Global food production and consumption trends—an energy-based approach

I have spent last month’s posts laying the ground-work for what I plan to be a series of posts examining food production, consumption and import-export trends.  This post kicks-off the series by considering world-wide food production and consumption trends from 1961 to 2007.

I am interesting in seeing whether or not there are signs of peaking food production and of countries shifting from being net food exporters to net food importers (or vice versa).  It stands to reason that if fossil fuels, and petroleum in particular, are important to the modern food production system, and if the world or various regions are at or near peak fossil fuel or petroleum production, then the world, regions or individual countries, may also be at or near peak food production.

I say may, because there does not appear to be a one-to-one relationship between petroleum production and population growth (see Part 10: Peak oil exports, peak oil and implications for population change) at least in part because relatively small amounts of petroleum are needed to maintain the food production system (see The relationship between hunger and petroleum consumption-Part 1 and subsequent posts). As an additional complication, there are other factors besides petroleum that may limit food production (see Export land model analysis of food production and consumption—Background).

Synopsis of my energy-based approach to food production and consumption

My approach is easy to describe and hard to implement.  It is not particularly useful to compare or add up different weights of multiple different food types, because different foods can have vastly different energy contents, as measured as units of energy per unit of weight.  Therefore, to make meaningful comparisons between the food production capabilities of different regions or different periods of time within the same region, I derived a set of conversion factors for the ~100 food items that are listed in the Food Balance Sheets published by the Food and Agriculture Organization (FAO) of the United Nations.  I used these conversion factors to convert each of the FAO-listed food items into a total energy amount (in units of Peta Joules, PJ) and then added up the total food energy products per year for the region of interest.  As you will see below, I also look at a number of subcategories of food items.  Plots of total food energy versus time, in years is, in essence, looking at the aggregated annual food production rate for the region of interest.  The interested reader can read my previous posts here, here and here if you want more details of the procedure that I have implemented.

This approach is analogous to looking at total fossil fuel production rates for different countries or regions, such as published by the EIA or BP, by converting barrels of petroleum, cubic feet of gas and tons of coal into common energy units.  It’s just that I have to deal with converting about 100 different food items instead of only three different types of fossil fuels.

World-wide food production trends from 1961 to 2007

You may ask, why only data up to 2007?  A lot has happened since 2007!  Unfortunately, the FAO is still partying like its 2007.  For instance, according to the FAOSTAT Release Calendar Food Balance Sheets up to 2009 will be published in June 30, 2012. 

Why the 3-year hold-up?  The Handbook provides some insights:

The accuracy of food balance sheets, which are in essence derived statistics, is, of course, dependent on the reliability of the underlying basic statistics of population, supply and utilization of foods and of their nutritive value. These vary a great deal both in terms of coverage as well as in accuracy. In fact, there are many gaps, particularly in the statistics of utilization for non-food purposes, such as feed, seed and manufacture, as well as in those of farm, commercial and even government stocks. To overcome the former difficulty, estimates can be prepared while the effect of the absence of statistics on stocks is considered to be reduced by preparing the food balance sheets as an average for a three-year period.
from 5. ACCURACY OF FOOD BALANCE SHEETS, page 10

The FAO appears to be holding back yearly data so that they can calculate a 3-year moving average of the food items reported in their food balance sheets. 

This doesn’t make a lot of sense to me, since there are obviously other ways to report data as a moving average but with the average updated as every new year of data is collected and reported.  Or, they could just release the raw data and let the end-use deal with it.  By choosing to delay the release of data by 3 years, the FAO pretty well eliminates itself as being a timely source of information.

Okay, that's enough ranting, with these limitation is mind, let’s look at the time series of data that is publically available.

Figure 1 shows the time course of the change in total world food energy production (referred to as “Domestic Supply” in the vernacular of the FAOSTAT Food Balance Sheets).

 The world-wide total annual food energy production rate (blue circles), expressed in energy units of Peta Joules per year (PJ/yr), has increased nearly 3 times, from about 19000 PJ/yr in 1961, to about 56000 PJ/yr in 2007. 

Figure 1 also shows the break-down of food energy production into two major components: plant-derived food energy production and animal-derived food energy production.  The bulk of food energy, about 90%, is derived from plants.  The remaining 10% of food energy production is derived from animal food energy production.

At least at first glance, the rate of food energy production appears to be increasing linearly at about 2.4%/yr over this 46-year time scale; indeed the linear regression coefficient, r² equals  0.996.  Plant food energy production has also been increasing over this 46-year period at an overall growth rate of 2.4 % per year, which is about the same as the total food energy production growth rate.  Similarly, food energy production is derived from animals has been increasing at about the same growth rate (2.2% per year). 

For comparison, I plotted the world’s population, as reported by the FAO, over this same time period (black Xs).  Since 1961, the population has increased about 2.2 times, while total food energy production has increased 3 times.

Figure 1 also shows that the food energy production rate has increased faster than the population.  That is, every year there was more food energy being produced per person.  For instance, from 1961 to 2007 the world’s population increased on average 1.7% ± 0.3% per year.  Over the same time period, food energy production increased on average 2.4% ± 1.1% per year.  

A more detailed analysis, however, shows some interesting time dependent-trends in the changes in both the population and food production data—the growth rate of both of these is decelerating.

Figure 2 show 5-year averages of the year-to-year percent changes in population and food energy production. 

The 5-year average year-to-year change in population (black cross hatched bars) has been declining from 2% per year in 1962-66 to 1.2% per year in 2002-07.   That is, the population is still growing but it is growing at a slower rate in 2002-2007 than in 1962-66.

Likewise, the total food energy production rate (blue bars) was growing at 3.8% per year in 1962-66 but in 2002-07 it was down to 2.4% per year.  The 2002-07 total food production growth rate is actually up from the previous 15 years (i.e., 1.6%/yr, 1.9%/yr, 1.8%/yr in 1987-91, 1992-96, 2002-07, respectively).  Once again,  the trend is that the food production rate is growing, but growing at an increasinly slower rate with time. 

Generally, the growth in the food production rate is always greater than the population growth.  The only 5-year period where food production growth was less than population growth was in 1987-91.  That period from 1987-91 is not likely a situation that could last too long—if the population were growing at a faster rate than food production growth, eventually a short fall in food would occur and population growth would have to fall back down to or below the same level as food production growth.   Of course, the short-fall could be temporarily averted if there was less food waste or if food used for purposes other than feeding humans were cut back.  As I will show you later, that doesn't appear to be the trend, however.

Figure 2 also shows the analogous 5-year average food production rates for plant derived food (red bars) and animal-derived food (green bars).  Because total food production is ninety-percent plant-derived, the percentage change in plant food production closely matches the change in total food production.  The trend in animal food production growth is also down from 3%/yr in 1962-66 to 2%/yr in 2002-07.  The 2002-07 growth rate in animal-derived food production is up from a minimum of 1.4%/yr in 1992-96.

Figure 3 presents the same data as in Figure 2 but this time as a scatter-plot of the 5-year averages of the year-to-year percent changes in population and food energy production as function of the mid-year of each 5-year averaging period.  The solid lines are the linear regression best fits, which I have extrapolated out to 0% growth. 

The deceleration in population growth is clearer than the deceleration in food production growth, but as discussed above, these are both on downward trends.  If population growth stays on the linear trend predicted by the linear regression best fit then it will reach zero growth in 2068.  If food production growth stays on the linear trend predicted by the linear regression best fit then it will reach zero growth in 2053.

However, because the slope of the declining food production growth rate is steeper than the declining population growth rate, an interesting thing happens in about 2032—the linear prediction lines cross.  That is, by 2032 both population growth and food production growth rates will both equal about 0.75 %/yr.  I predict that soon after 2032, the human population growth rate decline would go down to match the declining food production growth rate.  If this is the case, then the human population growth rate would reach zero about the same time as the food production growth rate reaches zero—about 2053.

Major Components of Food Energy Production

Figures 4-6 show the break-down of the major food items that contribute to the plant-derived and animal-derived food energy production rates.

Figure 4 illustrates that as a percentage of the total food energy production, plant-derived food energy has every-so-slightly increased from about 89% in 1961 to 90% in 2007. 

The six subcategories of plants (cereals, sugars, starches, oil crops, vegetable oils and pulses & tree nuts are based on FAO’s categories although I have combined some these into super categories (e.g., sugars are the sum of sugar crops, sugars & sweeteners; pulses and tree nuts are combined).  Together they account for about 84-85% of the total food energy production throughout the 1961-2007 reporting period.

Figure 4 also illustrates that it is the cereals that contribute half of the total food energy production.  The FAO defines cereals as the aggregate of nine different food items (as discussed previously, each item, in turn, can be the aggregate of several different types of food): Wheat, Rice, Barley, Maize, Rye, Oats Millet, Sorghum and Other Cereals.   Cereals accounted for more than 50% of the total food energy production until 2002, when it dropped below 50%—in 2007, it accounted for 47%.

The five other major food items presented in Figure 4 each account for between 2 and 10 percent of the total food energy production.  The sugars have provides a fairly steady 10% of the total food energy production.  Starchy roots (potatoes, sweet potatoes, cassava, yams) have steadily declined from contributing about 7.8% to 4.5% of the total food energy production.  Pulses (e.g., beans, peas) and tree nuts have also declined from about 3.2% to 1.8%.    

So what has gone up?  The oil crops and vegetable oils. 

Oil crops (soyabeans a 10x increase, but also ground nut and coconuts about 2-3x) have increased from 2.4% to 8.0% of total food energy.  Vegetable oils, which includes, and separately accounts for the oil extracted from the oil crops, is likewise increased from 3.6% to 8.8% of total food energy. 

Not shown in the Figure are fruits (2%), vegetable (1%), alcoholic beverages (1.5), stimulants (coffee, tea, etc...; 0.1%) that round out the remaining contributors of plant-derived food energy.

Figure 5 focuses on just four plant food items that make major contributions to the food energy supply, three of the top cereal products: maize (corn), wheat and rice, and, the top oil crop / vegetable oil: soyabean and oil from soyabean.  Keep in mind that the production all of these food items are increasing with time and we are focusing here  on the increasing relative contribution of these food items to total food energy production.

In the mid-1990s maize overtook wheat as the top food energy contributor: in 2007 maize accounted for 18% of total world food energy production—“King Corn” indeed. 

The relative contributions of wheat and rice have gone down slightly by 1-2%.  Soyabean and soybean oil has risen from obscurity from 2.4% in 1961 to 8.0% of total food energy in 2007. 

Since the mid 1980s, the aggregate sum of these four food items: maize, wheat, rice and soyabeans/soyabean oil, account for 50% of total food energy production.

Figure 6 shows that the total animal-derived food energy has slightly declined from 11% to 10% of the total food energy from 1961 to 2007.

Major animal-derived food items include milk (down from 4.8% to 2.8%),  meat (up from 2.9% to 3.7%), animal fats (down from 2.5% to 1.7%) and sea food (up from 0.7% to 0.9%).   Together these four food types account for nearly all of the 10% from animal-derived food energy.

Different Major Consumptive Uses of the Food Energy Supply

Figure 7 shows the relative proportions of major uses of food energy production.  Not surprisingly, the top food energy use is for human use, which has slightly declined from 55% to 50% of the total food energy produced from 1961 to 2007.

From 1961 to 2007, the proportion of plant-derived human food energy has gone down slightly, from 84% to 83%, with animal-derived human food energy increasing from 16% to 17%.

The proportion of food energy production devoted to seed and feed (e.g., livestock feed) has also declined over the reporting period.  The relative amount of food energy used in the form of seed has declined from 5% to 2.2%.  After peaking at 26% in 1972, the proportion of food energy devoted to feed has declined to 19% in 2007. 

What has increased is the "other uses" category which was stable at 5-6% until 1988, and then steadily increased to 10% by 2007.   

As discussed in the second post in this series, the Handbook defines "other utilization," as food consumed by tourists and for non-food purposes.  But, since we are looking at world-wide numbers here, there are no "tourists," and therefore this number should represent non-food uses of food production. 

The Handbook cites oils and soaps as examples of non-food uses, but I wonder if energy use, i.e., bio fuels, could be another growing non-food use. 

Unfortunately FAOSTAT does currently not collect any data on bio fuels (see FAOSTAT FAQs). 

There was a separate listing in the FAOs Food Balance Sheets under the category of "alcohol non-food," and this category has dramatically increased in production from 4.3 PJ/yr in 1961 to 95.6 PJ/yr—a 22x increase.  But, as a proportion of total food energy production, this is still only 0.17% in 2007. 

Food Imports and Exports

Figure 8 shows the relative proportions of global food energy production involved in imports and exports.

Discussing food imports and exports in the context of world-wide statistics might not seem to make much sense, but I think there is an interesting point to be made regarding the relative amounts of food energy involved in cross-boarder import/export trade. 

Of course, the net global movement (exports minus import) is substantially equal to zero (red triangles).  But, since about 1970, the proportion of food energy has been steadily increasing from about 10% to 20% in 2007.   That is, the world's produced food is increasingly involved in import/export activity. 

Summary and Conclusions

By expressing food production and consumption in terms of the food's energy content, I was able to aggregate many different types of food items to provide a broad overview of global total food energy production and examine the relative importance of various sources of food production the major consumptive uses of food.

Since 1961, the first year for which data is available, there has been an impressive 3 times increase in food energy production—well outstripping the 2.2 times increase in population.

The fact that the food energy production rate has out-striped the human population growth rate over the same period might imply an increase in food energy available per capita.  A closer look shows that the proportion of total food production going to humans has declined slightly from 55% to 50% of the total food energy supply.  Similarly, the proportions of food energy going to seed and feed have also declined.  On the other hand the non-food use of food production has gone up.  So—has food energy per capita up or down?  Well, overall the human food energy per capita has gone up: the food energy per capita in 2007 is about 1.25 times greater than in 1961. 

Of course, these statistics don’t reveal how that food energy production is distributed in different regions of the world.  One would need a country-by-country or region by region analysis to address this.

The growth rate in total food energy production, like the human population growth rate, has been decelerating.  

I interpret this as an early warning sign of peaking total food production, although the signs of an actual peak are not readily discernable from the data such as shown in Figure 1.  Like all production peaks, you can’t clearly see peak production until you are at the plateau, or even better, on the decline sign of production.

It should not be surprising that the growth in total food energy production generally out-strips population growth—if it didn’t for any length of time, there would be a food shortage and human population growth would have to drop down to match the growth rate of food production.  The linear regression trends shown in Figure 3 suggest that this situation might occur starting about 2032.

A whooping 90% of total food energy production is derived from plants and the remaining 10% is derived from animals.  Cereals are clearly the major sources of global food energy. 

It is interesting that four food items, three cereal crops (maize, wheat and rice) and one oil crop (soyabeans/soyabean oil) provided about half of the total food energy production in 2007, up from 41% in 1961 (Figure 5).  This illustrates, I think, that the world is reliant on a relatively small number of food types for its food energy.  I think that this is a reflection of the success, or consequence, of the industrial-style petroleum-driven green food revolution over the past half-century.  But, this also shows that the world's food supply is not very resilient.  For instance, if there were to be a major production decline in any one of the four food types, total food energy production could easily decline by 10%.

Food as a globally imported or exported commodity has doubled since 1961.  By 2007 about 20% of the total food energy production was involved in cross-border imports or exports.  Again, I think that this is a reflection of the success, or consequence, of the industrial-style petroleum-driven green food revolution over the past half-century.  But this also shows that countries (at least the net food importing countries) are increasingly dependent on food imports.  Consequently, a major disruption in the food import/export system would cause a larger disruption in the food supply now, than it would have in 1961. 

---------------------

Well, as the first in a series of similar posts that I hope to put out over the coming weeks—the world seemed to be a good place to start, to get on overview of the global food production system, and, have a basis for comparison when looking at smaller regions.   I am keen to look at the quantities and rate of change in food imports and exports because I think that this will play an important role in determining where shortfalls in the food supply are likely to occur.

The amount of data crunching for these studies is quite daunting, so I have to be selective about what to look at. 

Where to go next? 

Let's look at the USA.