Interpretation of comparative LCAs: external normalization and a method of mutual differences

Abstract

PurposeIdentification of environmentally preferable alternatives in a comparative life cycle assessment (LCA) can be challenging in the presence of multiple incommensurate indicators. To make the problem more manageable, some LCA practitioners apply external normalization to find those indicators that contribute the most to their respective environmental impact categories. However, in some cases, these results can be entirely driven by the normalization reference, rather than the comparative performance of the alternatives. This study evaluates the influence of normalization methods on interpretation of comparative LCA to facilitate the use of LCA in decision-driven applications and inform LCA practitioners of latent systematic biases. An alternative method based on significance of mutual differences is proposed instead.MethodsThis paper performs a systematic evaluation of external normalization and describes an alternative called the overlap area approach for the purpose of identifying relevant issues in a comparative LCA. The overlap area approach utilizes the probability distributions of characterized results to assess significant differences. This study evaluates the effects in three LCIA methods, through application of four comparative studies. For each application, we call attention to the category indicators highlighted by each interpretation approach.Results and discussionExternal normalization in the three LCIA methods suffers from a systematic bias that emphasizes the same impact categories regardless of the application. Consequently, comparative LCA studies that employ external normalization to guide a selection may result in recommendations dominated entirely by the normalization reference and insensitive to data uncertainty. Conversely, evaluation of mutual differences via the overlap area calls attention to the impact categories with the most significant differences between alternatives. The overlap area approach does not show a systematic bias across LCA applications because it does not depend on external references and it is sensitive to changes in uncertainty. Thus, decisions based on the overlap area approach will draw attention to tradeoffs between alternatives, highlight the role of stakeholder weights, and generate assessments that are responsive to uncertainty.ConclusionsThe solution to the issues of external normalization in comparative LCAs proposed in this study call for an entirely different algorithm capable of evaluating mutual differences and integrating uncertainty in the results.