Context mappings
Each source list's raw compartments, and the consensus context they become. These are the rules the transform applies to every list at once — not the per-release mappings the merge reads to place one list.
| Source | Its compartment | Consensus context | Why |
|---|---|---|---|
| ecoinvent-3.12 | water → unspecified |
Environmental → Water → Unknown
envi-wate-unkn
|
Set context_iri manually. |
| ecoinvent-3.8 | air → indoor |
Environmental → Air → Indoor
envi-air-indr-unkn
|
The one context 3.8 has that 3.12 lacks: ecoinvent dropped the indoor air subcompartment, and its four flows (Carbon dioxide / Carbon monoxide / Methane, from soil or biomass stock, and Dimethyl ether) reach no 3.12 flow at all. EF 3.1's 'Emissions to air, indoor' already maps onto this same IRI, so they land beside the EF 3.1 indoor flows rather than inventing a context. |
| ecoinvent-3.8 | air → low population density, long-term |
Environmental → Air → Long-term
envi-air-lote
|
Moved off envi-air-grle-ru10pesq for the reasons on the identical ecoinvent-3.12 row; the subcompartment is unchanged between the two versions (347 flows in each). |
| ecoinvent-3.8 | air → lower stratosphere + upper troposphere |
Environmental → Air → Aircraft cruise height
envi-air-aicrhe
|
— |
| ecoinvent-3.8 | air → non-urban air or from high stacks |
Environmental → Air → Medium stack, <150 meters → Rural (<1000 people/square mile)
envi-air-mest15me-ru10pesq
|
`non-urban air or from high stacks` names two situations at once -- a release in the countryside, or a release from a tall stack anywhere -- and no context can hold both. It was published as envi-air-hist15me, which says a dairy barn's ammonia left a 150-metre chimney; the other reading, Ground level + Rural, says a power-station plume was released at head height, and nothing in a flow says which it is. EF 3.1's factors settle it, and they pick neither: in all 40 (flow, method) pairs where this compartment and one of EF's narrow non-urban compartments are both characterised, EF gives this one exactly the factor it gives `Emissions to non-urban air high stack` -- particles (PM2.5) 3.01757e-06, against 1.13763e-05 non-urban close to ground and 1.62949e-06 non-urban very high stack. That compartment is envi-air-mest15me-ru10pesq here, so this keeps the rural qualifier the compartment does assert and takes the stack height EF characterises rather than the one the name suggests. What is given up: a release that really did leave a >150-metre stack is published as a sub-150-metre one, which EF's own numbers say is the smaller error. Decided once for the three lists that spell this compartment differently -- EF 3.1, ecoinvent in all five versions, and BAFU's `emissions to air / low. pop.`, SimaPro's spelling of ecoinvent v2's `low population density` -- because moving one alone would split them apart. (#411) |
| ecoinvent-3.8 | air → unspecified |
Environmental → Air → Unknown
envi-air-unkn
|
Was envi-air-grle-unkn, which asserted a stratum the source does not name. ecoinvent's `unspecified` says only that the release went to air; reading it as ground level split it from EF 3.1's `Emissions to air, unspecified` and BAFU's `emissions to air / unspecified`, which both sit on envi-air-unkn -- one compartment under three spellings landing in two places. It split ecoinvent from itself as well: the rows the correspondence tables match onto EF's unspecified air flow (383 in 3.8 up to 1,175 in 3.12) were already published in envi-air-unkn, with the guard reconciling that by a permitted coarsening in correspondence-context-routing.json, and only the rows that found no EF partner stayed behind. On the 2026-08-13 build those were all 54 flows in envi-air-grle-unkn -- every one an ecoinvent creation, none carrying a characterisation factor, in a context no other list can reach. The coarsening entry is removed with this change. Nothing is given up: the ground-level qualifier was never in the source. (#411) |
| ecoinvent-3.8 | air → urban air close to ground |
Environmental → Air → Ground level → Urban (>1000 people/square mile)
envi-air-grle-ur10pesq
|
— |
| ecoinvent-3.8 | economic → primary production factor |
Economic
econ
|
— |
| ecoinvent-3.8 | inventory indicator → output flow |
Inventory Indicator
inin
|
— |
| ecoinvent-3.8 | inventory indicator → resource use |
Inventory Indicator
inin
|
— |
| ecoinvent-3.8 | inventory indicator → waste |
Inventory Indicator
inin
|
— |
| ecoinvent-3.8 | natural resource → biotic |
Resource → Biotic
reso-biot
|
— |
| ecoinvent-3.8 | natural resource → fossil well |
Resource → Water → Confined aquifer with fossil groundwater
reso-wate-coaqwifo
|
ecoinvent's one water compartment that names its body -- EF 3.1 gives all six of its water bodies one compartment, so the rest need a per-flow rule. The subcompartment holds exactly one flow, `Water, unspecified natural origin` 2caa889e, in every version 3.8 to 3.12. ecoinvent publishes no definition of `fossil well`, and the name reads either as a well into a fossil aquifer or as a fossil-fuel well producing formation water -- but those converge on this body rather than diverging. Produced water from a hydrocarbon well is connate water in a confined formation with negligible modern recharge, which is fossil groundwater; the two readings differ in why somebody drilled, not in what the water is, and this axis is the body. Salinity does not argue against it either: formation water is typically hypersaline, and salinity is a material property carried on the other axis. ecoinvent's own correspondence table sends this flow to EF 3.1's `ground water`, so they read it as groundwater too; `Confined aquifer with fossil groundwater` is the more precise version of what they said. The emission side already maps `water / fossil well` here, and this makes the two sides agree. Repointed in all five registered versions at once: the subcompartment is unchanged between them. See #276. |
| ecoinvent-3.8 | natural resource → in air |
Resource → Air
reso-air
|
— |
| ecoinvent-3.8 | natural resource → in ground |
Resource → Ground
reso-grou
|
— |
| ecoinvent-3.8 | natural resource → in water |
Resource → Water → Unknown
reso-wate
|
— |
| ecoinvent-3.8 | natural resource → land |
Land Use → Occupation
laus-occu
|
ecoinvent puts occupation and transformation in this one compartment, so the compartment cannot say which a flow is; the name does, and the rules that read it are in `name_prefix_context_mappings`. They cover every flow ecoinvent ships here in every registered version and refuse a name they do not recognise, so this row is the compartment's declaration rather than the rule that places a land flow (#339). |
| ecoinvent-3.8 | social → unspecified |
Social
soci
|
The compartment ecoinvent 3.8 ships only in its APOS release, holding the three exchanges cutoff lacks (#238). Mapped by name, which is how 3.8's other non-environmental compartments are already mapped: 'economic / primary production factor' to econ, the three 'inventory indicator' subcompartments to inin. The flows are allocation accounting rather than social indicators, so econ or inin could be argued -- but ecoinvent's own label is the only evidence in the record, which carries no CAS, no synonyms and no comment, and all three targets are equally unreachable from EF 3.1, which maps to none of them. Gone in 3.12, which drops the compartment entirely. |
| ecoinvent-3.8 | soil → agricultural |
Environmental → Ground → Agricultural
envi-grou-agri
|
— |
| ecoinvent-3.8 | soil → forestry |
Environmental → Ground → Silvicultural
envi-grou-silv
|
— |
| ecoinvent-3.8 | soil → industrial |
Environmental → Ground → Industrial
envi-grou-indu
|
— |
| ecoinvent-3.8 | soil → unspecified |
Environmental → Ground → Unknown
envi-grou-unkn
|
— |
| ecoinvent-3.8 | water → fossil well |
Environmental → Water → Confined aquifer with fossil groundwater
envi-wate-coaqwifo
|
— |
| ecoinvent-3.8 | water → ground- |
Environmental → Water → Unconfined aquifer
envi-wate-unaq
|
— |
| ecoinvent-3.8 | water → ground-, long-term |
Environmental → Water → Long-term
envi-wate-lote
|
Moved off envi-wate-unaq for the reasons on the identical ecoinvent-3.12 row; the subcompartment is unchanged between the two versions (269 flows in 3.12, 268 in 3.8), and 261 consensus flows collide with `water / unspecified` in both. |
| ecoinvent-3.8 | water → ocean |
Environmental → Water → Ocean
envi-wate-ocea
|
— |
| ecoinvent-3.8 | water → surface water |
Environmental → Water → Surface water
envi-wate-suwa
|
— |
| ecoinvent-3.8 | water → unspecified |
Environmental → Water → Unknown
envi-wate-unkn
|
— |
| ecoinvent-3.9.1 | air → low population density, long-term |
Environmental → Air → Long-term
envi-air-lote
|
Moved off envi-air-grle-ru10pesq for the reasons on the identical ecoinvent-3.12 row; the subcompartment is unchanged across every release from 3.01 to 3.12, so it resolves the same way in all of them. The previous comment here read `Uses default 'Ground level' strata. Temporal differentiation is done by LCIA, not LCI`, which is the reading 3.8 and 3.12 moved away from. |
| ecoinvent-3.9.1 | air → lower stratosphere + upper troposphere |
Environmental → Air → Aircraft cruise height
envi-air-aicrhe
|
— |
| ecoinvent-3.9.1 | air → non-urban air or from high stacks |
Environmental → Air → Medium stack, <150 meters → Rural (<1000 people/square mile)
envi-air-mest15me-ru10pesq
|
`non-urban air or from high stacks` names two situations at once -- a release in the countryside, or a release from a tall stack anywhere -- and no context can hold both. It was published as envi-air-hist15me, which says a dairy barn's ammonia left a 150-metre chimney; the other reading, Ground level + Rural, says a power-station plume was released at head height, and nothing in a flow says which it is. EF 3.1's factors settle it, and they pick neither: in all 40 (flow, method) pairs where this compartment and one of EF's narrow non-urban compartments are both characterised, EF gives this one exactly the factor it gives `Emissions to non-urban air high stack` -- particles (PM2.5) 3.01757e-06, against 1.13763e-05 non-urban close to ground and 1.62949e-06 non-urban very high stack. That compartment is envi-air-mest15me-ru10pesq here, so this keeps the rural qualifier the compartment does assert and takes the stack height EF characterises rather than the one the name suggests. What is given up: a release that really did leave a >150-metre stack is published as a sub-150-metre one, which EF's own numbers say is the smaller error. Decided once for the three lists that spell this compartment differently -- EF 3.1, ecoinvent in all five versions, and BAFU's `emissions to air / low. pop.`, SimaPro's spelling of ecoinvent v2's `low population density` -- because moving one alone would split them apart. (#411) |
| ecoinvent-3.9.1 | air → unspecified |
Environmental → Air → Unknown
envi-air-unkn
|
Was envi-air-grle-unkn, which asserted a stratum the source does not name. ecoinvent's `unspecified` says only that the release went to air; reading it as ground level split it from EF 3.1's `Emissions to air, unspecified` and BAFU's `emissions to air / unspecified`, which both sit on envi-air-unkn -- one compartment under three spellings landing in two places. It split ecoinvent from itself as well: the rows the correspondence tables match onto EF's unspecified air flow (383 in 3.8 up to 1,175 in 3.12) were already published in envi-air-unkn, with the guard reconciling that by a permitted coarsening in correspondence-context-routing.json, and only the rows that found no EF partner stayed behind. On the 2026-08-13 build those were all 54 flows in envi-air-grle-unkn -- every one an ecoinvent creation, none carrying a characterisation factor, in a context no other list can reach. The coarsening entry is removed with this change. Nothing is given up: the ground-level qualifier was never in the source. (#411) |
| ecoinvent-3.9.1 | air → urban air close to ground |
Environmental → Air → Ground level → Urban (>1000 people/square mile)
envi-air-grle-ur10pesq
|
— |
| ecoinvent-3.9.1 | economic → primary production factor |
Economic
econ
|
— |
| ecoinvent-3.9.1 | inventory indicator → output flow |
Inventory Indicator
inin
|
— |
| ecoinvent-3.9.1 | inventory indicator → resource use |
Inventory Indicator
inin
|
— |
| ecoinvent-3.9.1 | inventory indicator → waste |
Inventory Indicator
inin
|
— |
| ecoinvent-3.9.1 | natural resource → biotic |
Resource → Biotic
reso-biot
|
— |
| ecoinvent-3.9.1 | natural resource → fossil well |
Resource → Water → Confined aquifer with fossil groundwater
reso-wate-coaqwifo
|
ecoinvent's one water compartment that names its body -- EF 3.1 gives all six of its water bodies one compartment, so the rest need a per-flow rule. The subcompartment holds exactly one flow, `Water, unspecified natural origin` 2caa889e, in every version 3.8 to 3.12. ecoinvent publishes no definition of `fossil well`, and the name reads either as a well into a fossil aquifer or as a fossil-fuel well producing formation water -- but those converge on this body rather than diverging. Produced water from a hydrocarbon well is connate water in a confined formation with negligible modern recharge, which is fossil groundwater; the two readings differ in why somebody drilled, not in what the water is, and this axis is the body. Salinity does not argue against it either: formation water is typically hypersaline, and salinity is a material property carried on the other axis. ecoinvent's own correspondence table sends this flow to EF 3.1's `ground water`, so they read it as groundwater too; `Confined aquifer with fossil groundwater` is the more precise version of what they said. The emission side already maps `water / fossil well` here, and this makes the two sides agree. Repointed in all five registered versions at once: the subcompartment is unchanged between them. See #276. |
| ecoinvent-3.9.1 | natural resource → in air |
Resource → Air
reso-air
|
— |
| ecoinvent-3.9.1 | natural resource → in ground |
Resource → Ground
reso-grou
|
— |
| ecoinvent-3.9.1 | natural resource → in water |
Resource → Water → Unknown
reso-wate
|
— |
| ecoinvent-3.9.1 | natural resource → land |
Land Use → Occupation
laus-occu
|
ecoinvent puts occupation and transformation in this one compartment, so the compartment cannot say which a flow is; the name does, and the rules that read it are in `name_prefix_context_mappings`. They cover every flow ecoinvent ships here in every registered version and refuse a name they do not recognise, so this row is the compartment's declaration rather than the rule that places a land flow (#339). |
| ecoinvent-3.9.1 | soil → agricultural |
Environmental → Ground → Agricultural
envi-grou-agri
|
— |
| ecoinvent-3.9.1 | soil → forestry |
Environmental → Ground → Silvicultural
envi-grou-silv
|
— |
| ecoinvent-3.9.1 | soil → industrial |
Environmental → Ground → Industrial
envi-grou-indu
|
— |
| ecoinvent-3.9.1 | soil → unspecified |
Environmental → Ground → Unknown
envi-grou-unkn
|
— |
| ecoinvent-3.9.1 | water → fossil well |
Environmental → Water → Confined aquifer with fossil groundwater
envi-wate-coaqwifo
|
— |
| ecoinvent-3.9.1 | water → ground- |
Environmental → Water → Unconfined aquifer
envi-wate-unaq
|
— |
| ecoinvent-3.9.1 | water → ground-, long-term |
Environmental → Water → Long-term
envi-wate-lote
|
Moved off envi-wate-unaq for the reasons on the identical ecoinvent-3.12 row; the subcompartment is unchanged across every release from 3.01 to 3.12, so it resolves the same way in all of them. The previous comment here read `Temporal differentiation belongs to LCIA, not LCI`, which holds for EF 3.1's uncharacterised long-term water flow but not for ecoinvent's landfill and tailings leachate. |