Multidimensional characterization of global food supply from 1961 to 2013


Food systems are increasingly globalized and interdependent, and diets around the world are changing. Characterization of national food supplies and how they have changed can inform food policies that ensure national food security, support access to healthy diets and enhance environmental sustainability. Here we analysed data for 171 countries on the availability of 18 food groups from the United Nations Food and Agriculture Organization to identify and track multidimensional food supply patterns from 1961 to 2013. Four predominant food-group combinations were identified that explained almost 90% of the cross-country variance in food supply: animal source and sugar, vegetable, starchy root and fruit, and seafood and oilcrops. South Korea, China and Taiwan experienced the largest changes in food supply over the past five decades, with animal source foods and sugar, vegetables and seafood and oilcrops all becoming more abundant components of the food supply. In contrast, in many Western countries the supply of animal source foods and sugar declined. Meanwhile, there was remarkably little change in the food supply in countries in the sub-Saharan Africa region. These changes led to a partial global convergence in the national supply of animal source foods and sugar, and a divergence in those of vegetables and of seafood and oilcrops. Our analysis generated a novel characterization of food supply that highlights the interdependence of multiple food types in national food systems. A better understanding of how these patterns have evolved and will continue to change is needed to support the delivery of healthy and sustainable food system policies.


The past half-century has seen economic growth, urbanization, advances in technologies for agriculture and food production, food processing and storage, and an increasingly powerful and globalized food industry—all of which has led to profound changes in national and regional food systems1,2,3. A number of studies have reported trends over time in the global supply and/or consumption of individual foods and nutrients and in the diversity of foods supplied at national, regional and global levels4,5,6,7,8,9,10,11,12,13,14. Few of these studies, however, represent the totality of food supply patterns; those that considered multiple foods7,11,12,15,16 have not accounted formally for the interdependencies between the demand for and supply of different foods. This is an important omission because national food supplies are modified simultaneously by a mix of socio-economic, ecological, technological and commercial factors, with complex impacts on the availability of different foods due to these interdependencies, which create multiple possible trajectories for food systems. For example, different patterns and speeds of urbanization, rising national income or more widespread use of food processing and restaurant sales alter the variety of food types available or demanded, their sources and the price of food (partly through infrastructural changes)9.

There is, therefore, a need for a coherent multidimensional measure of food supply that can be tracked over time, as has been argued previously for individual consumption17,18. Here we developed a novel data-driven approach to characterize national food supplies that quantifies multidimensional patterns over time. We applied this method to a global database of food supply, and demonstrate how patterns of food supply changed from 1961 to 2013 across the world. These novel characterizations will be valuable to track country-level food systems and their different trajectories to identify common drivers of healthier and/or more sustainable food systems. This will, in turn, enable the development of national and regional food production and trade policies to maximize health and to minimize negative impacts on the environment.


Food supply scores

We summarized the availability of 18 food groups into numerical scores that characterize food systems in different countries and years. Figure 1 and Supplementary Fig. 1 show how the scores relate to the availability of specific foods, that is, the proportion of total energy available for human consumption from each food group; Supplementary Table 1 lists the individual food items in each group. The first score represents food systems characterized by animal source and sugar-based foods, and is higher where meat, milk, animal fats, eggs, offals and sugar and sweeteners are a more abundant part of the food supply, and lower where cereals make up a larger share of the food supply. The vegetable score is higher in food systems characterized by an abundance of vegetables, vegetable oils, treenuts and eggs. The starchy root and fruit score is higher in food systems with an abundance of these two foods, and decreases with abundance of cereals. Finally, the seafood and oilcrops score is higher in food systems that have an abundance of these foods. Almost 90% of the cross-country variation in food availability from 1961 to 2013 is explained by these four scores, which demonstrates their ability to characterize national food supplies parsimoniously and coherently.

Fig. 1: Loadings of each food group for the four food supply scores.

Warm colours indicate that an abundance of a food type as a component of the total energy from the food supply increases the scores and an absence decreases the scores; cold colours indicate that absence increases the scores and abundance decreases the scores.

Current food supply patterns and change over time

Figure 2 and Supplementary Table 2 present mean food supply scores by country for the period 2009–2013, and changes from 1961–1965 to 2009–2013. Although a food system characterized by a high supply of animal source foods and sugar is viewed as representing a typical affluent Western population17,19, and the highest scores for this pattern in 2009–2013 were seen in Iceland and Denmark, the scores were also high elsewhere, for example, in Argentina, Kazakhstan and Mongolia. The animal source and sugar score was low in most countries in sub-Saharan Africa and south Asia, with the lowest values seen in Burundi and Rwanda, whereas Latin American countries had a mix of low and high scores. The animal source and sugar score increased most over the half-century in China, followed by countries in southern and eastern Europe, east Asia and parts of central Asia. Meanwhile, six of the nine largest decreases took place in high-income English-speaking countries (that is Australia, Canada, Ireland, New Zealand, the United Kingdom and the United States of America). The cross-country variations in the score were similar in 1961–1965 and 2009–2013 (Supplementary Table 3).

Fig. 2: Mean food supply scores by country.

a,b, Scores by country for the period 2009–2013 (a) and the change from 1961–1965 to 2009–2013 (b). No data were available for countries shown in grey. As described in Methods, the scores are presented on a scale of 0 to 100 in which 0 represents the lowest value observed in any country from 1961 to 2013, and 100 the highest.

The vegetable score was highest in the ‘Silk Road’ band that stretches from east Asia (China and South Korea), through west Asia (Iran) to the Mediterranean (Lebanon and Greece). The lowest vegetable scores were seen in parts of sub-Saharan Africa, for example, Chad and Lesotho, and some Pacific islands, for example, Solomon Islands; the scores were also consistently low across Latin America. The largest increases in the vegetable score over the past half-century occurred in east Asia and parts of the Middle East, with a change of +75 in South Korea. Decreases in the score were typically small, and occurred largely in sub-Saharan African countries, which included Guinea and Sierra Leone. The cross-country variation of this score increased between 1961–1965 and 2009–2013 (Supplementary Table 3).

The starchy root and fruit score was highest in tropical sub-Saharan Africa, with the seven highest scores appearing in this area. It was lowest in east and south Asia, particularly in South Korea. In contrast to the animal source and sugar and vegetable scores, there was little change in the starchy root and fruit scores over time, and their variation decreased. Finally, the seafood and oilcrops score was high in South Korea and Japan, and in diverse island nations in the Pacific, Indian and Atlantic Oceans (for example, Kiribati, Seychelles, Iceland, and Bermuda); it was lowest in landlocked Burundi and Mongolia. Over time, the share of seafood and oilcrops in the food supply increased in parts of east Asia, particularly in South Korea (+62) and China.

Relationships between changes in scores

Correlations between changes in the food supply scores from 1961–1965 to 2009–2013 ranged from close to zero to moderately positive (Table 1). The moderate correlations between the changes in vegetable scores and both animal source and sugar scores and seafood and oilcrops scores were driven by heterogeneous changes in different food groups across countries (Supplementary Figs. 2 and 3). For example, the vegetable score increased in both east Asia and high-income Western countries. However, although east Asia also experienced a large rise in the animal source and sugar score, many Western countries, especially high-income English-speaking countries, experienced declines.

Table 1 Correlations between changes in food scores from 1961–1965 to 2009–2013

Overall change in national food supply

The index of overall change in food supply, which combines changes in the four scores, shows clear regional patterns (Fig. 3). The greatest changes in food supply from 1961–1965 to 2009–2013 occurred in east and southeast Asia, especially in South Korea, China and Taiwan, and in parts of the former Soviet Union and the Middle East. In high-income Western countries, the largest changes took place in six southern European countries (Cyprus, Portugal, Greece, Spain, Malta and Italy), and in some high-income English-speaking countries (for example, Australia and Canada). The countries with the smallest changes in their food supply were in sub-Saharan Africa (for example, Mali, Chad and Senegal), Latin America (for example, Argentina) and south Asia (for example, Bangladesh).

Fig. 3: Overall change in national food supply from 1961–1965 to 2009–2013.

This index is a weighted sum of the absolute values of change in the four food supply scores. The weights are the proportion of the total variance explained by each score, normalized to add to one. No data were available for the countries shown in grey.

Strengths and limitations

The main strength of our work is its novel scope of developing data-driven scores that characterize national food systems and have clear interpretations. Furthermore, we used a comprehensive open-source dataset with global coverage over a long time period. However, our analysis also has some limitations. The Food and Agriculture Organization (FAO) Food Balance Sheet data are estimates of food availability, which may be substantially different from food consumption20, and do not capture food waste or subsistence production; nor do they account for food processing, which may have health consequences above and beyond differences in the availability of food groups21. The FAO Food Balance Sheet data are provided at the national level, and therefore do not account for within-country heterogeneity. Additionally, there were no data available for some countries and territories, which include a number of Pacific islands (for example, American Samoa and Nauru) in which major changes to dietary patterns have consequences, such as obesity and diabetes, that are of particular concern22,23,24,25. Where data are available, the FAO acknowledges that data quality varies among countries and items, and over time26.


We found that four data-driven scores characterize major patterns in national food supply across the world and explain approximately 90% of the variation in worldwide food supplies over a period of approximately half a century. With the notable exception of countries in sub-Saharan Africa, there were substantial changes in national food supply patterns over the past 50 years. South Korea, China and Taiwan experienced the largest changes, with animal source foods and sugar, vegetables and seafood and oilcrops all becoming a more abundant component of food supply. This contrasts with high-income English-speaking countries, in which the animal source and sugar score has declined substantially.

FAO food balance data have been used previously to investigate various features and implications of food systems at the global level, which include food and nutrient supply11,13, dietary adequacy15, human trophic levels (that is, the position of humans in the food chain)16 and food trade14. However, these studies either used data on individual foods or constructed scores that were predefined, based on criteria such as the mean of the trophic level of food items in the diet16 or the ratio of energy available from Mediterranean and non-Mediterranean food groups15. In comparison, our data-driven approach used the internal structure and interrelationships of different food groups, and identified coherent food supply patterns that are present in all 171 countries over 53 years, but with widely varying scores. Despite differing approaches, some commonalities in the findings appear, such as increasing trophic levels over time16 as populations (especially in Asia) increase their consumption of animal source foods, and an overall increase over time in global food supply diversity11.

Our findings highlight the importance of examining entire national food systems and accounting for internal interdependencies, rather than changes in the supply of individual foods and food groups. This will allow us to understand better how factors, such as increasing income, affect multiple food groups simultaneously, and how food systems act collectively as potential determinants of nutritional status and health. Major shifts towards more diverse food supplies in emerging economies, especially in east and southeast Asia, may be partly responsible for substantial improvements in nutritional status (for example, reductions in stunting, anaemia and other micronutrient deficiencies) in this region27,28,29,30. For example, we assessed the strength of a crude association of food supply scores in 2009–2013 with national data from the same years on adult body mass index and adult height23,27. We identified a strong positive association of animal source and sugar scores with body mass index and height, and a moderate positive association of vegetable scores with body mass index and height (Supplementary Table 4).

We also highlight the relatively small scale of changes in food supply in south Asia and sub-Saharan Africa, which is in clear contrast to the large changes in east and southeast Asia. Low values of food supply scores other than starchy roots and fruit in much of sub-Saharan Africa suggest that food systems in the region are failing to deliver diverse diets and may be particularly low in animal source foods31. This food insecurity and poor dietary quality may help to explain the co-existence of undernutrition and overweight in many African countries23,25,27,28,29,30. Parallel to trends in low- and middle-income countries, in many high-income countries the declines in animal source and sugar supply and commensurate increases in vegetable supply indicate a possible trend towards a more balanced and healthier diet composition. There is a need to understand the technical, economic, political and social determinants of these trends, and to develop policies that will make them healthier and more sustainable.

Food production and trade also affect the local, regional and global environments through their impact on soil nutrient and biotic properties, water systems and emissions of greenhouse gases31,32,33,34,35,36,37,38,39,40. Our multidimensional characterization of food supply will allow a more comprehensive assessment to be made of the environmental impacts at a global scale. However, detailed data on the country of origin and international trade of foods, and how these interact with the food supply scores, is needed to investigate these impacts in specific countries, as has been done for air pollution41.

Multidimensional descriptions of national food systems can both illustrate concurrent trends in food supplies and identify interdependencies between different constituents of population-level diets. Such data provide novel information that can be used to underpin agricultural and trade policies for a sustainable and healthy future.



We downloaded food balance data from the website of the FAO (; these data were updated on 12 December 2017. Food balance sheets have been published by the FAO since 1949 and describe the availability of different foods for human consumption. As described in detail in the Food Balance Sheets Handbook26, the FAO has used official and unofficial data, its own technical knowledge and feedback from national governments to create the series of food balance sheets; further details are available in the FAO archives ( The current data were assembled from a variety of sources, which include national statistics, farmer stock surveys and industrial censuses. For each food item, the domestic supply quantity comprises production and imports less exports, adjusted for variations in stocks (for example, food stored by governments). The quantity of food is the domestic supply quantity, less food losses and food used for feed and seed. The quantity of food was then used to calculate the food supply (kilocalories per capita per day), which are the data used in our analyses26.

We used data from 18 food groups for the years 1961–2013: cereals, starchy roots (for example, potatoes), sugar and sweeteners, pulses (for example, beans and peas), treenuts, oilcrops, vegetable oils, vegetables, fruits, stimulants, spices, meat, offals, animal fats, eggs, milk, fish and seafood, and aquatic products, which include aquatic mammals and plants (Supplementary Table 1). We excluded the miscellaneous category (which includes infant food and other unspecified items), sugar crops and alcoholic beverages.

Data for Burundi, Comoros, Eritrea, Libya, Seychelles, Somalia and Syria were not available in the most recent version of the food balance sheets. For Libya, Somalia and Syria, we used data from the previous version, which provided data for the period 1961–2011. For Burundi, Comoros, Eritrea and Seychelles, we used data from the next most recent version, which provided data for the period 1961–2009.

Cleaning and imputation

We examined time series for all the country–food type pairs and identified outliers and countries with implausible data. We removed data for the Occupied Palestinian Territory, as there were large discontinuities in the data, probably because the governance and reporting systems changed over time. We also removed data for Maldives, which were implausibly low for many food type–year combinations, which caused discontinuities in the time series. Data for the current Sudan were only available for 2012 and 2013, and no data were available for South Sudan, so we report estimates for former Sudan. Finally, we removed all data for the former Yugoslavia, owing to large and inconsistent discontinuities between Yugoslavia and its successors, and Serbia and Montenegro and its successors.

Three other countries for which data were available ceased to exist during the period of analysis: Union of Soviet Socialist Republics, People’s Democratic Republic of Ethiopia (modern Ethiopia and Eritrea) and Czechoslovakia. Furthermore, data for Belgium and Luxembourg were combined by the FAO from 1961 to 1999. We created complete time series for successor countries based on the time series for the original countries as follows. First, in the three years after dissolution, we calculated the availability for each food type in the original countries by weighting the availability in their successor countries by population share. We then calculated the ratio of mean per-capita availability in the successor country in those three years to the availability in the original country. We multiplied the per-capita availability in the original country by this ratio to create predissolution time series for successor countries. Finally, these estimates were rescaled, so that for each country–year–food type combination, the sum of the availability in the successor countries was equal to availability in the original country.

The final dataset comprised each combination of 18 food groups, 171 countries and 53 years. After cleaning, 3,714 data points (2.3% of the data) were missing. The item with the largest missingness was aquatic products, with 1,191 missing data points (13% of the total for that item). We imputed missing values using statistical models with a hierarchical structure, fitted using the integrated nested Laplace approximation method42. Separate models were fitted for each food type and region, in which subregions and countries formed the two levels of the hierarchy for each model (Supplementary Table 5). Estimates for each country and year were informed by data from other years in the same country and that from other countries, especially those in the same subregion with data for similar time periods. The model incorporated non-linear time trends, which comprised a combination of linear terms and a first-order random walk, all modelled hierarchically.

Statistical analysis

The data for the 18 food groups were provided in units of kilocalories per capita per day. To characterize the food supply patterns independently of the total energy from these 18 food groups available in each country, we divided each data point by the total sum of calories for that country, in units of kilocalories per person per day. Data on the energy available from each food group for each country–year were therefore expressed as a proportion of the energy available from all 18 food groups; that is, the values for each country–year summed to one.

We carried out a principal component analysis on the food supply composition data43; principal component analysis identifies patterns by finding weighted sums of variables that explain as much of the variance in the data as possible. The first four principal components explained 89.2% of the variance in the data. We applied a varimax rotation to the loadings of the four principal components44. This rotation aids interpretation by producing a small number of coefficients with large values, and many coefficients close to zero. For presentation, we scaled each varimax-rotated component score linearly to lie in the range 0 to 100; that is, the country–year with the lowest score was scaled to 0, and that with the highest score to 100.

We calculated an overall index of change in the national food supply. The absolute values of the changes in the scores were each weighted by the proportion of the total variance explained by its varimax-rotated principal component, normalized to add to one (0.46, 0.21, 0.18 and 0.15, respectively, for the absolute change in animal source and sugar score, absolute change in vegetable score, absolute change in starchy root and fruit score and absolute change in seafood and oilcrops score). These values were then summed to give the value of the index:

Index of change = (0.46 × absolute change in animal source and sugar score) + (0.21 × absolute change in vegetable score) + (0.18 × absolute change in starchy root and fruit score) + (0.15 × absolute change in seafood and oilcrops score)

Reporting Summary

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

Data availability

The data analysed in this study are published by the Food and Agriculture Organization of the United Nations, and are available from The results of this study (that is, the scores and change index) are available from the website of the NCD Risk Factor Collaboration at


  1. 1.

    Popkin, B. M. Relationship between shifts in food system dynamics and acceleration of the global nutrition transition. Nutr. Rev. 75, 73–82 (2017).

  2. 2.

    Pingali, P. Westernization of Asian diets and the transformation of food systems: implications for research and policy. Food Policy 32, 281–298 (2007).

  3. 3.

    Pinstrup-Andersen, P., Pandya-Lorch, R. & Rosegrant, M. W. The World Food Situation: Recent Developments, Emerging Issues and Long-Term Prospects (The International Food Policy Research Institute, 1997).

  4. 4.

    Smith, M. R., Micha, R., Golden, C. D., Mozaffarian, D. & Myers, S. S. Global expanded nutrient supply (GENuS) model: a new method for estimating the global dietary supply of nutrients. PLoS ONE 11, e0146976 (2016).

  5. 5.

    Micha, R. et al. Global, regional and national consumption of major food groups in 1990 and 2010: a systematic analysis including 266 country-specific nutrition surveys worldwide. Br. Med. J. Open 5, e008705 (2015).

  6. 6.

    Khatibzadeh, S. et al. A global database of food and nutrient consumption. Bull. World Health Organ. 94, 931 (2016).

  7. 7.

    Schmidhuber, J. et al. The Global Nutrient Database: availability of macronutrients and micronutrients in 195 countries from 1980 to 2013. Lancet Planet. Health 2, e353–e368 (2018).

  8. 8.

    Popkin, B. M. Urbanization, lifestyle changes and the nutrition transition. World Dev. 27, 1905–1916 (1999).

  9. 9.

    Kearney, J. Food consumption trends and drivers. Phil. Trans. R. Soc. B 365, 2793–2807 (2010).

  10. 10.

    Wolmarans, P. Background paper on global trends in food production, intake and composition. Ann. Nutr. Metab. 55, 244–272 (2009).

  11. 11.

    Khoury, C. K. et al. Increasing homogeneity in global food supplies and the implications for food security. Proc. Natl Acad. Sci. USA 111, 4001–4006 (2014).

  12. 12.

    Remans, R., Wood, S. A., Saha, N., Anderman, T. L. & DeFries, R. S. Measuring nutritional diversity of national food supplies. Global Food Secur. 3, 174–182 (2014).

  13. 13.

    Beal, T., Massiot, E., Arsenault, J. E., Smith, M. R. & Hijmans, R. J. Global trends in dietary micronutrient supplies and estimated prevalence of inadequate intakes. PLoS ONE 12, e0175554 (2017).

  14. 14.

    Wood, S. A., Smith, M. R., Fanzo, J., Remans, R. & DeFries, R. S. Trade and the equitability of global food nutrient distribution. Nat. Sustain. 1, 34–37 (2018).

  15. 15.

    da Silva, R. et al. Worldwide variation of adherence to the Mediterranean diet, in 1961–1965 and 2000–2003. Public Health Nutr. 12, 1676–1684 (2009).

  16. 16.

    Bonhommeau, S. et al. Eating up the world’s food web and the human trophic level. Proc. Natl Acad. Sci. USA 110, 20617–20620 (2013).

  17. 17.

    Hu, F. B. et al. Prospective study of major dietary patterns and risk of coronary heart disease in men. Am. J. Clin. Nutr. 72, 912–921 (2000).

  18. 18.

    Ioannidis, J. P. A. The challenge of reforming nutritional epidemiologic research. J. Am. Med. Assoc. 320, 969–970 (2018).

  19. 19.

    Cordain, L. et al. Origins and evolution of the Western diet: health implications for the 21st century. Am. J. Clin. Nutr. 81, 341–354 (2005).

  20. 20.

    Pomerleau, J., Lock, K. & McKee, M. Discrepancies between ecological and individual data on fruit and vegetable consumption in fifteen countries. Br. J. Nutr. 89, 827–834 (2003).

  21. 21.

    Hall, K. et al. Ultra-processed diets cause excess calorie intake and weight gain: an inpatient randomized controlled trial of ad libitum food intake. Cell Metab. 30, 67–77 (2019).

  22. 22.

    NCD Risk Factor Collaboration. Trends in adult body-mass index in 200 countries from 1975 to 2014: a pooled analysis of 1698 population-based measurement studies with 19.2 million participants. Lancet 387, 1377–1396 (2016).

  23. 23.

    NCD Risk Factor Collaboration. Worldwide trends in body-mass index, underweight, overweight, and obesity from 1975 to 2016: a pooled analysis of 2416 population-based measurement studies in 128.9 million children, adolescents, and adults. Lancet 390, 2627–2642 (2017).

  24. 24.

    NCD Risk Factor Collaboration. Worldwide trends in diabetes since 1980: a pooled analysis of 751 population-based studies with 4.4 million participants. Lancet 387, 1513–1530 (2016).

  25. 25.

    NCD Risk Factor Collaboration. Rising rural body-mass index is the main driver of the global obesity epidemic in adults. Nature 569, 260–264 (2019).

  26. 26.

    Food Balance Sheets: A Handbook (Food and Agriculture Organization of the United Nations, 2001).

  27. 27.

    NCD Risk Factor Collaboration. A century of trends in adult human height. eLife 5, e13410 (2016).

  28. 28.

    Stevens, G. A. et al. Global, regional, and national trends in haemoglobin concentration and prevalence of total and severe anaemia in children and pregnant and non-pregnant women for 1995-2011: a systematic analysis of population-representative data. Lancet Global Health 1, e16–e25 (2013).

  29. 29.

    Stevens, G. A. et al. Trends and mortality effects of vitamin A deficiency in children in 138 low-income and middle-income countries between 1991 and 2013: a pooled analysis of population-based surveys. Lancet Global Health 3, e528–e536 (2015).

  30. 30.

    Stevens, G. A. et al. Trends in mild, moderate, and severe stunting and underweight, and progress towards MDG 1 in 141 developing countries: a systematic analysis of population representative data. Lancet 380, 824–834 (2012).

  31. 31.

    Willett, W. et al. Food in the Anthropocene: the EAT-Lancet Commission on healthy diets from sustainable food systems. Lancet 393, 447–492 (2019).

  32. 32.

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

  33. 33.

    Monfreda, C, Ramankutty, N. & Foley, J. A. Farming the planet: 2. Geographic distribution of crop areas, yields, physiological types, and net primary production in the year 2000. Global Biogeochem. Cy. 22, GB1022 (2008).

  34. 34.

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

  35. 35.

    Vörösmarty, C. J., Green, P., Salisbury, J. & Lammers, R. B. Global water resources: vulnerability from climate change and population growth. Science 289, 284–288 (2000).

  36. 36.

    Kalnay, E. & Cai, M. Impact of urbanization and land-use change on climate. Nature 423, 528–531 (2003).

  37. 37.

    Matson, P. A., Parton, W. J., Power, A. & Swift, M. Agricultural intensification and ecosystem properties. Science 277, 504–509 (1997).

  38. 38.

    Vitousek, P. M. et al. Human alteration of the global nitrogen cycle: sources and consequences. Ecol. Appl. 7, 737–750 (1997).

  39. 39.

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

  40. 40.

    Foley, J. A. et al. Solutions for a cultivated planet. Nature 478, 337–342 (2011).

  41. 41.

    Zhang, Q. et al. Transboundary health impacts of transported global air pollution and international trade. Nature 543, 705–709 (2017).

  42. 42.

    Rue, H., Martino, S. & Chopin, N. Approximate Bayesian inference for latent Gaussian models by using integrated nested Laplace approximations. J. R. Stat. Soc. B 71, 319–392 (2009).

  43. 43.

    Pearson, K. LIII. On lines and planes of closest fit to systems of points in space. Lond. Edinb. Dublin Phil. Mag. J. Sci. 2, 559–572 (1901).

  44. 44.

    Kaiser, H. F. The varimax criterion for analytic rotation in factor analysis. Psychometrika 23, 187–200 (1958).

Download references


This study was funded by the Wellcome Trust Biomedical Resource & Multi-User Equipment Programme (101506/Z/13/Z). J.B. was supported by a Royal Society Research grant (RS/R1/180086). R.G. and A.D.D. were supported by Wellcome Trust grants 205200/Z/16/Z and 210794/Z/18/Z. M.D.C. was supported by an Academy of Medical Sciences Springboard Award (HOP001/1029).

Author information

M.E. and G.D. developed the study concept. J.B., G.M.S. and J.K.L. obtained the data, conducted the analyses and prepared the results. R.G., G.A.S., F.F., J.E.B., M.D.C. and A.D.D. contributed to the data, analyses and interpretation. J.B. and M.E. wrote the first draft of the paper with input from the other authors.

Correspondence to James Bentham.

Ethics declarations

Competing interests

M.E. received personal fees from Prudential, Scor and Third Bridge, and a charitable grant from the AstraZeneca Young Health Programme, outside the submitted work. The other authors declare no competing interests.

Additional information

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

Supplementary information

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Bentham, J., Singh, G.M., Danaei, G. et al. Multidimensional characterization of global food supply from 1961 to 2013. Nat Food 1, 70–75 (2020).

Download citation