0 votes

Dear all

This may come as stupid question but I just entered the Pedigree Matrix for uncertainty theme and struggling to understand the way it works. Also, I could not find any described application of it. Hence, my post here. 

For what I understood, the latest version uses normal distribution values (as opposed to lognormal as in the previous version). in the exemple hereunder, I extract values of the Pedigree matrix from "Data quality guideline for the ecoinvent database V 3.0 - 2013" 

If there is an updated version of these values, please let me know, that would be very helpful. 

Ok, so let say I collect data for a process and this process needs 10 kWh of electricity (I only have one data point, provided to me by the user guide of the machine). 

Based on the "default basic uncertainty" of the above document the "variance of the underlying lognormal distribution" is: 0.0006. 

Let say the Pedigree matrix is (2,2,2,2,2): "the variance of the underlying lognormal distribution" would be (according to the table in the above cited document): 0.0006;0.0001;0.0002;0.000025;0.0006 

If I add them all as said by the document the "variance of the summed final distribution" is 0.046. 

Now, what does that mean to my original 10 kWh data? Can I take the square root of the variance and use that as percent standard deviation of my 10 kWh data ? 10*0.046 = 0.46 --> 10 +- 0.46 (mean +- SD) ? 

Is that how it works? 

Thank you very much in advance for your time and comprehension,


in Miscellaneous by (120 points)

1 Answer

0 votes
by (92.9k points)

Hi, you find indeed updated factors in our older pedigree report, https://www.greendelta.com/wp-content/uploads/2017/03/Pedigree_report_final_May2012.pdf, or article,

Ciroth, A., Muller, S., Weidema, B. et al. (2016): Empirically based uncertainty factors for the pedigree matrix in ecoinventInt J Life Cycle Assess 21, 1338–1348. https://doi.org/10.1007/s11367-013-0670-5

In the report, you find also quite detailed explanations on how to calculate - but basically, if you "use as uncertainty value" in openLCA, openLCA does that: