Background: A carbon footprint is the sum of greenhouse gas emissions (GHGEs) associated with food production, processing, transporting, and retailing.
Objective: We examined the relation between the energy and nutrient content of foods and associated GHGEs as expressed as g CO2 equivalents.
Design: GHGE values, which were calculated and provided by a French supermarket chain, were merged with the Composition Nutritionnelle des Aliments (French food-composition table) nutrient-composition data for 483 foods and beverages from the French Agency for Food, Environmental and Occupational Health and Safety. Foods were aggregated into 34 food categories and 5 major food groups as follows: meat and meat products, milk and dairy products, frozen and processed fruit and vegetables, grains, and sweets. Energy density was expressed as kcal/100 g. Nutrient density was determined by using 2 alternative nutrient-density scores, each based on the sum of the percentage of daily values for 6 or 15 nutrients, respectively. The energy and nutrient density of foods were linked to log-transformed GHGE values expressed per 100 g or 100 kcal.
Results: Grains and sweets had lowest GHGEs (per 100 g and 100 kcal) but had high energy density and a low nutrient content. The more–nutrient-dense animal products, including meat and dairy, had higher GHGE values per 100 g but much lower values per 100 kcal. In general, a higher nutrient density of foods was associated with higher GHGEs per 100 kcal, although the slopes of fitted lines varied for meat and dairy compared with fats and sweets.
Conclusions: Considerations of the environmental impact of foods need to be linked to concerns about nutrient density and health. The point at which the higher carbon footprint of some nutrient-dense foods is offset by their higher nutritional value is a priority area for additional research.
Alex’s Notes: There is a common belief that vegetarian diets are more environmentally friendly than omnivorous diets. One measure of the environmental impact of a food is through its carbon footprint, estimated as greenhouse gas emissions (GHGEs). No doubt GHGEs are present at every corner of food production such as agricultural production, food processing, transport, distribution, and storage. Even at home we continue this process with food preparation, waste, and disposal. So the central question that cannot be avoided is what food patterns balance good nutrition with sustainability? To answer this, we must first understand the relation between the carbon footprint of food and their nutrient content.
Andrew Drewnowski from the University of Washington obtained the GHGE values for 661 foods and beverages provided by the Casino Group in France. These included fresh and processed meats and dairy products, grains, fats, sweets, and frozen or processed vegetables and fruits. Fresh produce data wasn’t available, and the GHGE data for all products assumed they were purchased from a store and did not include transportation, home preparation, disposal, or waste by the consumer. Water and diet beverages (<5 kcal/100g), oils, ground coffee, syrups, flour, and pastry mixes were excluded from the analysis, and the remaining 483 foods and beverages were analyzed for their GHGE values per 100g of food and per 100 kcal using the French government’s food database.
Carbon Footprint on a weight basis (per 100g of food)
For all the food categories except processed fruits and vegetables, agriculture was by far the most significant source of GHGEs, making up 60-72% of the total GHGEs per 100g of food. In fact, the agricultural GHGE for grains was 30% higher than the GHGE of its processing, transportation, packaging, and storage combined. For meat this rose to 160%. For fruits and vegetables it was only 26% the total, however. Overall, meats were the greatest source of GHGEs, followed by dairy products, grains, sweets, and lastly, fruits and vegetables.
When the GHGE per 100g was plotted against the kcal per 100g, some very strong and significant associations became evident. Values were r = 0.92 for dairy, r = 0.77 for meat, r = 0.32 for processed and frozen fruit and vegetables, and r = 0.77 for sweets, while grains actually had a negative association (r = -0.31), indicating that the more kcal they supplied, the less carbon cost they had. Some of the worst offenders were meat products, cheese, eggs, offal, and basically all the animal products, while the most environmentally friendly foods were sugars, honey, syrup, chips & snacks, chocolate, and biscuits.
Carbone Footprint on an energy basis (per 100 kcal of food)
It sounds like the vegan environmental argument was right… right? Well technically, yes, when you consider the above. But the weight of a food says nothing of its energy content, so the researchers went on to see what effect changing weight to energy had on the associations. Agricultural costs were still the lion’s share of the food categories, but processed fruits and vegetables provided the most GHGEs overall, and specifically in their processing, storage, and transportation. This one change from weight to calories made fruits and vegetables the worst offenders of all the food categories. The GHGE for meats was literally cut in half when compared to its value for per 100g, bringing it down to the level of dairy, which didn’t noticeably change. Grains were also cut in half, and sweets dropped to a third of their per weight values and became the lowest GHGE food category title.
When plotted against the same kcal per 100g scale as earlier, all those strong and significant associations really did a one-eighty. While chocolate, biscuits, sugars, honey, and syrup still performed the best; all food groups had a negative association on average between their GHGEs per 100kcal and their kcal per 100g (r = -0.35 for dairy; r = -0.33 for meat; r = -0.78 for grains; r = -0.72 for fruits/vegetables; r = -0.87 for sweets).
Beyond energy, what about nutrients?
Talk about a drastic change when food is plotted in a (in my opinion) more realistic manner (i.e. by energy value). But that still doesn’t provide a fair battleground for foods, as energy is only one part of the equation. To get still a more accurate answer for our question, we must look at the nutrient profile of these foods.
The first nutrient profile was based on six vital nutrients per 100 kcal of food (ND-6): protein, vitamin D, calcium, phosphorus, potassium, and magnesium. The most nutrient dense foods, by far, were fish, milk, poultry, eggs, cheese, and yogurt. Processed vegetables were also high on the ND-6 list, but had nearly double the GHGE output that the animal foods did. Grains and sweets had the lowest ND-6 and GHGE values. The most significant finding here, however, was that there was a very slight negative association between ND-6 and GHGE for only animal products. This means that as the ND-6 profile for animal foods increased, the GHGE decreased. Conversely, there was a steep positive association for all other food categories such as grains and sweets, suggesting that the more nutrients you want from plants and snacks, the greater the GHGEs.
The second nutrient profile took things a step further and was based on not six, but 15 nutrients per 100kcal of food (ND-15): protein, fiber, vitamins A, C, D, and E, thiamin, riboflavin, niacin, folate, calcium, iron, phosphorus, potassium, and magnesium. These nutrient values were expressed as a percentage of the RDA and summed together. For better or worse, the researchers made the maximum achievable RDA percentage 100% to avoid extreme values for a single nutrient from influencing the total score. With the ND-15, most every association mentioned in the ND-6 remained with few exceptions. The associations for grains and sweets were ever so minimally attenuate, while breakfast cereals and bars provided the same ND-15 sum as eggs, fish, and milk – except at one-third the GHGEs.
I believe the first important thing to point out is how big a difference reporting parameters can make in associations. Simply switching from a per weight basis to a per energy basis reverses the GHGE trend in most food categories.
Second, there is no doubt that sugar and sweets have the lowest GHGEs based on this study, but they also had the lowest nutritional value. Coming back to our question about what food patterns balance good nutrition with sustainability, these foods cannot be viewed as the most sustainable because they do not provide good nutrition. Similarly, the most nutrient-dense foods also had the greatest GHGEs, although the higher GHGE cost of some meat and dairy products were compensated for, to some extent, by their higher nutritional value. That said, the bottom line remains the same – tradeoffs in balancing nutrition with the environmental impact and cost of the diets need to be made.
Lastly, a handful of caveats need to be mentioned. The foods available in this study were from France, and can thus not be generalized easily to other countries. Moreover, the foods were not necessarily representative of what people would consume. There were a high number of processed meats, but only a few fresh meats, and literally no fresh produce. When looking at nutrient-density, protein quality was not taken into account, nor was calcium or iron bioavailability. The GHGE values did not account for consumer contributions, resulting, for example, in underestimations for foods that require refrigeration, cooking, and/or have a lot of inedible waste (this is of course subject to the individual’s food practices).
The foods were also subject solely to industrial production. A 2012 study from Washington State Universitycompared the environmental impact of conventional, natural, and grass-fed beef production systems, and concluded that – contrary to popular belief – conventional has less GHGEs than grass-fed. Keep in mind, however, that the paper is incomplete in that it does not account for the tremendous agricultural burden of growing and transporting corn and other grains to feed the conventional cattle. Finally, the GHGEs are only one measure to assess the environmental impact of food production. If these analyses were also to include the use of water, and the long-term impact of land-use on soil quality, then it would be hard to not imagine that things would appear quite differently. On this note, more studies are definitely needed to determine the relationships between nutrients, energy, and environmental impact of foods. What can be taken away from the current study is that – with regard to GHGEs – animal foods are not the devil they are made out to be, at least not when considered in relation to their nutrient and energy content.