Page 236 - Materials Chemistry, Second Edition

L1644_C05.fm Page 209 Monday, October 20, 2003 12:02 PM

variability in the waste input composition regarding trace elements. In comparison
with the LCI results, the uncertainties in the environmental damage estimations in
the form of external costs are higher and sum up to a geometric standard deviation
of 2.62 with filters and 2.81 without filters. These relatively high geometric standard
deviations are influenced less by the emissions (except for As with a s 3.4, for Ni
g
2.2 and for the PM 2.1) than by the impact of human health data, especially
dose–response functions for cancer (s 3) and damage factor of acute YOLL (s 4).
g
g
Moreover, monetary valuation for YOLL and cancer is an important source of
uncertainties with s of 2.1 in the same way as the dispersion model with s of 2.
g
g
To overcome incomparability related to the dispersion models, using an internation-
ally accepted reference model is proposed. Using homogeneous dose–response and
exposure–response functions approved by the World Health Organization (WHO)
could have the same effect in the future.

5.9 QUESTIONS AND EXERCISES

1. Explain the main differences between uncertainty and variability in envi-
ronmental systems analysis.
2. In which categories are the sources of uncertainty classified in tools?
3. Distinguish which of the following related situations would be sources of
uncertainties and which would be sources of variability in the tools pre-
sented in Chapter 5 of a coal power plant:
• Location of the factory
• Time of the study
• Data related to the prevalence of the winds
• Percentages of age, gender and diseases of population
• Extrapolation from animal studies to humans
• Sample sizes for animal and human studies
4. Associate the following situations in the same tool to the types of uncer-
tainties and variability listed in the column on the right side:

Use of data corresponding to emissions associated
with a former situation Uncertainty / choices
• Selection of kJ per kg of product as functional unit • Variability among sources and objects
• Use of data belonging to a system originally from a • Parameter uncertainty
geographical area different from that subject to study
• Human preferences involved in the system under study • Model uncertainty
• Use of simplified models for the calculation of factors • Spatial/temporal variability
• Presence of multiwaste processes • Uncertainty choices
5. What are the main sources of parameter uncertainty and how are they
reflected in the outcomes of an LCA?
6. Which implications for an LCIA have the nonconsideration of the model
spatial and temporal characteristics?
7. Would the use of a parameter uncertainty analysis in an LCA be suitable
when large model uncertainties have been detected?

© 2004 CRC Press LLC