World Food Production

So I was reading a post about Malthusian Crises on a web site and I went to recreate some numbers cited by a debunker in the comments. I used 1975 and 2010 because they're round numbers and 1975 is comfortably benefiting the Green Revolution. It turns out 2010 was disproportionally productive so I used the numbers from http://faostat3.fao.org/home/index.html#DOWNLOAD to make a graph of the results.

The numbers here are unitless because the actual units would be meaningless and we only care about the comparisons. If the world wasn't starving to death in 1965 then it won't be starving to death when people have comparable amounts of food at other times.

(I also may be on a dangerous graphing binge.)

The top graph is the amount of cereal grains produced per person. The bottom graph is the amount of cereal grains per person per unit area.

The trendlines are logarithmic and fit much better than linear trends. We seem to be coming out of the tremendous gains of the Green Revolution. Nonetheless we can stand to make dizzyingly less food without scumming to a mass population destroying famine.

Interesting thing that falls out of these numbers. If we hold our production of grain exactly steady until the world population reaches 10 billion we'll produce only slightly less grain per capita than we did in 1961.

Virtually every apocalyptic prediction I heard about resources when growing up (case in point: oil) has turned out to be nonsense- for the timelines given, anyway. I don't know if it is a an alarmist tendency or a general underrating of human ability to adapt, but in all cases we seem to be able to go on far longer than thought, and resources just won't be an issue n my lifetime.

I was watching an old episode of Doomwatch the other day, an early 70s British semi-realistic science drama about a team that looked into potential ways in which science could harm mankind. The series was heavily based upon hard scientific research. One episode dealt with over-population and had the line "By the year 2000, there will be 80 million people living in the United Kingdom, and that's just cold, solid fact". Well, now it is 2013 and we're nowhere close, holding at about 65 million right now. But people really did believe things like that.

Variants on Mycoprotein could feed the world now.... Easily, cheaply and safely. No one need be hungry.

-Mitch

Originally posted by Symmetric Chaos
The numbers here are unitless because the actual units would be meaningless and we only care about the comparisons. If the world wasn't starving to death in 1965 then it won't be starving to death when people have comparable amounts of food at other times.

sort of... It would be nice to know what level the average is at compared to daily food needs.

This also only covers grains as opposed to all nutritional needs?

Originally posted by Symmetric Chaos
The top graph is the amount of cereal grains produced per person. The bottom graph is the amount of cereal grains per person per unit area.

the most astonishing thing about these graphs:

so, it depends on how you want to interpret it, but the shift in value on the Y axis is amazing. So, average for all people is considerably lower, at all points, than is average by-area. There are two ways of interpreting this:

a) When looking at specific areas, there are a few outliers where extremely high levels of food output per person that are forcing the average up.

b) When looking at all people, there are so many people that are at the lower end of food production that it is pulling the average down.

It actually expresses the discrepancy in food distribution in one of the most magnificently simple ways ever.

Originally posted by Symmetric Chaos
The trendlines are logarithmic and fit much better than linear trends. We seem to be coming out of the tremendous gains of the Green Revolution.

If all you are going for is r-squared fit, you could add a third degree polynomial to the regression over the food/person graph and increase its fit. Neither seem to show particularly curved growth over the Green Revolution, with the food/person graph showing a regression nearly back to 60's-70's level production in the earliest years of the new millennium.

I'm not so sure that a line with a single is the best way to describe the food/person data, or at least, sums over an interesting change in production in the early 2000s that isn't explained by the Green Revolution. Additionally, if data could be pulled back further than 1960 in the by-area graph, whatever curve is seen in that data may wash out, as the by-area graph seems considerably more linear than the food/person one.

Originally posted by Symmetric Chaos
Nonetheless we can stand to make dizzyingly less food without scumming to a mass population destroying famine.

without units on the Y axis and knowledge of the average necessary human consumption, you aren't actually demonstrating this with the graphs.

Originally posted by soopercavell
Variants on Mycoprotein could feed the world now.... Easily, cheaply and safely. No one need be hungry.

the prairies in West/Central North America grow enough grain to feed the world over.

the vast majority of this grain is used to feed cattle, in bio-fuels, or for other reasons that are not specifically about increasing human caloric intake. Further, about 30-50% of food that does make it to market in North America is thrown away.

The issue has never been needing new or wonderful technology or proteins.

Originally posted by Oliver North
the prairies in West/Central North America grow enough grain to feed the world over.

the vast majority of this grain is used to feed cattle, in bio-fuels, or for other reasons that are not specifically about increasing human caloric intake. Further, about 30-50% of food that does make it to market in North America is thrown away.

The issue has never been needing new or wonderful technology or proteins.

The truth is simpler than that I agree, it's easier for the rich to have a poor. Mycoprotein would in it's manufacture reduce CO2 and CH4 footprints that increased traditional protein production methods would not and could be done on site almost anywhere, reducing transportation and associated Carbon footprints, NOX etc... Or so some smart guy once told me if I remember correctly, which I probably don't, not being very smart and all.

-Mitch

hey, Sym, where did you get these numbers from? there are some correlations I'd love to run, see if that drop in food/person is significantly different from the by-area numbers. Seemingly, for there to be a drop in food/person and not in the by-area numbers, the losses associated with per person production would have had to have been compensated for by increases in other areas, meaning that the poorest or largest groups of people lost food production in ways associated with minorities making huge gains...

Originally posted by soopercavell
The truth is simpler than that I agree, it's easier for the rich to have a poor. Mycoprotein would in it's manufacture reduce CO2 and CH4 footprints that increased traditional protein production methods would not and could be done on site almost anywhere, reducing transportation and associated Carbon footprints, NOX etc... Or so some smart guy once told me if I remember correctly, which I probably don't, not being very smart and all.

-Mitch

sure, and farmers around the world are the least likely to introduce new technologies, technologies that reduce the need of farming destroy local economies in the third world, and most third world farming is designed to serve first world food needs anyways, so it is only the rich who would benefit from such technologies

see: the green revolution

like I said, the answer has nothing to do with technology

Originally posted by Oliver North
hey, Sym, where did you get these numbers from?

http://faostat3.fao.org/home/index.html#DOWNLOAD

It's in his initial (is that the right word for starting?) post.

-Mitch

Originally posted by Oliver North
sure, and farmers around the world are the least likely to introduce new technologies, technologies that reduce the need of farming destroy local economies in the third world, and most third world farming is designed to serve first world food needs anyways, so it is only the rich who would benefit from such technologies

see: the green revolution

like I said, the answer has nothing to do with technology

Some bright guy I once knew once told me, that something called Monsanto (whatever that is) do bad stuff and keep people poor....

Do you think they really might let people starve?

I even heard countries have Food Mountains and stuff, although I have never seen one on a map.

-Mitch

Originally posted by Oliver North
This also only covers grains as opposed to all nutritional needs?

Yes, the FAO allows you to check other things as well. I picked cereal grains because they're an example that gets used a lot, probably because they're a staple crop. Obviously the graphs are vulnerable to variations in other areas. It could be that food production in other areas shows significantly different trends. I'm not sure how I would combine that information.

Originally posted by Oliver North
so, it depends on how you want to interpret it, but the shift in value on the Y axis is amazing. So, average for all people is considerably lower, at all points, than is average by-area. There are two ways of interpreting this:

a) When looking at specific areas, there are a few outliers where extremely high levels of food output per person that are forcing the average up.

b) When looking at all people, there are so many people that are at the lower end of food production that it is pulling the average down.

It actually expresses the discrepancy in food distribution in one of the most magnificently simple ways ever.

It would be interesting to look at this as a map or plotted against the proportion of starving people in the world.

Originally posted by Oliver North
If all you are going for is r-squared fit, you could add a third degree polynomial to the regression over the food/person graph and increase its fit. Neither seem to show particularly curved growth over the Green Revolution, with the food/person graph showing a regression nearly back to 60's-70's level production in the earliest years of the new millennium.

I'm not so sure that a line with a single is the best way to describe the food/person data, or at least, sums over an interesting change in production in the early 2000s that isn't explained by the Green Revolution. Additionally, if data could be pulled back further than 1960 in the by-area graph, whatever curve is seen in that data may wash out, as the by-area graph seems considerably more linear than the food/person one.

I was just going through the fits that Excel gives me. More complex trendlines do fit better but logarithmic is easy to read. I don't actually know much about the subtitles of regression analysis. You could argue that the trend was logarithmic from 1961 through 1975 or so and has been linear ever since but I think there's too much noise to make that kind of determination.

Originally posted by Oliver North
without units on the Y axis and knowledge of the average necessary human consumption, you aren't actually demonstrating this with the graphs.

I'd argue that if in 1961 2.85 food units (its 100 kg, other numbers needed scientific notation, it was easier to read this way) per person was enough to live on then 2.85 food units per person will be enough to live on in the future. The point is the avoid having to estimate the amount of food a person needs. This way we can rely on what we know about the past. 1961 didn't have planet destroying famines so if we drop back to that level we wouldn't have planet destroying famines.

Originally posted by Oliver North
hey, Sym, where did you get these numbers from? there are some correlations I'd love to run, see if that drop in food/person is significantly different from the by-area numbers.

I linked to the FAO site you'll have to navigate it yourself but its pretty intuitive.

Population numbers for each year came from some site advocating zero population growth. They might be slightly inflated.

Originally posted by Symmetric Chaos
(its 100 kg, other numbers needed scientific notation, it was easier to read this way)

This is incorrect. The numbers in the original post are "tons per person times ten" and "tons per person per hectare times ten billion" both are yearly numbers, I assume. If you convert them back to the original values they smooth about a bit and have new numbers that are about as meaningless to the reader (attached).

I tried applying this so Africa specifically and got very low numbers. I suspect this is because it only measure production not consumption. Imports and exports become important when you look at the world regionally.

I think the scale on the Y axis gets shifted on the second graphs relative to the first, which makes it seem smoother

Originally posted by Oliver North
the prairies in West/Central North America grow enough grain to feed the world over.

the vast majority of this grain is used to feed cattle, in bio-fuels, or for other reasons that are not specifically about increasing human caloric intake. Further, about 30-50% of food that does make it to market in North America is thrown away.

The issue has never been needing new or wonderful technology or proteins.

Indeed, and not just because Americans have a tendency to buy more than they need and food ends up spoiling before it can be consumed. Edible food from fresh grains to veggies to live stock is purposely destroyed to keep market prices at a given.

I find that revolting.

so, just some weird things that are popping up in what I'm doing.

So, the sort of target calorie intake for a healthy human (ignoring sources of nutrients) is roughly 2000 per day. Looking at cereals alone, the lowest weight available, per day, per person, was just shy of 780 grams, and was in 1961 (the rate quickly increased to ~1000 and has been consistent).

Assuming a similar weight to calorie conversion between all cereals and wheat (the only pasta I had on hand), each gram of cereal produces roughly 3.53 calories, meaning that at worst, the daily production of just cereals could provide the caloric needs of all humans on the planet, and has been able to for at least 50 years. The step forward from this, if I were interested to (and I might actually) would be to break down what % of cereals are maize, wheat, etc, figure out a more accurate conversion and come up with a more solid number, but this is a good first pass, and doesn't include meats, fruits, vegetables, roots/nuts/etc.

Basically, what everyone has been saying thus far is not only true, it is incredibly true: we produce way more food than we should need to, especially given how rampant hunger is, even in the first world.

****, I mean, the numbers I ran are for foods that are non-perishable. Why don't we just export pasta and rice everywhere?

Also, the dip in cereals production in the early 2000s is also a period where there is a massive increase in fruit and vegetable production, and may reflect diversification of crops as markets become more prosperous. Additionally, while cereals may have been good because it is a staple, this type of analysis is so sensitive to the variables used. If you went with vegetables, it would look like we have 10000s of times more than necessary, if you went with roots & potatoes, it would look like we face imminent starvation. In the end, I'd love a graph that plotted available calories vs population need, but the effort may not be worth the measure; from the data we have already we know it will show we have too much for there to morally be so many hungry people.

Originally posted by Oliver North
so, just some weird things that are popping up in what I'm doing.

So, the sort of target calorie intake for a healthy human (ignoring sources of nutrients) is roughly 2000 per day. Looking at cereals alone, the lowest weight available, per day, per person, was just shy of 780 grams, and was in 1961 (the rate quickly increased to ~1000 and has been consistent).

Assuming a similar weight to calorie conversion between all cereals and wheat (the only pasta I had on hand), each gram of cereal produces roughly 3.53 calories, meaning that at worst, the daily production of [b]just cereals could provide the caloric needs of all humans on the planet, and has been able to for at least 50 years. The step forward from this, if I were interested to (and I might actually) would be to break down what % of cereals are maize, wheat, etc, figure out a more accurate conversion and come up with a more solid number, but this is a good first pass, and doesn't include meats, fruits, vegetables, roots/nuts/etc.

Basically, what everyone has been saying thus far is not only true, it is incredibly true: we produce way more food than we should need to, especially given how rampant hunger is, even in the first world.

****, I mean, the numbers I ran are for foods that are non-perishable. Why don't we just export pasta and rice everywhere?

Also, the dip in cereals production in the early 2000s is also a period where there is a massive increase in fruit and vegetable production, and may reflect diversification of crops as markets become more prosperous. Additionally, while cereals may have been good because it is a staple, this type of analysis is so sensitive to the variables used. If you went with vegetables, it would look like we have 10000s of times more than necessary, if you went with roots & potatoes, it would look like we face imminent starvation. In the end, I'd love a graph that plotted available calories vs population need, but the effort may not be worth the measure; from the data we have already we know it will show we have too much for there to morally be so many hungry people. [/B]

You are sooo smart are yous like a scientist or sumfin?

-Mitch