Page 330 - Materials Chemistry, Second Edition

L1644_C07.fm Page 298 Monday, October 20, 2003 12:10 PM

chosen because data are usually available for these administrational units; population
data are available for municipalities, districts and provinces.

7.7.7.3 Fate and Exposure Assessment
Comparing the parameter uncertainty and variability involved in the LCI in relation
to the impact pathway analysis (IPA; see Chapter 5) shows that uncertainties in the
fate and exposure assessment as well as in the consequence/effect analysis are more
important than those in the LCI analysis. This is due to the dose–response and
exposure–response functions, weighting schemes, and fate models. Examples of fate
models used in this study include dispersion software, multimedia models and long-
range transport programs (as included in integrated impact assessment models).
The program BEEST, as well as EcoSense, was primarily developed for the
calculation of concentration increments due to power plant emissions. This work,
however, applies these programs also for lower stack heights and volume flows (in
the case of BEEST) in order to make the method applicable to a wider range of
industrial process chains. In this case particularly, in order to include transport
processes, several assumptions are needed. Unfortunately, the uncertainties cannot
be quantified.
In general, the uncertainties related to Gaussian dispersion models are important;
they are rather simple descriptions of quite complicated natural processes. Because
substance data influence the dispersion of pollutants, uncertainties in these data are
directly introduced into the results of the damage estimations. Another major source
of uncertainty related to the fate models used is the choice of the modeling area and
the grid size of the dispersion programs. In the long-range transport program WTM,
the overall modeling area is confined to Europe (Eurogrid as implemented in
EcoSense); a contribution to the population exposure from outside is neglected. An
outside contribution is related to North Africa, which lies close to Spain, but which
is not included in the Eurogrid. The outside contribution cannot be quantified,
however. Because the long-range exposure does not depend very much on local
variations, the resolution of 100 ¥ 100 km seems to be appropriate for the use of

Gaussian dispersion models. The grid for BEEST calculations is chosen so that the
proximity of the stack (where the concentration increment is very sensitive to the
distance from the stack) is described quite well without increasing the uncertainties
for greater distances from the stack up to 100 km.
In the case of the site-dependent impact assessment study, with respect to the
BEEST program, it has been seen that the outcome for I near is highly sensitive to the
volume flow and the mass flow chosen. Although EcoSense allows the introduction
of any mass or volume flow and still leads to the same results for I , the volume
far
flow and the mass flow must be determined carefully for BEEST calculations. The
volume flow is derived as a function of the stack height from the evaluation of
several industrial processes. The mass flow is derived from the volume flow and the
thresholds for emissions of waste incinerators that apply in Catalonia. It must be
stated, though, that the assumption for the volume flow is based only on a small
number of industrial processes and that the used threshold is a political value.

© 2004 CRC Press LLC