ecoinvent Version 3.4

ecoinvent v3.4 was released in October 2017 featuring over 1'000 new and updated datasets in four different sectors: plastics recycling, chemicals, natural gas and electricity.

Oct 24th, 2023

ecoinvent v3.4 updated the electricity markets for both attributional and consequential system models, partitioning the Indian electricity sector by states and grids. Further, data for the European natural gas supply, for chemical products, and for the recycling of PE and PET were added and updated. 

As we wish to ensure clarity and transparency, we openly document the changes made between the latest update and the previous version of the database. Therefore, we offer various files that document and clarify the changes that have been made with each version of the database. These include detailed the report of changes, correspondence files, as well as files that include all activities of each version. Further, we document all known issues of each version to alert database users and to support their understanding and use of the database. 

New and Updated Data

Updates and Extensions of  Electricity Market Data 

In the attributional system models, market compositions are available for all 142 countries which are part of the IEA statistics: 100% of (statistically represented) global electricity generation is covered by country-specific market data. The consequential system model contains new electricity market compositions for 40 countries, based on projections from the EU and the IEA. These new data eliminate important shortcomings of the consequential system model and provide a consistent basis for electricity supply in consequential LCA. 

Indian Electricity by State and Grid 

The electricity production in the ecoinvent database is not just updated on a regular basis, but new geographies are added and existing geographies are divided into smaller geographical regions to provide more detail. 

India is the seventh biggest country in the world (by area), with a booming economy. As such, the role of India is becoming increasingly important when assessing the environmental impacts of the supply chains of many products. That is why more than 160 new electricity production datasets and markets for the 29 states and 5 electricity grids in India were introduced in ecoinvent v3.4. 

Natural Gas 

More than 100 datasets composing the European supply chains (all imports of natural gas to European countries) were updated. With this version, the database consistently reflects the most current technology and supply mixes in all the European countries and their major suppliers. 

Chemical Products 

With the update of more than 120 datasets and the introduction of 50 new products, ecoinvent significantly extended the coverage of the chemical sector. Version 3.4 includes more recent data for main exchanges, such as heat and electricity, as well as production volumes, CAS numbers, formulas, etc. 

Recycling of PE and PET 

New datasets covering the sorting and recycling of post-consumer polyethylene (PE) and polyethylene terephthalate (PET) packaging materials are now included in version 3.4. These new datasets broaden the range of waste fractions and recycling activities covered by the database. 

Relevant Documents and Files 

Report of Changes

This report covers all changes made in the ecoinvent database between version 3.3 (2016) and version 3.4 (2017). It reflects both the database-wide changes, as well as the sector-specific changes. 

Report of Changes

This automatically generated excel file provides an overview of changes made on the undefined unit process datasets between versions 3.3 and 3.4 of the ecoinvent database. The Report of Changes XLS can be found in the «Files» section of ecoQuery, the online version of the ecoinvent database version 3. It is available to licensees of the ecoinvent database. 

Correspondence File 

The correspondence file is a spreadsheet that lists all datasets in versions 3.3 and 3.4 and matches corresponding datasets between versions. 

Activity Overview for ecoinvent v3.4, Undefined 

An excel file listing all the activities and products present in the ecoinvent version 3.4 database including most of the metadata (geography, ISIC classification, activity type, technology level, product type, unit, etc.). This file also includes a list of all geographies, elementary and intermediate exchanges present in the database. 

Activity Overview for ecoinvent v3.4, APOS 

An excel file listing all the activities present in the ecoinvent version 3.4 database including metadata related specifically to the Allocation at the point of substitution. 

Activity Overview for ecoinvent v3.4, Consequential 

An excel file listing all the activities present in the database including metadata related specifically to the Substitution, consequential, long-term system model. 

Activity Overview for ecoinvent v3.4, Cut-Off 

An excel file which lists all the activities present in the database including metadata related specifically to the Allocation, cut-off by classification system model. 

LCIA Methods in ecoinvent v3.4 

List of all LCIA methods including all the impact categories as well as list of selected LCI cumulative results provided by ecoinvent. 

Known Issues

This section lists all known data errors in version 3.4 of the ecoinvent database. 

Data Issues 

  • The markets for electricity, medium voltage are missing inputs of electricity from municipal waste incineration. These will be reintroduced with the next database release. 
  • The markets for electricity in the geography Great Britain in the attributional system models are incorrect. Inputs from the following supplying technolgies are missing: wind, photovoltaic, hydropower, nuclear, waste. 
  • The “market for natural gas, high pressure, PL” is missing; that means that suppliers to that market (imports) are currently supplying (with very little shares) the RoW “market for natural gas, high pressure” . The distortion on the RoW market is minor. 
  • The product MDD is currently modeled in two supply chains: “methylene diphenyl diisocyanate” (older data) and “methylene diphenyldiisocyanate” (more recent data). Only the older supply chain is used by consumers, so there is no double counting. 
  • The product epoxy resin is currently modeled in two supply chains: “epoxy resin, liquid” (older data) and “epoxy resin” (more recent data). Only the older supply chain is used by consumers, so there is no double counting. 
  • The activities “ground granulated blast furnace slag production”, GLO and US have overestimated CO2 emissions, the value should be 0.020 kg CO2 per kg GGBFS rather than 0.2 kg CO2 per kg GGBFS. Corrected versions of both activities can be found in the ecoEditor. 
  • The value of the exchange “maize seed, for sowing” is incorrect in the datasets “maize grain production, US, 2006” and “maize grain production, GLO, 2006”. The value should be 0.003220612 kg/kg of maize grain, which corresponds to 30 kg/ha. Corrected versions of both activities can be found in the ecoEditor. 
  • In the dataset “styrene production”, the benzene emissions to water and to air are too high. With support of PlasticsEurope, it could be established that the benzene emissions to water can be removed completely from the dataset and the benzene emissions to air were decreased to 2.0E-07 kg. A corrected inventory was introduced in version 3.8 

Consequential System Model Issues 

  • Three elementary exchanges are found in the compartment “social”. These exchanges can be ignored, both at the unit process and the inventory level, as ecoinvent does not yet account for social impacts. 
  • The table below shows dataset making a demand of a product that is not supplied by the specified supplying activity. These two exchanges can be ignored. 

LCIA Method Implementation Issues 

In the method “ecological scarcity 2013”, categories “main air pollutants and PM” and “total”, the characterization factors for 3 substances were not added when updating from 3.3. to 3.4. The substances are: Ethylene, Trichloroethylene and Dimethyl carbonate. The CF for all 3 substances is 140000. In the category “main air pollutants and PM”, the score of 33 datasets would increase more than 5%. The largest differences are for datasets directly emitting one of the substances, with a change of 179%. For the “total category”, only 7 datasets see their score change more than 5%, with a maximum of 38%. 

In the method “IMPACT 2002+ (Endpoint)”, exchanges quantified in kBq received CFs from terrestrial and aquatic ecotoxicity, even if those should not apply to radioactive emissions. Ionising radiation CFs were also accidentally attributed to the human toxicity indicator. These two mistakes drastically change the score of many datasets. 

A full report of the CF mistakes and their effect on dataset scores is available upon request. For more details, please contact us