Negative Impact Scores with USDA's NREL database?

Question

This is a similar question to the one posed as "Why do I get negative results in all impact categories?", however it deals with the LCA Commons database from USDA.

Are you aware of issues using the "Auto-link processes" function with NREL data? 

As an example, I created a simple product system with 1 input (1MJ of Electricity, anthracite coal, at power plant - RNA). If I auto-link the processes when generating the product system, many impact assessment categories have negative values. 

Any tips would be great. Thanks!

Comments

In digging deeper into the auto-linking, the issues seems to be with the Process: "Crude oil, extracted - RNA"; once its supply chain is removed, the emissions are positive.

When the supply chain for the process is manually added, the issue occurs with the process: Residual Fuel Oil, combusted in industrial boiler - RNA

---

Hi, I seem to have similar issues. How did you figure out which process has to be removed? Also, are we supposed to remove a process just like that? Thanks in advance

---

To troubleshoot where the error was occurring, I unticked the auto-link function when generating the product system; I manually linked them.

From my generated Product System, I then used the "Build Supply Chain" tool to construct it one Tier at a time. After a new tier was created, I "Calculated" to see what the Flows looked like.

Once the tier generating negative flows was found, I investigated each of the processes in that tier to identify the culprit.

As was mentioned elsewhere in this post - make sure you're using the most up-to-date version of USLCI.

---

Hi, Thanks a lot for your reply. So I did the same as you said and figured out for which upstream process I was getting negative results and once I remove it from the supply chain, I can see the positive results.
Now my concern is what does it imply by removing the supply of an upstream process? I mean does it somehow alter the boundary that I have chosen for my study? I am sorry if my question sounds silly, I do not have an extensive understanding of this.

Answer 1

Make sure you are using the latest version of the US LCI. I believe this issue has been resolved.

https://github.com/uslci-admin/uslci-content/blob/dev/docs/release_info/press-release.md#2020-summer-quarter-scheduled-june-30-occurred-july-14

Comments

I downloaded the USLCI from the Federal LCA Commons in January 2022 and am still seeing negative results as reported in these comments.

Answer 2

@byoung is correct; this issue was resolved with the corrections released as of the USLCI 2020 (Summer Quarter) version. 

The Nexus paid-for version does not reflect this correction. If you use the latest version of USLCI from the Federal LCA Commons, you will also need methods adapted to the Federal Elementary Flow List (FEDEFL), to which the USLCI converted in 2019. The FEDEFL and the FEDEFL-adapted methods are available on the Federal LCA Commons for download and import, e.g., into your openLCA desktop application. Please see the USLCI GitHub Support Content (https://github.com/uslci-admin/uslci-content/wiki) and/or the NREL USLCI Quick Help YouTube video series (https://www.youtube.com/playlist?list=PLmIn8Hncs7bFUOyXZNGXwG4LtdoTfLz6Q) for details on USLCI Database versioning, using FEDEFL-adapted methods, and running calculations on USLCI in openLCA.

Alternatively, you can manually implement relevant corrections from the USLCI Change Log into the Nexus paid-for verison: https://github.com/uslci-admin/uslci-content/blob/dev/docs/release_info/change-log.md.

Comments

Thank you Rebe - it is a bit embarassing for us of course and we should try to get the versions faster on Nexus!