1. The problem solved

Since launch, EcoCalc has consolidated 22+ ecotoxicology and chemical regulation databases (USEtox, EFSA OpenFoodTox, HESI EnviroTox, JRC, NORMAN, ECHA CHEM, etc.). These databases are designed to assess pure chemical substances with a defined molecular structure.

But natural ingredients (not chemically modified) — olive oil, shea butter, plant extracts — are Natural Complex Substances (NCS) or UVCB sub-type 3: complex mixtures of hundreds of molecules, identified by their botanical source and manufacturing process. Pure-substance databases have no data for these ingredients (CAS numbers such as 8001-31-8, 8001-25-0, etc. do not appear in USEtox, EFSA, HESI…).

Outcome before June 2026: a cosmetic formulator looking up the footprint of a vegetable oil ended up on empty tabs.

Today: as soon as EcoCalc detects a natural ingredient, a 🌱 Footprint LCA & WFN tab is displayed first in the Search Ingredients module, powered by two combined sources.

2. Source A — AGRIBALYSE® 3.2 (ADEME)

Indicators provided: the 16 official midpoint indicators of the European PEF / EF 3.1 method (climate IPCC AR6, water AWARE, Posch acidification, freshwater/marine/terrestrial eutrophication, aquatic ecotoxicity, cancer and non-cancer human toxicity, particulate matter, ozone, fossil/mineral resources, land use, ionising radiation) + the aggregated EF single score + the climate emission breakdown (biogenic / fossil / land-use change).

Scope: cradle-to-farm-gate — from upstream agriculture to the farm gate. Does not cover downstream chemical transformations (saponification, hydrogenation, ethoxylation, formulation, packaging, distribution, end-of-life).

DQR quality index (Data Quality Rating): every ingredient displays its LCA quality rating on the official PEF scale (1 = excellent, 5 = very low), with a clickable in-app explainer.

3. Source B — Water Footprint Network (Mekonnen & Hoekstra 2011)

Indicators provided: the raw water footprint per kilogram of product, decomposed into three components:

• 🟢 Green water — rainwater actually consumed by the crop (evapotranspired or incorporated into the harvest)
• 🔵 Blue water — surface water and groundwater mobilised (mainly irrigation)
• ⚫ Grey water — theoretical water volume required to dilute the pollution generated (nitrogen fertiliser leaching) down to ambient water quality standards

Complementarity with AGRIBALYSE: AGRIBALYSE provides m³ deprived/kg weighted by AWARE (the official PEF method, integrating regional scarcity). WFN provides the raw volume in L/kg (absolute volume engaged). The two readings complement each other for eco-design: the first tells "how much pressure on a scarce resource", the second tells "what absolute volume of water".

4. How it shows up in EcoCalc

In the 🔍 Search Ingredients module, as soon as a user searches for a natural ingredient by CAS, INCI name or common name, the system automatically detects that it is an NCS and:

1. Displays the 🌱 Footprint LCA & WFN tab in first position.
2. Disables (with an explanatory banner) the Substance Profile and Analysis tabs, which rely on pure-substance models not applicable to NCS.
3. Presents the two sources as two distinct sections of equal weight:Section A — 🌱 AGRIBALYSE 3.2: highlights (EF score, climate, AWARE water) + full table of the 16 PEF indicators + DQR + attribution.Section B — 💧 Water Footprint Network: highlights (total + green/blue/grey in L/kg) + detailed table + CC-BY 3.0 attribution.
4. Section A — 🌱 AGRIBALYSE 3.2: highlights (EF score, climate, AWARE water) + full table of the 16 PEF indicators + DQR + attribution.
5. Section B — 💧 Water Footprint Network: highlights (total + green/blue/grey in L/kg) + detailed table + CC-BY 3.0 attribution.

Current coverage: 90 unique INCI ingredients mapped to 51 AGRIBALYSE processes, with 80 % also covered by WFN raw water footprint data (see breakdown below).

5. Coverage by category

🛢️ Vegetable oils and fats (16 AGRIBALYSE products)

Virgin olive oil, coconut oil (refined and copra), palm oil (refined and generic), sunflower oil, rapeseed oil, sesame oil, soybean oil, peanut oil, linseed oil, maize oil, cottonseed oil, grapeseed oil, hazelnut oil, avocado oil, cocoa butter.

The 3 entries without WFN: avocado oil, hazelnut oil, grapeseed oil (not listed as derived products in Table 3 of Mekonnen & Hoekstra).

🌰 Nuts and oilseeds (16 AGRIBALYSE products)

Almond, hazelnut, nuts (fresh coconut, dried coconut, Brazil nut, macadamia, pecan), peanut, sesame seed, sunflower seed, whole soya bean, linseed, chia, pine nut.

Covers ingredients used in cosmetics as extracts, mechanical exfoliants, whipped butters or powders.

🌶️ Spices (16 AGRIBALYSE products → 41 INCI thanks to synonyms)

Cinnamon, cardamom, clove, cumin, turmeric, ginger (powder and root), bay leaf, nutmeg, paprika, poppy seed, black pepper, white pepper, cayenne pepper, saffron, vanilla (alcoholic and aqueous extract).

A category particularly relevant for natural fragrances, oleoresins and cosmetic extracts. The 9 entries without WFN are mainly turmeric, bay leaf and saffron (absent from the WFN 2011 reference).

🍯 Sugars, honey and similar (3 AGRIBALYSE products)

Honey (cosmetic INCI MEL), cane molasses (SACCHARUM OFFICINARUM), saccharin sweetener (excluded in practice from the NCS scope as a pure CAS substance).

This category is deliberately narrow — refined white sugar and fructose, being pure-molecule CAS substances (57-50-1and 57-48-7), fall outside the "complex natural ingredient" scope of this tab.

Priority ingredients not yet covered — enrichment leads

The following cosmetic ingredients are not covered today by AGRIBALYSE (nor by WFN 2011) because they are not produced on French/European soil:

• Exotic butters and oils: argan, shea (Butyrospermum), jojoba, marula, baobab, monoi, castor
• Essential oils and hydrosols: lavender, rose, mint, rosemary, etc. — what is missing is the industrial extraction, not the plant
• Plant-derived surfactants: Sodium Cocoate, Decyl Glucoside, Cocamidopropyl Betaine — the raw material (coconut oil, glucose) is available, but the downstream chemical process is not covered

Request the addition of a priority ingredient from the EcoCalc team — enrichment is done case-by-case from peer-reviewed open-access scientific publications.

6. Three concrete examples

Example 1 — Coconut oil (Cocos Nucifera Oil)

Interpretation: coconut oil mobilises about 4,500 L of water per kg of product, but 99.9 % comes from rainwater consumed by the coconut palm (no irrigation). This is consistent with the production zones (humid tropical South-East Asia). The AWARE-weighted score remains moderate (10.4 m³ deprived) because production happens in low-water-stress regions.

Example 2 — Virgin olive oil (Olea Europaea Fruit Oil)

Interpretation: very favourable climate footprint (1.63 kg CO₂eq/kg, the lowest of the three examples) because Mediterranean olive trees are a low-mechanised perennial crop. But the water footprint is the heaviest of the three: 14,431 L/kg, 16 % of which is irrigation water drawn from already-stressed Mediterranean resources — hence also the highest AWARE-weighted score (22.2 m³ deprived/kg, the highest of the three). A textbook example of a carbon-vs-water discrepancy that a multi-indicator LCA helps to objectify.

Example 3 — Refined palm oil (Elaeis Guineensis Kernel Oil)

Interpretation: palm oil shows a high climate footprint (6.65 kg CO₂eq/kg, the highest of the three — CH₄ emissions from drained peat soils in South-East Asia) but moderate water consumption (1.96 m³ deprived because no irrigation in the humid equatorial zone). Lowest AWARE score of the three. The environmental discussion around palm oil therefore plays out on climate, land use and biodiversity — not on water.

7. Comparative reading — what these three examples tell the formulator

No "miracle" ingredient. Each has its weak point:

• Olive = excellent climate but heavy water (Mediterranean irrigation)
• Coconut = balanced profile, almost exclusively rainfall-based water
• Palm = lowest aggregated EF score, least critical water, but high climate impact

This is precisely what a multi-indicator PEF analysis + two-dimensional water reading (raw volume WFN + scarcity-weighted AGRIBALYSE) brings to light — and what a simple "kg CO₂ / kg ingredient" figure taken out of context would never say.

8. What these data do not say

EcoCalc remains a pre-screening / eco-design decision-support tool:

• No certification. The values displayed do not constitute an ISO 14040/14044 LCA, an ISO 14067 product carbon footprint declaration, or a certified Product Environmental Footprint (PEF).
• Cradle-to-farm-gate scope only. Downstream transformations (saponification of coconut oil into soap, hydrogenation, ethoxylation, blending into a formulation, packaging, distribution, end-of-life) are not included.
• No raw LCI. AGRIBALYSE distributes in its open Etalab format only the aggregated LCIA indicators (already multiplied by the EF 3.1 characterisation factors). Disaggregated inventory flows (kg of CH₄ emitted, kg of N emitted, kWh consumed) require an Ecoinvent licence.
• No essential oils, no non-European exotic ingredients (argan, shea, jojoba, marula, castor), no derived surfactants — these ingredients fall outside the AGRIBALYSE scope.
• Limited geographical granularity. All values are either global averages (WFN) or French averages(AGRIBALYSE), not per specific country of origin.

9. Open access and source transparency

All data added to EcoCalc is open access, under open licences guaranteeing commercial reuse:

• AGRIBALYSE 3.2 under Etalab Open Licence 2.0 (French government, equivalent to CC-BY)
• WFN paper Mekonnen & Hoekstra 2011 under Creative Commons Attribution 3.0

No value is invented or computed by EcoCalc. Every row of the internal dataset contains:

• The DOI of the source publication
• The table number from which the value is extracted
• The exact product label as used in the source
• A methodological note when an approximation is used (for example, cumin is proxied via the "anise, fennel, coriander" group in the WFN classification)

The attribution and DOI are displayed in every EcoCalc UI view, so that any user can trace back to the original source and verify the value.

10. Further reading

• EcoCalc: ecocalc.eu
• AGRIBALYSE 3.2 (ADEME): agribalyse.ademe.fr — dataset DOI 10.57745/XTENSJ
• Mekonnen & Hoekstra 2011 paper: hess.copernicus.org/articles/15/1577/2011 — DOI 10.5194/hess-15-1577-2011
• PEF / EF 3.1 method: eplca.jrc.ec.europa.eu (JRC, European Commission)
• Ingredient suggestions for future additions: contact via ecocalc.eu

Document prepared in June 2026 by Net-Zero Impact SAS (NZI), publisher of EcoCalc. For the current version, see ecocalc.eu/about.