Chemistry Reference
In-Depth Information
models have to be completed by an assessment of regional background levels and
evaluations/modelling of small scale local processes (e.g. [
4
]). Dispersion modelling
on the urban scale is possible using special Gaussian plume models which may
include canyon-plume-box approaches for street canyons (as implemented, e.g. in
the IMMIS - air quality models [
14
,
15
], a widely applied tool for air quality
assessment in Germany)
Even higher computation capacities are needed for chemical transport models
(CTMs) which extend the pollution transport by modules simulating the atmo-
spheric multiphase chemistry and aerosol physics. CTMs are usually operated in a
nested mode, with large grid cell sizes (e.g. 125 km
125 km) on hemispheric or
European scale and small grids for the region of interest (down to 1 km
1 km).
Several CTMs have been developed in Europe, in Germany mainly REM-Calgrid
[
16
] and EURAD [
17
] have been used. Still, CTM models appear to have problems
with accurate modelling of PM concentrations, particularly at high PM levels [
18
].
Nevertheless they may be used for assessing contribution of sources, in the most
simple case by comparing model runs with the emissions of a target source category
either switched on or off. Considerable computational resources are needed since
these runs have to be done sequentially (or on parallel systems). To overcome this
situation methods have been recently developed [
19
] to “tag” or “label” emitted
compounds in such models and hence to detect the various contributions in the
modelled PM concentrations. The latter approach may also be viewed as more
realistic since it also includes all non-linear effects.
2.6 Back-Trajectory Modelling
Source apportionment studies are frequently complemented by air mass origin
analyses to identify where high pollution levels or certain compounds of interest
come from. For small to meso scales this can be done by, e.g. wind rose analysis,
whereas long-range transport is better assessed by using back trajectories calculated
by a suitable dispersion model [
20
].
Most of the CTM models mentioned in the previous section allow back-
trajectory calculations; however, access is possible only through the research
groups operating these models. Therefore, the most widely used tool is the open
access NOAA Hysplit model [
21
].
In a basic approach for each measurement day several trajectories are calculated
varying in the time of arrival and height above ground level for backward periods of
usually 3 or 4 days. Such single-day calculations have been proven to be particularly
useful in case of short-term dust events caused, e.g. by long-range dust intrusions
from arid regions or wildfires. It should be noted that due to spatial resolution limits
(typically grid size is 1
1
) back trajectories cannot be used for the identification
of sources located close to the receptor site.
To get a temporally more representative picture of the distant regions associated
with, e.g. episodes with high PM levels computation of trajectory data is needed for
