Which foods deliver the most nutritional value for the smallest environmental footprint?
Each food's Nutritional Value Score is set against its environmental footprint. Foods in the green “win-win” quadrant deliver high nutritional value at relatively low impact. Switch the impact category or functional unit to explore the trade-offs.
Impact
Groups
NVS against environmental footprint
Each food is plotted by its Nutritional Value Score against its environmental footprint per 100 NVS. Foods in the green quadrant deliver high nutritional value at relatively low impact.
mPt per 100 NVS
Methods, definitions and sources
Log scale. 83 of 83 foods shown.
Indonesia: 83 foods from 90 production/origin records, across 5 food groups. Individual impacts remain available alongside the aggregate score.
Each food averages its eligible production and origin variants. Its aggregate is available only when every included variant has a verified estimate.
Aggregate source: Manuscript results 22 April 2026; adapted EF 3.1 V1.01. EF 3.1 normalization and weighting, expressed in millipoints (mPt). The full published score includes uncertain toxicity estimates and categories beyond the seven individual impacts. Lower values indicate lower impact. Not numerically comparable with LENS scores.
Land use (m²·year crop eq): Land occupation and transformation, in square metre-years of annual crop equivalent (ReCiPe 2016).
Water use (m³e): Water use weighted by water scarcity, expressed as cubic metres of deprivation equivalent. The manuscript’s water stage table does not reconcile with its published total; the stage breakdown is unavailable.
Particulate matter (g PM2.5e): Fine particulate matter formation potential, expressed as grams of PM2.5 equivalents.
Terrestrial acidification (kg SOâ‚‚e): Potential for emissions to acidify land, expressed as sulphur dioxide equivalents.
Freshwater eutrophication (kg Pe): Potential for nutrient enrichment of freshwater, expressed as phosphorus equivalents.
Marine eutrophication (kg Ne): Potential for nutrient enrichment of marine water, expressed as nitrogen equivalents.
The source results are already per 100 NVS. No second nutritional conversion is applied in that view. Indonesia’s NVS and impacts use the published manuscript results together. Kenya and Rwanda use the LENS dashboard inventories. These historical nutritional inputs may differ from the current NVS Explorer.
The per-kg and per-1,000-calorie views are estimates: multiply the per-100-NVS result by NVS ÷ 100, then by 100 ÷ kcal per 100 g for calories. These are NVS-weighted estimates, not conversions using the model’s measured food quantities. Foods without energy data remain unavailable only in the calorie view.
Calorie estimates use a supporting food-composition lookup. These energy values are not verified model reference flows.
Kenya and Rwanda use LENS v1.1: each impact is measured against local or regional sustainability thresholds, then summed. Indonesia uses EF3.1 normalization and weighting and includes impact categories beyond the seven shown individually, including uncertain toxicity estimates. Their numerical scores are not directly comparable.
Scatterplots and rankings use a log scale, adjusted where needed to keep zero estimates visible. This does not change the underlying values. A zero estimate is not evidence of no environmental effect.