Editorial NexoScience
Imagine you empty your shopping bags onto the kitchen counter tonight and someone asks you to price the contents in greenhouse gases, hectares of farmland, cubic metres of water and species pushed towards extinction. Most of us have no idea what that receipt would look like. A large British study did the arithmetic for 55,504 real people, and the gap between the top and the bottom of the scale turned out to be far wider than most food debates assume.
The study measured what people actually eat, not what a model assumes they eat. The dietary data come from the EPIC-Oxford cohort, which recruited adults across the UK between 1993 and 1999 partly through vegetarian and vegan society mailouts, which is why the sample contains an unusually large number of people on plant-based diets: 2,041 vegans, 15,751 vegetarians, 8,123 fish-eaters and 29,589 meat-eaters split into low (under 50 g of meat a day), medium (50 to 99 g) and high (100 g or more). Each participant filled in a validated 130-item food frequency questionnaire and self-classified their own diet, so the categories include real-world eaters who occasionally breach the strict definition rather than idealised textbook diets.
The environmental half of the equation came from thousands of farms, not from a single average. Researchers linked the 289 food items in the questionnaire to the Poore and Nemecek database, a review of 570 life-cycle assessments covering more than 38,000 farms in 119 countries. Processed foods were matched against 4,015 real UK supermarket products so that the ingredient composition of a ready meal or a biscuit was properly accounted for. A two-stage Monte Carlo simulation with 1,000 iterations then carried the enormous variation in how and where food is produced all the way through to the final diet-group estimates, with every diet standardised to 2,000 kcal a day and adjusted for age and gender.
Every single environmental indicator rose in step with animal-based food consumption. Compared with high meat-eaters, vegan diets accounted for 25.1% of greenhouse gas emissions (95% uncertainty interval 15.1 to 37.0%), 25.1% of land use (7.1 to 44.5%), 46.4% of water use (21.0 to 81.0%), 27.0% of eutrophication, the nutrient pollution that triggers algal blooms (19.4 to 40.4%), and 34.3% of biodiversity impact (12.0 to 65.3%). Broken down by gas, the contrasts sharpen: methane emissions from high meat diets were 15.3 times higher than from vegan diets, a direct consequence of ruminant digestion, while nitrous oxide, driven mostly by fertiliser used to grow animal feed, was 3.6 times higher.
You do not have to go vegan to move the needle. This is arguably the most policy-relevant part of the paper. Low meat-eaters, people still eating meat but under 50 g a day, produced 57.2% of the carbon dioxide of high meat-eaters, 57.4% of the eutrophication and 43.8% of the land use. For most indicators the difference between low and high meat consumption reached at least 30%, which places a substantial chunk of the available benefit within reach of someone who simply eats meat less often.
The findings describe an association, not a controlled experiment. Nobody was randomly assigned to a diet. People chose their own eating patterns and reported them themselves, and the groups differ in ways that matter: vegans and vegetarians were younger than fish-eaters and meat-eaters, and vegans reported lower total energy intake, which is precisely why the authors standardised everything to 2,000 kcal. That standardisation is conservative, since meat-eaters genuinely consume more calories on average; without it, vegan footprints fell to between 5% (methane) and 38% (water use) of the high meat group. Two further caveats deserve to be flagged honestly. The dietary data are nearly three decades old, so today’s meat and dairy alternatives are barely represented, and the biodiversity metric only counts terrestrial vertebrates, leaving plants, invertebrates and marine life out of the picture. The water use and biodiversity comparisons also carry wide uncertainty, to the point where the vegetarian biodiversity estimate (64.8% of high meat-eaters) has an interval running from 24.5 to 102.3%, overlapping with no difference at all.
The variation in how food is produced does not dissolve the pattern. That was the central question the researchers set out to test, and the answer was no. Uncertainty for any individual food is genuinely large, but once dozens of foods are combined into a whole diet, the ranking between groups holds. Independent cohorts point the same way: in France, vegan diets produced 24.5% of the emissions of meat-eaters, almost exactly matching the British figure.
Scaled to the country, the numbers become a policy problem. UK adult diets add up to roughly 120 million tonnes of CO2 equivalent, 230,000 km2 of farmland, 15 km3 of agricultural water and 690 kilotonnes of phosphate equivalents a year. Against a legal commitment to cut emissions 78% by 2035 and halt biodiversity loss by 2030, and with technological fixes, yield gap closure and waste reduction offering perhaps 15% together, the authors argue that dietary change is not an optional extra. Environmental labelling, restrictions on promotions and taxes on high-carbon foods are the levers they name, borrowed from tools already used in healthy-eating policy.
Source: “Vegans, vegetarians, fish-eaters and meat-eaters in the UK show discrepant environmental impacts”, published in Nature Food on 20 July 2023. DOI: 10.1038/s43016-023-00795-w. Open access under a CC-BY licence.
