Why do Monte Carlo results greatly disagree with Quick results?

Question

I ran a process for 1 m3 of concrete and I get a 'Quick result' of 647 kg CO2eq/m3. When I run 10 Monte Carlo iterations I get a range of results with mean = 363 kgCO2eq , median = 361 kgCO2eq , 5%tile = 342 kgCO2eq , 95%tile = 401 kgCO2eq. I tried more iterations and I get a similar range of results.

Why would the 'Quick result' (which makes more sense) be so vastly different than the Monte Carlo results?

Comments

Hi Geoff, I do not have  a straightforward answer to this but another user experienced a similar issue: https://ask.openlca.org/1186/monte-carlo-giving-widely-different-results-to-analysis

But of course in theory, the results can differ a lot if the uncertainty parameters are very high.

Answer 1

Thanks Jonas, I found my issue and it was an oversight on my part. I did not realize the geometric mean doesn't get automatically updated to the input in the Amount column. Since I copied the process and then changed the Amount input, I did not, in turn, change the Gmean in the uncertainty column. Wouldn't it be best to auto-update to the deterministic Amount entered? I see gsigma is updated automatically when I change the Data quality entry column.

Comments

So is this a bug or am I missing something?