Author Topic: Revised implementation of IPCC2013 in the ecoinvent v3.2 database  (Read 18144 times)

Offline pbeilschmidt

  • Administrator
  • Hero Member
  • *****
  • Posts: 586
    • View Profile
While reading the ecoinvent v3.2 Change Report [Moreno Ruiz E., Lévová T., Bourgault G., Wernet G., (2015). Documentation of changes implemented in ecoinvent database 3.2. Ecoinvent Centre, Zürich, Switzerland], section 2.9 on changes in impact assessment methods caught my attention.

There was no general update for LCIA methods with the changeover from ecoinvent v3.1 to v3.2, but there have been changes to the IPCC2013 implementation, that deserve a mention here, as it will affect the LCA results of our users who work with the IPCC2013 method.

The report on page 16 explains that "ecoinvent has decided to exclude near-term climate forcers and to opt for CFs without the carbon-climate feedback effects". Also, some characterization factors (CF) had to be corrected. Therefore, "due to these significant mistake corrections and  change in methodological choices, we recommend not to use the impact assessment scores calculated for 3.1 for IPCC2013".

Apparently, when the IPCC Fifth Assessment Report (AR5) was released in 2013, due to some changes and differentiations in the method, the extraction of the most appropriate characterization factors (CFs) wasn't really trivial. Values with and without carbon feedbacks were given. For example, for methane, a GWP100a value of "28" was suggested if not considering carbon-climate feedback (CCFB), and a value of "34" was proposed including the carbon-climate feedback. ecoinvent at the time of releasing ecoinvent v3.1 had opted to go for the higher value with CCFB (see screenshot).

This decision has now been rectified and the implementation of IPCC2013 has been corrected.

Umberto NXT users who install the new ecoinvent v3.2 database and then update their existing LCA projects should be aware that the results of their LCIA may change, if they are using the IPCC2013 method. The characterization factors are attached to the elementary flows (e.g. methane, fossil) and thus when updating from v3.1 to v3.2, this may eventually lead to new LCIA results.

Please see the attached screenshots to see the difference for the elementary flow "methane, fossil". In IPCC2007 the GWP100a value was "25" (from AR4, 2007). In the implementation in ecoinvent v3.1 the CF was changed to "34". In the new release ecoinvent v3.2 it was changed to "29,7", based on the unrounded original value of "28,5" from source table 8.SM.17 in AR5 and an addition of "1,2" to account for the difference in impact on climate change between "methane, non-fossil" and "methane, fossil" as described by Boucher (2009). The GWP100a CF for "methane, non-fossil" is at "28,5" rather than using the rounded value "28" given in source table 8.A.1 given on page 731 of the AR5.

ecoinvent has published a 14-page report on the issue that explains in detail the reasoning for the changes in the implementation. This report [Bourgault G (2015). Implementation of IPCC impact assessment method 2007 and 2013 to ecoinvent database 3.2] is available to users logged in to the database in the 'Files' section (file 'IPCC implementation_v3.2.pdf'). Make sure you read it should you be using the IPCC2013 method in your LCA project.

Global Temperature Potential GTP20a and GTP100a characterization factors are now also given in the IPCC2013 implementation and can be used in addition to GWP.

The CFs are the ones without climate-carbon feedback (CCFB). According to the above report "the values of GWP and GTP with CCFB are not published for all gases [in the AR5]. Only the values without CCFB are available for CO, NOx, SO2, VOC and fossil methane. Until all CFs are available with CCFB, only the metrics without CCFB are implemented."

ifu Hamburg can assist you with your LCA project. Should you need help or wish to discuss the implication of the changes described above, please contact our Umberto Support Team.

Click on the below screenshots to enlarage them. The first screenshot is from ecoinvent v3.1 in Umberto NXT, the second one for ecoinvent v3.2 in Umberto NXT (currently being tested, release is imminent). Boths show the CFs for 'methane, fossil' from the IPCC2007 and IPCC2013 LCIA methods (other LCIA methods are collapsed and values not visible).
« Last Edit: December 17, 2015, 14:00 by pbeilschmidt »
Peter Müller-Beilschmidt
myUmberto Moderator
ifu Hamburg GmbH