Method and Data Sources

1. At a glance

All results are indicative. They use national average data for each ingredient, not data from a particular farm, supplier or kitchen, and they are not a verified PEF study.

2. Scope and boundaries

2.1 Functional unit

Every figure is per serving. The user enters each ingredient's weight for one serving; batch recipes are divided by the number of servings. Larger portions therefore score higher, which is intended: portion size is one of the main levers a kitchen has.

2.2 Life cycle stages

Agribalyse models each food from farm to consumption: agriculture, processing, packaging, transport, retail and distribution, and the consumer stage. CHEF uses each food's total over all stages. Products that Agribalyse lists as cooked (for example boiled pasta or rice) include household cooking energy; dry and raw products do not. Restaurant cooking energy, kitchen waste and the venue's own operations are not included.

2.3 Calculation

For each ingredient, the 16 indicator values per kg are multiplied by the weight used, and the results are added up for the dish.

2.4 Geography

Agribalyse represents food consumed in France, including imports in their French market shares. Its electricity is the French mix, which is mostly nuclear, so ionising radiation is higher and fossil resource use lower than they would be for a UK or US kitchen.

2.5 Source of beef and lamb

Where beef and lamb are farmed changes their climate footprint more than for any other ingredient. Agribalyse's beef and lamb are French: beef with no deforestation, and lamb from a flock with high emissions per kg. Plain beef and lamb cuts in the recipe table therefore have a Select source menu. An origin scales the ingredient's climate figure by a ratio taken within one published study, so different weight units and system boundaries are not mixed. The farm stage is 97–99% of these ingredients' climate footprint, so the ratio is applied to the whole climate figure. The other 15 indicators are unchanged. Choosing organic loads the organic version of the ingredient instead (section 6.2), which changes all 16 indicators.

Beef from newly cleared Amazon land can be far higher still (more than 700 kg CO₂e/kg carcass, Cederberg et al., 2011), but the data are too uncertain to offer as a choice. Shrimp has no source menu: farmed shrimp ranges from about 2 to 7 kg CO₂e/kg at the farm, but can be hundreds of times higher if the pond was dug in cleared mangrove (Kauffman et al., 2017), and the country of origin does not reveal this. A note is shown on shrimp rows instead.

3. Environmental grade

3.1 The 16 indicators

CHEF uses all 16 impact categories of EF 3.1. Each is normalised by the yearly impact of an average person worldwide and weighted with the EU's official weights. CHEF groups them into four positions for chefs and guests; every indicator belongs to exactly one position, so the four add up to the single score.

Position shares of the EU weighting: Climate 21.1%, Land 17.9%, Water 16.2%, Energy and resources 44.9%.

3.2 Single score

One millipoint (mPt) is one-thousandth of an average person's total yearly footprint. With these factors, CHEF reproduces Agribalyse's own published EF 3.1 single score for each food to within 1%.

3.3 Grade boundaries

The grade compares the dish with a main meal's share of today's average footprint:

• An average person's footprint is 1,000 mPt a year, about 2.74 mPt a day.
• Food is 46% of the environmental footprint of EU consumption, measured with the same 16 indicators and weights (European Environment Agency, Consumption footprint indicator, 2023 data). That gives food about 1.26 mPt per person per day.
• Taking a main meal as about a third of a day's food gives about 0.4 mPt. CHEF calls this "an average meal", and it is the boundary between grades C and D.

Each grade below that halves the impact and each grade above doubles it, because food impacts span a very wide range (a lamb dish can have 20 to 40 times the impact of a lentil dish).

The 46% food share comes from EU consumption data, while the per-person reference (1,000 mPt a year) is the EF method's global average. Combining the two is an approximation. Every dish is currently rated against a main meal, so starters, sides and desserts will usually score better than a main course of the same ingredients.

3.4 The four positions

Climate, Land, Water, and Energy and resources are each graded on the same bands scaled by that position's share of the EU weighting. For example, Climate carries 21.1% of the weight, so its boundaries are 21.1% of the overall ones; converted to kg CO₂e per serving these are the values in the last column above. Land, Water, and Energy and resources combine indicators in different units, so they are shown as grades rather than physical amounts.

3.5 Ingredient colours in the recipe table

Each ingredient has a coloured dot per position for its impact per kg, in the style of a nutrition traffic light. The bands are the 50th and 90th percentiles of each position across 2,099 raw ingredients in the database (composite dishes, ice cream, baby food and organic rows excluded): green for the lower half, amber for the middle, red for the top 10%.

3.6 Impact share and hotspots

The impact share is an ingredient's share of the dish's single score (not its share of the weight). On the Pro certificate, the hotspot marker flags the largest ingredients that together reach 80% of the dish's impact, up to three ingredients and each at least 10%.

4. Nutrition

4.1 Nutrient data

Nutrients come from the Ciqual 2025 food composition table (ANSES), matched to each environmental ingredient. Where Ciqual gives no energy value, energy is calculated from the macronutrients with the conversion factors of Regulation (EU) No 1169/2011, Annex XIV. Values that Ciqual does not report are counted as zero, so totals for some micronutrients may be understated.

4.2 Raw and cooked weight

Ingredient weights are raw (as bought or prepared). Users can enter the weight of a finished serving; per-100 g figures, traffic lights and the Nutri-Score then use the served weight, while per-serving totals and environmental impacts are unchanged. When a recipe gives a cooked weight for pasta, rice, grains or pulses, CHEF converts it to the dry weight using a yield from the water content of the dry and cooked foods in Ciqual.

4.3 Reference Intakes

%RI figures are per serving against the adult Reference Intakes of Regulation (EU) No 1169/2011, Annex XIII (energy 8,400 kJ / 2,000 kcal, fat 70 g, saturates 20 g, carbohydrate 260 g, sugars 90 g, protein 50 g, salt 6 g), and the Nutrient Reference Values for vitamins and minerals. Fibre has no Reference Intake, so none is shown. Vitamins and minerals are listed only when 100 g of the dish provides at least 15% of the Nutrient Reference Value (Annex XIII, Part A 2), with %NRV per 100 g and per serving.

4.4 UK front-of-pack traffic lights

Fat, saturates, sugars and salt are coloured per 100 g of the dish using the UK thresholds: fat over 17.5 g red, over 3 g amber; saturates over 5 g red, over 1.5 g amber; sugars over 22.5 g red, over 5 g amber; salt over 1.5 g red, over 0.3 g amber. For servings over 100 g, a nutrient is also red when one serving gives more than 30% of the Reference Intake (fat over 21 g, saturates over 6 g, sugars over 27 g, salt over 1.8 g), as the guide requires. Energy is not coloured. In markets without a colour scheme of their own (for example the EU, Australia and New Zealand, Japan and China), the same colours are shown as a warning guide.

4.5 Nutri-Score

The indicative Nutri-Score uses the 2023 updated algorithm for general foods, per 100 g of the dish, including the fruit, vegetable and legume percentage and the red meat protein cap. It has not been validated by the scheme's owner and is not a licensed use of the Nutri-Score logo.

4.6 USA Nutrition Facts

The indicative US panel uses the rounding rules and Daily Values of 21 CFR 101.9. US total carbohydrate includes fibre. Ciqual does not record added sugars or trans fat, so CHEF estimates them from the ingredients: added sugars from sugars, honey and syrups and the sucrose in sweetened products; trans fat as the natural trans fat in dairy, beef and lamb fat. It is not a compliant US pack label.

4.7 UK nutrient profile (HFSS)

The UK Nutrient Profiling Model 2004/05 is applied as set out in the Department of Health technical guidance (2011), per 100 g of the dish as served: points for energy, saturated fat, total sugars and sodium, minus points for fruit, vegetables and nuts (from raw ingredient weights, dried fruit and tomato purée counted double), AOAC fibre and protein. Protein is not counted when a food has 11 or more A points and fewer than 5 fruit, vegetable and nut points. A food scoring 4 or more, or a drink scoring 1 or more, is "less healthy". The calculation reproduces the guidance's worked examples. The HFSS promotion and placement rules apply only to prepacked food in the 13 categories of the Food (Promotion and Placement) (England) Regulations 2021, sold by businesses with 50 or more employees, so CHEF shows the score, not a legal verdict.

4.8 Labelling by market

The Nutrition tab, the Nutritional Profile table, the Lite label and page 1 of the Pro certificate follow the chosen market. Each market's rules were checked against its official source in October 2026. All outputs are indicative: they are not compliant pack or menu labels, and businesses should confirm them against the current rules before use.

Where a market applies a rule only to added sugars, sodium or fat (Chile, Mexico), CHEF decides what was added from the ingredient names. The Gulf (GCC) standards are sold by subscription and the Abu Dhabi documents could not be reached, so that market was built from an unofficial English translation of the GSO 2233 panel (USDA FAIRS Saudi Arabia 2025) and Dubai Municipality's calorie guide. Its rounding (FDA rules), carbohydrate basis (available carbohydrate), traffic light thresholds and Nutri-Mark grade are indicative stand-ins until the official texts are checked.

5. Allergens

CHEF reads each ingredient name against keyword rules for the 14 allergens of Regulation (EU) No 1169/2011, Annex II, tested against every ingredient name in the database. Prepared products such as stocks and sauces are flagged "check label", because their recipes vary by supplier. Users can add or remove allergens for each dish. Each market's list (section 4.8) is mapped from these results, naming the cereal, nut or species where its rules require it. CHEF also flags pork and alcohol from the ingredient names, for halal and other dietary needs; these can be corrected by the user and shown on labels, and they are not a halal certification. Allergen information is indicative and must be confirmed with supplier specifications.

6. Ingredient matching and data gaps

6.1 Matching

Typed or imported ingredients are matched to database foods by name, with synonym rules and kitchen units (cups, spoons, cloves, pieces and so on) converted to grams. When an item can't be matched, for example a composite product or an ingredient in another language, an AI model (Google Gemini) may suggest a translation or a breakdown into basic foods. Photos of dishes or written recipes are also read by the AI. Every match is shown to the user and can be changed before and after it is added.

6.2 Stand-ins and estimates

• Similar foods: a small number of ingredients that Agribalyse doesn't cover use the values of a close equivalent (for example allspice uses cinnamon).
• Spices: Agribalyse gives most spices one generic value, so their figures are estimates.
• Yeast: Agribalyse has no yeast; baker's and nutritional yeast use the environmental values of baking powder (fresh yeast scaled for its dry matter), with their own Ciqual nutrients.
• Organic foods: built from the conventional food by scaling its farming stage, indicator by indicator, by Agribalyse's organic-to-conventional farm-gate ratio for the matching commodity. They are matched only when the user asks for organic; in the recipe table, ingredients with an organic version have a Conventional/Organic switch (beef and lamb: the Select source menu, section 2.5). Nutrients are those of the conventional food, except three chicken items that have Ciqual organic values.

7. Limitations

• Results use French national averages from Agribalyse, not the user's own supply chain, origin or farming system. Only beef and lamb can be adjusted for origin (section 2.5), and only for climate.
• Restaurant cooking energy, kitchen food waste and venue operations are outside the boundary.
• Ingredient matching relies on names; a wrong or approximate match changes the result. Users should check the matches shown.
• The grade boundaries combine EU consumption data with the EF method's global per-person reference (section 3.3).
• Nutrient values reflect raw ingredients; vitamin losses in cooking are not modelled.
• The Nutri-Score, UK traffic lights, US panel and allergen statements are indicative and are not compliant labels on their own.

8. Version history

When the method changes, the version number changes, so results in reports can be matched to the method used.

9. References

1. ADEME (2024) Agribalyse 3.2: French environmental database of agricultural and food products. Documentation: https://doc.agribalyse.fr/
2. ANSES (2025) Ciqual French food composition table, version 2025. https://ciqual.anses.fr/
3. European Commission (2021) Commission Recommendation (EU) 2021/2279 on the use of the Environmental Footprint methods to measure and communicate the life cycle environmental performance of products and organisations.
4. Andreasi Bassi, S. et al. (2023) Updated characterisation and normalisation factors for the Environmental Footprint 3.1 method. EUR 31414 EN. Luxembourg: Publications Office of the European Union.
5. Sala, S., Cerutti, A.K. and Pant, R. (2018) Development of a weighting approach for the Environmental Footprint. EUR 28562 EN. Luxembourg: Publications Office of the European Union.
6. European Environment Agency, Consumption footprint indicator (2023 data). https://www.eea.europa.eu/en/analysis/indicators/consumption-footprint-based-on-life
7. Instituto Escolhas (2020) From pasture to plate: subsidies and the environmental footprint of the beef chain in Brazil. São Paulo: Instituto Escolhas.
8. Cederberg, C. et al. (2011) Including carbon emissions from deforestation in the carbon footprint of Brazilian beef. Environmental Science & Technology 45(5), 1773–1779.
9. Rotz, C.A. et al. (2019) Environmental footprints of beef cattle production in the United States. Agricultural Systems 169, 1–13.
10. Wageningen Livestock Research (2022) GHG emission in beef and leather production systems: a literature review.
11. Ledgard, S. et al. (2011) Carbon footprinting of New Zealand lamb from an exporting nation's perspective. EAAP annual meeting; AgResearch (2022) Carbon footprint of New Zealand beef and sheepmeat, for Beef + Lamb New Zealand and the Meat Industry Association.
12. Wiedemann, S. et al. (2014) Environmental impacts and resource use of Australian lamb. Meat & Livestock Australia, project B.CCH.2072.
13. Kauffman, J.B. et al. (2017) The jumbo carbon footprint of a shrimp: carbon losses from mangrove deforestation. Frontiers in Ecology and the Environment 15(4), 183–188.
14. ISO 14040:2006 and ISO 14044:2006, Environmental management — Life cycle assessment.
15. Regulation (EU) No 1169/2011 on the provision of food information to consumers, Annexes II, XIII and XIV.
16. Department of Health and Social Care and Food Standards Agency (2016) Guide to creating a front of pack (FoP) nutrition label for pre-packed products sold through retail outlets.
17. Santé publique France (2023) Nutri-Score: update of the algorithm for general foods.
18. US Code of Federal Regulations, 21 CFR 101.9, Nutrition labeling of food, and 21 CFR 101.11, Nutrition labeling of standard menu items.
19. Department of Health (2011) Nutrient Profiling Technical Guidance; The Food (Promotion and Placement) (England) Regulations 2021; The Calorie Labelling (Out of Home Sector) (England) Regulations 2021.
20. Canada, Food and Drug Regulations (C.R.C., c. 870), section B.01.401; Health Canada, Table of Daily Values and Front-of-package nutrition symbol labelling guide for industry.
21. Secretaría de Economía / Secretaría de Salud, NOM-051-SCFI/SSA1-2010, modification published in the Diario Oficial de la Federación, 27 March 2020.
22. Chile, Ministerio de Salud, Decreto 977 Reglamento Sanitario de los Alimentos, articles 115 and 120 bis.
23. Food Standards Australia New Zealand, Australia New Zealand Food Standards Code, Standards 1.2.3 and 1.2.8, Schedules 9 and 12; Health Star Rating System Implementation Guide v9.1 (2026).
24. Consumer Affairs Agency, Japan, Food Labelling Standards (食品表示基準), Cabinet Office Ordinance No. 10 of 2015.
25. National Health Commission and State Administration for Market Regulation, China, GB 28050-2025 and GB 7718-2025, with official explanatory notes.
26. GCC Standardization Organization, GSO 2233:2021 Requirements of Nutritional Labeling and GSO 9 Labeling of Prepackaged Food Stuffs, as summarised in USDA Foreign Agricultural Service (2025) Saudi Arabia: Food and Agricultural Import Regulations and Standards Country Report (SA2025-0013); Dubai Municipality (2020) Guide for Calorie Declaration in Food Establishments, v3.0; Abu Dhabi Quality and Conformity Council, Nutri-Mark (ADG 44).
27. Food Allergen Labeling and Consumer Protection Act (FALCPA, 2004) and Food Allergy Safety, Treatment, Education, and Research (FASTER) Act (2021).

Questions about the method, or found a figure that looks wrong? Contact us. See also the FAQ.

Return to CHEF