Foreground vs background data: getting the distinction right

Every data point in your LCA model falls into one of two categories: foreground (under your control, varying with your decisions) or background (from external sources, taken as fixed). Setting this split correctly is essential for interpreting the model — and it is the first structural question a verifier asks.

Definitions

Foreground (specific) data is measured data from the manufacturer's own plant or close suppliers. Annual energy consumption, raw-material inputs, production waste, direct emissions are all foreground. Background (generic) data comes from global databases (ecoinvent, GaBi, Sphera DB) and represents averages — not the individual plant but the broader sector. An LCA model always mixes the two; the question is the proportions.

What does EN 15804+A2 require?

EN 15804+A2 mandates transparent reporting of the "specific data" share — your EPD must say (for example) 62% of total GWP comes from manufacturer-specific data, 38% from generic. Most PCR Part Bs require a minimum 50% specific share. Low specific share is a verifier-rejection trigger, especially when raw materials or energy use lack supplier-specific data.

Edge cases

• Your supplier has an EPD — does this count as specific data? Yes, supplier-specific EPDs are foreground
• Plant electricity consumption is measured but you use the EU average grid mix — Consumption is foreground, mix is background
• Product mass is measured but raw materials come from ecoinvent — Mass is foreground, raw-material profile is background
• Air-conditioning system usage data — specialised equipment is often missing from ecoinvent, sector average applies → background
• Complex product's sub-components from suppliers — modelling sub-component with ecoinvent is background; supplier-specific data is foreground

Foreground-expansion strategies

Raising the foreground share boosts EPD quality but costs effort. Most effective tactics: 1) request measured data from suppliers for major raw materials (cement, steel, plastics); 2) use meter readings for energy consumption; 3) measure waste and emissions on site; 4) use supplier EPDs when available; 5) for remaining items, pick the most specific ecoinvent version (and localise via custom activity creation).