Eating oil

Revisiting the energy - food nexus

“For industrial man no longer eats potatoes made from solar energy, now he eats potatoes partly made of oil” (Howard T Odum)

Perhaps you heard some statement about how much oil, or energy, is used to produce our food. Often I hear figures in the range of that it takes 10 units of external energy to produce 1 unit of food energy. But one can also read that modern farming is energy positive, at least the proponents of biofuel make claims that they produce three times more energy than they use.

The whole discussion is admittedly quite complex and I will try here to sort things out – a bit at least. Let’s start with some definitions and boundaries. On the output side the question is if we discuss the situation at the farm-gate, i.e. gross output or if we discuss food (including or excluding alcoholic beverages) brought to the market, or if we discuss the total agriculture commodities, including biofuels, industrial crops (cotton, corn and cane for plastics etc.). We also need to determine how we equal food calories with energy contents of the crops not eaten or the part of crops not eaten.

On the input side, we first need to determine what kinds of energy that are included and what kinds are not included. The main energy supply to farming is sun light, but it is just a small share of that energy reaching the field that is converted by the photosynthesis into carbohydrates or other energy dense products. And of that only a smaller part is present in the crops we harvest. If one calculates backwards from the crops actually harvested, we find that the energy in harvested products represents only 0.4% of the solar energy that reaches the fields. Of this 0.4%, only 61% is actually used. Thus, in reality we use only around 0.25% of the solar energy reaching the ground. In my discussions here I don’t include solar energy.

Human labour energy is another input that sometime is included, especially in research taking a life cycle analysis approach and those wishing to compare “primitive agriculture” with industrial agriculture. But to include human energy would in many cases lead to a double counting, as the human energy is derived from the agriculture system. The input and the output would be the same for a self-sufficient farm family working only manually and with no sales. Some of the analyses use the energy consumed for work of a human labourer, which might be just half or a third of the total food energy and some use the actual output of human work which might be less than 10 percent of the food intake. One can make a similar argument about animal draught power. To the extent the draught animals are fed from the agriculture system their energy is just circulating in the system in a similar way as animal feed for the other livestock.

The boundaries are also very important when discussing the inputs. A farm gate perspective loses the considerable amounts of energy that is used in the processing, distribution and cooking of food. The latter is rarely included, but is an essential part of the agri-food system as it is in its prepared and cooked forms that food enters into our bodies.

Back of the envelope

I always find it valuable to make some back-of-the-envelope calculations to put any figures into context. The world’s total energy production was 622 EJ in 2022 (up 50% since 2000), according to the IEA. The global food supply the same year, counted on energy in food made available to consumers (which includes waste) according the FAO:s database 8,658,000,000,000 kcal = 36 EJ. Assuming that the agrifood system as a whole consumes approximately one fourth of the total energy use (i.e. just above 150 EJ), that means that less than 5 units of energy are used to produce one unit of food. However, that figure will also include energy use for non-food crops, roughly 12 percent of the agriculture output is used for biofuels or industrial purpose, although they don’t use as much energy in the post farm gate stage. If we, again, assume that the farming part of this take 20% of the energy used, we land approximately on a 1:1 ratio on a farm gate level.

The total energy use of the US agrifood system

The excellent USDA ERS report, Energy Use in the U.S. Food System, by Patrick Canning and colleagues from 2010 states that the whole food chain consumes around 16% of the total energy use in the United States and that delivering the average American’s 2,000 calorie diet requires nearly 32,000 calories of energy inputs. Farm operations in the United States consume only 14% of the total energy used in the food chain, while handling, processing and retail on the one hand and preparation and consumption on the other, use more or less equal shares of the rest. This greatly influences the energy ratio of foods we eat, and comparisons of what we eat look very different than those made of produce at the farm gate.

Delivering the average American’s 2,000 calorie diet requires nearly 32,000 calories of energy inputs. 

Different kinds of agricultural production and different foods have different energy ratios. The energy efficiency of deep-sea fishing, meat production from feedlots and vegetables grown in heated green­houses is very low. It is much higher for grains, grazing animals and root crops. One third of all energy in the food system in the United States is used for snacks, convenience foods and beverages. Food that requires cold chains and a lot of processing consumes a lot of energy post-harvest. At the farm level, grains are most efficient (which is why they were traditional staple foods in the first place), but processing them demands a lot of energy.

When we add in what happens at the consumption level things get even more complicated. Here we need to add the energy use in house­holds and food eaten ‘out’. This includes the energy spent on frying and boiling, and storage in refrigerators and freezers etc. More than a quarter of the energy in the food chain in United States is used in households and one sixth in the food service sector, restaurants, cafés and catering.

Similar, but lower figures from Sweden

A mapping of the energy use of the Swedish food system, 2000, by Christine Wallgren and Mattias Höjer also shows that 25% of the energy in the food system is used in household electricity and 19% in the farms. The total energy use of the food system in Sweden corresponded to around 9 percent of the total energy use. The energy use per capita and day equals 9,000 kcal, a figure very much lower than the 32,000 kcal in the US research. I have not been able to clarify why there is such a difference, but I am quite certain that it is mostly a reflection of different methodology and boundaries.

Dramatic decreasing energy return in the Spanish agriculture system

An analysis of the metabolism of the Spanish agriculture system between 1960 and 2008 by Gloria Guzman and colleagues shows a remarkably deteriorating energy efficiency, primarily as a result of the intensification of livestock production based on imported feed. Note that this analysis ends at the farm gate.

Fertilizers represent the biggest use of energy on the farm level

Zeke Marshall and Paul E Brockway published A net energy analysis of the global agriculture, aquaculture, fishing and forestry system in 2020. Their analysis includes human labour and also animal labour, and as the title show also forestry, aquaculture and fisheries. According to their research the net Energy Return on Energy invested of the whole system increased from 2.9 in 1971 to 4.0 in 2017. A lot better than my napkin figures. How come? To begin with, the calculation doesn’t include the post-harvest processing, distribution and consumption stages. In addition, their analysis includes forestry and their figures show that a quarter of the energy output is from forestry while a minor share of the energy use is for forestry. The inclusion of aquaculture and fishery makes just a small difference. They didn’t include the energy embedded in machinery, buildings or the transports of inputs to the farms, which can be considerable in the case of animal feed and fertilizers. Taking all that into account their figures are not much different from my napkin’s.

Perhaps surprisingly, the amount of human labour increased during the period. That is a result of that rapid population growth still “outweighs” the reduced share of the population engaged in farming. Human and animal labour, direct energy use and indirect use embedded in fertilizers and pesticides (of which fertilizers represent 80%) stands for roughly one third each.

Another paper by Paul Steenwyk et al, analysing the muscle and machine work in agriculture since 1800, states that human muscle work remained fairly constant over the whole period, while animal work decreased considerably. As late as 1920, 97% of all work on the farm was fuelled by feed or food, by 2012, they represent less than 20%. Of course, year 1800 the global population is estimated to have been 1 billion and today we are 8 billion, so the share of people working in agriculture has been reduced, according to their figures from 30% to 10% expressed as a percentage of total population and not as a percentage of the work force. Note that this is only for the physical work on farms, not fertilizers or embedded energy etc. and not post farm gate, all of which represent a bigger share than the physical work on farms.

Comparing apples and apple pie

As you can see, one can use all kinds of figures to support various statement and argument. When comparing earlier periods with today, I believe it is highly misleading to only calculate farming operations’ energy use, even if we include fertilizers and energy embedded in machinery. The “efficiency” of the farming operations is a result of the global market economy and the specialization and scale associated with that. But it is also based on the corresponding global food system and consumption pattern. Before the industrialization and commodification of farming, a very high proportion of farm output was processed and consumed on the farms themselves or in the near vicinity. This means that the energy use for processing, transport, wholesale, retail and consumption, which is a bigger part of total energy use, should be included in order to compare two systems in a meaningful way.

Already cooking is wasting energy

Examining the energy ratios in the food system gives an interesting per­spective but conclusions shouldn’t be taken to the extreme. When we started cooking we improved the efficiency of our own metabolism considerably, while the energy efficiency went down. Farming with fire, slash and burn, has an appalling energy ratio, but is an efficient way of farming when land is abundant and forests regrow.

When energy prices rise, agriculture prices follow suit. This was seen when food prices rose following the oil price hike in 2007-2008, and again in Europe after the energy crunch following Russia’s invasion of Ukraine. Increased energy prices influence food prices through at least five mechanisms. Higher energy prices make production, at all stages, more expensive; it becomes more interesting to use biomass for biofuels, thereby reducing food production leading to higher prices; it will increase transport costs that are directly reflected in food prices; and competition will be reduced in the food sector as increased trans­port costs re­duce global competitive pressure; and solar farms are now also competing with food production.

Rising energy costs (whether they are caused by climate policy or fossil fuel depletion) will make those parts of our food supply which are most energy-inefficient, such as air-freighting food and global cool chains, obsolete. The price of nitrogen fertilizers and pesticides will increase considerably and they will therefore be used less. It is an open question if it will be cheaper to produce nitrogen fertil­izers from solar energy than to bind nitrogen from crop rotations with biological nitrogen fixation. I would bank on the latter. The fossil free nitrogen fertilizers now brought to the market are 2-3 times more expensive than the fossil derived version. In Sweden and Denmark two such projects are just put on hold.

An energy scarce future will also shift diets back to more grain and pulses and less processed and ready-made foods. Meat and milk from grazing animals would increase, while the recent growth of crop-fed beef, chicken and pork will be reversed.

Increasing energy prices will also realign the balance between the urban and the rural to some degree and will most certainly pose a big challenge to megacities of thirty million people in areas without food production. Their situation will be precarious (their situation is already precarious from a lot of perspectives). In general, the rift between where food is produced and where it is consumed will have to narrow and long distance trade in food will be dramatically re­duced. Sea transportation of quite large quantities of grain is possible, even with sailing ships. As late as the beginning of the 20^th century there were still thousands of great windjammers shipping grain between Australia and Europe.