Stochastic multi-attribute analysis (SMAA) as an interpretation method for comparative life-cycle assessment (LCA)

Purpose

Comparative life-cycle assessments (LCAs) today lack robust methods of interpretation that help decision makers understand and identify tradeoffs in the selection process. Truncating the analysis at characterization is misleading and existing practices for normalization and weighting may unwittingly oversimplify important aspects of a comparison. This paper introduces a novel approach based on a multi-criteria decision analytic method known as stochastic multi-attribute analysis for life-cycle impact assessment (SMAA-LCIA) that uses internal normalization by means of outranking and exploration of feasible weight spaces.

Methods

To contrast different valuation methods, this study performs a comparative LCA of liquid and powder laundry detergents using three approaches to normalization and weighting: (1) characterization with internal normalization and equal weighting, (2) typical valuation consisting of external normalization and weights, and (3) SMAA-LCIA using outranking normalization and stochastic weighting. Characterized results are often represented by LCA software with respect to their relative impacts normalized to 100 %. Typical valuation approaches rely on normalization references, single value weights, and utilizes discrete numbers throughout the calculation process to generate single scores. Alternatively, SMAA-LCIA is capable of exploring high uncertainty in the input parameters, normalizes internally by pair-wise comparisons (outranking) and allows for the stochastic exploration of weights. SMAA-LCIA yields probabilistic, rather than discrete comparisons that reflect uncertainty in the relative performance of alternatives.

Results and discussion

All methods favored liquid over powder detergent. However, each method results in different conclusions regarding the environmental tradeoffs. Graphical outputs at characterization of comparative assessments portray results in a way that is insensitive to magnitude and thus can be easily misinterpreted. Typical valuation generates results that are oversimplified and unintentionally biased towards a few impact categories due to the use of normalization references. Alternatively, SMAA-LCIA avoids the bias introduced by external normalization references, includes uncertainty in the performance of alternatives and weights, and focuses the analysis on identifying the mutual differences most important to the eventual rank ordering.

Conclusions

SMAA-LCIA is particularly appropriate for comparative LCAs because it evaluates mutual differences and weights stochastically. This allows for tradeoff identification and the ability to sample multiple perspectives simultaneously. SMAA-LCIA is a robust tool that can improve understanding of comparative LCA by decision or policy makers.