Dealing with Postharvest Losses
This is Part 2 of a three part series about food loss and waste. Last week, I discussed preharvest losses, and next week, I will discuss food waste, which occurs after the food has reached the intended consumer. This week is postharvest losses, which occur after harvesting and in the distribution system.
There is the story, illustrating what is known as the Streetlight Effect, of the drunk man who is looking for his keys under a streetlight. A policeman comes upon him and helps search for the keys. After a few minutes, the policeman asks, “are you sure you lost them here?” The drunk man says, “no, I lost them at the park. But here is where the light is”. So is the case with postharvest losses. The statistics available for postharvest losses are much more thorough than for preharvest losses, and so this is where NGOs place greater attention.
Today, we will examine the historical perspective on humanity’s millennia-old battle against postharvest losses, how postharvest loss occurs today, and how they can be reduced.
For most of human history, the primary lifeway was that of the hunter-gatherer. People had a limited capacity to store things, such as food, and for the most part the food that was available was that which could be hunted or gathered then and there.
With agriculture came one of the central civilizational needs: to transport food across space and across time. By “across space”, I mean to transport food from where it is grown to where it is needed, particularly cities, and by “across time”, I mean to store food from when it is harvested to when it is eaten, as the harvest must last the full year, and longer when one considers preparation for the inevitable bad harvest.
Food storage and transport coevolved with agriculture, and in fact appeared even before full domestication of crops. Granaries have been discovered from around 11,300 to 11,175 years ago in the Jordan Valley. At that time, humans were in a transitional period between the hunter-gatherer lifeway that had long prevailed and the agrarian lifeway that was to come, by deliberately cultivating wild cereals but not having yet domesticated them. These granaries featured suspended floors to protect against moisture and rodents, things that we’ll talk about in a moment.
There are several functions needed of an effective granary. First, it needs to control humidity, as too much moisture fosters mold growth, which causes grain to spoil. Grain also needs to be kept cool, as warmth also encourages mold growth and also insect infestation. These two goals are in tension, as warmth facilitates drying, and so the ideal granary design varies by climate. Aeration, or the use of air flow through small openings in the stored grain, is necessary to keep stored grain dry.
Grain is dried before being placed into a granary. Although grain drying is an ancient process, doing so efficiently remains an active area of research.
The earliest records of insecticides are sulfur compounds used by the Sumerians around 2500 BC and mercury and arsenic compounds used by the Chinese around 1200 BC. Synthetic insecticides were developed from the 19th century, including the infamous DDT. In AD 304, the Chinese scholar Ji Hang described the use of ants to prevent citrus tree damage by preying on pest insects.
Ancient farmers wielded an arsenal of weapons against rodents, another class of pests that threatened stored food. The ancient Egyptians used traps, fumigation, and prayers to the gods against rodents. The Egyptians were also aided by one of humanity’s closest friends.
Domestication of cats appears to have begun around 10,000 years ago in the Middle East, coinciding with the development of agriculture. As permanent settlements appeared, cats may have been attracted by rodents who gathered at garbage heaps, and they found employment in rodent control. Although cat domestication probably did not begin in Egypt, it is in Egypt that we see the first evidence of full domestication—the word “domestication” literally means to put into a house, deriving from the Latin domus.
Rat catching became a profession in the early Middle Ages, which may be best known to us in the tale of the Pied Piper of Hamelin. According to the legend, the Pied Piper used his magic flute to lure the town’s rats away. But the townsfolk did not pay the Pied Piper as promised, and so he retaliated by leading the children of the town to their deaths in a river with the same magic flute. This is where we get the idiom “to pay the piper”.
Preservation is another potent tool for managing losses and enabling farther trade and longer storage. This brief article discusses drying, freezing, fermenting, pickling, curing, jam and jelly, and canning. We can’t discuss all of these things today, but let’s consider curing. There are records of Egyptians using salt preservation in 2000 BC, and it may well have been done much earlier. Salt is effective in preserving foods by drying them out, and the bacteria that cause spoilage can’t live in a saline environment. Salting was heavily used by the UK Royal Navy before refrigeration. With ships at sea for months at a time, salt pork, salt beef, and biscuits with generous amounts of salt would last for a long voyage. Mark Kurlansky has written Salt: A World History to document in detail this topic whose importance can be easily underestimated. Aside from preservation, humans generally like the taste of salty food.
I don’t think that any food technology of the Industrial Age is even close to being as important as refrigeration. Assessments of what constitutes “practical” vary, but this article places the first practical refrigerated railway car at 1867. Now, it was possible to ship milk and butter across hundreds of miles into cities. Now, it was feasible to make oranges available year-round in non-tropical environments, a luxury that was previously only available to royalty, if they were lucky. With refrigerated ships, American beef exports to the British Isles increased nearly a thousand-fold in a few years.
In the late 19th century, New York City experienced a banana craze, thanks to imports of the Gros Michel and other varieties of bananas from around the world, something that would not have been possible without refrigeration and fast steamships. Discarded banana peels on the street were a major problem, and then-police commissioner Theodore Roosevelt zealously enforced laws against this behavior. The City’s public service campaign against banana peel litter led to the ubiquitous comedy trope of slipping on a banana peel. When the banana blight set in that ultimately led to the near-extinction of the Gros Michel bananas, one reaction was the 1923 Frank Silver and Irving Cohn novelty song, Yes! We Have No Bananas.
This article goes on to argue that refrigeration is one of three core technologies—the others being elevators and the telegraph/telephone—that enabled large metropolises to emerge in the 19th century. It’s not just that air conditioning made hot Southern cities bearable, or personal refrigerators replacing iceboxes for consumer convenience, though those things are certainly important too. It’s that cities are hungry, and for the logistics of bringing a varied, nutritious diet into cities on a reliable basis, refrigeration is the key. The article further argues that the refrigerated rail car was instrumental in setting the American West, with tragic consequences for the indigenous populations there.
The technologies of food storage, preservation, and transportation play a central and under-appreciated role in the long transition from subsistence hunter-gathering to the safe, reliable, varied, nutritious, and cheap globalized food economy that we enjoy today. And as we’ll see, there is still a long way to go.
Our data for postharvest losses is much better than for preharvest losses. Our World in Data, citing the Food and Agricultural Organization of the United Nations, reports that world postharvest losses are about 13%, varying regionally from 8% in East Asia to 24% in West Africa. This paper explains what one would suspect: losses are higher in areas that lack the capital to invest in more efficient systems.
Today’s first step is harvesting. In poorer countries, harvest is mostly done manually, rather than with mechanical combines as in wealthy countries. Agricultural labor shortages during the harvest season cause delays, which exposes crops to shattering losses due to too low a moisture content, attacks in the field from birds and rodents, and other problems. This study finds wheat losses in India at the harvest phase to be 1.84%, a figure that has improve dramatically over the last few decades, and this study finds an increase in losses due to lack of harvesting equipment.
The next phase is threshing and cleaning. Again, delays, which may result from a lack of equipment, result in losses from pests. Mechanical threshing was found to result in lower loss than manual threshing for rice in Ghana.
After that comes drying. A study by Hodges et al., as cited by Atkar et al., found that sun drying, the most common method of natural drying, resulted in a loss of 3-5%, whereas mechanical drying had a loss of 1-2%. Sun drying requires leaving crops exposed to birds, and there is a risk of the weather not being as sunny as hoped.
Then comes storage, which is the phase with the greatest share of losses. This paper finds that loss for 90+ days of storage of maize in Uganda can exceed 59% with traditional storage. No modern storage facility examined had a 90+ day loss rate exceeding 3%.
After storage is transportation. In low-income countries, losses tend to be higher as a result of bruising and spillage resulting from poor roads and vehicles, whereas such losses are near zero in wealthy countries. The unloading of crops at the processing facility also tends to be more efficient in wealthy countries, resulting in fewer losses. Mechanisms of loss include bruising or spillage as just noted, rejection of shipments, poor temperature control, inspection delays when food is shipped internationally, and problems with packaging. I haven’t found many good numbers on transportation losses.
Last month, a paper made the extraordinary claim that full development of cold chain infrastructure for food distribution could save the equivalent of 1.8 billion tons of CO2 and save 620 million tons of food loss. I haven’t looked at the paper’s calculations in detail, which are supplied as supplementary information. But I do note that Our World in Data reports, drawing on figures from Climate Watch, that world agriculture is responsible for 5.87 billion tons CO2-equivalent of emissions. Given that all food loss worldwide comprises 13% of all food production, and that 13% of 5.87 billion tons is 763 million tons, and that not all losses are relevant to the cold chain, I find this claim to stretch plausibility, even considering that of the 13% of food loss, a disproportionate amount is highly carbon-intensive meat, seafood, fruits, and vegetables.
Next, we get to milling. The paper reports that milling of rice has a theoretical yield of 71-73%, defined as the ratio of milled rice mass to rough rice mass, but the average yield from small mills in five rice-producing countries in Southeast Asia is 57%. FarmProgress reports that the yield in the United States is typically 68-72%.
Finally comes processing and packaging. This article states that 12% of all food loss and waste occurs at the processing and packaging state, which is around 2-3% of all food harvested. Packaging can get a bad wrap—pun intended—for contributing to plastic waste, but such worries should be balanced against the potential for good packaging to extend shelf life, protect food, and thus reduce food waste. See also, this article, which raises the issue but does not definitely calculate if reduction of food waste offsets the environmental costs of more packaging.
Distribution and retail are the next phases, but I will save those for next week’s discussion of food waste.
It is clear that when it comes to reducing food waste, the big numbers are found with bringing modern machinery and infrastructure to where it is lacking. With mechanical combines, mechanical threshing and drying, modern grain storage, developed transportation infrastructure with adequate refrigeration, and mechanical milling, I reckon that postharvest losses could be cut by at least half. I would also expect at least some room for improvement at the frontier; those supply chains that use modern machinery at all points are probably not the best possible.
But this does leave one important piece of unfinished business. As discussed at the beginning, a central function of efficient food distribution is to allow for more complex supply chains. Farmers can, and often do, feed megacities on the other side of the world. To what extent does efficiency result in further complexifying supply chains, rather than a reduction of food losses? An astute reader may recognize this question as a reformulation of the rebound effect.
I won’t comment on whether highly complex agricultural supply chains are desirable, as such a digression into a heavily value-laded question would lengthen what is already a long post, nor do I wish to contribute to the overdone food-miles debate. The same paper I discussed above on the potential of refrigeration states,
Developing more localized, less industrialized ('farm-to-table') food supply chains in both industrialized and non-industrialized contexts may save greater quantities of food than optimized cold chains.
I also find this statement implausible, granted without having looked at the calculations in detail. I find that slogans such as farm-to-table, locavore, food sovereignty, etc. are attractive slogans to many people, but they lack rigor. The paper may be an attempt to bring such rigor, but advocates of a localized food system often do so for reasons that are not reducible to straightforward environmental metrics.
Quick Hits
The Breakthrough Institute, working with other organizations, celebrates this week Congress passing the ADVANCE Act by overwhelming margins. The bill will increase the capacity of the Nuclear Regulatory Commission and update the NRC’s mission statement to consider the risks of rejecting nuclear power as well as approving it. The bill now heads to President Biden, who is expected to sign it.
I want to comment at greater length soon on the advances in automated coding in the last few years, but one recent interesting paper is on a system called SWE-agent. The system moves beyond single function generation from a docstring, a task at which state-of-the-art language models now excel, and toward complex software engineering. The paper also introduces an agent-computer interface, analogous to the human-computer interface that has been the subject of an extraordinary amount of research over the last few decades. The paper promulgates a few principles of good ACI design, which are again analogous to principles of good HCI design, but interfaces that work well for humans, such as Visual Studio Code, are not necessarily suitable to artificial intelligence agents.
I recently played the Pixel Remaster version of Final Fantasy II. I have similar comments to those that I made a few months ago after playing Final Fantasy I. Pixel Remaster did a better job (if that’s what you want) of making a game faithful to the original than most remakes, but rebalancing and innovations such as the minimap took a game that was frustratingly hard and made it too easy. As for the game itself, FF2 has a complex and well-written lore, and in that regard may be the best on the NES/Famicom (even more so than the sequel, Final Fantasy III). One thing I always found remarkable about FF2 is the ambitious range of gameplay features: the keyword system, the highly open overworld, the complex progression system, and the large set of spells. I have the sense that the developers had ambitions that exceeded what they could do with a Famicom cartridge, and it is a shame that few of these ideas were developed for future games in the series.