Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

# Options for keeping the food system within environmental limits

## Abstract

The food system is a major driver of climate change, changes in land use, depletion of freshwater resources, and pollution of aquatic and terrestrial ecosystems through excessive nitrogen and phosphorus inputs. Here we show that between 2010 and 2050, as a result of expected changes in population and income levels, the environmental effects of the food system could increase by 50–90% in the absence of technological changes and dedicated mitigation measures, reaching levels that are beyond the planetary boundaries that define a safe operating space for humanity. We analyse several options for reducing the environmental effects of the food system, including dietary changes towards healthier, more plant-based diets, improvements in technologies and management, and reductions in food loss and waste. We find that no single measure is enough to keep these effects within all planetary boundaries simultaneously, and that a synergistic combination of measures will be needed to sufficiently mitigate the projected increase in environmental pressures.

## Main

The global food system is a major driver of climate change1,2, land-use change and biodiversity loss3,4, depletion of freshwater resources5,6, and pollution of aquatic and terrestrial ecosystems through nitrogen and phosphorus run-off from fertilizer and manure application7,8,9. It has contributed to the crossing of several of the proposed ‘planetary boundaries’ that attempt to define a safe operating space for humanity on a stable Earth system10,11,12, in particular those concerning climate change, biosphere integrity, and biogeochemical flows related to nitrogen and phosphorous cycles. If socioeconomic changes towards Western consumption patterns continue, the environmental pressures of the food system are likely to intensify13,14,15,16, and humanity might soon approach the planetary boundaries for global freshwater use, change in land use, and ocean acidification11,12,17. Beyond those boundaries, ecosystems could be at risk of being destabilized and losing the regulation functions on which populations depend11,12.

Here we analyse the option space available for the food system to reduce its environmental impacts and stay within the planetary boundaries related to food production. We build on existing analyses that have advanced the planetary-boundary framework in terms of systemic threats to large-scale ecosystems11,12,18,19,20, discussed the role of agriculture with respect to those pressures10,21, and analysed the impacts on individual environmental domains22,23, including selected measures to alleviate those impacts22,23,24. The planetary-boundary framework is not without criticism, particularly because of the heterogeneity of the different boundaries and their underlying scientific bases, including the difficulty of defining global ecosystem thresholds for local environmental impacts25,26,27. Despite these limitations, we consider the planetary-boundary framework to be useful for framing, in broad terms, the planetary option space that preserves the sustainability of key ecosystems. We acknowledge the ongoing debate by quantifying the planetary boundaries of the food system in terms of broad ranges that reflect methodological uncertainties (see Methods), and by reporting the environmental impacts in absolute terms (for example, emissions in tonnes of carbon dioxide equivalents), which allows for comparisons to other measures of environmental sustainability.

We advance the present state of knowledge by constructing and calibrating a global food-systems model with country-level detail that resolves the major food-related environmental impacts and includes a comprehensive treatment of measures for reducing these impacts (see Methods). The regional detail of the model accounts for different production methods and environmental impacts that are linked by imports and exports of primary, intermediate and final products. We use the food-system model and estimates of present and future food demand to quantify food-related environmental impacts at the country and crop level in 2010 and 2050 for five environmental domains and the related planetary boundaries: greenhouse-gas (GHG) emission related to climate change; cropland use related to land-system change; freshwater use of surface and groundwater; and nitrogen and phosphorus application related to biogeochemical flows.

To characterize pathways towards a food system with lower environmental impacts that stays within planetary boundaries, we connect a region-specific analysis of the food system to a detailed analysis of measures of change, including reductions in food loss and waste, technological and management-related improvements, and dietary changes towards healthier, more plant-based diets (Extended Data Table 1). The scenarios regarding food loss and waste align with and exceed commitments made as part of the United Nations’ Sustainable Development Goals28,29,30. The scenarios concerning technological change account for future improvements in agricultural yields and fertilizer application, increases in feed efficiency, and changes in management practices31,32,33,34. Finally, the scenarios around dietary change include changes towards dietary guidelines and more plant-based dietary patterns that are in line with present evidence on healthy eating35,36,37.

In our baseline trajectory, we account for different socioeconomic pathways of population and income growth33, and project future demand for environmental resources in the absence of technological changes and dedicated mitigation measures. Although some of the measures of change considered here can be expected to be implemented by 2050, their level of ambition is uncertain and implementation will not happen automatically. We therefore analyse each measure of change explicitly and differentiate between two degrees of implementation: medium and high ambition. Measures of medium ambition are in line with stated intentions (for example, reducing food loss and waste by half), and measures of high ambition go beyond expectations but can be considered attainable with large-scale adoption of existing best practices (for example, reducing food loss and waste by 75%).

## Environmental impacts of the food system

Our analysis indicates that current and projected levels of agricultural production, in the absence of targeted mitigation measures, will greatly affect the Earth’s environment. We estimate that, in 2010, the food system emitted roughly the equivalent of 5.2 billion tonnes of carbon dioxide in GHG emissions in the form of methane and nitrous oxide; the food system also occupied 12.6 million km2 of cropland, used 1,810 km3 of freshwater resources from surface and groundwater (bluewater), and applied 104 teragrams of nitrogen (TgN) and 18 teragrams of phosphorus (TgP) in the form of fertilizers (see Methods, ‘Data availability’). Our estimates are comparable to previous estimates of food-related GHG emissions1,38 of 4.6–5.8 billion tonnes of carbon dioxide equivalents, global cropland use39 of 12.2–17.1 million km2 in 2000, bluewater use5,20 in 2000 of 1,700–2,270 km3, and nitrogen40 and phosphorus40,41 application in 2010 of 104 TgN and 15.8–18.8 TgP.

Food production and consumption are projected to change between 2010 and 2050 (Extended Data Table 2) as a result of expected socioeconomic developments (Supplementary Table 1). Those developments include the growth of the global population by about a third (with a range of 23–45%, from 6.9 billion in 2010 to 8.5–10 billion in 2050) and a tripling of global income (with a range of 2.6–4.2, from US$68 trillion in 2010 to US$180–290 trillion in 2050)33. Because of these changes, we predict the environmental pressures of the food system to increase by 50–92% for each indicator in the absence of technological change and other mitigation measures (Fig. 1). The greatest increases along this baseline pathway are projected for GHG emissions (87%, range 80–92%), then for the demand for cropland use (67%, range 66–68%), bluewater use (65%, range 64–65%), phosphorus application (54%, range 51–55%) and nitrogen application (51%, range 50–52%).

Specific food groups vary in their environmental impacts (Fig. 1). The production of animal products generates the majority of food-related GHG emissions (72–78% of total agricultural emissions), which is due to low feed-conversion efficiencies, enteric fermentation in ruminants, and manure-related emissions42; the feed-related impacts of animal products also contribute to bluewater use (around 10%) and pressures on cropland, as well as nitrogen and phosphorus application (20–25% each). By comparison, staple crops have generally lower environmental footprints (impacts per kg of product) than animal products (Extended Data Table 3), in particular for GHG emissions, but they can have high total impacts because of their higher production volumes (Extended Data Table 2). According to our estimates, staple crops grown for human consumption are responsible for a third to a half (30–50%) of cropland use, bluewater use, and nitrogen and phosphorus application. The projected population growth between 2010 and 2050 contributes to a general increase in the impacts of each food group, and the projected income growth changes the relative contribution of each, with a shift towards a larger proportion of impacts from animal products (7–16% increase across environmental domains) and fruits and vegetables (2–28% increase), and a smaller proportion from staple crops (7–19% reduction).

## Changes in food management, technology and diets

Reducing food loss and waste is one measure for reducing food demand and the associated environmental impacts. At present it is estimated that more than a third of all food that is produced is lost before it reaches the market, or is wasted by households28. For our analysis, we evaluated the impacts of reducing food loss and waste to one half—a value in line with pledges made as part of the Sustainable Development Goals29—and we also considered a reduction in food loss and waste by 75%, which is probably close to the maximum theoretically avoidable value30. We estimate that halving food loss and waste would reduce environmental pressures by 6–16% compared with the baseline projection for 2050, and that reducing food loss and waste by 75% would reduce environmental pressures by 9–24% (Fig. 2). Relatively more staple crops and fruits and vegetables are wasted than animal products28, which explains why the impacts of changes in food loss and waste are smaller for the livestock-dominated domains, such as GHG emissions, than for the staple-crop-dominated ones, such as cropland and bluewater use and nitrogen and phosphorus application.

Technological changes increase the efficiency of production and reduce the environmental impact per unit of food produced. We analysed the most commonly considered technological advances and changes in management practices with respect to their environmental impacts (Extended Data Table 1). The measures include: increases in agricultural yields, which reduce the demand for additional cropland32,33; rebalancing of fertilizer application between overapplying and underapplying regions32, as well as increasing nitrogen-use efficiency34,43 and phosphorus recycling7, which reduce demand for additional nitrogen and phosphorus inputs; improvements in water management that increase basin efficiency, storage capacity, and better utilization of rainwater33; and agricultural mitigation options, including changes in irrigation, cropping and fertilization that reduce methane and nitrous oxide emissions from rice and other crops, and changes in manure management, feed conversion and feed additives that reduce enteric fermentation in livestock31. We estimate that implementing these measures could reduce the environmental pressures of the food system by 3–30% compared with the 2050 baseline projection in medium-ambition scenarios, and by 11–54% in high-ambition scenarios (Fig. 2). In each case, the higher-end estimates are for the staple-crop-dominated environmental indicators (cropland and bluewater use, and nitrogen and phosphorus application), for which general improvements in water management, agricultural yields, phosphorus-recycling rates and nitrogen-use efficiencies are particularly effective. The lower-end estimates are for GHG emissions, for which the contribution from livestock-related emissions is, to a large extent, an inherent characteristic of the animals and therefore cannot be reduced more substantially through existing mitigation options31,44 (Extended Data Table 4).

Dietary changes towards healthier diets can reduce the environmental impacts of the food system when environmentally intensive foods, in particular animal products, are replaced by less intensive food types15,16. For our analysis, we analysed dietary changes towards diets in line with global dietary guidelines for the consumption of red meat, sugar, fruits and vegetables, and total energy intake35,36; as well as to more plant-based (flexitarian) diets that more comprehensively reflect the current evidence on healthy eating37,45 by including lower amounts of red and other meats and greater amounts of fruits, vegetables, nuts and legumes (Extended Data Tables 1 and 5). We estimate that, compared with the baseline projection for 2050, dietary changes towards healthier diets could reduce GHG emissions and other environmental impacts by 29% and 5–9%, respectively, for the dietary-guidelines scenario, and by 56% and 6–22%, respectively, for the more plant-based diet scenario (Fig. 2). The changes are in line with the dietary composition of the diets and the environmental footprints of each food group (Fig. 1, Extended Data Table 1 and Supplementary Table 2). Changes in meat consumption dominate the impacts on GHG emissions, while for the other domains the environmental pressures associated with greater consumption of fruits, vegetables, nuts and legumes are more important but outweighed by the environmental benefits associated with lower consumption of meat, staple crops and sugar, and a generally lower energy intake in line with healthy body weights and recommended levels of physical activity35 (Extended Data Table 6).

To understand how the combined implementation of some or all of the discussed measures could influence the environmental pressures of the food system, we constructed an environmental option space by combining all measures of medium ambition and all measures of high ambition. Our analysis indicates that much of the increase in environmental pressures that is expected to occur by 2050 could be mitigated if measures were combined (Fig. 2). Combining all measures of medium ambition could reduce environmental pressures by around 25–45% compared with the baseline projection for 2050, resulting in total environmental impacts that are within 15% above and below present impacts. Combining all measures of high ambition could deliver reductions of 30–60%, resulting in environmental impacts that are 20–55% less than the current ones. In line with the differentiated impacts of the different measures of change, dietary change contributes the most to the reductions in GHG emissions, and technological and management-related changes contribute the most to reductions in the other environmental impacts, while reductions in food loss and waste contribute up to a third to the overall reductions (Extended Data Fig. 1).

## Planetary option space

What level of reduction in environmental pressures should be aimed for? We can explore this question through comparison to the associated planetary boundaries that are intended to describe a safe operating space for humanity. For our analysis, we adapted or newly quantified the food-related planetary-boundary values, including upper and lower limits (Extended Data Table 7, Extended Data Fig. 2 and Methods). According to our quantification, the planetary boundaries define a space around the present values for most environmental domains, with a mean value slightly below present values for food-related GHG emissions, at current values for cropland use, slightly above present values for bluewater use, and substantially below present values for nitrogen and phosphorus application (Fig. 2). Following the baseline trajectory of population and income change, and the related changes in food consumption and production, would lead to all mean values of the planetary boundaries being crossed. The environmental impacts of the food system would exceed the planetary boundaries for food-related GHG emissions by 110%, for cropland use by 70%, for bluewater use by 50%, for nitrogen application by 125%, and for phosphorus application by 75%.

Our analysis indicates that staying within planetary boundaries is possible with a combination of measures of high ambition for GHG emissions and nitrogen and phosphorus application, and with a combination of measures of medium ambition for cropland and bluewater use (Fig. 2). An analysis of the planetary option space details the possible combination of measures (Fig. 3). It shows that staying within the mean value of the GHG boundary requires ambitious dietary change towards more plant-based, flexitarian diets, in combination with either reductions in food loss and waste or technological improvements; staying within the mean values of the cropland and bluewater boundaries requires technological improvements in combination with reductions in food loss and waste; and staying within the mean values of the nitrogen and phosphorus boundaries requires ambitious technological improvements combined (for the nitrogen boundary) with dietary changes towards more plant-based diets, reductions in food loss and waste, and, in some combinations, a more optimistic socioeconomic development pathway that includes lower population and higher income growth than is expected at present. Combining those measures synergistically results in adoption of different measures of technological change for each environmental domain, coupled in each case to dietary changes towards more plant-based diets, reductions in food loss and waste, and an optimistic socioeconomic development pathway (Fig. 4).

## Uncertainties

Our estimates are subject to several uncertainties. Some of the planetary-boundary values have a large uncertainty range, which reflects the difficulties of scaling up local environmental pressures to global levels12,20, in particular regarding bluewater use and nitrogen and phosphorus application (see Methods). The planetary-boundary framework can therefore provide only a very broad measure of the sustainability of the food system. Our analysis indicates that using the upper bound of the planetary-boundary range increases the option space (Fig. 3) and, for example, does not require reductions in food loss and waste or a more optimistic socioeconomic development pathway; however, meeting the lower bound of the planetary-boundary range would not be possible for bluewater use and nitrogen application with the mitigation options considered here. Using different control variables to measure the state of planetary boundaries could also affect the option space. However, assessing the impacts of nitrogen pollution by using a measure of nitrogen surplus that accounts for all inputs and offtakes of nitrogen had little influence on the option space (Extended Data Fig. 3).

Other uncertainties are related to the set-up of our modelling framework. Although we did consider some feedback effects between the different measures of change—particularly between changes in yields and the demand for bluewater, nitrogen and phosphorus use—this was limited to the scenarios of medium ambition (see Methods). This method allowed for the differentiated adoption of ambitious technological change for domains other than cropland use without also requiring such levels for the latter. In a sensitivity analysis, we assessed the feedback effects that very high yield increases could have on nitrogen and phosphorus application32, and found that the demand for nitrogen and phosphorus could increase across the different scenario combinations with large yield-gap closures by 8–14% and 25–32%, respectively, which would moderately reduce the planetary option space for those scenarios (Extended Data Fig. 3). In line with our focus on mitigation measures, we did not assess the impacts that climate change could have on crop yields and freshwater availability46. While economic responses might be able to mitigate some proportion of the biophysical impacts of climate change47, such responses could reduce the availability and effectiveness of additional mitigation and adaptation measures, and thereby reduce the planetary option space.

Additional research would reduce the uncertainty of our scenario analysis. In our scenarios of change, we chose to focus on changes—technological, dietary, and in food loss and waste—that are considered realistic or attainable, or have been set as goals. This means that we did not include technologies or mitigation measures that have large uncertainties at present, such as soil carbon sequestration, nitrogen-fixing cereals, or landless biomass production. Some of those measures have shown some prospect in certain regions, but it is not yet clear whether they are scalable and what their relationship to existing technologies and environmental targets would be48. For example, land-based carbon sequestration, while reducing GHG emissions, could put additional pressures on croplands or pastures, with implications for land-use and biodiversity targets. Other areas for further research include the quantification of co-benefits of food-system change, for example, on health15,49, biodiversity50, and the economy47, as well as context-specific metrics of sustainability and a greater focus on livelihood, for example in terms of food security51.

## Policy implications

Our analysis suggests that staying within the planetary boundaries of the food system requires a combination of measures: GHG emissions cannot be sufficiently mitigated without dietary changes towards more plant-based diets; cropland and bluewater use are best addressed by improvements in technologies and management that close yield gaps and increase water-use efficiency; and reducing nitrogen and phosphorus application will require a combination of measures to stay below the mean values of the planetary boundaries, including dietary change, reductions in food loss and waste, improvements in technologies and management that increase use efficiencies for nitrogen and recycling rates for phosphorus, and efforts in global socioeconomic development.

Implementation of these measures will depend on the regulatory and incentive framework in each region. In particular, practical options exist for improving technologies and management practices (Extended Data Table 1), but adoption of those options will require investment in public infrastructure, the right incentive schemes for farmers (including support mechanisms to adopt best available practices), and better regulation (for example, of water use and quality). Concrete options also exist for improving socioeconomic development in developing countries, including investments in education, particularly for women, and improving access to general and reproductive health services52. Meaningfully reducing food loss and waste will require measures across the entire food-supply chain30, with possible emphasis on investments in agricultural infrastructure, technological skills, storage, transport, and distribution in developing regions; and education and awareness campaigns, food labelling, improved packaging that prolongs shelf life, and changes in legislation and business behaviour that promote closed-loop supply chains (in which waste is recycled back into the system) in developed areas. For dietary change, the available evidence suggests that providing information without additional economic or environmental changes has a limited influence on behaviour, and that integrated, multicomponent approaches that include clear policy measures might be best suited for changing diets53,54. Those can include a combination of media and education campaigns; labelling and consumer information; fiscal measures, such as taxation, subsidies, and other economic incentives; school and workplace approaches; local environmental changes; and direct restriction and mandates54. An important first step would be to align national food-based dietary guidelines with the present evidence on healthy eating and the environmental impacts of diets55,56.

Our analysis suggests that the environmental impacts of the food system could increase markedly owing to expected changes in food consumption and production, and, in the absence of targeted measures, would exceed planetary boundaries to the extent that key ecosystem processes could become at risk of being destabilized. Synergistically combining improvements in technologies and management, reductions in food loss and waste, and dietary changes towards healthier, more plant-based diets, with particular attention to local contexts and environmental pressures, will be a key challenge in defining region-specific pathways for the sustainable development of food systems within the planetary option space. We hope that the country-specific data and suite of scenarios produced for this study (see Methods, ‘Data availability’) can provide a good starting point for this endeavour.

## Methods

### Food-system model

For our analysis, we constructed a food-systems model that connects food consumption and production across regions (Supplementary Information). We distinguished several steps along the food chain: primary production (including non-food uses, for example, in industry, seed banks, and as biofuels); trade in primary commodities; processing to oils, oil cakes and refined sugar; use of feed for animals; and trade in processed commodities and animals (Extended Data Table 2). We parameterized the model with data from the International Model for Policy Analysis of Agricultural Commodities and Trade (IMPACT)33 on current and future food production, processing factors, and feed requirements for 62 agricultural commodities and 159 countries. Projections of future food consumption and production were based on statistical association with changes in income and population, and were in line with other projections57.

To assess the environmental impacts of the food system, we paired the food-system model with a set of country-specific environmental footprints related to GHG emissions, cropland use, bluewater use, and nitrogen and phosphorus application (Extended Data Table 3; data available upon request). In line with projections of the allowable agricultural emissions budget58, and our separate treatment of land use, we focused on the non-CO2 emissions of agriculture, in particular methane and nitrous oxide. Data on GHG emissions were adopted from country-specific analyses of GHG emissions from crops59 and livestock38. Non-CO2 emissions of fish and seafood were calculated on the basis of feed requirements and feed-related emissions of aquaculture60, and on projections of the ratio between wild-caught and farmed fish production61,62. Our baseline emissions estimate agrees well with existing ones that follow the same methodology1,63.

Data on cropland and consumptive bluewater use were adopted from the IMPACT model33. To derive commodity-specific footprints, we divided use data by data on primary production, and we calculated the footprints of processed goods (vegetable oils, refined sugar) by using country-specific conversion ratios33, and splitting co-products (oils and oil meals) by economic value to avoid double counting. We used country-specific feed requirements for terrestrial animals33 to derive the cropland and bluewater footprints for meat and dairy, and we used global feed requirements for aquaculture60 and projections of the ratio between wild-caught and farmed fish production61,62 to derive the cropland and bluewater footprints for fish and seafood.

Data on fertilizer application rates of nitrogen and phosphorous were adopted from the International Fertilizer Industry Association40. In line with the planetary boundaries, we focus on application rates as the control variables in our main analysis. However, we note that regional environmental impacts often depend on the surplus of reactive nitrogen, a measure that accounts for all inputs and offtakes of nitrogen64. For a sensitivity analysis, we therefore constructed a region-specific nitrogen-budget module and linked it to the food-system model. Therein, we define the nitrogen surplus as the sum of fertilizer use, fixation by crops, manure application, human excreta and atmospheric deposition, minus nitrogen offtake by crops22,43,65 (Supplementary Information). The results of the sensitivity analysis are reported in Extended Data Fig. 3.

### Scenario analysis

We used the food-system model to estimate the environmental impacts of the food system in 2050 on GHG emissions, cropland use, bluewater use, and nitrogen and phosphorus application. To estimate the environmental impacts in the absence of dedicated mitigation measures (a scenario we term ‘baseline projection’), we paired the footprints of current intensity to future projections of food demand along several socioeconomic pathways that were developed by the climate-change research community (Supplementary Table 1), including a middle-of-the-road development pathway (SSP2), a more optimistic pathway with higher income and lower population growth (SSP1), and a more pessimistic pathway with lower income and greater population growth (SSP3)66,67,68. Underlying the pathways are data and projections of the age, sex and educational structure of populations, as well as age-specific fertility, mortality and migration67.

We then analysed the option space for reducing the environmental pressures of the food system by constructing scenarios of changes in food loss and waste, technological change and dietary change (Extended Data Table 1). For each measure, we differentiated between changes of medium and high ambition. Estimates of food loss and waste were based on percentage values reported by the UN Food and Agriculture Organization (FAO)28. In the standard scenario (waste/2), we assumed that food losses at the production side and food waste at the consumption side are reduced by half—a goal in line with the UN Sustainable Development Goals for 2030. In the ambitious scenarios (waste/4), we assumed reductions in food loss and waste of 75%, which is probably close to the maximum value that can be theoretically avoided30.

The scenarios of technological change (tech and tech+) include projected efficiency gains in emissions intensities, agricultural yields, feed conversion, water use, and nitrogen and phosphorus application (Extended Data Table 4). For the scenarios describing changes in emissions intensities of foods, we incorporated the mitigation potential of bottom-up changes in management practices and technologies by using marginal abatement cost curves31 and the value of the social cost of carbon (SCC) in 205069. The mitigation options included changes in irrigation, cropping and fertilization that reduce methane and nitrous oxide emissions for rice and other crops, as well as changes in manure management, feed conversion and feed additives that reduce enteric fermentation in livestock. We used SCC values of 72 US dollars per metric ton of CO2 equivalents (US$/tCO2 equivalents) associated with a rate of discounting future climate damages by 3% for the scenario of medium ambition (tech), and implemented all available mitigation options (equivalent to using a SCC of above 99 US$/tCO2 equivalents) for the scenario of high ambition (tech+). No marginal abatement curves were available for some crops, such as fruits, vegetables, nuts, sugar crops and oilseeds. Adopting the average mitigation potential for staple crops for these crops would increase the total mitigation potential by 1%.

Efficiency gains in agricultural yields, water management and feed conversion were based on IMPACT projections33. For water management, we relied on an integrated hydrological model within IMPACT that operates at the level of watersheds and accounts for management changes that increase basin efficiency, storage capacity and better utilization of rainwater33. For most crops, improvements in water management exceed increased water demand associated with yield improvements, except for soybeans. For agricultural yields, the gains in land-use efficiency matched estimates of yield-gap closures of about 75% between present yields and yields that are feasible in a given agricultural-climatic zone32. The potential efficiency gains in nitrogen and phosphorus application rates included rebalancing of fertilizer application rates between overapplying and underapplying regions in line with closing yield gaps32. In the ambitious technology scenario (tech+), we increased yield-gap closures to 90% on the basis of data from a previous study32, and assumed additional improvements in nitrogen-use efficiency of 30% (in line with targets suggested by the Global Nitrogen Assessment34) and a recycling rate of phosphorus7 of 50%. No further changes in efficiency were assumed for water use in the tech+ scenario. For most crops, land-use efficiencies increase in the ambitious technology scenario, except in the case of soybeans, which are assessed on a more conservative basis in a previous study32 than by the IMPACT team.

The scenarios of dietary change include shifts towards diets that are in line with global dietary guidelines (guidelines), and towards dietary patterns that are more specialized but nutritionally balanced (flexitarian). For the former, we followed suggestions to limit the intake of red meat to less than 300 g per week70 and the intake of added sugar to less than 5% of total energy intake (about 31 g per day)71, to consume five portions (400 grams per day) or more of fruits and vegetables36, and to balance energy intake (and physical activity levels) to maintain a healthy body weight35. Estimates of energy intake were based on the calorie needs of a moderately active population of US characteristics for height, divided into five-year age groups72—something that can be seen as an upper bound. Calorie needs reach a maximum of 2,500 kcal per day for ages 19–25 (averaged between men and women), but are reduced to 2,000 kcal per day for ages 66 and older. The average calorie needs differed by region according to its age composition, and averaged around 2,100 kcal per day. In a sensitivity analysis, we implemented changes in dietary composition only, without restricting energy intake. Baseline intakes of food and energy were calculated from food-availability projections of the IMPACT model by using region-specific factors of food waste and ratios of the edible portions of foods28.

In scenarios of ambitious dietary change, we increased the stringency of the global recommendations and defined more plant-based (flexitarian) dietary patterns that reflect current evidence on healthy eating37,46,73 (Extended Data Table 5 and Supplementary Table 2). The flexitarian diets included: at least 500 g per day of fruits and vegetables of different colours and groups (the composition of which is determined by regional preferences); at least 100 g per day of plant-based protein sources (legumes, soybeans and nuts); modest amounts of animal-based proteins, such as poultry, fish, milk and eggs; and limited amounts of red meat (one portion per week), refined sugar (less than 5% of total energy), vegetable oils that are high in saturated fat (in particular palm oil) and starchy foods with a relatively high glycaemic index. We aimed to preserve the regional character of dietary patterns by maintaining the regional composition of specific foods within broader categories, such as preferences for specific staple crops (wheat, maize, rice and so on) and fruits (temperate or tropical).

### Planetary boundaries

The planetary-boundary framework attempts to define a safe operating space for humanity that is characterized by a stable Earth system10,11,12. Above planetary boundaries, it is suggested that ecosystem processes are at risk of becoming destabilized11,12. To contextualize the environmental impacts of the food system, we critically reviewed and adapted planetary-boundary values for GHG emissions, cropland use, bluewater use, and nitrogen and phosphorus application (Extended Data Table 7). For the climate-change boundary, we adopted an emissions budget for food-related (non-CO2) GHG emissions that is in line with having a 66% chance of limiting global warming to below 2 °C (Representative Concentration Pathway (RCP)2.6); we derived this budget from a model comparison of three integrated assessment models58, normalized to the marker scenario of the associated emissions pathway63. The resulting budget of 4.7 GtCO2 equivalents (range 4.3–5.3 GtCO2 equivalents). focuses on the non-CO2 emissions related to agriculture (methane and nitrous oxide), in line with previous assessments58 and methodology followed by the International Panel on Climate Change. However, we note that agriculture and land use also act as source and sink for CO2, for example through deforestation and carbon sequestration in soils74. How those flows should be balanced vis-à-vis the emissions from other sectors, and how additional pressures from land-based CO2 sequestration contribute or counteract other sustainability targets and planetary boundaries, are important questions for future research.

Large uncertainties exist as to what an appropriate planetary boundary for land use should be12. From an analysis of forest biomes, a boundary value12 was previously suggested in line with maintaining (not increasing pressure on) present forest cover. Such a target is in line with the strongly correlated target for biosphere integrity if nonagricultural land is placed under protection of biodiversity-compatible land use12,75,76. Because our modelling framework explicitly tracks cropland use, we translate the suggested target to a value of keeping current cropland use at 12.6 million km2 (range 10.6–14.6 million km2), given our own model calculations using the IMPACT model33. In future work, it will be desirable to include the role of pastures, an explicit treatment of forest cover, and further differentiation of other forms of land cover. However, a complication with switching from land use to forest cover is that the latter depends not only on agriculture, but also on wood harvesting, urbanization, and other socioeconomic variables. More than two-thirds of agricultural land is used for grazing. Converting highly productive grazing land into cropland could therefore be a conservation strategy that would relax the boundary value for cropland without affecting forest cover. However, estimates of feasible conversion ratios are still a matter of debate23.

Two basin-level assessments of the environmental flow requirements of river systems have been used to suggest planetary boundaries for the consumption of bluewater12,20. We adopt the more stringent values of the more detailed standalone analysis (2,800 km3; range 1,100–4,500 km3)20, which includes the other suggested values in its uncertainty range12,77. Because not all bluewater is used in agriculture, we scale from total consumptive bluewater use (2,550 km3)5 to the consumptive bluewater used in agriculture (1,810 km3) as assessed with our hydrological model33, which yields a boundary of 1,980 km3 (range 780–3,190 km3) of bluewater used in agriculture. We note that uncertainties persist about the concrete assumptions on environmental flow requirements12,78, and about which methodology would be best suited79.

To inform the boundary value for reactive nitrogen, a previous study19 calculated global risk values for eutrophication on the basis of region-specific estimates of current nitrogen concentration in run-off and concentrations that would stay below ecological and toxicological thresholds of inorganic nitrogen pollution. The original boundary value for nitrogen was calculated by multiplying the global risk value by an estimate of current anthropogenic nitrogen fixation (fertilizer use plus fixation by crops)19. Here we apply the risk values to nitrogen application from fertilizers—in line with the focus in the planetary-boundary literature on anthropogenic disruptions of ecosystems11,12—and we use the nitrogen surplus (the sum of fertilizer use, fixation by crops, manure application, human excreta and atmospheric deposition, minus nitrogen offtake by crops) as a control variable in a sensitivity analysis (Extended Data Fig. 3). The resulting estimate of 52–69 TgN per year (67–90 TgN when using nitrogen surplus as a control) might be considered conservative, because the previous study19 maintained regions that currently apply less than the critical load of nitrogen at that value, which in some cases can be much lower than needed from an environmental and food-security perspective80. For that reason, we adopted an upper boundary value in line with a scenario32 that balanced nitrogen application between overapplying and underapplying regions and closed yield gaps to 75%, which yielded a final boundary value of 69 TgN (range 52–113 TgN) of nitrogen application from fertilizers (90 TgN (range 67–146 TgN) of nitrogen surplus).

Unlike nitrogen, phosphorus can build up in the soil and is washed out as run-off during erosion7. Existing estimates of boundary values for phosphorus18 have several shortcomings in that they are based on constant erosion rates and do not take into account critical sources of phosphorus, such as human waste/excreta. In the previous study19 a global phosphorus-flow model was developed that focused on added phosphorus assuming steady-state surface pools, critical phosphorus concentrations of 50–100 mg per litre to prevent eutrophication, and flexible recycling rates (Extended Data Fig. 2 and Supplementary Information). Under no-waste recycling, the long-term phosphorus boundary amounted to 6–12 TgP per year, increasing to 8–16 TgP per year at a recycling rate of 50%. In line with our focus on scenarios of change, we adopted the latter values. As with nitrogen, there are great regional imbalances of phosphorus application81, so we again infer an upper tolerable value from a scenario32 that rebalanced phosphorus application between overapplying and underapplying regions and closed yield gaps to 75%. The resulting internally derived phosphorus boundary is 16 TgP (range 8–17 TgP) of phosphorus application.

### Reporting summary

Further information on research design is available in the Nature Research Reporting Summary linked to this paper.

## Data availability

The data that support the findings of this study are available from the Oxford University Research Archive (ORA; https://ora.ox.ac.uk) at https://ora.ox.ac.uk/objects/uuid:d9676f6b-abba-48fd-8d94-cc8c0dc546a2.

## References

1. 1.

Smith, P. et al. in Climate Change 2014: Mitigation of Climate Change (eds Edenhofer, O. et al.) 811–922 (Cambridge Univ. Press, 2014).

2. 2.

Vermeulen, S. J., Campbell, B. M. & Ingram, J. S. I. Climate change and food systems. Annu. Rev. Environ. Resour. 37, 195–222 (2012).

3. 3.

Foley, J. A. et al. Global consequences of land use. Science 309, 570–574 (2005).

4. 4.

Newbold, T. et al. Global effects of land use on local terrestrial biodiversity. Nature 520, 45–50 (2015).

5. 5.

Shiklomanov, I. A. & Rodda, J. C. World Water Resources at the Beginning of the Twenty-First Century (Cambridge Univ. Press, Cambridge, 2004).

6. 6.

Wada, Y. et al. Global depletion of groundwater resources. Geophys. Res. Lett. 37, L20402 (2010).

7. 7.

Cordell, D. & White, S. Life’s bottleneck: sustaining the world’s phosphorus for a food secure future. Annu. Rev. Environ. Resour. 39, 161–188 (2014).

8. 8.

Diaz, R. J. & Rosenberg, R. Spreading dead zones and consequences for marine ecosystems. Science 321, 926–929 (2008).

9. 9.

Robertson, G. P. & Vitousek, P. M. Nitrogen in agriculture: balancing the cost of an essential resource. Annu. Rev. Environ. Resour. 34, 97–125 (2009).

10. 10.

Campbell, B. et al. Agriculture production as a major driver of the Earth system exceeding planetary boundaries. Ecol. Soc. 22, 8 (2017).

11. 11.

Rockström, J. et al. A safe operating space for humanity. Nature 461, 472–475 (2009).

12. 12.

Steffen, W. et al. Planetary boundaries: guiding human development on a changing planet. Science 347, 1259855 (2015).

13. 13.

Davis, K. F. et al. Meeting future food demand with current agricultural resources. Glob. Environ. Change 39, 125–132 (2016).

14. 14.

Jalava, M., Kummu, M., Porkka, M., Siebert, S. & Varis, O. Diet change—a solution to reduce water use? Environ. Res. Lett. 9, 074016 (2014).

15. 15.

Springmann, M., Godfray, H. C. J., Rayner, M. & Scarborough, P. Analysis and valuation of the health and climate change cobenefits of dietary change. Proc. Natl Acad. Sci. USA 113, 4146–4151 (2016).

16. 16.

Tilman, D. & Clark, M. Global diets link environmental sustainability and human health. Nature 515, 518–522 (2014).

17. 17.

Hoekstra, A. Y. & Wiedmann, T. O. Humanity’s unsustainable environmental footprint. Science 344, 1114–1117 (2014).

18. 18.

Carpenter, S. R. & Bennett, E. M. Reconsideration of the planetary boundary for phosphorus. Environ. Res. Lett. 6, 014009 (2011).

19. 19.

de Vries, W., Kros, J., Kroeze, C. & Seitzinger, S. P. Assessing planetary and regional nitrogen boundaries related to food security and adverse environmental impacts. Curr. Opin. Environ. Sustain. 5, 392–402 (2013).

20. 20.

Gerten, D. et al. Towards a revised planetary boundary for consumptive freshwater use: role of environmental flow requirements. Curr. Opin. Environ. Sustain. 5, 551–558 (2013).

21. 21.

Gordon, L. J. et al. Rewiring food systems to enhance human health and biosphere stewardship. Environ. Res. Lett. 12, 100201 (2017).

22. 22.

Bodirsky, B. L. et al. Reactive nitrogen requirements to feed the world in 2050 and potential to mitigate nitrogen pollution. Nat. Commun. 5, 3858 (2014).

23. 23.

Erb, K.-H. et al. Exploring the biophysical option space for feeding the world without deforestation. Nat. Commun. 7, 11382 (2016).

24. 24.

Conijn, J. G., Bindraban, P. S., Schröder, J. J. & Jongschaap, R. E. E. Can our global food system meet food demand within planetary boundaries? Agric. Ecosyst. Environ. 251, 244–256 (2018).

25. 25.

Lewis, S. L. We must set planetary boundaries wisely. Nature 485, 417–418 (2012).

26. 26.

Montoya, J. M., Donohue, I. & Pimm, S. L. Planetary boundaries for biodiversity: implausible science, pernicious policies. Trends Ecol. Evol. 33, 71–73 (2018).

27. 27.

Schlesinger, W. H. Planetary boundaries: thresholds risk prolonged degradation. Nat. Rep. Clim. Change 3, 112–113 (2009).

28. 28.

Gustavsson, J., Cederberg, C., Sonesson, U., Van Otterdijk, R. & Meybeck, A. Global Food Losses and Food Waste: Extent, Causes and Prevention (FAO, 2011).

29. 29.

United Nations General Assembly. Resolution Adopted by the General Assembly on 25 September 2015. 70/1 Transforming Our World: the 2030 Agenda for Sustainable Development (United Nations, 2015).

30. 30.

Parfitt, J., Barthel, M. & Macnaughton, S. Food waste within food supply chains: quantification and potential for change to 2050. Phil. Trans. R. Soc. B 365, 3065–3081 (2010).

31. 31.

Beach, R. H. et al. Global mitigation potential and costs of reducing agricultural non-CO2 greenhouse gas emissions through 2030. J. Integr. Environ. Sci. 12, 87–105 (2015).

32. 32.

Mueller, N. D. et al. Closing yield gaps through nutrient and water management. Nature 490, 254–257 (2012); corrigendum 494, 390 (2013).

33. 33.

Robinson, S. et al. The International Model for Policy Analysis of Agricultural Commodities and Trade (IMPACT)—Model Description for Version 3. IFPRI Discussion Paper 1483 (IFPRI, 2015).

34. 34.

Sutton, M. A. et al. Our Nutrient World: The Challenge to Produce More Food and Energy with Less Pollution (NERC/Centre for Ecology and Hydrology, Edinburgh, UK, 2013).

35. 35.

World Health Organization. Human Energy Requirements. Report of a Joint FAO/WHO/UNU Expert Consultation, Rome, Italy, 17–24 October 2001 (World Health Organization, 2004).

36. 36.

World Health Organization. Diet, Nutrition and the Prevention of Chronic Diseases. Report of the Joint WHO/FAO Expert Consultation (World Health Organization, 2003).

37. 37.

Willett, W. C. & Stampfer, M. J. Current evidence on healthy eating. Annu. Rev. Public Health 34, 77–95 (2013).

38. 38.

Tubiello, F. N. et al. The FAOSTAT database of greenhouse gas emissions from agriculture. Environ. Res. Lett. 8, 015009 (2013).

39. 39.

Ramankutty, N., Evan, A. T., Monfreda, C. & Foley, J. A. Farming the planet: 1. Geographic distribution of global agricultural lands in the year 2000. Glob. Biogeochem. Cycles 22, GB1003 (2008).

40. 40.

Heffer, P., Gruère, A. & Roberts, T. Assessment of Fertilizer Use by Crop at the Global Level 2010–2010/1. Report A/17/134 rev (International Fertilizer Association and International Plant Nutrition Institute, 2013).

41. 41.

Zhang, J. et al. Spatiotemporal dynamics of soil phosphorus and crop uptake in global cropland during the 20th century. Biogeosciences 14, 2055–2068 (2017).

42. 42.

Gerber, P. J. et al. Tackling Climate Change through Livestock: a Global Assessment of Emissions and Mitigation Opportunities (Food and Agriculture Organization of the United Nations, 2013).

43. 43.

Mueller, N. D. et al. Declining spatial efficiency of global cropland nitrogen allocation. Glob. Biogeochem. Cycles 31, 245–257 (2017).

44. 44.

Herrero, M. et al. Greenhouse gas mitigation potentials in the livestock sector. Nat. Clim. Change 6, 452–461 (2016).

45. 45.

Katz, D. L. & Meller, S. Can we say what diet is best for health? Annu. Rev. Public Health 35, 83–103 (2014).

46. 46.

Rosenzweig, C. et al. Assessing agricultural risks of climate change in the 21st century in a global gridded crop model intercomparison. Proc. Natl Acad. Sci. USA 111, 3268–3273 (2014).

47. 47.

Nelson, G. C. et al. Climate change effects on agriculture: economic responses to biophysical shocks. Proc. Natl Acad. Sci. USA 111, 3274–3279 (2014).

48. 48.

Garnett, T. et al. Grazed and Confused? Ruminating on Cattle, Grazing Systems, Methane, Nitrous Oxide, the Soil Carbon Sequestration Question—and What it All Means for Greenhouse Gas Emissions (Food Climate Research Network, 2017).

49. 49.

Springmann, M. et al.  Health and nutritional aspects of sustainable diet strategies and their association with environmental impacts: a global modelling analysis with country-level detail. Lancet Planet. Health 2, e451–e461 (2018).

50. 50.

Tilman, D. et al. Future threats to biodiversity and pathways to their prevention. Nature 546, 73–81 (2017).

51. 51.

Springmann, M. et al. Global and regional health effects of future food production under climate change: a modelling study. Lancet 387, 1937–1946 (2016).

52. 52.

Abel, G. J., Barakat, B., Kc, S. & Lutz, W. Meeting the Sustainable Development Goals leads to lower world population growth. Proc. Natl Acad. Sci. USA 113, 14294–14299 (2016).

53. 53.

Mozaffarian, D. Dietary and policy priorities for cardiovascular disease, diabetes, and obesity: a comprehensive review. Circulation 133, 187–225 (2016).

54. 54.

Mozaffarian, D. et al. Population approaches to improve diet, physical activity, and smoking habits: a scientific statement from the American Heart Association. Circulation 126, 1514–1563 (2012).

55. 55.

Ritchie, H., Reay, D. S. & Higgins, P. The impact of global dietary guidelines on climate change. Glob. Environ. Change 49, 46–55 (2018).

56. 56.

Gonzales Fischer, C. & Garnett, T. Plates, Pyramids and Planets. Developments in National Healthy and Sustainable Dietary Guidelines: a State of Play Assessment (Univ. Oxford, 2016).

57. 57.

Alexandratos, N. & Bruinsma, J. World Agriculture Towards 2030/2050: The 2012 Revision. ESA Working Paper No. 12-03 (Food and Agriculture Organization of the United Nations, 2012).

58. 58.

Wollenberg, E. et al. Reducing emissions from agriculture to meet the 2 °C target. Glob. Change Biol. 22, 3859–3864 (2016).

59. 59.

Carlson, K. M. et al. Greenhouse gas emissions intensity of global croplands. Nat. Clim. Change 7, 63–68 (2017).

60. 60.

Troell, M. et al. Does aquaculture add resilience to the global food system? Proc. Natl Acad. Sci. USA 111, 13257–13263 (2014).

61. 61.

Chan, C. Y. et al. Fish to 2050 in the ASEAN Region (WorldFish Center and International Food Policy Research Institute, 2017).

62. 62.

Rosegrant, M. W. et al. Quantitative Foresight Modeling to Inform the CGIAR Research Portfolio (International Food Policy Research Institute, 2017).

63. 63.

van Vuuren, D. P. et al. The representative concentration pathways: an overview. Clim. Change 109, 5–31 (2011).

64. 64.

Fowler, D. et al. The global nitrogen cycle in the twenty-first century. Phil. Trans. R. Soc. B 368, 20130165 (2013).

65. 65.

Lassaletta, L., Billen, G., Grizzetti, B., Anglade, J. & Garnier, J. 50 year trends in nitrogen use efficiency of world cropping systems: the relationship between yield and nitrogen input to cropland. Environ. Res. Lett. 9, 105011 (2014).

66. 66.

Chateau, J., Dellink, R., Lanzi, E. & Magne, B. Long-Term Economic Growth and Environmental Pressure: Reference Scenarios for Future Global Projections (Organisation for Economic Co-operation and Development, 2012).

67. 67.

Samir, K. C. & Lutz, W. The human core of the shared socioeconomic pathways: population scenarios by age, sex, and level of education for all countries to 2100. Glob. Environ. Change 42, 181–192 (2014).

68. 68.

O’Neill, B. C. et al. A new scenario framework for climate change research: the concept of shared socioeconomic pathways. Clim. Change 122, 387–400 (2014).

69. 69.

Interagency Working Group on Social Cost of Greenhouse Gases. Technical Update on the Social Cost of Carbon for Regulatory Impact Analysis—Under Executive Order 12866 (United States Government, 2013).

70. 70.

World Cancer Research Fund/American Institute for Cancer Research. Food, Nutrition, Physical Activity, and the Prevention of Cancer: A Global Perspective (American Institute for Cancer Research, Washington DC, 2007).

71. 71.

World Health Organization. Guideline: Sugars Intake for Adults and Children (World Health Organization, 2015).

72. 72.

US Department of Health and Human Services & US Department of Agriculture. Dietary Guidelines for Americans 2015-2020 8th edn (Skyhorse Publishing, 2017).

73. 73.

Mozaffarian, D., Appel, L. J. & Van Horn, L. Components of a cardioprotective diet: new insights. Circulation 123, 2870–2891 (2011).

74. 74.

Tubiello, F. N. et al. Agriculture, Forestry and Other Land Use Emissions by Sources and Removals by Sinks: 1990–2011 Analysis. Working Paper Series ESS 14/02 (Food and Agriculture Organization Statistical Division, 2014).

75. 75.

Dinerstein, E. et al. An ecoregion-based approach to protecting half the terrestrial realm. Bioscience 67, 534–545 (2017).

76. 76.

Newbold, T. et al. Has land use pushed terrestrial biodiversity beyond the planetary boundary? A global assessment. Science 353, 288–291 (2016).

77. 77.

Rockström, J. et al. Planetary boundaries: exploring the safe operating space for humanity. Ecol. Soc. 14, 32 (2009).

78. 78.

Gerten, D. et al. Response to comment on “Planetary boundaries: guiding human development on a changing planet”. Science 348, 1217 (2015).

79. 79.

Pastor, A. V., Ludwig, F., Biemans, H., Hoff, H. & Kabat, P. Accounting for environmental flow requirements in global water assessments. Hydrol. Earth Syst. Sci. 18, 5041–5059 (2014).

80. 80.

Galloway, J. N. et al. Transformation of the nitrogen cycle: recent trends, questions, and potential solutions. Science 320, 889–892 (2008).

81. 81.

MacDonald, G. K., Bennett, E. M., Potter, P. A. & Ramankutty, N. Agronomic phosphorus imbalances across the world’s croplands. Proc. Natl Acad. Sci. USA 108, 3086–3091 (2011).

82. 82.

Oppenheimer, M. et al. in Climate Change 2014: Impacts, Adaptation, and Vulnerability (eds Field, C. B. et al.) 1039–1099 (Cambridge Univ. Press, 2014).

## Acknowledgements

This research was funded by the EAT as part of the EAT–Lancet Commission on Healthy Diets from Sustainable Food Systems, and by the Wellcome Trust, Our Planet Our Health (Livestock, Environment and People (LEAP)), award number 205212/Z/16/Z. In addition, K.W. and D.M.-D. acknowledge support from the CGIAR Research Programs on Policies, Institutions, and Markets (PIM) and on Climate Change, Agriculture and Food Security (CCAFS). K.M.C. acknowledges support from the USDA National Institute of Food and Agriculture Hatch project HAW01136-H, managed by the College of Tropical Agriculture and Human Resources. J.F. thanks Bloomberg Philanthropies and USAID for support. B.L.B. acknowledges support from the European Union’s Horizon 2020 research and innovation programme under grant agreement 689150 SIM4NEXUS; the SUSTAg project; and the German Federal Minister of Education and Research (BMBF) under reference number FKZ 031B0170A. L.L. acknowledges support from MINECO, Spain, co-funded by European Commission ERDF (Ramon y Cajal fellowship, RYC-2016-20269). M.T. and M.J. thank FORMAS (grant 2016-00227). M.C. acknowledges a Balzan Award Prize and the Grand Challenge Curriculum at the University of Minnesota–Twin Cities. M.S. and H.C.J.G. acknowledge support from the Wellcome Trust, Our Planet Our Health (Livestock, Environment and People (LEAP)), award number 205212/Z/16/Z. P.S. acknowledges support from a British Heart Foundation Intermediate Basic Science Research Fellowship, FS/15/34/31656. M.R. thanks the British Heart Foundation, grant number 006/PSS/CORE/2016/OXFORD. J.R. acknowledges support from the ERC-2016-ADG 743080 (ERA).

### Reviewer information

Nature thanks K. J. Boote, G. Robertson, P. Smith and the other anonymous reviewer(s) for their contribution to the peer review of this work.

## Author information

Authors

### Contributions

M.S. designed the study, compiled the models, conducted the analysis, interpreted the results and wrote the manuscript. K.W., D.M.-D. and M.C. contributed data and model components for the food-systems model. B.L.B., L.L. and W.d.V. contributed data and model components for the analysis of nitrogen and phosphorus. S.J.V., M.H. and K.M.C. contributed data for the analysis of GHG emissions. M.J. and M.T. contributed data for the analysis of fish and seafood. W.W. designed the flexitarian diet and contributed to the discussion on the health aspects of dietary change. F.D. contributed to the discussion on the planetary boundary related to land use. L.J.G. and R.Z. contributed to the discussion on water use. P.S. and M.R. contributed to discussion on the health aspects of dietary change. B.L. facilitated discussions and contributed to the discussion on the planetary boundaries related to the food system. J.F. contributed to the discussion and background of the study. J.R., H.C.J.G. and D.T. contributed to discussion on the planetary boundaries related to the food system. All authors commented on the manuscript draft and approved the submission.

### Corresponding author

Correspondence to Marco Springmann.

## Ethics declarations

### Competing interests

The authors declare no competing interests.

Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

## Extended data figures and tables

### Extended Data Fig. 1 Reduction in environmental impacts when measures are combined.

Shown are combinations of all measures of medium ambition (comb(med)) and of all measures of high ambition (comb(high)). The mitigation measures include changes in food loss and waste (loss&waste), technological change (technology) and dietary change (diets) for a middle-of-the-road development pathway. The differences to development pathways that are more optimistic (higher income and lower population growth) and more pessimistic (lower income and higher population growth) are indicated by the uncertainty range around the markers (socio-econ).

### Extended Data Fig. 2 Overview of major flows of phosphorus at the global scale.

The external acceptable phosphorus (P) input is determined by the acceptable long-term accumulation of phosphorus in the soil (P soil) and sediment (P sediment) at a phosphorus concentration in surface waters (P surface water) that equals a critical threshold. The phosphorus boundary is affected by the fraction of phosphorus that is taken up by humans (P human; frPuptake being the P-use efficiency, PUE, of the complete food chain, from mined phosphorus (P mine) to P intake) and the fraction of phosphorus excreted by humans (P waste) that is not recycled to land (1 − frPrec), which becomes a point source for water pollution. This phosphorus can only be stored in sediment at a given phosphorus-retention fraction (frPret,sed), while the recycled phosphorus can additionally be stored in soil (at a retention fraction frPret,soil). The critical phosphorus input (Pin(crit)) can be calculated as the sum of critical phosphorus retention in the soil and sediment, and a critical input to surface water (oceans) that is due to run-off and leaching. The Supplementary Information contains a full derivation of phosphorus flows and quantitative estimates of critical phosphorus inputs.

### Extended Data Fig. 3 Planetary option space related to different control variables of nitrogen and yield-related feedback effects.

The control variables include nitrogen inputs related to synthetic fertilizers as used in the main analysis, and the more comprehensive measure of nitrogen surplus that accounts for all inputs and offtakes of nitrogen. The types of feedback effects include changes in nitrogen and phosphorus application associated with closing yield gaps by 75%, as modelled in the tech scenario for cropland use (main), and changes associated with closing yield gaps by 90%, as modelled in the tech+ scenario for cropland use (high yields). Colours and numbers indicate combinations that are below the lower bound of the planetary-boundary range (dark green, 1), below the mean value but above the minimum value (light green, 2), above the mean value but below the maximum (orange, 3), and above the maximum value (red, 4).

## Supplementary information

### Supplementary Information

This file contains Supplementary Methods which provide additional detail about the food systems model, the planetary boundary estimate for phosphorus application, and the nitrogen budget model used in a sensitivity analysis.

### Supplementary Table 1

This file contains an overview of income and population changes in the socioeconomic development pathways.

### Supplementary Table 2

This file contains estimates of global food consumption in the dietary scenarios.

## Rights and permissions

Reprints and Permissions

Springmann, M., Clark, M., Mason-D’Croz, D. et al. Options for keeping the food system within environmental limits. Nature 562, 519–525 (2018). https://doi.org/10.1038/s41586-018-0594-0

• Accepted:

• Published:

• Issue Date:

### Keywords

• Model Food Systems
• Safe Operating Space
• Plant-based Diet
• Planetary Boundaries Framework
• Reducing Food Loss

• ### Risk assessment of soil erosion by using CORINE model in the western part of Syrian Arab Republic

• Alaa Khallouf
• , Swapan Talukdar
• , Endre Harsányi
• , Hazem Ghassan Abdo
•  & Safwan Mohammed

Agriculture & Food Security (2021)

• ### Economic policy instruments for sustainable phosphorus management: taking into account climate and biodiversity targets

• Beatrice Garske
•  & Felix Ekardt

Environmental Sciences Europe (2021)

• ### China’s food loss and waste embodies increasing environmental impacts

• Li Xue
• , Xiaojie Liu
• , Shijun Lu
• , Guangyan Cheng
• , Yuanchao Hu
• , Junguo Liu
• , Zhengxia Dou
• , Shengkui Cheng
•  & Gang Liu

Nature Food (2021)

• ### Small targeted dietary changes can yield substantial gains for human health and the environment

• Katerina S. Stylianou
• , Victor L. Fulgoni
•  & Olivier Jolliet

Nature Food (2021)

• ### Food systems are responsible for a third of global anthropogenic GHG emissions

• M. Crippa
• , E. Solazzo
• , D. Guizzardi
• , F. Monforti-Ferrario
• , F. N. Tubiello
•  & A. Leip

Nature Food (2021)