PNAS (Proceedings of the National Academy of Sciences) is one of the world’s most-cited and comprehensive multidisciplinary scientific journals, publishing more than 3,500 research papers annually. The Proceedings of the National Academy of Sciences (PNAS), a peer reviewed journal of the National Academy of Sciences (NAS), is an authoritative source of high-impact, original research that broadly spans the biological, physical, and social sciences. The journal is global in scope and submission is open to all researchers worldwide.
PNAS was established in 1914 in honour of the semicentennial anniversary of the National Academy of Sciences. Since then, they have worked to publish only the highest-quality scientific research and to make that research accessible to a broad audience. In addition, PNAS publishes science news, Commentaries, Perspectives, Special Features, podcasts, and profiles of NAS members. In 1995, PNAS began accepting Direct Submissions from researchers without an NAS affiliation. While they retain strong ties with the NAS, whose members oversee the journal’s rigorous three-tier peer review process, they now receive more than 17,000 Direct Submissions each year. Direct Submissions account for more than 75% of the research we publish.
The approach
One barrier to enabling transitions to more environmentally sustainable food systems is the lack of detailed environmental impact information. PNAS provides an initial approach to overcome this barrier using publicly available information to derive first estimates of the environmental impact of >57,000 food products across four indicators: greenhouse gas emissions, land use, water stress, and eutrophication potential. Pairing it with a measure of nutrition shows a tendency for more nutritious foods to be more environmentally sustainable, and that like-for-like substitutes can have highly variable environmental and nutritional impacts. By estimating the environmental impacts of food products in a standardized way, their approach provides a step to enable informed decision making by end users such as consumers and policy makers.
Understanding and communicating the environmental impacts of food products is key to enabling transitions to environmentally sustainable food systems. While previous analyses compared the impacts of food commodities such as fruits, wheat, and beef, most food products contain numerous ingredients. However, because the amount of each ingredient in a product is often known only by the manufacturer, it has been difficult to assess their environmental impacts. Here, they develop an approach to overcome this limitation. It uses prior knowledge from ingredient lists to infer the composition of each ingredient, and then pairs this with environmental databases to derive estimates of a food product’s environmental impact across four indicators: greenhouse gas emissions, land use, water stress, and eutrophication potential.
Using the approach on 57,000 products in the United Kingdom and Ireland shows food types have low (e.g., sugary beverages, fruits, breads), to intermediate (e.g., many desserts, pastries), to high environmental impacts (e.g., meat, fish, cheese). Incorporating NutriScore reveals more nutritious products are often more environmentally sustainable but there are exceptions to this trend, and foods consumers may view as substitutable can have markedly different impacts. Sensitivity analyses indicate the approach is robust to uncertainty in ingredient composition and in most cases sourcing. This approach provides a step toward enabling consumers, retailers, and policy makers to make informed decisions on the environmental impacts of food products.
One key step to enabling transitions to an environmentally sustainable food system capable of meeting international environmental targets is to estimate and then communicate the environmental impacts of food products available for purchase. This information is increasingly desired. Consumers increasingly want to make decisions on the environmental sustainability of foods, food corporations are setting ambitious net zero greenhouse gas targets, and food retailers are beginning to implement front-of-pack eco labels on their food products. While previous analyses were a step toward providing environmental impact information on foods, they focused on food commodities such as fruits, red meat, or nuts. This leaves a major information gap, as the majority of the tens of thousands of food products for purchase at food retail stores contain multiple ingredients.
This means the environmental impacts of most food products are not readily known. There are at least two reasons for this: First, the exact amount of each ingredient and their supply chain in each food product are often considered a trade secret, and thus the quantitative composition of a product’s ingredients is not often provided on a food’s ingredient list. Second, the sheer number of food products makes the task daunting, as an individual retailer often markets tens of thousands of food products. Although environmental certification labels such as the Roundtable on Sustainable Palm Oil and the Marine Stewardship Council for seafood are an initial step to communicating the environmental impacts of foods, these certifications cover a small set of foods and do not report a quantitative measure of a food’s environmental impact. This makes it difficult to compare the sustainability of foods labelled with different environmental certifications and foods not labelled with any certification.
To begin addressing this information barrier, they developed and tested the accuracy of an algorithm that uses publicly available information to derive first estimates of the environmental impacts of food products. Using these results, they investigated trends in environmental sustainability across types of food products. They further illustrated two potential applications of this approach, first by examining the correlation between the environmental and nutritional impacts of food products, and second by investigating the variation in environmental and nutritional impacts of similar and potentially substitutable foods.
Results
To visualize the results of our analyses, and because time-restricted consumers may prefer simpler ecofriendly labels, they derived a single estimated composite environmental impact score per 100 g of product that ranges from 0 (no impact) to 100 (highest impact). This composite score condensed information from the four environmental indicators, placed equal weight on each indicator, and is on a linear scale. This means, on average, across the 4 environmental indicators, a product with an estimated environmental impact score of 10 has 5 times the impact of a product with a score of 2, but half the impact of a product with a score of 20. Pairwise correlations between the ordered environmental impacts of each indicator showed that foods with a low environmental impact for one indicator on average have low impacts for other indicators, although there are some exceptions to these trends.
For example, almond production results in relatively few greenhouse gas emissions but typically results in high levels of water stress, whereas fishery-caught crustaceans can result in high amounts of greenhouse gas emissions but require little to no land use. The environmental impact estimates for each indicator remain available for situations that are better suited to the disaggregated estimates—for instance, when companies have targets focusing on a single environmental outcome (e.g., as in net zero greenhouse gas emissions policies).
Environmental impact
They examined the estimated environmental impact of foods categorized into the different food classifications used by food retailers. These food classifications increase in specificity from Department (e.g., “Bakery”), to Aisle (e.g., “Croissants, Brioche, and Pastries”), to Shelf (e.g., “Croissants”). Here, they focus on products available at Tesco because it is the largest food retailer in the United Kingdom and because each retailer has its own classification system. To better enable comparison between similar and potentially substitutable foods, they additionally sorted Tesco products into one of eight broad food types based on the product’s Aisle and Shelf: Beverages; Fruits, Vegetables, and Nuts; Cereals and Bread; Snacks; Desserts; Kitchen Accessories; Prepared foods; and Dairy, eggs, meat, and plant-based alternatives.
At Tesco, Aisles with the lowest estimated environmental impacts are often sugary drinks and other beverages composed primarily of water. This is because these products contain small quantities of sugar and other ingredients (e.g., flavouring, syrups, fruit) per 100 g, with the majority of the product composed of water. Vegetables, snacks (e.g., chips, crisps, popcorn), dairy and meat alternatives, some cereal grains, and breads had an estimated environmental impact score below 2. Many desserts (e.g., cakes, biscuits, pies), other cereals and breads, and prepared foods (e.g., pizzas, ready meals) had an estimated environmental impact score that ranged from 2 to 5. Higher impact Aisles with an average estimated score from 5 to 10 included nuts, sweet and savoury spreads, cheese, fish, and some meats (pork and poultry). The highest impact Aisles with estimated scores >10 primarily contained beef and lamb products. Trends at other retailers follow similar patterns.
Correlations between Environmental and Nutritional Impacts
One potential use of the environmental information is to pair it with a measure of nutrition quality to illustrate potential trade-offs between environment and nutrition. Previous analyses focusing on single-ingredient foods found a general trend for healthy foods to have low environmental impacts and for less healthy foods to have high environmental impacts. However, while useful to investigate broad trends, these analyses are limited because most of the products available for purchase in UK food retail stores that provided an ingredient list contained more than one ingredient (96.4% of products in the data sample). It is therefore unclear whether this tendency is also observed across the array of products available in food retail stores.
They assessed the nutrition quality of products using NutriScore, which is a food health profiling method used in multiple countries that has improved population health outcomes. NutriScore gives a numeric score to food products for each of seven food components: calories; salt; saturated fats; sugar; protein; fibre; and fruits, nuts, vegetables, and certain oils (olive oil, nut oils, and rapeseed oil). These numeric scores are then summed and converted into an A (most nutritious) to E (least nutritious) ranking, which they converted to a numeric score ranging from 1 (most nutritious) to 5 (least nutritious) to allow for averaging across products. However, NutriScore is known to have limitations. For example, it does not account for how processing or home preparation (e.g., frying and sautéing) may affect the nutrition impact of a food. However, a product’s estimated NutriScore is highly correlated with a product’s estimated nutrition impact under alternative approaches that incorporate these aspects of food preparation. Due to this and other limitations, NutriScore is being revised to better reflect evidence from epidemiological and public health literature.
Across all retailers, comparing the mean estimated environmental and nutritional impact of retail Aisles containing only food products suggests a tendency for more environmentally sustainable Aisles to be more nutritious than less sustainable Aisles, but with large variation around this general trend. This correlation was also significant for 3 of the 8 retailers in this analysis. Correlations including only drinks were significant across all of the retailers and for one individual retailer, while correlations including both foods and drinks were significant across all of the retailers but for none of the 8 individual retailers.
Impacts of Replacing Meat, Dairy, and Eggs with Alternatives
Replacing meat, dairy, and eggs with plant-based alternatives could have large environmental and health benefits in places where consumption of these foods is high. There are multiple ways to achieve this dietary change, including direct and large substitutions (e.g., beans instead of beef), or smaller transitions between like-for-like products. In some cases, large substitutions may be difficult because of taste preferences, cultural norms, or access to appropriate alternatives. Instead, smaller transitions could be more palatable. They therefore examined specific types of food—sausages, pesto sauces, lasagna, and cookies—to investigate how the environmental and nutritional impacts of direct substitutes may vary and how sourcing may affect the environmental impacts of these products. To identify smaller differences in nutrition quality that may not be possible to identify in NutriScore’s A to E ranking system, they report nutrition quality by scaling the numeric algorithm underlying NutriScore so that it ranges from 0 (best nutrition quality) to 100 (worst nutrition quality).
The algorithm and associated data inputs are available at Oxford Research Archives (https://ora.ox.ac.uk/objects/uuid:4ad0b594-3e81-4e61-aefc-5d869c799a87 ). Due to legal constraints, the product-level data available at the above link is anonymized. A nonanonymized version of the product level data are available under license upon request from R.H. and P.S. To request a nonanonymized version of the product-level data used in the analyses for the purpose of replicating findings, please email foodDBaccess@ndph.ox.ac.uk.
