Geoscience Reference
In-Depth Information
described by Mie theory (Bohren and Huffman 1983 ). Key governing parameters
include , r and the complex refractive index of the particle, m ,( m
k i ), where
the real part n represents the amount of scattering and the imaginary part k represents
the amount of absorption, both of which vary with . Thus, for a given particle
size (or size distribution) and refractive index that varies with wavelength, spectral
properties of scattering, absorption and extinction can be calculated. These can then
be applied in a radiative transfer model to calculate irradiances and/or radiances
throughout an atmospheric column, subsequently determining the radiative effect
(RE) of the dust.
Observations of dust particles increasingly show that the assumption of sphericity
is unrealistic, even at small sizes. Optical properties for non-spherical dust can be
calculated using a range of other techniques (e.g. Otto et al. 2009 ;Yangetal. 2007 ).
Oblate spheroids with an effective axis ratio of 1.16 were indicated by lidar measure-
ments during the Saharan Mineral Dust Experiment (SAMUM) campaign. However,
the simulated optical properties differed by only 1 % in SSA and 4 % in asymmetry
parameter from those calculated using a spherical particle (Otto et al. 2009 ). The
extinction optical depth was within 3.5 % of the spherical assumption case, but the
backscattering towards the TOA was substantially enhanced (and therefore the dust
is more cooling when non-spherical properties are used). Haapanala et al. ( 2012 )
consider several approaches to compare dust of different shapes and conclude that
there is no simple relationship between particle shape and modification of the
direct RE in comparison with spheres. They also acknowledge that any uncertainty
introduced here is likely to be substantially smaller than other sources of uncertainty
concerning dust in climate models, for example, spatial variability or differences
in optical properties due to source region. However, non-sphericity is likely to be
much more important for applications that measure the impact of dust on a beam
oriented along a single direction, such as upwardly scattered radiation measured
by aircraft or satellites, and particularly for lidar measurements where backscatter
at 180 ı is crucial. Non-spherical scattering codes are considerably more expensive
than Mie code, and many single non-spherical codes do not cover the whole size
and wavelength ranges required for use with dust.
D
n
C
11.2.3
Sensitivity of Optical Properties to Size
and Composition: An Illustrative Example
Dust optical properties are highly variable due to the sensitivity to particle size and
composition. An overview of dust size distributions is given in Chap. 2 . Figure 11.1
shows two contrasting examples of dust size distributions. The red line shows
a size distribution retrieved from AErosol RObotic NETwork (AERONET; see
also Chap. 7 ) ground-based sun photometer measurements at Cape Verde in the
Atlantic Ocean over 8 years (Dubovik et al. 2002 ), representative of transported
dust. Contrastingly, the black line shows a size distribution from recent aircraft
measurements (Ryder et al. 2013 ) over the remote Sahara, much closer to dust
Search WWH ::




Custom Search