Chemistry Reference
In-Depth Information
Table 1 Data sets
Country
Name and type of site
Mass fractions
Period
Belgium
(CHEMKAR)
Zelzate (ub), Borgerhout (ub)
Houtem(rb), Aarschot (rb)
PM10
September 2006-
September 2007
The Netherlands (BOP) Schiedam (ub),
Hellendoorn (rb)
PM10, PM2.5,
PM10-2.5
January-August 2008
Germany
(IUTA/LANUV)
Styrum (ub), Eiffel (rb)
PM10
April-September
2008
ub
urban background,
rb
rural background
used to estimate the contributions from specific sources like sea salt or mineral dust.
In order to make a sound comparison between the three sets, contributions were (re)
calculated using the same algorithms. The presence of
secondary inorganic aerosol
(SIA) is calculated as the sum of NO3, SO4 and NH4 and results from direct
measurements. Sulphate concentrations are corrected for a small amount of sea salt.
Carbonaceous compounds in PM contain other elements (e.g. oxygen)
augmenting the organic mass. Organic matter (OM) concentrations are usually
calculated from organic carbon, but the conversion factor remains rather uncertain.
Factors vary between 1.2 and 2 and probably vary between lower values near sources
and higher values after processing in aged air masses. Here, a factor of 1.4 is used
which seems most common in literature. Elemental carbon (EC) can be measured
straightforward. In the case of sea salt (SS), there are two tracers: sodium and
chloride. Several algorithms are in practice; one of them calculates the SS part
from both Na and Cl. However, in the case of chloride, reactions with HNO3 may
occur in the atmosphere as well as on the filter thereby releasing HCl (under the
formation of NaNO3). In addition, the presence of chloride may suffer from the
evaporation of NH4Cl (from quartz filters). For these reasons, sodium is selected as
the sole tracer: SS
3.26*Na (the factor follows from the composition of sea water).
Various algorithms are also in use to estimate the contribution of mineral dust
(MD). With MD is meant all fugitive windblown and mechanically resuspended
dust with a composition comparable to the earth's crust. Since chemical analyses of
PM samples measure elements directly, the approach here is to sum over those
elements known to be present in the earth's crust: Al, Si, CO3, Ca, Fe, K, Mn, Ti
and P [
10
]. Weights were first recalculated to correct for their oxidised form (e.g. Si
is usually present as SiO
2
). MD is a parameter difficult to estimate. The use of other
algorithms in the estimation of MD results in different values, e.g. the one
formulated by Denier van der Gon et al. [
11
]. Also, local anthropogenic sources
may contribute (e.g. metallurgical industry).
Finally, the concentrations of all constituents were summated and compared with
the gravimetrically measured mass to establish the unaccounted mass denoted
unknown. The assumption here is that the various data sets accurately describe the
PM characteristics. The filtration devices employed at the sites were all equivalent to
the reference method. However, corresponding studies use different procedures with
respect to filter handling, data treatment and selection, analytical techniques, etc. In
addition, temporal variation, local site characteristics, artefacts like volatilisation
and particle-bound water as well as varying distances from major sources will affect
ΒΌ
