ecoinvent Forum Archive

Hi gang,

I am having trouble with the ReCiPe water depletion indicator in the Ecoinvent LCIA. I have been using version 3.4 and I observed that the CF for "Water, unspecified natural origin" is 0.001, while the CFs for all the other natural water sources was 1.0.  The value of 1.0 makes sense, because both the LCIA indicator and the inventory flows are measured in m3. Why is the "unspecified natural origin" only valued at 0.001?

I thought it might be an error so I downloaded the version 3.6 LCIA indicators and I see ReCiPe has been updated to version 1.13 - that's great, except all the water depletion CFs are gone in this version! All we have is water emissions, to air, which are given CFs of 1.0. Is this an error? Or am I somehow mistaken?

Thanks in advance

Dear Brandon,

To your first question, why the characterization factor for the midpoint indicator “water depletion, unspecified natural origin” is 0.001: This is because some elementary exchanges are in kg while others are in m3. If you check the documentation on CFs for ReCiPe (2008, Page 107) the unit for unspecified natural origin is given in m3/kg, thus the value of 0.001.

To what happened to the the water depletion indicator: Because the ecoinvent database is not completely water balanced, the best approximation for the water depletion indicator is to use water-to-air exchanges, as these are the main contributors to the indicator and deliver the best results. We are aware that this is not a perfect solution, but it is intentional and no error.

Kind regards,

Johannes MüllerJunior Data Analyst, ecoinvent

Johannes,

Thanks for the reply. However, I do think there are error in both cases.

It may be that ReCiPe measures water from undisclosed natural sources in kg, but ecoinvent does not. In the MasterData/ElementaryExchanges.xml the reference unit for "Water, unspecified natural origin", which has UUIDs 831f249e-53f2-49cf-a93c-7cee105f048e (resources/in water) and 478e8437-1c21-4032-8438-872a6b5ddcdf (resources/in ground), is clearly stated to be m3. I am looking at 3.6 cutoff but it has been true for years. 

In the second case, that seems quite surprising, and it is certainly not consistent with ReCiPe. Can you provide any support for this? how was it disclosed?

-Brandon

Johannes,

I see two things- (1) the ReCiPe 2008 method is now obsolete so data errors are no longer corrected and (2) water withdrawals are correctly measured with the "Selected LCI results, additional, resource, water" indicator.

I would still be interested to see documentation of the new air-transpiration-as-water-depletion indicator. 

-Brandon

Dear Brandon,

You are currently bumping your head on an aspect of the issue of LCI-LCIA interface.  Some LCIA indicators have been developed with little consideration about the characteristics of the LCI coming out of a software. 

LCI’s are not water balanced.  This is the result of some of our datasets not being balanced, but also because of the allocation procedure.  It is not possible to conserve mass balance with allocation, unless you are using a mass-based allocation factor.  So if you simply apply positive factors to water consumptions, and negative factors to water emitted back to water or soil, you will get unreliable results. 

The best solution we found at ecoinvent is to use water to air as the closest approximation of the water no longer available to the ecosystem.  In the absence of a large-scale water balance project and a mass-correction algorithm, this was the quickest think we could do to give meaningful results.  We are open to other suggestions. 

Best,

Guillaume Bourgault, project manager at ecoinvent