Parametrization to solve the issue of correlation in uncertainty and sensitivity analysis in life cycle assessment

Abstract Purpose Several papers that address the topic of uncertainty analysis briefly mention the fact that the correlation between uncertain inputs is ignored. Indeed, there are several reasons why including correlations in large-scale life cycle assessment (LCA) studies is challenging. Yet, it would increase the realism of such studies, and it may also lead to changed conclusions. Here, following several suggestions in the literature, we address the topic with the technique of parametrized LCA. Methods We develop a general computational framework for parametrized LCA, and show how it can also incorporate stochastic variables. A simple system illustrates the ideas, also in a comparative setting with more traditional uncorrelated and correlated uncertainty analyses. Results and discussion In addition to uncertainty analysis, the parametrized approach can be further extended to perform sensitivity analyses, of different kinds, and to perform uncertainty apportioning. This is also illustrated in a comparison with the non-parametrized model. Conclusion Parametrized LCA is an important form of LCA which has so far been primarily applied, without a proper presentation of the formal structure. It appears to be especially useful in the context of uncertainty analysis, sensitivity analysis, and uncertainty apportioning.