Indicator selection in life cycle assessment to enable decision making: issues and solutions

With an ever increasing list of indicators available, life cycle assessment (LCA) practitioners face the challenge of effectively communicating results to decision makers. Simplification of LCA is often limited to an arbitrary selection of indicators, use of single scores by using weighted values or single attribute indicators. These solutions are less attractive to decision makers, since value judgments are introduced or multi-indicator information is lost. Normalization could be a means to narrow the list of indicators by ranking indicators vs. a reference system. This paper shows three different normalization approaches that produce very different ranking of indicators. It is explained how normalization helps maintain a multi-indicator approach while keeping the most relevant indicators, allowing effective decision making.

The approaches are illustrated on a hand dishwashing case study, using ReCiPe as the impact assessment method and taking the European population (year 2000) as the reference situation. Indicators are ranked using midpoint normalization factors, and compared to the ranking from endpoint normalization broken down by midpoint contribution.

Endpoint normalization shows Resources as the most relevant area of protection for this case, closely followed by Human Health and Ecosystem. Broken down by their key driving midpoints, fossil depletion, climate change and, to a lesser extent, particulate matter formation and metal depletion, are most relevant. Midpoint normalization, however, indicates Freshwater Eutrophication, Natural Land Transformation and Toxicity indicators (marine and freshwater ecotoxicity and human toxicity) are most relevant.

A three-step approach based on endpoint normalization is recommended to present only the most relevant indicators, allowing more effective decision making instead of communicating all LCA indicators. The selection process breaks out the normalized endpoint results into the most contributing midpoints (relevant indicators) and reports results with midpoint level units. Bias due to lack of data completeness is less of an issue in the endpoint normalization process (compared to midpoint normalization), while midpoint results are less subject to uncertainty (compared to endpoint results). Focusing on the relevant indicators and key contributing unit processes has proven to be effective for non-LCA expert decision makers to understand, use, and communicate complex LCA results.