66 Aerosol Intensive Optical Properties …
415
The online aerosol-radiation coupling has been activated for all the 5 global
species, using the default particle parameterization of the model (REF case) mainly
based on the OPAC dataset [3]. The new refractive indexes used for the PTB case
(real: n r , and imaginary: n i parts) are reported in Table 66.1 (references indicated in
the last row). Also the percent variations with respect to the reference values and the
wavelengths (λ) of the optical measurements are reported in the table (the acronym
VIS stands for “visible range”). Note that a separate parameterization for primary
organic matter (POM) and SOA has been adopted in the PTB case. Moreover, in the
PTB case the new refractive indexes for dust and sea salt have been applied only in
stations dominated by natural aerosols, while that for organic matter and sulfate only
in stations dominated by anthropogenic aerosols.
66.4 Results
Simulated (both REF and PTB cases) and observed ω and g are plotted in Fig. 66.1
for Cairo_EMA_2 and La_Parguera (stations dominated by natural aerosols), Beijing
and GSFC (stations dominated by anthropogenic aerosols).
Cairo_EMA_2 (top-left panel) is close to dust sources (dominated by dust in
simulations). In the REF case the model strongly underestimates the observed ω and
overestimates the observed g. These model biases seem to be related to a too high
imaginary index for the reference dust. The new n i value (PTB case), indeed, causes
the simulated ω to significantly increase and the simulated g to slightly decrease,
as expected, reducing thus both the mean biases. However a further correction is
needed, above all for g, which could be generated for example by a higher fine to
coarse ratio for mineral dust. La_Parguera (top-right panel) is a pure coastal station
(with intrusions of transported dust during summer in simulations). In the REF case,
we note a strong underestimation of the observed ω during summer (in conjunction
with the dust intrusions) and a general overestimation of the observed g (above all
during winter months). The need of a lower imaginary index for dust is confirmed,
since the new n i value (PTB case) causes the simulated ω to increase and the simulated
g to slightly decrease in summer months. However, it seems that a missing of fine
particles (for example fine sea salt or white organics) is the main contribution to the
positive g mean bias in winter months.
Beijing (bottom-left panel) is a pure urban station (with very high fractions of
carbonaceous aerosols in simulations). In the REF case, the model clearly underestimates the observed ω but shows a good agreement with the low observed g
(coherently with the dominant presence of small species). The negative ω mean bias
can be attributed to a too high imaginary index for the reference organic matter (0.006
at 0.550 μm) but also to the very high simulated fractions of black carbon (which
also cause g to strongly decrease). GSFC (bottom-right panel) is characterized by an
anthropogenic background not directly affected by urban emissions (higher sulfate
and lower black carbon fractions in simulations). In the REF case, we note a less
415
The online aerosol-radiation coupling has been activated for all the 5 global
species, using the default particle parameterization of the model (REF case) mainly
based on the OPAC dataset [3]. The new refractive indexes used for the PTB case
(real: n r , and imaginary: n i parts) are reported in Table 66.1 (references indicated in
the last row). Also the percent variations with respect to the reference values and the
wavelengths (λ) of the optical measurements are reported in the table (the acronym
VIS stands for “visible range”). Note that a separate parameterization for primary
organic matter (POM) and SOA has been adopted in the PTB case. Moreover, in the
PTB case the new refractive indexes for dust and sea salt have been applied only in
stations dominated by natural aerosols, while that for organic matter and sulfate only
in stations dominated by anthropogenic aerosols.
66.4 Results
Simulated (both REF and PTB cases) and observed ω and g are plotted in Fig. 66.1
for Cairo_EMA_2 and La_Parguera (stations dominated by natural aerosols), Beijing
and GSFC (stations dominated by anthropogenic aerosols).
Cairo_EMA_2 (top-left panel) is close to dust sources (dominated by dust in
simulations). In the REF case the model strongly underestimates the observed ω and
overestimates the observed g. These model biases seem to be related to a too high
imaginary index for the reference dust. The new n i value (PTB case), indeed, causes
the simulated ω to significantly increase and the simulated g to slightly decrease,
as expected, reducing thus both the mean biases. However a further correction is
needed, above all for g, which could be generated for example by a higher fine to
coarse ratio for mineral dust. La_Parguera (top-right panel) is a pure coastal station
(with intrusions of transported dust during summer in simulations). In the REF case,
we note a strong underestimation of the observed ω during summer (in conjunction
with the dust intrusions) and a general overestimation of the observed g (above all
during winter months). The need of a lower imaginary index for dust is confirmed,
since the new n i value (PTB case) causes the simulated ω to increase and the simulated
g to slightly decrease in summer months. However, it seems that a missing of fine
particles (for example fine sea salt or white organics) is the main contribution to the
positive g mean bias in winter months.
Beijing (bottom-left panel) is a pure urban station (with very high fractions of
carbonaceous aerosols in simulations). In the REF case, the model clearly underestimates the observed ω but shows a good agreement with the low observed g
(coherently with the dominant presence of small species). The negative ω mean bias
can be attributed to a too high imaginary index for the reference organic matter (0.006
at 0.550 μm) but also to the very high simulated fractions of black carbon (which
also cause g to strongly decrease). GSFC (bottom-right panel) is characterized by an
anthropogenic background not directly affected by urban emissions (higher sulfate
and lower black carbon fractions in simulations). In the REF case, we note a less