Discrepancy in Water Consumption (ReCiPe 2016) Results Between openLCA and ecoinvent

Question

Hello everyone,

I’m working with ecoinvent 3.10.1 datasets in openLCA to assess water consumption according to ReCiPe 2016 Midpoint (H). However, I’ve noticed a discrepancy between the ecoinvent LCIA methods implementation and the openLCA LCIA methods implementation for several landfill processes:

• treatment of waste polyethylene terephthalate, sanitary landfill | waste polyethylene – RoW
• treatment of waste polyethylene, sanitary landfill | waste polyethylene | Cutoff, U – RoW
• treatment of waste polypropylene, sanitary landfill | waste polypropylene | Cutoff, U – RoW

When I use the openLCA LCIA method for water consumption (ReCiPe 2016) in these datasets, I obtain negative results (see image on the left). On the other hand, when I switch to the ecoinvent LCIA method for water consumption (ReCiPe 2016), I get positive results (image on the right).

I suspect this is because the ecoinvent method does not include characterization factors for water outputs (thus, no subtraction from the total), while the openLCA method does, causing a net negative water consumption if outputs exceed inputs.

My questions:

1. Can anyone confirm that this difference (negative vs. positive water consumption) stems from the different treatment of water outputs in these two methods?
2. Which of these results is considered the “correct” representation of water consumption in ReCiPe 2016?

Any insights or guidance would be greatly appreciated. Thank you!

Answer 1

https://ask.openlca.org/8073/water-use-recipe-calculation-with-ecoinvent

Comments

In case you do not look at regionalized water use (which is then only a water use amount indicator multiplied by 42.95) maybe the ecoinvent approach is ok because it will not have allocation distortions from water inputs in processes that are allocated.

In case you are looking at regionalized water use impacts (which makes sense if water use is an impact category of focus) you cannot use the ecoinvent approach, since you need a regionalized impact method.

Answer 2

Hi, did you set the impact direction for the categories? As this influences the sign.

Then, you can see, when you expand the view and better when you have first flows and then processes (so when you check "flow" in the sheet you are showing) how much each flow contributes, negative or positive. But we will check here as well.

Comments

Hi Andreas, thanks so much for the feedback and suggestions.  I’ll double-check the impact direction for each category and look at the detailed contributions of individual flows.

Thanks again for taking a look at this!