Chemistry Reference
In-Depth Information
In addition to organic compounds also inorganic components may change during
the transport. Potassium chloride (KCl) occurs in young smoke, whereas increased
amounts of potassium sulfate (K
2
SO
4
) and potassium nitrate (KNO
3
) are present in
old smoke. This is due to the rapid substitution of chloride by sulfate and nitrate in a
smoke plume. That behavior was seen in Helsinki in March 2002 when sulfur-rich
particles collected during the smoke episode contained very little chloride because
it had already depleted during the transport [
12
].
The size distribution of LRT smoke particles may change during the transport
[
28
]. By comparing the data from Spitsbergen to that from the fire area (Minsk in
Belarus and Toravere in Estonia), it was found that the average median radii for the
fine mode decreased from the source area (0.16
m)
during the episode in spring 2006. Same behavior was also seen for the coarse
mode; however, the main difference between the size distributions was for the
mode centered in 5-7
m) to Spitsbergen (0.12
m
m
m that was not found at Spitsbergen at all. In fire areas, the
fine and coarse mode concentrations were approximately ten times higher during
the smoke events than in the reference days, whereas at Spitsbergen the fine mode
concentration increased by a similar factor but the coarse mode was only 1.5-2
times higher during the most intense episode. Thus, even though both fine and
coarse mode aerosols were emitted during the fires or were produced by dynamic
processes in the fresh smoke, coarse aerosols were deposited during the transport to
Spitsbergen whereas smaller aerosols had longer atmospheric residence times and
therefore were drifted in the plume into the Arctic region. Deposition of coarse
particles during the transport was also seen in optical properties of smoke aerosol.
Single scattering albedo (SSA) increased with the distance from the fire source [
28
].
The increase in SSA indicated that the oxidation and transport of the aerosols from
the source area towards the Arctic reduced the absorption of particles, whereas the
scattering process became more dominant. Reduced absorption was explained by
the deposition of coarse particles that probably had some absorbing aerosols.
m
5 Modeling
In order to show the dispersion of plumes originated from the fires, LRT smoke
plumes have been modeled. FLEXPART particle dispersion model was used to
indicate that the smoke was transported from detected source region to Spitsbergen
and Iceland in spring 2006 smoke [
10
]. In addition to the dispersion of smoke, the
fire-related PM concentrations in the smoke plumes can be modeled. For example
in Helsinki during the episode in April to May 2006, the Finnish emergency and air
quality modeling system SILAM [
44
] and during August 2006 episode the fire
assimilation system (FAS; [
45
]) were used to forecast the PM
2.5
concentrations
generated by biomass burning [
18
,
19
]. In general, SILAM and FAS were able to
reproduce PM concentrations, even though gaps were obtained in the modeled time
series due to diverse reasons. Receptor modeling (e.g., positive matrix factoriza-
tion; PMF) can be used to find out the contribution of smoke to the observed PM
