Uncertainty due to methodological choices (e.g. when defining the system boundaries, allocation vs. system expansion) has not been addressed in this study, because it is considered to be beyond the scope. However, the uncertainty in the estimation of environmental impacts associated with LCI raw input data variances was assessed.
In the SimaPro LCI databases, the quantities of required raw materials are stated for each basic process. Paired with each data entry is an estimate of the standard deviation in these raw material quantities. These standard deviations are used to define the range of uncertainty in each quantity. Monte-Carlo simulation, which is a function built into SimaPro 8.0 (LCS 2014), is used to propagate data base value uncertainty to overall uncertainty across each environmental impact category. In the Monte-Carlo approach, 10,000 runs using random LCI data, generated within 95% confidence interval for each raw material input quantity, are calculated. Uncertainty distributions for each overall impact category are derived from these results. The uncertainty analysis performed is just a first approximation to a more robust analysis (i.e. using primary data) that would more substantially improve the assessment reliability. However, preliminary analysis demonstrates that the uncertainty in the comparative assessment is negligible.