myUmberto - Forum

Umberto LCA+ => Data (ecoinvent v3, GaBi) => Topic started by: pbeilschmidt on October 30, 2012, 08:09

Title: Differences ecoinvent v2.2/v3 data - the new EcoSpold data format
Post by: pbeilschmidt on October 30, 2012, 08:09
ecoinvent v3: Why is a new data format necessary?

The "old" ecoinvent v2.2 was based in the EcoSpold v1 data format, while for the new ecoinvent v3 database the EcoSpold v2 data format has been developed.

A full documentation of the changes made and the new items in the database can be found in the ecoinvent Overview and Methodology report (Data Quality Guidelines) ( Below are some of the changes. For questions, please reply to this post. We will try to answer your questions and add more information over time. We also offer ecoinvent v3 courses.

This post tries to explain why a new data format (EcoSpold v2) has become necessary before the ecoinvent v3 database and datasets could be created.

The new EcoSpold v2 data format is the result of a development over the last 15 years: Initially based in the SPOLD97 and SPOLD99 formats, developed by the Society for Promotion of Life Cycle Assessment Development (SPOLD), it was developed further by the ecoinvent Centre and launched as EcoSpold v1 data format in 2000, when it became the basis for the ecoinvent 2000 database. New requirements have arisen in the LCA field in recent years. To be able to provide datasets that cater for the new data needs, it was decided to update this format significantly.
The process for a new format definition was launched in 2008. The new data format was publicly released in May 2009 as beta version, and in 2010 in the final version. Just as the predecessor the EcoSpold v2 data format is an open format, which is provided along with a freeware tool (EcoEditor) for detailed analysis and creation of datasets by editors.

In other words, the ecoinvent v3 data contains a number of elements that were impossible to collect, manage, and store in the datasets based on the EcoSpold v1 data format.

The definition of the new data format according to the ecoinvent Centre was guided by the following principles:
- Facilitating database maintenance and extension
- Better support for alternative modelling options
- Data exchange with the ILCD format
- Support for mass balances, energy balances and monetary balances
- Support for language versions
- Support for better documentation

This sounds somewhat abstract to the common user of the data, so here are some examples of what is really new in these datasets. It must be stressed at this point, that even if the new datasets that are developed based on the new EcoSpold v2 format may contain all the data, it is not automatically also available and usable in any LCA software. LCA software will have to be extended over the time to provide support for the new elements in the datasets.

Datasets in ecoinvent v3 now have the following new/additional characterstics.
- System Models: Formerly only one system model for attributional modeling with fixed allocation rules for multi-output processes was possible. With the new data format, several system models can be provided to allow for use in different contexts.
- Parent/Child datasets and Linking into Activity datasets: This concept is especially important for the creation and maintenance of datasets for activities that differ in geographical and/or technical scope. A parent dataset is defined and can have numerous child datasets which describe the differences to the parent dataset. An activity dataset relating to a specific geographic location (e.g. country) and technology level hence is created by linking to one or more parents.
- Geography: The geographic reference of a process was formerly described by the ISO 3166-1 alpha-2 country codes (e.g. CH, DR, FR) or a three-letter region code (e.g. RER, GLO). In the new version of the data format the geographical validity of a dataset is expressed by a reference to an external location file. This file in KML format (OpenGIS KML Encoding Standard, a public format maintained by the Open Geospatial Consortium) already contains default values for all countries and most regions, but other geographical areas can be supplemented. Equipping an activity dataset with a geographic reference opens up the way for impact assessments which in the future may depend on specific location of the resource consumption or release to nature (e.g. in Water Footprinting)
- Mathematical Definition of Flow Values: Rather than having to define coefficients for exchanges in an activity dataset, it is now also possible to define a mathematical formula that is assessed to determine the value for the exchange. The mathematical formula can have references to values of other flows, to parameters and even flow properties (e.g. dry mass, calorific value).
- Flow Properties, e.g. Production Volume: A field for annual production volume has already been present in the EcoSpold v1 data format, but has seen a significant expansion and is now be entered for all output intermediate exchanges. Combining production volumes with the artificial identifiers for geographical location "Global" (GLO) and "Rest-Of-World" (ROW) opens the way for a worldwide coverage of LCI datasets through extrapolation. Flow properties in general are a very future-oriented concept, that will later allow to manage local material properties.

The ecoinvent v3 database was then created based on this new data format. The old datasets were converted at first without modifications, so that they can be used in the new database. Of course some of the new fields will have to be populated for these old datasets over the time.

New datasets that were created for this database version that were created from scratch based on the new data format do of course contain all the required information. And then, while tackling this conversion from ecoinvent v2 to v3, there were numerous modifications and additions: flows (material names) were renamed, new basic units and unit types were introduced, old ones abolished. Errors were corrected.

Further information on the new data format can be found on the website of the ecoinvent Centre ( and in the presentations held on the 48th Swiss LCA Discussion Forum (