ecoinvent 3.3

Known Data issues | ecoinvent 3.3

This page lists all known data errors in ecoinvent version 3.3. We are sorry for any inconvenience these cause and are working hard to eliminate these issues in an upcoming update to the database.

We document these issues openly to make our users aware of these problems and to help them understand the errors so they may work around them while we work on the update and error fix.

Current Known Issues | ecoinvent 3.3

Data issues


  • The activity “cotton production” (GLO, US and CN) contains an error related to the electricity input. The value represents the amount in MJ, missing the conversion to kWh. The error is currently subjected to correction and will be updated for the release of ecoinvent version 3.4.


  • The casting of aluminium seems to contain an error related to energy consumption. We are investigating this issue and will give all details, and fix it as soon as possible.


  •  The activity "sheep production, for meat" GLO was created as copy of the same activity for the geography US. When this was done the amount of reference product was recalculated but the other exchanges were left unchanged. The difference between the live weight of US versus GLO sheep is correct. It is possible to access this type of data on the faostat website, in the data section for "yield" of "livestock primary". Nevertheless the exchanges for the GLO dataset are incorrect in relation to the reference product. Users should use the US dataset, keeping in mind that if their geography of interest is not US they should look into the validity of the exchanges for the scope of their study.


  • The activity "natural gas production" US seems to contain some errors. We are investigating this issue and will give all details, and fix it as soon as possible.


  • The activity "anaerobic digestion of manure" for both geographies CH and GLO are mass imbalanced. The amount of digester sludge produced is too high and it is now under correction. The excess amount of digester sludge is responsible for negative GHG emissions for electricity in certain geographies in the consequential system model. The error is currently subjected to correction and will be updated for the release of ecoinvent version 3.4.


  • The activity "electrolysis of lithium chloride, GLO, 2006" contains an inaccurate calculation related to its electricity input. The datasets will be corrected for version 3.4 of the database based on the information reported in the source cited in the dataset. The electricity input for this type of process and technology ranges from 28 to 32 kWh per kg of lithium produced. 30 is the average of 28 and 32 kWh. Given the fact that the process yields 0.15 kg of lithium the electricity input was scaled accordingly (30 kWh per kg lithium times 0.15 kg of lithium yields 4.5 kWh).


Comment display issues


  • In the datasets with activity name "electricity production, natural gas, combined cycle power plant", the general comment elements contain the following string of text: "['text', '", ending with "']". The issue can be solved by removing these two character strings.


  • Some electricity or combined heat and electricity production datasets have another general comment display problem. Comments display for example "{{children_only_technology}}", which should be replaced by a string of character specified by the data provider. For a list of affected datasets and the value these character string should take, please contact


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 

Ecological scarcity 2013


  • The method calls for the characterization of the "dissipative use" of material extracted by mining operation. This means that the CFs in this impact category should be applied to amounts of taking into account that a fraction of the extracted material may be reused or recycled. The ecoinvent database does not allow to calculate easily this quantity. The scores currently calculated by ecoinvent in this category do not represent what the method intends. These scores are nevertheless useful in an LCA, but should not be used directly to support conclusion about this particular aspect of a supply chain. Rates of recycling and reuse of the material with the highest contribution to the score should be assessed before reaching any conclusion.


  • The method calls for the characterization of the "dissipative use" of water. This is well captured by implementing the CF for evaporated water to air, and represents 100% of the dissipated water in the majority of the datasets in ecoinvent. However, water is also dissipated when it chemically reacts, for example in the case of concrete and concrete. If large quantity of cement or concrete are included in the system at hand, it is possible that the water impact score be underestimated by the implementation proposed by ecoinvent.


  • Emissions of metal to the industrial soil should be zero, as they the method does not assume transfer to water and ingestion through other mechanism. Also, there is no legal limit of dioxins to water in the subcompartment groundwater, groundwater long term and ocean, so their CFs should have been zero. The exchanges below are currently wrongly characterised.