Chemistry Reference
In-Depth Information
longer periods (up to several years) and multiple sites. The obtained trajectory
cloud can be further processed by means of statistical methods like clustering [
22
]
to identify the most relevant types of air mass transport to the sites or areas under
consideration.
Moreover, advanced evaluation methods exist to apportion PM levels measured
at the receptor site to the trajectory segments (e.g. the residence time weighted
concentration method [
23
]) or to regions (grid cells) hit by trajectory ensembles
(e.g. the potential source contribution function, PSCF [
24
,
25
]). Software tools are
available which facilitate such calculations and visualisations of the results [
26
].
3 Source Apportionment Studies Carried Out in Germany
3.1 Overview
Source apportionment studies have been carried out since the late 1990s in a
number of the German Federal States (see Table
1
and Fig.
4
). Source attributions
were done by chemical analysis and tracer assignments in basic approaches, often
extended by basic Lenschow analyses (spatial increment calculations); the full
Lenschow approach including emission register assignments has been used less
frequently. Also receptor modelling using multivariate factor analyses (either PCA
or PMF) were applied several times, sometimes supported or complemented by
results of dispersion or chemical transport models and back-trajectory analyses.
In Fig.
4
an overview of the “chemical composition”, already grouped according
to source processes, of PM10 in Germany is shown. The unaccounted PM mass in
these mass closures is frequently around 10-20% which can reasonably be assigned
to particle bound water and analytical uncertainties. Higher unaccounted mass
fractions in most cases also comprise the mineral dust fraction since analyses of
silicon and aluminium are scarcely done in routine. In some studies only the main
ionic PM constituents were analysed, and the unaccounted mass then became the
major fraction and comprises carbonaceous particles as well.
Apparently there is only a limited spatial variability in the overall chemical
composition. This indicates that a large fraction of PM, in particular the secondary
inorganic and organic aerosol, is formed and dispersed on a rather large spatial scale
and can be found everywhere. The increase of PM concentrations in urban back-
ground compared to rural conditions thus can be (at least partly) explained by less
favourable atmospheric dispersion conditions within the cities leading to accumu-
lation of the particles advected from the regional surroundings. However, in case of
influences by strong local sources e.g. high-trafficked street canyons, a clear shift in
the chemical profile towards higher shares of related compounds (carbonaceous
aerosols and metals) can be seen.
