Source Apportionment of Airborne Dust in Germany: Methods and Results - Urban Air Quality in Europe - page 200

Fig. 3 Statistical receptor model categories and specific models ( italics / dashed arrows ) (reprinted
from [ 1 ] as modified from [ 12 ] with permission from Elsevier)
contributions based on data intrinsic correlations between the different compounds.
Trace compounds (heavy metals, polycyclic aromatic hydrocarbons, organic tracers
like levoglucosan) in the obtained source profiles, that is a source group-specific
chemical composition of PM, are important auxiliary indictors for source identifi-
cation. Temporal variation of factors and tracer compounds (seasonal, weekday/
week-end and diurnal) may provide additional information to verify the source
assignments (e.g. [ 13 ]).
A further complication is the basic assumption of the statistical methods that
source profiles neither change during air transport nor with time. Therefore they
cannot be applied strictly to secondary aerosol constituents formed in the atmo-
sphere by gas-to-particle conversion processes. Still, the secondary aerosol
constituents tend to be grouped into one source group since they have a common
“source”, i.e. formation in air triggered by solar irradiation.
2.5 Dispersion and Chemical Transport Models
Small scale dispersion models are traditionally used in environmental assessment
studies for new industrial plants to evaluate the add-on impact on air pollution levels.
State-of-the-art Lagrange models calculate the trajectories of particles released by
the source within three-dimensional wind fields which are generated from typical
meteorological data including information on orography, surface roughness and
obstacles (buildings). Although the computational effort is significant such models
are a valuable complementary approach in a PM source apportionment, particularly
addressing large point sources. The add-on concentrations delivered by dispersion