Page 125 - Materials Chemistry, Second Edition

108 LIFE CYCLE ASSESSMENT HANDBOOK

study's goal and scope. In making such a choice, the following factors should
be taken into consideration:

• Physical delimitation of activities such as principal process type
(e.g., from site-specific to industry-average type) and the specific
size of the process to be modeled;
• Impact categories to be evaluated during the impact assessment;
• Technology covered;
• Time period covered;
• Geographical area covered;
1
• Cut-off rules for data, if any, are applied (these rules should pro-
vide a rationale for the significance of the various flows of the unit
process dataset);
• Provision of uncertainty information for inputs and outputs of
the process to allow for uncertainty analysis;
• Targeted databases for unit process datasets that are considered
to be high priority; and
• Intended use of the dataset in general (applications, modeling
situations including attributional or consequential modeling,
comparative assertions).

A well-defined scope helps answer questions, which, in turn, help the ana-
lyst determine the level or type of information that is required. For example,
even when the analyst can obtain actual industry data, in what form and to
what degree of specificity should the analyst show the data (i.e., the range
of values observed, industry average, plant-specific data, best available con-
trol techniques, etc.)? Recommended practice for external life-cycle inventory
studies includes the provision of a measure of data variability in addition to
averages. Frequently, the measure of variability will be a statistical parameter,
such as standard deviation (EPA 2006).

5.2.2 A Word on Consequential Life Cycle Assessment

LCA was initially developed to assess industrial systems related to consumer
products. Since then, there has been a distinct shift in applying it to larger scales
of industrial operations. By 2005, LCA practitioners began making a distinc-
tion between how LCAs that accounted for stoichiometric-like relationships
between physical flows to and from a product or process in an attributional
style, to a ones that were more encompassing of the consequences of change
in response to decisions, in a consequential LCA (Curran et ah, 2005). As a
result, the process of system expansion (to avoid or deal with the allocation

1 Non-reference product flows, waste flows, and elementary flows that can safely be labeled as
"irrelevant" can be ignored (i.e. "cut-off"). However, care must be taken to not cut off more flows
and related impacts than are acceptable to still meet the goal and scope, and the datasets used to
model a system meet the required completeness (UNEP/SETAC 2011).