Page 294 - Materials Chemistry, Second Edition
P. 294
280 R.K. Rosenbaum et al.
of (unquantified or unperceived/unconscious) uncertainty due to their lower envi-
ronmental relevance compared to endpoint indicators. As depicted in Fig. 11.4,
overall uncertainty may increase or decrease from a midpoint to an endpoint
indicator of a given impact category, depending on the uncertainty of models and
parameters used for endpoint modelling.
However, Weidema (2009) pointed out that this “figure implies that it is possible
to make a trade-off between relevance and uncertainty, in which the overall error is
minimised … [and] … that the consequences of the decision will be less uncertain if
the decision is taken at the point where the overall error is minimised—that is, at a
midpoint […] (e.g., at the level of CO 2 -equivalents)”, which is a common per-
ception among LCA practitioners and clients. Weidema then rightfully argues that
“When the decision is implemented, however, the consequences occur not only at
the level of the midpoint but also at the level of the endpoint (the decision will
result in lost species and lost lives). This implies that the apparently low uncertainty
of the decision at midpoint does not reduce the uncertainty of the consequences of
the decision at endpoint level, which are still as uncertain as indicated at the bottom
of [the] figure […]. If the consequences at endpoint level (e.g., lost species and lost
lives) are what we really are interested in (as implied by the maximum relevance),
then taking the decision at the midpoint level (e.g., CO 2 -equivalents) is simply the
same as ignoring the true uncertainty of the consequences of the decision.” In other
words, if minimal or avoided environmental consequences are the objective of a
Fig. 11.4 Conceptual representation of how overall uncertainty may decrease (middle) or increase
(right) from midpoint to endpoint (damage) in an impact pathway (left); uncertainty of
interpretation and uncertainty of models and parameters contribute to different extents to overall
uncertainty on midpoint (early in the impact pathway) and on endpoint/damage level (end of the
impact pathway) while environmental relevance increases [taken from Hauschild and Potting
(2005)]