Introduction to the Nutrient Modeling Website

This website accompanies the paper Income growth and climate change effects on global nutrition security to mid-century (Nature Sustainability; see reference below) and should be viewed in conjunction with that paper. It presents detailed, country-specific information on affordability, food and nutrient availability by food group and adequacy and dietary diversity for scenarios to 2050.

The website has a tabbed interface; one tab for each topic with sub-tabs to provide detail.

Paper reference: Nelson, Gerald, Jessica Bogard, Keith Lividini, Joanne Arsenault, Malcolm Riley, Timothy B. Sulser, Daniel Mason-D’Croz, Brendan Power, David Gustafson, Mario Herrero, Keith Wiebe, Karen Cooper, Roseline Remans, Mark Rosegrant, 2018. “Income Growth and Climate Change Effects on Global Nutrition Security to Mid-Century.” Nature Sustainability 1 (12). doi:10.1038/s41893-018-0192-z>. The paper can also be viewed at https://rdcu.be/bdq0B.

The data and R code used to produce the results for this paper and this website are available for download. The reference for it is: Nelson, Gerald, Jessica Bogard, Keith Lividini, Joanne Arsenault, Malcolm Riley, Timothy B. Sulser, Daniel Mason-D’Croz, Brendan Power, David Gustafson, Mario Herrero, Keith Wiebe, Karen Cooper, Roseline Remans, Mark Rosegrant, 2018. The nutrient modeling R project, Available from https://github.com/GeraldCNelson/nutmod. doi: 10.5281/zenodo.2280474.

Choose a country and scenario as the initial choice for the information shown in each tab. See the Glossary for more information on the scenario choices.

Introduction to the Nutrient Modeling Website

This website accompanies the paper Income growth and climate change effects on global nutrition security to mid-century (Nature Sustainability; see reference below) and should be viewed in conjunction with that paper. It presents detailed, country-specific information on affordability, food and nutrient availability by food group and adequacy and dietary diversity for scenarios to 2050.

The website has a tabbed interface; one tab for each topic with sub-tabs to provide detail.

  • Affordability
  • Food Availability (by food group)
  • Nutrient Availability (by food group)
  • Nutrient Adequacy
  • Nutrient Quality
    • AMDR—Acceptable Macronutrient Distribution Range
    • NB—Nutrient Balance Score
    • MRV—Maximum Recommended Intake
  • Dietary diversity
    • Shannon diversity index
    • Nonstaple share of dietary energy
    • Rao's quadratic entropy metric
  • Glossary
  • Developer info
    • Data
    • Files
    • IMPACT Metadata
    • Food group lookup table
    • File documentation
  • Acknowledgements
  • For further information

Paper reference: Nelson, Gerald, Jessica Bogard, Keith Lividini, Joanne Arsenault, Malcolm Riley, Timothy B. Sulser, Daniel Mason-D’Croz, Brendan Power, David Gustafson, Mario Herrero, Keith Wiebe, Karen Cooper, Roseline Remans, Mark Rosegrant, 2018. “Income Growth and Climate Change Effects on Global Nutrition Security to Mid-Century.” Nature Sustainability 1 (12). doi:10.1038/s41893-018-0192-z>. The paper can also be viewed at https://rdcu.be/bdq0B.

The data and R code used to produce the results for this paper and this website are available for download. The reference for it is: Nelson, Gerald, Jessica Bogard, Keith Lividini, Joanne Arsenault, Malcolm Riley, Timothy B. Sulser, Daniel Mason-D’Croz, Brendan Power, David Gustafson, Mario Herrero, Keith Wiebe, Karen Cooper, Roseline Remans, Mark Rosegrant, 2018. The nutrient modeling R project, Available from https://github.com/GeraldCNelson/nutmod. doi: 10.5281/zenodo.2280474.

Download

Food expenditure, per capita income and affordability

Average expenditure on an individual commodity is its consumer price times the quantity available to the average consumer. Summing across all commodities gives total expenditure on modelled food items. The share of food expenditure in per capita income is an indicator of the affordability of the diet.

Note:

Per capita GDP and expenditures—thousand PPP dollars.

Share of income—Percent of per capita income that total expenditure accounts for.

Consumption of fish and alcoholic beverages is not included in this calculation because the current modeling environment doesn’t generate prices for the food items in these food groups.

Average daily food availability by food group

This tab presents information on average food availability for 14 food groups, in grams per day. Explicit diversity measures are presented in the Dietary Diversity tab, but this tab provides indirect evidence about the diversity of the average diet.

Mouse over a point on the radar chart to see the quantity of foods in the food group available per day in grams.

Details on the food group composition are available in the Data and Developer Info tab. Scenario definitions are in the glossary. Data can be downloaded from the link at the bottom of the page.

Download

Nutrient availability

Diversity of food consumption is considered by nutritionists to be an important component of a healthy diet. Elements of diversity include diversity of food intake within a day, across several days, for the average diet. And it includes diversity of individual food items as well as from food groups and staples and non-staples. The data available for this analysis only allows for selected diversity measures.

Nutrients sourced from different food groups are likely to be accompanied by more diverse availability of other needed nutrients.

This tab presents bar graphs showing nutrient availability by food group.

Note: Nutrient units are based on standard usage. The values in the bar graph and table are the average amount available to a consumer per day.

Macronutrients—carbohydrate, protein, fat, and total fiber. Units = grams.

Micronutrients

  • Vitamins
    • Vitamins in milligrams—Niacin, Riboflavin, Thiamin, Vitamin B6, Vitamin C, Vitamin E
    • Vitamins in micrograms—Folate, Vitamin A RAE, Vitamin B12, Vitamin D, Vitamin K
  • Minerals
    • Minerals in milligrams—Calcium, Iron, Magnesium, Phosphorus, Zinc
    • Minerals in grams—Potassium

Nutrient adequacy and kilocalorie availability

The human body needs many essential nutrients from the diet. On the other hand, some nutrients have negative health effects if consumed in excess. Nutrient needs for good health are collectively called Dietary Reference Intake (DRI) measures. Nutrients with negative health consequences when consumed in excess are called disqualifying nutrients.

This tab presents adequacy ratios—the ratio of average daily availability of a beneficial nutrient to the requirement for a representative consumer. This analysis uses the Recommended Daily Allowance (RDA) or the Adequate Intake (AI) for nutrients that do not have an RDA (fiber and potassium for all population groups; all nutrients for infants aged 0-6 months; all nutrients for infants 6-12 months except protein, iron and zinc for which an RDA is available). Since RDA and AI values are by age group and gender, the adequacy ratio uses an average value weighted by the share of the age and gender group in total population. Since the population groups change in composition over time, the requirements for a representative consumer do so as well.

The radar graphs below can be resized by making the web page larger. Each of the four graphs shows the adequacy ratios for the nutrients arrayed around the circle. The legend to the left of the graphs shows the values for each of the rings in the graph. The closer to the center, the smaller the adequacy ratio for that nutrient. The legend at the bottom of the four graphs shows the colors for the three years represented - 2010, 2030, and 2050.

This page also includes results for the quantity and share of total energy from carbohydrates, protein, sugar, and fat. The radio buttons to the left of the bar charts change the scenario for which these data are displayed.

Data for all the graphs on this page are displayed at the bottom of the page and can be downloaded from the side panel.

Download

Acceptable Macronutrient Distribution Range (AMDR)

This tab presents Acceptable Macronutrient Distribution Range (AMDR) results.

The AMDR gives suggested upper and lower limits for the contribution of carbohydrates, protein, and fat to total dietary energy. The AMDR values used here are for males and females 4 years and older. The high values (red) are 35 percent for fat, 65 percent for carbohydrates and 30 percent for protein. The low values (green) are 25, 45, and 10 percent.

Note:

Values in the last three columns are percent of total dietary energy from the macronutrient in the third column.

Download

Nutrient Balance Score

Nutritionists consider diversity of food consumption an important component of a healthy diet. Elements of diversity include diversity of food intake within a day, across several days, for the average diet. And it includes diversity of individual food items as well as from food groups and staples and non-staples.

This tab shows the nutrient balance (NB) score. It captures nutritional adequacy across the range of qualifying nutrients. It uses adjusted adequacy ratios, where the maximum value for each ratio is set to one. It is the sum of the adjusted adequacy ratios divided by the total number of nutrients. A diet that includes all qualifying nutrients in sufficient amounts has an NB of 100.

Download

Maximum Recommended Intake (MRV) ratios

This tab presents Maximum Recommended Intake (MRV) results.

Some nutrients have negative health consequences if consumed in excess. For this study, three are included—sugar, total saturated fatty acids, and ethanol. The maximum recommended daily intakes are:

  • sugar—10 percent of dietary energy
  • total saturated fatty acids—10 percent of dietary energy
  • ethanol—20 gm for adults except pregnant and lactating women; zero otherwise.

Analogous to the adequacy metric presented in the Adequacy ratios tab, the data below show a ratio of actual daily availability to the maximum recommended amount for the average consumer. Ideally, this ratio should be less than one.

Download

Shannon diversity index

The Shannon diversity index (SD) based on mass of food consumed from different sources:

Shannon Diversity Index

where si is the share (by mass) of the ith food item of the 61 available in IMPACT. If consumption of an item is zero then its contribution to the index is set to zero.

When all foods are available in equal amounts, the index is equal to ln(N), where N is the total number of foods considered. The more unequal the distribution, the smaller the indicator value. SDNorm is the normalized version of SD (SD *100/ln(N) where 100 is equal consumption of all food items and 0 is consumption of only 1 food item.

Download

Nonstaple share of dietary energy

Nutrients sourced from non-staples are likely to be accompanied by more diverse availability of other needed nutrients.

This tab shows the share of non-staples in total dietary energy. This metric is a useful simple indicator of diet diversity. A greater proportion of energy available from non-staples typically indicates dietary patterns with higher consumption of nutrient-dense foods such as vegetables, fruits and animal-sourced foods. This metric of dietary diversity is most relevant in lower-income countries which have a low share of energy from non-staples.

Note: Values in the table are percent of the total dietary energy provided by nonstaple food sources. Foods that are considered staples are identified in the Food Group Lookup Table in the Data and Developer Info tab.

Download

Rao's quadratic entropy metric

This tab shows Rao's quadratic entropy metric. This measure of diversity is based on the Euclidian distance between nutrient adequacy for pairs of food items, weighted by the product of the shares of each food item in total quantity of food. It is superior to the Shannon Diversity metric because it takes into account relative diversity by weighting availability shares; however, like the Shannon Diversity metric, it implicitly values all nutrients equally.

Rao's QE

Euclidean distance

  • Raos QE = Rao's quadratic entropy
  • diij = Euclidean distance between adequacy ratios (r) for each of the N nutrients in food items i and j
  • S = number of food items
  • Raos QE = Rao's quadratic entropy
  • N = total number of nutrients
  • pi= ratio of weight of food item i to total food weight

In the figures in this tab, Rao's quadratic entropy values are scaled from 0 to 100 across all countries, years and scenarios using the following formula.

Scaled Rao's QE

Glossary

This tab provides a glossary of terms and more detailed discussion of some of the concepts used in other tabs.

SSP—Shared Socioeconomic Pathways

SSPs include both qualitative and quantitative elements. For this paper we use per capita GDP and population from three SSPs.

  • SSP1—A future with faster and more even income growth.
  • SSP2—The reference scenario. A future with middle of the road outcomes.
  • SSP3—A future with with slower and more uneven income growth.

This research uses three scenarios that combine SSP data with climate data to 2050.

  • SSP1_NoCC—Population and income data from SSP1 with no effects of climate change on crop productivity.
  • SSP2_NoCC—Population and income data from SSP2 with no effects of climate change on crop productivity.
  • SSP3_NoCC—Population and income data from SSP3 with no effects of climate change on crop productivity.
  • SSP2_HGEM—Population and income data from SSP2 with crop productivity altered by weather data derived from the climate change results from the HadGEM2-ES implementation of CMIP5 centennial simulations using RCP8.5. This combination of climate model and RCP gives the largest negative effects on nutrient availability of the three climate models we have results for.

RCP—Representative Concentration Pathway. A greenhouse gas concentration trajectory that results in a radiative forcing value in 2100 relative to pre-industrial values. The RCP used in this analysis is RCP8.5. It would result in a temperature increase between the pre-industrial period and late 21st century of between 2.6 and 4.8 degrees C.

Representative consumer—Nutrient requirements are based on age and gender, and in the case of women whether pregnant or lactating. The availability data are for the average consumer (total availability divided by total population). The population data set used does include age and gender. The requirements used in this analysis are for a representative consumer, weighting each age and gender group by its population. Hence the requirements for the representative consumer change over time as the demographics of the population also change.

RDA—Recommended Dietary Allowance. The average intake level of a nutrient that would meet the needs of 97.5 percent of healthy individuals in a group.

AI— Adequate intake is the recommended average daily nutrient intake level, based on experimentally derived intake levels or approximations of observed mean nutrient intake by a group (or groups) of apparently healthy people that are assumed to be adequate.

AMDR—The Acceptable Macronutrient Distribution Range (AMDR) is defined as a range of intakes for a particular energy source. An AMDR is expressed as a percentage of total energy intake.

NBC—The Nutrient Balance Concept (NBC) has been proposed as a measure of overall nutritional quality of foods or meals which distinguishes between qualifying (required for good health) and disqualifying (harmful in excess) nutrients Source.

Disqualifying nutrient—A term from the U.S> Food and Drug Administration in determining permissable health claims. "Disqualifying nutrient levels means the levels of total fat, saturated fat, cholesterol, or sodium in a food above which the food will be disqualified from making a health claim." Source.

MRV—Maximum Recommended Intake Value. Some nutrients have negative health consequences if consumed in excess. For this study, three are included—sugar, total saturated fatty acids, and ethanol. The MRV is the maximum amount that should be consumed of one of these nutrients. For sugar and total saturated fatty acids it is 10 per cent of dietary energy. For ethanol it is 20 grams per day for adults except pregnant and lactating women; zero otherwise.

Download

This tab makes it possible to view, sort, and download data sets used in the country-specific tabs.

Some key features:

  • Search field—type a text string, click Return, and only those rows that contain the string are retained.
  • Sort columns—click a column name to sort the row by ascending or descending value. If you have filtered the data, the sorting is only for the selected rows.
  • Show entries—choose a value in this dropdown to reduce or increase the number of rows shown.
  • Download button—Writes a gzip file (not a zip file) to the Downloads directory of the table shown.

Data download

Information on file content for developers

Region names and codes

Food group lookup table

Nutrient lookup table

This table shows the country-specific nutrient composition of the IMPACT model's 61 commodities.

Country-specific varieties are used for maize, rice, and wheat and the nutrient composition of composite commodities are determined on a country-specific basis based on the nutrient content and consumption of the individual items that make up the composite. The analysis includes individual or categories of: cereals (seven); animal species (five); roots and tubers (six); marine species (six); fruits and vegetables (six); as well as beverages, nuts, fats and oils

File documentation

Acknowledgements

The authors would like to thank their respective institutions for support. CSIRO authors acknowledge funding from the CSIRO Science Leaders Programme, the CGIAR Research Programme on Climate Change Agriculture and Food Security (CCAFS) and the Bill and Melinda Gates Foundation. IFPRI authors acknowledge the financial support of the CGIAR Research Program on Policies, Institutions, and Markets and CCAFS. Joanne Arsenault acknowledges the World Food Center at UC Davis for providing initial support for her involvement in this effort. David Gustafson acknowledges the financial contributions provided by the ILSI Research Foundation and related partners for its initial support of his participation.

The authors would also like to thank James W. Jones for first suggesting the methodological approach used here, Edward Fern for guidance in the use of the nutrient balance score, Stephan Wood for insights into choice of diversity metrics, Zheyuan Li for R code for Rao's quadratic entropy measure, and Alona Bunning and Laurian Unnevehr for helpful comments on earlier drafts.

Any errors are the responsibility of the authors.

Author names and affiliations

  • Gerald C. Nelson—Department of Agricultural and Consumer Economics, University of Illinois, Urbana-Champaign, Champaign, IL, USA
  • Jessica Bogard—Commonwealth Scientific and Industrial Research Organisation, St. Lucia, QLD, Australia
  • Keith Lividini—International Food Policy Research Institute, HarvestPlus, International Food Policy Research Institute, Washington, DC, USA
  • Joanne Arsenault—Program in International and Community Nutrition, Department of Nutrition, University of California, Davis, Davis, CA, USA
  • Malcolm Riley—CSIRO, Commonwealth Scientific and Industrial Research Organisation, Health and Biosecurity Flagship, Clayton South Vic, Australia
  • Timothy Sulser—International Food Policy Research Institute, Environment and Production Technology Division, International Food Policy Research Institute, Washington, DC, USA
  • Daniel Mason-D'Croz—Commonwealth Scientific and Industrial Research Organisation, St. Lucia, QLD, Australia
  • Brendan Power—Commonwealth Scientific and Industrial Research Organisation, St. Lucia, QLD, Australia
  • David Gustafson—Independent scientist, St. Louis, MO, USA
  • Mario Herrero—Commonwealth Scientific and Industrial Research Organisation, St. Lucia, QLD, Australia
  • Keith Wiebe—International Food Policy Research Institute, Environment and Production Technology Division, International Food Policy Research Institute, Washington, DC, USA
  • Karen Cooper—Nestlé Research Centre, Lausanne, Switzerland
  • Roseline Remans—Bioversity International, Maccarese, Italy and Faculty of Bioscience Engineering, Ghent University, Ghent, Belgium
  • Mark Rosegrant—International Food Policy Research Institute, Environment and Production Technology Division, International Food Policy Research Institute, Washington, DC, USA

Further information

For further information on the IMPACT model, please see this link.

Please contact the IFPRI IMPACT modelling team for specific questions and issues.