A study led by members of the Mandala Consortium at the University of Cambridge, has revealed that the impact of food production on species extinction risk can differ by up to a thousand times depending on the type of food and where it is produced. These findings, published in Nature Food in September 2025, provide fresh insights into how changing food policies could help reduce biodiversity loss.

By far the greatest threat to most of the world’s wild species is the loss and degradation of natural habitat. Around 75% of threatened or near-threatened land-based vertebrates are facing increased extinction risk due to reductions in the size, quality, and connectivity of their habitats. This is driven primarily by the expansion of agricultural land used to produce food, arising from the growing number of people to feed, and an increase in demand for foods that have a disproportionately greater impact on habitats.

To better quantify the impacts of different foods on species extinction risk, researchers at the University of Cambridge Zoology Department have developed Land-cover change Impacts on Future Extinctions (LIFE) metric. This assesses the impact of land use on the extinctions of some 30,000 land vertebrate species (mammals, birds, reptiles and amphibians) and can estimate the impact of producing 1 kilogram of different food types on the chances of extinctions.

Dr Thomas Ball, Lead author on the study said:

The LIFE metric allows users, for the first time, to directly assess the impacts of different land uses across the globe on species extinction risks. Linking this to data on crop and livestock production means that we can investigate the potential extinction impacts of almost all food products.”

Beef and beetroot, north and south

The study finds significant differences in the impacts of different food types:

  • Ruminant meat (largely beef and lamb) has a per-kilogram impact often roughly 1000 times greater impact than vegetables, and about one hundred times greater than plant-derived proteins like legumes.
  • Even within food types (e.g. dairy, legumes, coffee), there is large variation in impact depending on where the food is produced, sometimes over ten times as much.

Where the food is produced and consumed is also crucial:

  • In high income countries such as the United Kingdom, most of the extinction risk comes from imported commodities. In the UK, around 45% of the food eaten is imported, and the overseas biodiversity impact of UK food consumption is roughly 20 times greater than the impact of UK’s own domestic production.
  • A large portion of this impact is driven by imported beef and lamb, as well as tropical products such as coffee, tea, cocoa, bananas, and other imported fruits.
  • In contrast, for countries such as Brazil, Uganda and India, most of the biodiversity risk arises from domestic production. In particular, animal products and commodities grown in the tropics are generally much more impactful than staple crops and vegetables.

LIFE changes

These findings are directly relevant to Mandala’s mission of understanding how food systems can better balance human and planetary health. The research shows that, while different mitigation strategies are likely needed in different settings, even moderate shifts in food production and diets could lead to large reductions in extinction risk.

The authors modelled alternative diet scenarios for the United States (including the EAT-Lancet Planetary Health Diet, vegetarian and vegan models) and found that reducing ruminant meat consumption markedly lowers species extinction risks.

Improvements in domestic production methods, sustainable intensification, and sourcing from lower-impact regions may also reduce impacts. The study emphasises that policy, trade, and supply-chain decisions matter, not simply individual dietary choices.

The researchers acknowledge that the LIFE metric focuses on terrestrial habitats and vertebrates. Further works is needed to extend this understanding to the impacts on aquatic species, plants, invertebrates, as well as other threats such as pollution, pesticides or greenhouse-gas effects of food systems. Data availability and quality still also pose limitations, but the model is designed so that, as data improves, LIFE values can easily be recalculated.

Find out more