LCA result Validation

Question

Hello. Can anyone provide good link or any other resources related to validation of LCA results?
I have tried to do LCA with 3 different cases. Most of the inventory used was obtained from ecoinvent database 3.6. It give results,yes. But how will I know that this result is true?

Thanks in advance.

Answer 1

Hello, this is a really good question, not easy to answer. Basically, the review, of an LCA study or also of individual datasets, tries to perform this check whether the results obtained are "good". You can do this yourself, but typically it is useful to ask someone else to perform this check.

Basically, you can compare to what you expect (e.g., a coal power plant has high GHG contributions), to other, similar studies, etc.. In openLCA you have the possibility to check elementary flows that are not contributing to specific LCIA categories, and find those that maybe should contribute, and then adapt the LCIA method if needed. And so on.

If you want you can even have a look at my old editorial about validation in LCA: Ciroth, A. Validation – The Missing Link in Life Cycle Assessment. Towards pragmatic LCAs. Int J Life Cycle Assessment 11, 295–297 (2006). https://doi.org/10.1065/lca2006.09.271

Comments

Thank you for the response and providing additional journal. However, from what I understand in the text, at the moment, validation of lca result is currently not possible. The recommendation is that each material/processes inputted in the boundary system must be individually checked.

In the response you sent me: "You can do this yourself, but typically it is useful to ask someone else to perform this check.", is this referring to pidegree matrix?
If yes, what is the acceptable GSD (sigma g)?
I observed that some of the data I used in my cases study (Database from ecoinvent) has  5:5:5:5:5 values with sigma g of 2.51, does this mean the input data is not reliable?
What does it signify?
What sigma g value is acceptable?

Thank you so much

---

My editorial was taking a very high level view, and says that in the end, strictly speaking, noone can validate LCA results in the sense that you see "in reality" how the environmental impacts over the life cycle of a specific product are (and compare this with the calculated / modeled result. That said, there are practices to check if datasets or studies are correct, compliant with ISO for example, and these work. I would not make the intitial evaluation too technical (so not start with a specific GSD) but it is often first important to see whether flows are as you expect etc., or, for studies, impact assessment results. Also here, for datasets, I can refer to a an article: Ciroth, A., Foster, C., Hildenbrand, J. et al. Life cycle inventory dataset review criteria—a new proposal. Int J Life Cycle Assess 25, 483–494 (2020). https://doi.org/10.1007/s11367-019-01712-9.
To your last point: A five time five quality assessment is of course not good, but this assessment reflects the ecoinvent default assessment (fitting to the goal and scope of ecoinvent) which can be different from goal and scope of your study. Not for reliability of the source, typically, but for location and time, possibly.