This website is for healthcare professionals only.


We use cookies to collect information about how you use this site. We use this information to make the website work as well as possible and continue to improve it.

Scientific Updates /

Environmental impact of 57,000 new UK & Irish foods

11 September 2022


Original research

Members of the EAT Lancet commission and UK diet experts have come together to expand the data available for the environmental impact of foods, by producing an algorithm which provides the impact of composite foods in the UK and Ireland.

Building on the internationally renowned Poore and Nemecek database which focused mainly on food commodities, this study provides the environmental impact of foods most commonly found in retail supermarkets and most commonly consumed by individuals in the UK and Ireland. Most of these foods are composite foods, made up of more than a couple of ingredients.

To calculate the environmental impact of composite foods, it is necessary to first identify the impact of each individual ingredient based on its volume using the established Poore and Nemecek data, and then summing up the values to provide the total impact of the food. As most recipes of composite foods is proprietary, the exact volume of each ingredient has to be estimated. The group developed an algorithm to estimate the weight of individual ingredients based on current ingredient labelling requirements and aligning with equivalent recipes of just over 1,500 foods whose full recipes were made available.

Algorithm provides an environmental impact score of 0-100 for composite foods

The algorithm calculated the environmental impact as a score of 0 to 100, with 0 indicating the lowest environmental impact and 100 indicating the highest environmental. The values were linear, thus a product with a score of 10 would have an environmental impact five times greater than a product scoring 2, and half the environmental impact of a product with a score of 20.

The environmental impact score takes into consideration four environmental indicators: greenhouse gas emissions (GHGe), land use, water scarcity water use and eutrophication. The score placed equal weighting across all four indicators.

Algorithm found to have reasonable accuracy

The algorithm’s accuracy was assessed by using the 1,547 foods whose full recipes were known and was found to have reasonable accuracy.

The data is split by retail category / aisle placement. The highest impact scores were meat, cheese and some ready meals, whilst the lowest environmental scores were for beverages as the main volume of ingredient was water.

Environmental impact score correlates well with nutritional quality of foods but NOT drinks

The lower the environmental score for foods, the better the nutritional quality as assessed by Nutriscore. This was not the case for beverages, as the main volume of sweetened beverages is of course water.

Scores per 100g - serving sizes may yield different scores

The algorithm scores for environmental impact and Nutriscore are based on 100g volume of foods or drinks. This need to be taken into consideration when interpreting the data. Some foods such nuts and breakfast cereals are often consumed at around 30g servings – hence the environmental score would be significantly lower than calculating per 100g. Similarly for tea and coffee, which are shown to have a higher environmental impact per 100g than sweetened beverages. However, tea and coffee is consumed in quantities less than 5g. Conversely, beverages are often consumed at volumes 2.5 to 3.3 times greater than the current score provided for 100g of product. Hence the scores provided in this publication at per 100g could be significantly higher or lower depending on usual serving size. The authors have also calculated per servings and this data is available in the supplementary material.

A summary of alignment and misalignment with Nutriscore

Figure 1: Environmental impact score and nutrition impact score per 100 g of multi-ingredient food products

PNAS Clark Environ and Nutrition 57K foodsDownload figure Colours indicate food types. Points indicate the environmental impact and nutrition impact scores of aisles with similar foods condensed together. This figure is sourced from the original publication: Clark et al. PNAS USA. 2022;119(33):e2120584119

Win-win aisles for Nutriscore and environmental score

  • Fruit & vegetables

  • Salads

  • Breakfast cereals

  • Some breads

  • Meat alternatives (e.g., tofu, vegan sausages)

Lose-lose aisles for Nutriscore and environmental score

  • Cheese

  • Chocolate

  • Savoury pies

  • Quiches

Win for Nutriscore but Lose for environmental score

  • Fish & seafood

  • Nuts*

  • Some ready meals

    *The authors make a note that although nuts score high for environmental impact, this is based on 100g consumption which is significantly higher than an average serving of 30g, thus the serving size needs to be considered.

Lose for Nutriscore but win for environmental score

  • Sweets

  • Cakes

  • Sweetened beverages including fruit juice

  • Frozen desserts

  • Condiments

Switching within category can produce significant environmental reductions

The team compared the scores between food categories that could be swapped due to their similarity. This was seen as important with regard to behaviour change, as people would be more likely to swap like with like rather than out of category. For example, switching meat for beans and pulses, would produce the biggest environmental reduction, however, consumers are more likely to switch to a plant-based meat alternative option such as soya mince or mycoprotein. The group even found switching the type of meat source in a sausage would produce significant environmental reductions: beef and lamb sausages having a greater environmental score compared to pork sausages which have a greater impact compared to chicken or turkey sausages which have a greater impact than plant-based alternatives.

In summary

This study has expanded our knowledge of the environmental impact of composite foods. Additionally, the algorithm score of 0-100, would be an easier to understand concept for most consumers rather than comparing actual environmental data on different foods. Although, based on UK and Ireland data, there is no reason why the algorithm cannot be adapted for use across different countries who use similar food labelling criteria.

Most importantly, this study demonstrates that, in the main, no trade-off is need when choosing foods that are good for the environment and health. Drinks and a handful of food items, may need special consideration where the nutrition quality does not align with the environmental footprint e.g. sweetened beverages and foods consumed at significantly lower serving sizes than 100g.


  1. Clark M, Springmann M, Rayner M, et al. Estimating the environmental impacts of 57,000 food products. Proc Natl Acad Sci USA. 2022 Aug 16;119(33):e2120584119. doi: 10.1073/pnas.2120584119

Scientific review

Plant-based eating and carbon footprint

Scientific review

Plant-based eating

Original research

The healthiness and sustainability of national and global food based dietary guidelines

Share this article on social media.