In the second stage of uncertainty analysis, the parameters which had a high sensitivity based on the perturbation analysis
were used in a Monte Carlo simulation. The Monte Carlo was implemented on top of the matrix LCA in excel using Simulacion 4.0 add-in, with 2000 samples. As is common practice in LCA, the input variable distributions were assumed to be lognormal and the spread of the distribution was estimated using a pedigree matrix (Frischknecht et al., 2005; Huijbregts, 2001; Lloyd and Ries, 2007).