**Source:** HESTIA Team, [ELCD database](https://nexus.openlca.org/database/ELCD) database.

**Licence:** [ELCD III data set licence](https://nexus.openlca.org/ws/files/8651). Free to use, in commercial and non-commercial studies and to develop derived data sets, provided the data set owner and the ELCD database are credited and nothing suggests endorsement by the European Commission or the data set owner. HESTIA publishes only the emission and resource-use factors it calculated from the ELCD processes; the life cycle inventories themselves are not published.

**Citation:** ELCD III data sets, http://lct.jrc.ec.europa.eu. (c) European Commission 1995-2011. ELCD database 3.2; data set owner PE INTERNATIONAL GmbH.

The `emission` indicators of each ELCD 3.2 process mapped to a HESTIA Input term: one row per process, one column per HESTIA term.

Every value is per one reference unit of the process — typically 1 kg, 1 MJ or 1 m³ — calculated in openLCA with EF 3.1 Method (adapted), economic allocation and regionalisation, and converted to HESTIA terms with [hestia-convert-base](https://gitlab.com/hestia-earth/hestia-convert-base) and the [hestia-flowmaps](https://gitlab.com/hestia-earth/hestia-flowmaps). A flow with no mapping contributes nothing.

## Changes from the source

The inventory is grouped into few columns on purpose: each column becomes one Emission or resourceUse indicator per mapped Input on every Cycle that uses this lookup. How each group is formed is in `elcd32.md`, linked below.

- Values rounded to 3 significant figures; nothing is dropped for being small.
- Regionalisation summed away: a cell is the total over every location.
- The columns are the same in every dataset. A column this dataset has no value for holds `0`, which means none reported, not impossible.
- Only emissions that can move an EF v3.1 result are published — 96 columns. A column is kept when it is needed to cover 99% of the contributions to some process in some category, as HESTIA's own EF v3.1 models characterise it. A term with no EF factor is absent: `co2ToAirInputsProductionShortCycleBiogenic` is the notable one.
- Pesticide emissions are rolled up into the keys `herbicideUnspecifiedAi`, `fungicideUnspecifiedAi`, `insecticideUnspecifiedAi`, and `pesticideUnspecifiedAi` for everything else. The fallback has no characterisation factor and is published for the inventory it reports.
- Radioactive isotopes are each characterised with their own EF v3.1 ionising radiation factor, then summed into one column per compartment keyed `ionisingCompoundsU235EqEf31`, in kBq U-235 eq.
- Most of the mass in the `pesticideTo*` columns (95.8% when measured) is keys that are not pesticide active ingredients, so summing them does not give a pesticide load.

How each choice was made and measured, and this dataset's known limits, are in `elcd32.md` in the newest [release folder](https://gitlab.com/hestia-earth/hestia-glossary/-/tree/develop/background-emissions/releases). `elcd32-emission-changelog.md` says which values each release moved.