Page 331 - Materials Chemistry, Second Edition

L1644_C07.fm Page 299 Monday, October 20, 2003 12:10 PM

As performed by Nigge (2000), the calculation of the effective emission height
h , which includes the physical stack height as well as information about stack
eff
temperature and volume flow, may resolve these problems of uncertainties. For this
reason, h is provided in this study as well. If the impact indicators are given for
eff
h , the results may be applicable to all kinds of industries under the condition that
eff
the stack height, volume flow and temperature are known. Moreover, it should be
stated that a large variety of algorithms can calculate the effective emission height
for different purposes. Generally speaking, it seems to be difficult to find one
procedure of resolving these problems in defining source characteristics. This
requires further research.
Another important source of uncertainty for the numerical results of the site-
dependent human health impact factor for Catalonia is the fact that the models
ISCST-3 in BEEST (short-range) and WTM in EcoSense (long-range) may be in
poor accordance in terms of dispersion. As stated earlier, the dispersion results for
BEEST are highly sensitive to the source characteristics. Moreover, these charac-
teristics must be chosen so that the results at the outer boundary comply with the
results calculated by the EcoSense modeling area that begins there. This is especially
difficult because the EcoSense program calculates concentration increments for the
grid cells as a whole, i.e., the actual concentration at the modeling boundary with
the BEEST program cannot be determined. This means that, if the source charac-
teristics for BEEST are chosen “wrong,” the overall outcome of I total would be
erroneous because the EcoSense model is not adapted to the source characteristics
introduced in BEEST.
A related problem is that higher wind speeds usually lead to lower values of
I near . One may suppose that at least parts of the substances are therefore going into
long-range transport covered by the Eurogrid of EcoSense. Thus, I should increase
far
but EcoSense is not able to take this problem into account because it provides a
coherent set of meteorological data for every grid cell in Europe and is usually not
applied in combination with other models. This leads to lower values of I total for
higher wind speeds, which is accepted because one may suppose that the population
density outside Catalonia is smaller. For this reason an enhanced long-range transport
due to higher wind speed leads to lower values of I total, because the increase in long-
range exposure is smaller than the decrease in short-range exposure. However, this
is only a theoretical reasoning because no change in I can be observed.
far
The same problem applies to the stack height. Although the height has a great
influence on the outcome for the short-range modeling (because it is assumed that
the higher the stack, the more pollutants go into the long-range transport), the results
for long-range modeling are quite insensitive to stack height. The longer the atmo-
spheric residence time of the pollutant is, the greater the uncertainties to which
problems related to wind speed and stack height lead. A pollutant with a short
atmospheric residence time is deposited and decays mostly in the short-range mod-
eling area, while a pollutant with a longer atmospheric residence also accounts for
a significant long-range exposure and is therefore subject to greater uncertainties
with respect to wind speed and stack height.
Moreover, the local population exposure subtracted from the population exposure
in the EcoSense area cannot be compared to the results of I near because I near is much

© 2004 CRC Press LLC