Two-dimensional Monte Carlo simulations in LCA: an innovative approach to guide the choice for the environmentally preferable option

Uncertainty and variability need to be taken into account in life cycle assessment (LCA) studies to make robust decisions. We introduce a novel approach in LCA that allows to decide if either uncertainty or variability is dominating in the results: two-dimensional Monte Carlo simulations (2DMC). We aim to do so in a pedagogical and transparent way, allowing interested readers to fully grasp all technical details for their own potential use in future studies.

In 2DMC, an approach from quantitative risk assessment, the model parameters are divided into four categories: deterministic, variable, uncertain, and uncertain ánd variable; and appropriate distributions are selected. These distributions are sampled separately, so they can be assessed separately in the output as well. Firstly, the approach was translated to the LCA context with an illustrative proof of concept model, freely available on our website. Further, two variants of the post-harvest apple chain in Belgium (bulk versus pre-packed) are worked out as a real life comparative LCA case study. This real-life case study is also analyzed in a classical, deterministic way and by performing a more often used one-dimensional Monte Carlo simulation (1DMC), allowing a comparison with the 2DMC results and associated interpretations.

Deterministic results do not reflect the complexity of reality. 1DMC results provide an indication on the robustness and conclusiveness of the result of a comparative LCA, but do not provide a way to guide further decisions. 2DMC results do provide this as results typically belong to one out of three possibilities. Firstly, the 2DMC results may confirm the result of the deterministic results. Secondly, the 2DMC curves may show proof that the two products are equivalent when it comes to environmental impact. One may then decide to analyze the variability causes further or that other reflections, like cost, should be considered as well. Thirdly, the 2DMC curves may indicate that more detailed and accurate information is needed to come to conclusive results.

1DMC results give a first indication on the need for a 2DMC analysis. If that is the case, 2DMC can be used in a comparative LCA to take uncertainty and variability separately into account. 2DMC results can guide decisions to obtain more conclusive results. We recommend to consider a 2DMC analysis when comparing two products or processes if needed, hereto, our proof of concept model fully documented available online may be a starting point.