their support comes from rural areas and imagine how those voters will react to a technical development that makes them as cutting edge as an 8-track.

Re: Blackleaf

[mishalak.livejournal.com]

Currently it is sustainable, though more expensive, to grow crops like mustard and soybean for oil to be converted into biodiesel. You get out roughly 200% of the energy you put into the farming and the conversion. But this is expensive relative to drilling for oil and refining it.

So if there were a doubling of yields on American farmland it would indeed mean a reduction in prices, but not all the way down. At current cost levels for oil prices would fall 17% and then stabilize at the level where it becomes practical to replace standard petroleum diesel with bio diesel. If world demand for energy increased it could actually increase crop prices depending upon the exact details of the transition with the usual caveats about this only being a very simple calculation and all.

Also after having done some more reading I've learned that the real losses in efficiently are due to oxygen. Higher amounts of the waste gas in the system mean that sometimes instead of doing the right reaction it does one termed photorespiration that produces junk. So if you found a way to more efficiently exclude and remove O2 from plant cells it would greatly increase yields. So the problem isn't that plants reflect green light or whatever, though that could bump up yields a bit.

[autopope.livejournal.com]

Unfortunately it's not one molecule -- it's this insane Heath-Robinson contraption with a total atomic weight measured in megaDaltons and about thirty different enzymes and peptides playing pass-the-parcel with electrons. There are two different metabolic pathways for photosynthesis in chlorophyl-based plants, the common C3 pathway and the more recently evolved, more efficient C4 pathway (used by Maize, among other plants), but the difference in relative efficiency is bollixed up by the grotesque inefficiency of the first step.

[mishalak.livejournal.com]

Okay, but for story purposes where might some massive research project in, say, China attack the problem to get the greatest increase in plant production? Where I'm going with this is to imagine a field growing advanced biotech crops. The first thing I thought of would be that they would be darker than present plants and possibly would need more water than unmodified counterparts.

[(Anonymous)]

Charlie, basal human metabolism is about 100 Watts at rest. Parts of photosynthesis are remarkably efficient -- the photosystems themselves push close to 100% -- but it takes about ten photons to fix one molecule of CO2. (James, here energy = h * frequency raises its ugly head.) Because of that, the theoretical max is about 25%. Turns out of cultivated plants, sugar cane is the winner, at 8% efficiency. Corn (maize) is about 2%, and random wild plant life on average is rather below 1%.

(And wash your mouth out with soap for saying 'advanced nanotech', Charlie. Plant life _is_ advanced nanotech. Well, except for the ribulose bisphosphate carboxylase glitch, but that probably couldn't be helped even with Drexlerian nanomagic.)

Carlos

[autopope.livejournal.com]

Yeah, well, I'm going from underused decades-old memory. Yes, enzyme chemistry is nanotech by most definitions -- but I suspect the prevalence of hydrogen bonds in determining tertiary and quaternary structure, and the requirement for water, tends to limit the efficiency of the reactions it can mediate by putting unpleasant constraints on things like temperature and pressure.

[twoeleven.livejournal.com]

i think that's backwards. we use heat and pressure cause our catalysts aren't as good as enzymes (generally). compare, say, a field of beans happily fixing nitrogen at ambient conditions to the haber process, which runs at 400 C and 200 atm.

otoh, the requirement to run in water does limit the choice of substrates somewhat.

[james-nicoll.livejournal.com]

Where did you get the various efficiencies for plants?

Call the human diet 400 watts, call the superplant 10% efficient at turning light into energy humans can use and set insolation at about 100 watts, then we need 40 m^2 per person. Any _obvious_ math errors this time?

That's about 25,000 people fed per km^2 or one million per 40 km^2. Canada needs 30x40 = 1,200 km^2, the US needs 300x40 = 12,000 km^2 and the world needs 6,000 x 40 = 240,000 km^2 to power everyone.

Definitely room to improve, there. Not that feeding people is an insurmountable problem right now but the footprint of agriculture is a pain.

[(Anonymous)]

Looks OK to me. This is the all-carb diet, though. But I think the 400 Watts can incorporate fats&proteins&vitamins as a fudge factor.

Googling for "photosynthetic efficiency" will give you various tables, but be careful with the definitions they use. The 8% figure for sugarcane can also be found in Lovelock in his discussion of C3/C4/CAM plants, if memory serves. I can dig up more from Hall & Rao later, if you'd like.

Charlie, I've never seen a Drexlerian design that didn't look like a high explosive about to go off. (Merkle's are a little better.) Poor guy just isn't a chemist, that's all.

Carlos

Hasn't Drexler been more or less marginalized?

[james-nicoll.livejournal.com]

That was my impression.

It's a shame the term "nanotech" has been irredeemably tainted by the more extreme proponents. I don't really want to use the term "biochemistry" because what if I'm talking about something designed with real biochem in mind but not using anything that is actually used on Earth?

Re: Hasn't Drexler been more or less marginalized?

[twoeleven.livejournal.com]

so call it "bioengineering". "protein engineering" and "metabolic engineering" are already accepted jargon in the field.

drexler is largely marginalized by serious chemists, but afaik, he still has a large following among nano-ninnies and random lay people.

Re: Hasn't Drexler been more or less marginalized?

[(Anonymous)]

Yeah, Drexler is pretty much out of the running. Personal problems and some stubbornness/willed ignorance. (When Nobel Prize-winning chemists suggest you have made a mistake in your proposed molecular design, you might want to take them at their word.) As a result, 'nanotechnology' in the real world means something very different from what Drexler proposed.

(There's a whole strange subclass of MIT-affiliated whoopsies like that. Norbert Wiener's wife claiming McCulloch seduced their daughter, setting back cybernetics a decade. Minsky and Papert stomping on perceptrons, setting back neural net research a decade. Chomsky and linguistics, setting back linguistics two, three decades. Various Media Lab things.)

For a neologism, might I suggest "artificial biochemistry"?

Carlos

[del-c.livejournal.com]

The 8% for sugarcane is when they're at the peak of their abilities; you'll need to average over the entire lifecycle to work out what humans can do with a plot of land, and that takes you back down to a couple of percent again.

Re: Hall and Rao's Photosynthesis, p.67 of the fifth edition shows efficiencies of up to 12% in certain wavelengths in Chlorella, but the operative word is "in certain wavelengths". Hall and Rao are all about the photosystems and the chloroplasts, but that's not a measure of the final productivity of even an ideal farm.

[(Anonymous)]

Del, Chlorella was used because it was easiest to manipulate in a spectrometer in 1960, and that 12% is an illustration of the Emerson effect at far-red wavelengths. The pages of Hall and Rao you want are page 108:

Lastly, we can discuss the quantum efficiency of CO2 fixation. Each mole quantum of red light at 680 nm contains 17.61 * 10^4 J of energy. Thus, at least three (48 * 10^4 / 17.6 * 10^4 = 2.7) mole quanta of 680 nm light will be required for one CO2 molecule to be fixed. However, experimentall, it is found that 8-10 quanta of absorbed light are required for each molecule of CO2 fixed or O2 evolved. From our knowledge of non-cyclic photosynthetic phosphorylation we deduce that there are two different light reactions required to reduce NADP with the electrons from H2O:

2NADP + 2H2O (4e- & 2 light reactions in chloroplasts) -> 2NADPH2 + O2

Thus we need at least 8 quanta (4 quanta per 4e (1 O2 molecule) * 2 light reactions) to reduce NADP and produce the necessary ATP at the time.

Nevertheless, photosynthetic CO2 fixation itself is only about 30% efficient (2.7 quanta / 8-10 quanta) as we can measure it. Taken in conjunction with an average efficiency of less than 1% for whole plants capturing and utilizing photosynthetically active sunlight (see Chapter 1), this reinforces the concept that these energy exchanges are necessary but wasteful and could be improved in artificial photosynthetic systems.

and page 4:

Energy losses:

47% loss due to solar photons outside the photosynthetically active region (400-700 nm) [the remainder is in the lower energy IR]

30% loss due to incomplete absorption or absorption by components other than the chloroplast

24% loss due to degradation of absorbed photons to excitation energy at 700 nm

68% loss due to conversion of excitation energy at 700 nm to chemical energy of D-glucose

35-45% loss due to dark and photorespiration

These are cumulative, multiplicative losses. About half is simply because the photons are not energetic enough to make the reaction go (the energy = h * frequency thing I alluded to before). After that, the coupling from the reaction center to carbon fixation. The confusion RuBisCO makes between CO2 and O2 is due to the similar charge and size of the two molecules, and is largely insuperable.

8% efficiency, incidentally, is the maximum rate for sugarcane under cultivation. For its full life-cycle, it's more like 4%.

Carlos

[penguinicity.livejournal.com]

400 watts is a tad shy of 350 kcal/hr, which, for a typical human, is brisk walk territory. The last time I had my RMR measured, my 90kg body was cranking through 1900 kcal/day -- just over 90 watts. Subtract some, since the average human still weighs in at less than 90, then add some since at least some of these humans are going to be walking around tending to the sunlight->food devices. Say 2000-2500 kcal/day, depending on how often the machines break and to what degree our subjects are couch potatoes.

The other interesting factor to estimate is the amount of energy required for temperature control and/or transportation. Either the population is dispersed amongst the fields, in which case some of them are going to need to heat/cool their dwellings, or else you stack everyone in northern California, and ship in food.

[del-c.livejournal.com]

Humans dissipate between 100W and 200W.

Plants are about 1-3% efficient over their whole life cycle (saying that parts of the process are close to 100% efficient is sort of true, but irrelevant, as the 100% efficient parts are about 1-3% of the total, in ways that can't be increased)

The equator gets about 420W/m2 averaged over the year. You'd think the poles would get zero, but since they get zero in winter anyway, and >zero in summer, the average for our 23° tilt planet comes to a surprisingly high 160W m-2 or so, according to a quick average of one of my old spreadsheets. That seems wrong somehow, but I don't see where I might have screwed up.

latitude
0 	420
10 	414
20 	396
30 	368
40 	330
50 	283
60 	233
70 	186
80 	161
90 	158

[del-c.livejournal.com]

I did make a mistake, but in the other direction: by averaging over thirteen months instead of twelve, I cheated the poles of 10 watts per m². The maths is horrifyingly hairy everywhere else^*, but it's relatively easy to show that at the pole itself, the value is

1,370W m^-2 ÷ pi × sin(23°) = 170W m^-2

*_{I don't know enough about elliptical functions to average over a year except by brute force, or even to know if it can be done analytically at all}