Page 79 - Materials Chemistry, Second Edition

64 M.Z. Hauschild

6.3 The Iterative Nature of LCA

In Fig. 6.1 a number of arrows indicate that rather than a linearly proceeding
process, LCA involves many feedback loops between the different phases of the
LCA. Insights from the impact assessment are used in refining the inventory
analysis and insights from both of these phases may feed back to the scope defi-
nition, e.g. in the setting of the boundaries of the product system, what to include
and what to exclude. Sensitivity and uncertainty analysis are thus not just per-
formed in the interpretation at the end but throughout the study as part of both
inventory analysis and impact assessment in order to identify the key figures or key
assumptions of the study and the data that are associated with the largest uncer-
tainties (see Chap. 11). Each phase of the methodology provides feedback to the
previous phases of the study and helps target the next iteration of the LCA. The best
precision is obtained with minimum work effort if the focus is on improving the key
figures wherever possible and needed, and on reducing the largest uncertainties.
In practice, the first iteration will often be a screening that covers the full life
cycle, but in terms of inventory data largely is based on easily accessible data from
available databases. Following the impact assessment, the parts of the product
system that contribute most strongly to the total results can be identified, and the
chosen boundaries of the product system can be tested. As a consequence, the
scoping may have to be refined. The impact assessment results also allow identi-
fying those inventory data or assumptions made in the inventory analysis that have
the largest influence on the overall results or for which the uncertainties are so large
that they potentially could be key figures. These data should be the target of the
next iteration, where effort should be focused on testing and refining these
assumptions or data and get more representative or recent data. Based on the revised
inventory a new impact assessment is performed, and the sensitivity analysis is
performed once more to see which are now the key figures and key assumptions.
Large uncertainties may also accompany the factors applied in the characterisation
of some of the inventory flows in the impact assessment, and if the sensitivity
analysis indicates that such uncertainties may have a decisive influence on the
results, these factors will also be the target of a consecutive iteration. Figure 6.2
illustrates the iterative approach to performing an LCA.
As illustrated by the narrowing spiral in Fig. 6.2, the uncertainty of the LCA
results is reduced through the repeated iterations, and these are carried on until the
remaining uncertainty of the results is sufficiently small to meet the goal of the
study. If the goal is to identify which among several alternatives has the lowest
environmental impacts, the number of needed iterations may be low if the alter-
natives show large differences in their impacts, while a higher number of iterations
will be needed if the alternatives are more similar. An LCA performed to support an
environmental product declaration with a general requirement to the uncertainty of
the impact scores can require a high number of iterations before all impact scores
are determined within the stipulated level of uncertainty.