Page 206 - Materials Chemistry, Second Edition

192 R.K. Rosenbaum et al.

product system. As a general principle, the larger the reference system, the lesser
the risk of such bias when normalising against the background activities of
society.
• While supporting comparison of results across impact categories, normalised
LCA results cannot be interpreted as reflecting a weight or importance of one
impact category relative to others. Normalisation helps to identify the impacts
from the product system that are large compared to the chosen reference system,
but large is not necessary the same as important. It is therefore not suitable as the
only basis for identification of key issues/impacts in a product system, unless
explicitly required by the goal and scope definition (e.g. evaluating the envi-
ronmental impact contribution of a product system to a reference system which
it is part of).
• Unless (a) the reference system is global or (b) all environmental interventions
of the product system assessed take place in the same region as those of the
reference system, the direct interpretation of normalised impacts as contributions
to or fractions of the reference system is misleading because parts of the life
cycle of the product or service take place in different regions of the world,
including outside the reference system.
By expressing the different impact scores on a common scale, normalisation can
also help checking for potential errors in the modelling of the product system. If the
results are expressed in person equivalents, it is possible to spot modelling errors
leading to extremely high or low impacts in some of the impact categories—like
frequent unit errors when emissions are expressed in kg instead of g. Looking
across the impact category results in a normalised impact profile, it is also possible
for the more experienced LCA practitioner to check whether they follow the pattern
that would be expected for this type of product or service.
Although characterisation and aggregation at endpoint level leads to fewer im-
pact scores (typically three), normalisation may still be useful with the same pur-
poses as normalisation at midpoint level. The calculation and application of the
endpoint normalisation references follows the same procedure as for midpoint
normalisation, just applying combined midpoint and endpoint characterisation
factors in Eq. 10.2.

10.3.2 Weighting (and Aggregation)

Weighting can be used to determine which impacts are most important and how
important they are. This step can only be applied after the normalisation step and
allows the prioritisation of impact categories by applying different or equal weights
to each category indicator. It is important to note that there is no scientificor
objective basis for this step. This means that, no matter which weighting method or
scheme is applied, it will always be based on the subjective choices of one person
or a group of individuals. Weighting can be useful for: