Dust Deposition - Mineral Dust: A Key Player in the Earth System - page 185

the presence of a charged interface between the particle surface and the air) may
cause E predictions to differ significantly from what happens in reality (Wang et al
2010 ).

8.2.3 Particle Size Distribution and Deposition in Dust Models

Since deposition is highly size dependent, the knowledge of the dust size distribution
is a prerequisite to estimate dust deposition fluxes (dry or/and wet). Then, in
addition to dust emission fluxes, the dust size distribution in source region must
be documented. However, direct measurements of the size distribution of dust
particles in or in the vicinity of dust source regions are rare (see Chap. 5 ) . Thus,
most of the knowledge on dust size distributions representative of source regions
is derived from wind-tunnel experiments (e.g., Alfaro and Gomes 2001 ). Only
recently, measurements of the dust size distributions have been performed over the
Sahara (Sow et al. 2009 ) and the Australian desert (Shao et al. 2011 ).
Early global dust aerosol models only used a unique aerosol particle size (e.g.,
Joussaume 1990 , 1993 ) to represent the dust size distribution. Later, a limited
number of modes have been used. A “mode” is characterized by a mean or median
diameter, a predefined mathematical “shape” (often a log-normal function in the
case of aerosols) and a constant geometric standard deviation for each mode.
For example, Tegen and Fung ( 1994 ) use four constant log-normal modes, while
Guelle et al. ( 2000 ) uses three log-normal semi-constant modes (e.g., median
diameter is changing during transport but having constant geometric standard
deviations).
To represent the dust size distributions during transport with a better accuracy,
models now commonly use a binning of the particle size distribution. The advantage
of this approach is that it does not need an assumption on the conservation of the
particle size modes (e.g., Ginoux et al. 2001 ; Gao et al. 2003 ; Gong et al. 2003 ;
Zhao et al. 2003 ; Meskhidze et al. 2005 ). Because the number of the particle size
bins directly controls the computing time, the particle size range or the number
of size bins (or both) is limited in models. For instance, Ginoux et al. ( 2001 )and
Meskhidze et al. ( 2005 ), respectively, use 7 and 5 bins between 0.2 and 12 m
in diameter, Zender et al. ( 2003 )use4binsbelow10m, Gao et al. ( 2003 ) 4 bins
below 12 m, while the model developed by Lu and Shao ( 2001 ) had 6 bins between
less than 2 and 125 m. One of the most sophisticated representations of the dust
size distribution in a 3D dust transport model is that by Zhao et al. ( 2003 ), who use
12 size bins over a large size range (0.01-40.96 m).
In 3D models, these size bins are often defined to have equal ranges in log D p
(e.g., Dulac et al. 1989 ; Schulz et al. 1998 ; Gong et al. 2003 ;Zhaoetal. 2003 ).
This splitting is thus called iso-log and each size bin is generally characterized by
its geometric mean diameter. This approach leads to uncertainties, which depend
on the number of bins used to represent the whole particle size distribution. Indeed,
Gong et al. ( 2003 )andForêtetal.( 2006 ) show that the number of bins directly