Glossary

The terms you meet on a factor page, in factor search's columns and filters, and across these docs, each defined in a few sentences with a link to the page that explains it in full. The column or code name is given where the term has one.

Emission factor#

An emission factor is the amount of greenhouse gas released per unit of an activity: per kWh of electricity, per kg of a product, per passenger-km, per euro spent. Multiply a quantity by the factor and you get emissions in kg CO2e. A factor is only meaningful with its unit, boundary, region, years and GWP method, which is why every factor here states them. See Your first emission calculation.

Unit basis#

The unit the factor is per: its denominator, recorded as unit_code (for example kwh, kg, l, p.km or eur) and grouped in unit_family. Two factors are only comparable on the same basis: per kg is not per litre, and per euro is neither. Some units are qualified on purpose, such as kg_n, a kilogram of nitrogen rather than of the product, so that they never match a plain kilogram by accident. A km row is one vehicle's kilometre whether the publisher wrote plain "km" or "vehicle-km": both spellings count the same thing and land on one code. See Units and unit conversions and the five transport bases.

Boundary#

The system boundary, system_boundary, says which life-cycle stages the value covers: burning the fuel only, the fuel's supply chain too, or a product from raw material to factory gate. Boundaries are grouped in boundary_preset (cradle to gate, full life cycle, one slice to add up) so you can compare like with like, and boundary_role says whether a number may be reported alone. It is the publisher's own declaration where they state one, and read from the structure of their data where they do not. See System boundaries.

GWP methodology#

The GWP methodology names the IPCC assessment report whose global warming potentials convert each gas to CO2e: ar2 (the Second Assessment Report, 1995), ar4, ar5 or ar6. Each report gives methane and nitrous oxide a different weight, so the same measured gases give a different CO2e under each. Two factors on different reports are not strictly comparable. One more names a table outside the IPCC reports: wmo2022, the WMO's 2022 ozone assessment, which some refrigerant figures are weighted on. Three further codes say that no single report applies: none on a pure-CO2 figure that needs no weighting, mixed on a total open-climate.ai summed from parts the publisher weighted on different reports, and unknown where the publisher never stated one. Each value names two: gwp_method_native, the report the publisher printed it on, and gwp_method_normalized, the report its normalised value stands on. The factor API serves the normalised value, so the method beside it is gwp_method_normalized; the publisher's own is in the factor's passport. See GWP values and IPCC report versions.

Calculation approach#

For electricity, calculation_approach is location_based or market_based. Location-based uses the average mix of the grid where the power is consumed; market-based uses the electricity you contracted for, or the residual mix when you hold no contract. The two can differ by an order of magnitude on the same kilowatt-hour, and GHG Protocol reporting asks for both side by side. See Market-based vs location-based electricity.

GHG scope#

The scope says where an activity's emissions belong in a greenhouse gas inventory: Scope 1 for the fuel and refrigerants an organisation burns or leaks itself, Scope 2 for the electricity, heat and steam it buys, and one of the fifteen Scope 3 categories for everything else in its value chain. No publisher prints it, because it is not a property of the factor but of who is reporting: factor search works it out from the normalised category and the stages the boundary covers, read for the category's default reporter, such as the office that buys the electricity, and serves it in ghg_scopes, one claim per scope or Scope 3 category. The filters are ghg_scope and, for a Scope 3 category, ghg_scope_category. One factor can carry two scopes at once, since bought electricity is Scope 2 for the power itself plus an upstream leg in category 3.3, and a figure that belongs in no scope, such as a biogenic or an avoided emission, is served as reported outside them. See Market-based vs location-based electricity.

Price basis#

A spend factor is per unit of money, and price_basis says which price that money is. basic excludes taxes on products and trade margins, producer adds the taxes on products, and purchaser adds trade and transport margins on top, which is closest to what a buyer pays. The same purchase gives a different result on each, so match the basis to how your spend is recorded. See Price basis.

GWP native#

The native value, value_co2e_native, is the number exactly as the publisher printed it, under their own GWP methodology. It is the value to cite when you quote a publisher. It is not the value to add across libraries that use different IPCC reports. See Native and normalised values.

GWP normalised#

The normalised value, value_co2e_normalized, is the same factor converted to AR6 GWP100 warming potentials, so that factors from different libraries can be compared and added. It is converted gas by gas from what the publisher printed. Where no exact conversion exists, gwp_normalization_basis reads unconverted and the column repeats the publisher's own number, which is not an AR6 value and must not be added to one. See Why some rows stay unconverted.

Year window#

The accounting years a value applies to, from applies_from_year to applies_to_year: a range such as 2019 to 2022, a single year, or an open end. Filtering on years keeps the factors whose window covers at least one year you picked, and shows the value for the latest of them. A value with no window at all is the only edition on record for its factor, so it applies to any year. See Applicable years.

Sums to parent#

A child factor with parent_relation sums_to_parent is one part of its parent's total, such as the well-to-tank part of a fuel's well-to-use factor. The parts partition the total: together they add up to it. Counting a part and the total on the same line is double counting. See Check that a decomposition sums to its parent.

Contained by parent#

A child factor with parent_relation contained_by_parent sits inside its parent's total but is not one of a set of parts that add up to it. The publisher breaks out some of what the total covers, not all of it. Never add these children together to rebuild the parent. See Read a factor page.

Derived factors#

A factor flagged is_derived is one open-climate.ai added to a library, not a number the publisher printed. Most are derived totals, added up from parts the publisher printed separately. A derived factor can also be a part the publisher never printed, which open-climate.ai estimates from published figures and flags is_estimated: the upstream energy supply and grid losses behind an electricity factor published for generation only (Estimated electricity legs), or the supply chain of a fuel whose publisher prints its combustion only (Estimated fuel legs). A total that includes such a part is flagged too. In factor search, the Derived factors toggle shows only these or hides them. A derived total you hide still heads its entry, marked as outside your filters, with the parts the publisher printed listed under it.

Derived total#

A derived total is a derived factor assembled by adding up parts the publisher printed separately; the publisher did not print the total itself. DESNZ, for example, publishes burning a fuel and the fuel's supply chain in separate tables, and the well-to-use total is their sum. Use the parts for reporting that keeps them apart, and the total where a method asks for the sum. See Check that a decomposition sums to its parent, or view all derived factors in factor search.

Licence tier#

license_tier sorts each publisher's licence into open, copyleft, restricted or prohibited. Open and copyleft factors are shown and exportable in full, copyleft ones with a share-alike condition. Restricted factors are listed with their description but their values are withheld, and prohibited ones are not served. See Licences and redistribution.

Data quality#

data_quality_normalized puts every factor on one scale, high, medium, low or very_low, derived from the publisher's own rating where they ship one, else from a stated uncertainty or a reviewed library default. It is the only quality measure you can compare across libraries. Values that cannot be used as they stand, such as a zero that is an artefact of how the source is built, are not served at all. See Data quality ratings.

Region from the publisher's default#

When region_is_publisher_default is true, the publisher states no geography on the row and it is filed under the library's own default territory, such as the United States for a U.S. library. The region is then a property of the source, not a boundary the factor was measured within. Check it before comparing the factor with one whose region the publisher chose deliberately. See Choosing between similar factors.

Category#

category and sub_category are our own normalised activity axis: what the factor measures, on the same tree for every library. Filtering on one compares like with like across publishers, which the publishers' own trees cannot do. It is not the publisher's classification, which is kept separately. See Find a factor.

Activity#

activity_full is the publisher's own name for what the factor measures: the product, process or activity, in their words and at their level of detail. Two libraries rarely name the same thing the same way. Use the category to line them up, and the activity to check that a factor is the thing you bought. See Read a factor page.

Text origin#

A served text says who wrote it, in emission_factor_text.origin, and it can have three origins: source is the publisher's own words, translated is a machine translation of them, and generated is a paragraph open-climate.ai wrote from the factor's own metadata. On a factor page, a chip beside a passage says Written by open-climate.ai or, for the publisher's own words in another language than the page's, In French. A machine translation wears no chip: the licence card says the publisher's text on the page is translated by open-climate.ai, and the Show ADEME's original toggle under the passage opens the publisher's own words, once the page has found them. Factor search's description preview says Translated by open-climate.ai under a translation. Quote a translated note as a translation, never as the publisher's wording, and go back to the publisher's own record when you need their exact words. See Whose words they are and Provenance markers.

Library#

A library is a published dataset of emission factors under its publisher's name, such as ADEME's Base Carbone or the UK government's conversion factors, identified by a short code in library. Each library has its own licence, release cycle and conventions. The same activity can therefore carry a different value in two libraries, and both can be right. See Choosing a library.

Public libraries#

The public libraries are the ten these docs write their examples against: ADEME, AGRIBALYSE, AIB, DESNZ, EXIOBASE, MITECO, Open CEDA, ÖKOBAUDAT, U.S. EPA and USEEIO. Every count and quoted value on these pages filters to them, so a query you copy returns the number the page states. They are the example set, not factor search, which serves many more. See Choosing a library for what each one covers.

Region#

region_code is the geography the factor describes, as a short code: ISO 3166 for countries and sub-regions, except five Chinese provinces filed under ecoinvent's codes (see Regions), and named codes such as GLOBAL for wider areas. In factor search's Region column, hovering a code shows its full name; the factor page shows the code alone. A factor measured for one country is not a stand-in for its neighbour, whose energy mix and practices can differ. A code can also be defined by exclusion: a complement such as EXIOBASE_ROW_EUROPE is the region it names minus the countries the publisher models separately, so it never stands in for one of those countries. See Choosing between similar factors and Regions.

Value#

The value is kilograms of CO2e per one unit of the factor's denominator, under the GWP method named beside it. Each value has two versions: value_co2e_native as published, and value_co2e_normalized, converted to AR6 where an exact conversion exists. The factor API serves one, value_co2e_normalized, and shows the native number only in the factor's passport, and only when the value was recharacterised to AR6. The open data files and factors_flat carry both. Compare two values only on the same unit, boundary and CO2e basis. See Read a factor page.

CO2e basis#

co2e_basis says which carbon pools the CO2e counts. Fossil CO2, fossil methane, nitrous oxide and the F-gases are in under every code; the codes part on the other three pools:

co2e_basis Fossil CO2, CH4, N2O and F-gases Biogenic CH4 Land-use change Biogenic CO2 Convention
fossil_only in out out out EN 15804 GWP-fossil
biogenic_co2_excluded in in out out GHG Protocol and IPCC inventories
net_total in in in, net of removals in, net of removals full carbon balance
io_total in not separated not separated not separated input-output model

For a fossil fuel fossil_only and biogenic_co2_excluded give the same number; they part on biomass, food and land use, where biogenic_co2_excluded still counts the methane from cattle, manure, landfill and wastewater. net_total subtracts the CO2 the biomass took up, so it can sit below the fossil figure, or below zero. io_total is an input-output model's total for a whole supply chain and does not separate the pools. A few libraries print land-use change inside a biogenic_co2_excluded total; where they also print it on its own, it is the land-use leg value_co2_luluc_native. The same activity on two bases is two different numbers. See Biogenic carbon, land use and the CO2e basis.

Radiative forcing#

radiative_forcing says whether an aviation factor adds the extra warming of emissions released at altitude, contrails and nitrogen oxides, on top of the CO2 itself: with or without, and empty for anything that is not a flight. The uplift can nearly double a flight's figure. A value with it and one without are never interchangeable, so state which one you report. See Flights: radiative forcing.

Grid mix#

grid_mix says which electricity a kWh is assumed to be. Location-based mixes are the grid's production or consumption mix, or the marginal plants that follow demand; market-based mixes are the residual mix, the total_supplier_mix every supplier in a country discloses together, one supplier's product (supplier_specific) or a certified green tariff. The first two are national figures; the last two apply only if you hold that contract. See The grid mixes.

Currency year#

currency_year is the price level the publisher priced a spend factor in. It is a price level, not a date of the data, and it does not change when you pick another year. You do not adjust for it yourself: a spend factor is served once per year, already brought to that year's prices, so tick the year of your spend and multiply. See spend factors and the price level and Currency year.

Supplier#

supplier_name and supplier_code name the counterparty a factor is specific to: the declarant of an EPD, or the electricity retailer behind a supplier-specific tariff. A generic factor names none, and has_supplier lets you filter on the difference. A value with a named supplier describes that supplier's product, not the market average. See Find a supplier-specific electricity factor.

EPD#

An Environmental Product Declaration is a third-party-verified life-cycle assessment that one manufacturer or trade association publishes for one product or product group, in EN 15804 modules. is_epd marks factors that come from one. An EPD describes that product, not the average of its category, so compare it with generic factors carefully. See EN 15804 modules.

Data type#

data_type says whose number the value is. specific is one supplier's own product, average is an average over a group of products or producers, and generic is modelled background data with no producer attached. A supplier link alone does not make a factor specific: an industry EPD's declarant is a trade association. See Choosing between similar factors.

Reference year#

reference_year is the year the publisher attaches to a value: depending on the publisher, the year the data describes or the year of the edition. It identifies the edition you used when you cite a factor. To pick a value for a reporting year, use the year window instead. See Reference year.

Gas legs#

Beside its headline CO2e, a factor can carry the amounts it is made of, one column per gas or carbon origin, such as value_ch4_native or value_co2_biogenic_native. An empty leg means the publisher did not report that part; it never means zero. The legs are what makes a conversion to AR6 exact. See The gas legs.

Top-level factor#

A factor flagged is_top_parent has no parent: it is a whole, not a part of another factor's total. Counting only top-level factors is how you avoid counting the same emissions twice. Its children, where it has any, show how the total breaks down. See Read a factor page.

Default value#

One factor can carry several values: one per year, and sometimes several methods for the same year. The default is the one you get when you choose neither, flagged is_default_indicator for the method and is_latest_activity_year for the year. See Choosing a value.

Spend-based factor#

A spend-based factor is per unit of money (unit_family monetary), usually from an input-output model of a whole economy. It estimates the emissions behind a purchase from what you paid, when you do not know the physical quantity. It is coarser than an activity-based factor, and it must be matched on currency and price basis; the price level is handled for you by the year you pick. See Spend-based vs activity-based emission factors.

Residual mix#

The residual mix is the grid electricity left after every certified green or supplier-specific claim has been removed. It is the market-based factor for a buyer with no contract of their own, and it is usually higher than the average grid mix. See The residual mix.

Slug#

A slug is a factor's readable, permanent identifier, such as ef-ademe-anthracite-fr-kg-well_to_wheel-7825b0a9: the library, activity, region, unit and boundary, then a short code. It is set once and never changes, and a slug that was ever published keeps resolving, so it is safe to store and to cite. See Identifiers and stability.

Release#

A release is one published edition of a library, such as a year's conversion factors. Each release keeps its own licence tier, and older releases stay on record so that a figure can be reproduced. To keep a number stable over time, record the release you took it from. See Pin a factor value to a release.

System model#

A system model is the life-cycle question a release answers. Nearly every library is attributional: its factor is a share of the emissions of the system as it is, which is what a GHG Protocol inventory adds up. A consequential release, CONCITO's Big Climate Database today, answers another question: what one more unit of demand causes, credits for the production it displaces included, which is why one of its stages can be negative. The factor page marks such a value Consequential. Do not mix it with attributional factors in one inventory: automatic matching never picks a consequential factor, and you can still find and read one on the portal.

Dataset date#

The Dataset date in a factor citation copied from an earlier version of the factor page was a fixed date written into the site's own code. It stamped the site build, not the data: it read the same on every factor of every library, and it did not move when a library was loaded or an edition refreshed. So it is not a freshness guarantee for the value it was cited with, and it is not evidence that the factor was checked that day. Today's citation prints no date: its address pins the value. The dates that do track the data are the publisher's edition and the Loaded date on the library page. See Dates, and the old Dataset date.

Factor search#

Factor search is the published emission factors themselves, from every library, put on one model, and every surface that serves them: the search page, the factor pages, the exports, the downloads, the read API and the SQL model. It is open to read, and it is what nearly every page in these docs is about. See Working with factors.

Matching agent#

The matching agent is the separate product that takes a file of your own purchase or activity lines and picks a factor from factor search for each line, with the reasoning behind the choice. It is not part of factor search: it consumes it. Run it from the matching page, which needs no account, or from the /v2 endpoints in the API reference, which need a key.