The scale of the stakes is enormous. Food production occupies roughly 40 percent of all habitable land on Earth, drives about 80 percent of deforestation, and accounts for some 70 percent of terrestrial biodiversity loss. Livestock production sits at the center of this crisis, consuming one-third of the global cereal harvest, requiring about 40 percent of global arable land, and claiming all of the world’s pastures. Cattle production is particularly damaging because it combines direct on-farm emissions from enteric fermentation and manure with indirect emissions from feed production and land-use change. At the same time, the food system contributes roughly a third of total global greenhouse gas emissions, and climate change in turn accelerates biodiversity loss by altering habitats, disrupting migration patterns, and intensifying extreme weather events.

Previous life cycle assessments have generally concluded that meat analogues outperform beef across environmental indicators, but most of those studies relied on fixed or simplified supply chains and global-average environmental intensity data. That approach overlooks a crucial reality: the same ingredient can carry wildly different biodiversity consequences depending on the country and ecosystem it comes from. The new study addresses this gap by combining environmentally extended multi-regional input-output models with the GLAM life cycle impact assessment framework, which provides spatially and taxonomically explicit characterization factors for both land use and climate change. The GLAM land-use factors jointly account for land-use intensity and habitat fragmentation within a species-habitat relationship framework, while its climate factors model pixel-based temperature niche limits for individual species.

The technical machinery behind the analysis is considerable. The team employed a tiered hybrid model called FABEXIO, which integrates the Food and Agriculture Biomass Input-Output database, covering 191 countries and 123 food commodities, with EXIOBASE to capture non-biomass inputs such as fuels, fertilizers, and energy that conventional biomass models miss. The functional unit was a single raw burger patty of 113 grams, with recipes drawn from earlier experimental work on meat substitutes. Four alternatives were compared against beef: patties based on pea protein, soy protein, mealworm insects, and mycoprotein, a fungal protein grown on sugar-beet molasses. The ten consuming countries, Belgium, the Czech Republic, France, Germany, Greece, Italy, the Netherlands, Poland, Romania, and Spain, jointly represent about 80 percent of EU beef consumption and 82 percent of the associated biodiversity impacts.

The headline result is that land occupation dominates terrestrial biodiversity impacts, accounting for more than 96 percent of the total across all patty types and all countries, with climate change contributing less than 4 percent. This finding aligns with the IPBES global assessment, which identifies land-use change as the leading cause of terrestrial biodiversity loss, but it contradicts earlier life cycle studies that had identified climate change as the dominant driver using older impact assessment methods. Beef patties generally showed the highest land-driven biodiversity impacts, driven jointly by pastures and cropland, with pastures dominating in Spain, Greece, and Italy. On climate-driven impacts, beef also ranked highest everywhere, primarily due to methane emissions from the cattle themselves.

But the most surprising result concerns the pea patty. In Belgium, the Czech Republic, Germany, and Poland, the pea-based burger actually exceeded beef in terrestrial biodiversity impact. The culprit is coconut oil. Although it makes up only 5.7 grams of a pea patty, coconut oil sourced largely from the Philippines, Indonesia, and Vanuatu dominates the biodiversity impacts of pea and mycoprotein patties, contributing roughly 75 percent and 58 percent of their totals respectively. Tropical ecosystems carry extremely high biodiversity sensitivity in the GLAM framework because of high global extinction probabilities, small ecoregion sizes, and fragmented land-use patterns, with cropland characterization factors in the Philippines and Indonesia reaching up to 2.2 times ten to the minus fourteen PDF per square meter. By contrast, beef supply chains in the countries where pea patties performed worst are highly diversified, relying primarily on domestic production with modest imports from Australia, India, and Italy, all of which have comparatively low biodiversity sensitivity.

The decoupling between land-use footprints and biodiversity impacts is one of the study’s most consequential insights. Pea protein, the main ingredient at 20 grams per patty, dominates the pea patty’s land-use footprint at about 46 percent but contributes only around 14 percent of its biodiversity impact, because it is sourced mostly within the EU. Similarly, molasses, the main substrate for mycoprotein patties at 339 grams per patty, accounts for roughly 40 percent of the land footprint but only about 23 percent of biodiversity impacts on average. Conventional mass-based life cycle assessment methods, and land-use footprints more broadly, systematically fail to capture these spatial differences in biodiversity sensitivity, meaning high-impact, low-mass ingredients like coconut oil can slip through the environmental accounting unnoticed.

Not all alternatives fared badly. Soy patties consistently delivered the lowest biodiversity impacts across all ten countries, with their impacts driven mainly by soy protein and rape oil. Replacing beef patties with soy patties in food service, assumed to represent about 5 percent of national beef consumption, could reduce land-driven biodiversity impacts by 71 to 97 percent per patty, corresponding to an estimated 3.6 to 4.9 percent reduction in national biodiversity loss associated with beef. The largest gains would come in Spain, Greece, and Italy, where beef’s impact is highest at around 96 to 97 percent reduction, while Belgium would see the smallest benefit at 71 percent. Even here, geography matters: rape oil sourced from Australia, with a cropland factor of 9.1 times ten to the minus fifteen PDF per square meter, drove up impacts in Belgium, France, Germany, Italy, and the Netherlands, while domestic sourcing kept impacts lower in the Czech Republic, Poland, and Romania.

The sensitivity analyses tested the robustness of these conclusions against alternative allocation ratios, protein-based and energy-based functional units, and beef patty composition. The best- and worst-performing patties remained largely unchanged under most scenarios, though under an energy-based functional unit soy patties lost their top position in Belgium, Germany, and Poland due to their relatively low energy content of 191 kilocalories per patty. The most dramatic shift came when coconut oil was replaced with rape oil on a mass-equivalent basis: biodiversity impacts fell by approximately 74 percent for pea patties and 62 percent for mycoprotein patties, transforming the rankings. Under that scenario, beef patties consistently exhibited the highest biodiversity impacts across all countries, demonstrating that a single reformulation decision can outweigh the choice of protein source itself.

The authors frame their findings as a warning against biodiversity leakage, the displacement of harmful activities from one region to another through trade and consumption shifts. Local environmental gains in Europe may be offset, or more than offset, by impacts displaced to more biodiversity-sensitive tropical regions. They also caution that avoided biodiversity impact means preventing further habitat occupation, not restoring degraded ecosystems, since spared land can be reoccupied for urban expansion, alternative agriculture, or biofuel production. Their recommendations are concrete: manufacturers should adopt biodiversity-aware sourcing and reformulate away from high-impact ingredients, while policymakers should embed biodiversity metrics into sustainability standards, public procurement, and labelling schemes for alternative-protein products. The message for consumers is nuanced but clear: the plant-based burger is not automatically the biodiversity-friendly choice, and the geography of the supply chain matters as much as the patty on the plate.

Subject of Research: Terrestrial biodiversity impacts of replacing beef patties with alternative-protein burgers across ten EU countries

Article Title: Terrestrial biodiversity impacts of replacing beef with alternative-protein burger patties across ten European Union Countries

Article References: Zhang, Y., Rasul, K., Dorber, M., Stadler, K., Hertwich, E. G., & Verones, F. (2026). Terrestrial biodiversity impacts of replacing beef with alternative-protein burger patties across ten European Union Countries. Journal of Industrial Ecology, 30(4), 1761-1777. https://doi.org/10.1007/s44498-026-00120-1

Image Credits: AI Generated

DOI: 10.1007/s44498-026-00120-1

Keywords: biodiversity, alternative protein, beef, plant-based burgers, life cycle assessment, land use, coconut oil, soy protein, mycoprotein, supply chains, European Union, food system

The research team set out to answer a deceptively simple question: how are quantitative modelling approaches being used to support One Health in the context of agri-food systems, and how could they be better organized within governmental institutions? The question matters because the food system sits at the center of some of the most pressing health challenges of the century, from zoonotic disease emergence and antimicrobial resistance to foodborne illness and the nutritional consequences of environmental change. Public health authorities, the authors argue, cannot prioritize interventions or evaluate trade-offs across these domains without quantitative models that capture the dynamics of the animal-human-environment interface. Intuition and siloed expertise, however sophisticated, are not enough when a single policy decision ripples simultaneously through livestock production, ecosystems, and human diets.

To establish what governments are actually doing, the researchers first defined an operational scope of One Health by examining the One Health Joint Plan of Action, the strategic framework developed by the quadripartite alliance of the World Health Organization, the Food and Agriculture Organization, the World Organization on Animal Health, and the United Nations Environmental Programme. They then cross-referenced the plan’s proposed food-system activities against the publicly available One Health resources of the Group of 20 countries, the economies with the greatest institutional capacity to act. The result was sobering. The activities that governments have actually implemented cluster tightly around coordination, interoperability, and data sharing, and they concentrate on a short list of established threats: antimicrobial resistance, zoonoses, and vector-borne diseases. Broader food-system concerns, including nutrition, environmental sustainability, and the structural dynamics of agricultural production, remain largely outside the institutional One Health agenda.

That institutional narrowness, however, contrasts sharply with what is happening in the research literature. The team conducted a PRISMA-compliant scoping review of One Health quantitative modelling with relevance to the food system, covering publications from 2021 to 2026. The search yielded 129 relevant articles, a body of work that reveals how far methodological ambition has outrun governmental practice. Beyond the epidemiological data that one might expect, these studies drew on agricultural, geospatial, and ecological data streams to address food-system questions. Researchers were combining crop and livestock statistics with satellite-derived land cover information, wildlife distribution data, and environmental indicators, weaving together evidence from domains that traditionally have never shared a analytical framework, let alone a database.

The methodological landscape that emerged from the review is strikingly heterogeneous, and the authors identify three broad strategies that dominate the field. The first involves models that ingest heterogeneous data, fusing inputs of different types and resolutions into a single analytical structure. The second relies on mixed-effects approaches, statistical designs that explicitly account for variation at multiple levels, whether that means differences between farms, regions, or ecological zones, and can therefore represent the nested structure of real food systems rather than flattening it. The third and perhaps most consequential strategy is model integration: coupling separate models, each specialized for one domain, into combined frameworks capable of scenario analysis. A disease transmission model might be linked to an agricultural production model and an environmental impact model, allowing analysts to ask what happens to human health, animal welfare, and ecosystems under different policy futures simultaneously.

This is where the technical heart of the paper lies, and it is worth pausing on why integrated modelling is so difficult and so valuable. Each component model carries its own assumptions, spatial and temporal scales, and data requirements. An epidemiological model of pathogen spread may operate on daily time steps across networks of farms and markets, while an ecological model of land-use change may work in yearly increments across entire watersheds. Reconciling these scales, propagating uncertainty across model boundaries, and ensuring that outputs from one model are meaningful inputs to another are nontrivial engineering and statistical challenges. The fact that 129 recent studies attempt this integration at all signals a maturing field, but the review also makes clear that the approaches remain fragmented, with little standardization in how interdisciplinary collaboration is structured or how data infrastructures support it.

The conceptual contribution of the paper is a modular framework designed to operationalize modelling within One Health initiatives that involve the food system. Rather than prescribing a single monolithic model, the framework treats modelling as a set of interoperable modules that institutions can assemble according to their needs, connecting data acquisition, domain-specific models, and scenario evaluation in a flexible architecture. The design is explicitly aimed at supporting interdisciplinary collaboration between food-systems research and One Health practice, providing a common structure through which epidemiologists, agricultural scientists, ecologists, and data specialists can contribute components without needing to master one another’s entire methodological repertoire. For governmental institutions, the appeal is practical: a modular approach lowers the barrier to entry and allows capacity to be built incrementally rather than requiring an impossible upfront investment in a fully integrated system.

Underpinning all of this is a data infrastructure problem that the authors place at the center of their conclusions. The review found that genomic, geospatial, and agricultural data are the raw material of interdisciplinary One Health modelling, and that the availability and interoperability of these data streams largely determines what models can achieve. Genomic surveillance data allow pathogens to be traced across species boundaries and supply chains, revealing transmission pathways that would otherwise remain invisible. Geospatial data connect disease events to environmental drivers such as land conversion, water systems, and climate variables. Agricultural data anchor the models in the economic and biological realities of food production. Yet these data types are held by different ministries, agencies, and research communities, often in incompatible formats and under different governance regimes. The coordination and data-sharing activities that dominate current G20 One Health initiatives are, in this light, a necessary foundation, but the review suggests they must extend beyond the narrow disease-focused remit to support genuinely food-system-wide modelling.

The implications reach well beyond academic methodology. As climate change alters agricultural zones, as demand for animal protein reshapes livestock systems, and as antimicrobial resistance erodes the foundations of modern medicine, the trade-offs facing policymakers will become increasingly difficult to evaluate with single-domain analysis. A policy that reduces antimicrobial use in livestock may have consequences for animal welfare, farm economics, food prices, and ultimately human nutrition. A land-use change that boosts crop yields may expand habitat overlap between wildlife and humans, raising zoonotic spillover risk. Only quantitative models that span these domains can make such trade-offs explicit, and only institutions equipped with appropriate data infrastructure and collaborative frameworks can deploy those models where decisions are actually made. The gap between the 129 studies of the research frontier and the narrow institutional initiatives documented in the G20 review is therefore not merely an academic observation; it is a measure of how much analytical capacity remains untapped in exactly the places where it is most needed.

The study, funded by the German Federal Ministry of Research, Technology and Space under the WeAreOne project on public health synergies in the Anthropocene, arrives at a moment when One Health is being institutionalized worldwide, with national plans, dedicated agencies, and international frameworks proliferating. Its central message is a constructive tension: One Health is holistic in principle but narrow in practice, and quantitative modelling is both a symptom of that narrowness and a potential cure. By mapping the methodological state of the art, documenting the institutional shortfall, and proposing a modular framework grounded in genomic, geospatial, and agricultural data infrastructure, Burke and Risius have provided a concrete agenda for turning a unifying philosophy into an operational science. Whether governments seize that agenda will determine whether One Health remains a compelling slogan or becomes the analytical engine that the food systems of the twenty-first century urgently require.

Subject of Research: Interdisciplinary quantitative modelling for One Health applications in agri-food systems

Article Title: Interdisciplinary modelling approaches for One Health in agri-food systems: a scoping review and conceptual analysis

Article References: Burke, T. A., & Risius, A. (2026). Interdisciplinary modelling approaches for One Health in agri-food systems: a scoping review and conceptual analysis. BMC Public Health, 26(1), Article 2760. https://doi.org/10.1186/s12889-026-29666-2

Image Credits: AI Generated

DOI: 10.1186/s12889-026-29666-2

Keywords: One Health, agri-food systems, quantitative modelling, scoping review, antimicrobial resistance, zoonoses, food system, public health, geospatial data, genomic surveillance, interdisciplinary collaboration, G20