Numerical Dust Models - Mineral Dust: A Key Player in the Earth System - page 203

The aerodynamic surface roughness length z 0 plays a significant role in com-
puting the dependence of dust emissions on surface wind speeds. This surface
parameter is needed for the computation of surface friction velocity from surface
wind speed and influences the threshold friction velocity necessary to initiate dust
emission. Marticorena et al. ( 1997 ) developed a map of the distribution of z 0 in
the northwestern Sahara based on a detailed geomorphological analysis. The z 0
values used in global and regional models to compute large-scale atmospheric
dynamics contain information about vegetation and subgrid-scale topography in
order to simulate realistic grid-scale wind fields. These do not reflect the roughness
of the ground for the usually subgrid-scale dust sources and are not appropriate
for modelling dust emissions. Global models of the dust cycle often compute dust
emissions as a function of surface wind speed directly, implicitly assuming constant
surface roughness and neutral stability. Alternatively, for global and regional models
of dust emissions, datasets of satellite-derived z 0 are used. For use in global-
scale models, estimates of z 0 in arid and semi-arid regions are provided at 25-km
resolution from European remote sensing (ERS) scatterometer observations (Prigent
et al. 2005 ). For regional model application in the Sahara, a surface roughness
dataset with a 10-km resolution was retrieved from surface bidirectional reflectance
products provided by passive multidirectional measurements in the solar spectrum
of the Polarization and Directionality of the Earth Reflectance (POLDER-1) sensor,
which has been used in several regional models of Saharan dust (e.g. Laurent et al.
2005 ).
Not all surfaces in deserts or semi-deserts are effective sources of dust aerosol
particles (see Chap. 5 ) . The absorbing aerosol index (AI) derived from the total
ozone mapping spectrometer (TOMS) satellite instrument (Herman et al. 1997 )
has been widely used in order to evaluate large-scale patterns of atmospheric dust
loading and the extent of source areas (see Chap. 7 ) . From observations of the
TOMS AI, enclosed topographic depressions were suspected to be particularly
active sources of dust, the so-called hot spots of dust emission (Prospero et al.
2002 ). Such preferential dust sources have been prescribed in global dust emission
models in different ways (Ginoux et al. 2001 ; Tegen et al. 2002 ; Zender et al.
2003 ). Figure 9.1 shows two examples of widely used parameterizations of such
enclosed depressions. The first parameterization by Ginoux et al. ( 2001 ) derived the
distribution of topographic depression from a 1 ı 1 ı topography dataset, and dust

emissions were only allowed to occur in these areas. Tegen et al. ( 2002 ) prescribed
the global distribution of the potential extent of paleolakes (Fig. 9.1 ) as preferential
dust sources by using results from a water routing and storage model (Coe 1998 ).
The use of this parameterization was based on the assumption that climate variations
during the Quaternary have been sufficiently large to have allowed lakes to form in
closed basins at some time in the past, like the paleolake Mega-Chad that covered
the Bodélé Depression in Chad about 6,000 years ago (Leblanc et al. 2006 ). Possible
errors of this parameterization include the misrepresentation of a preferential source
that does not contain deflatable sediments and thus does not act as a dust source. It
may also miss preferential source areas that contain fine deflatable dust particles
from riverine transport, which are not located within an enclosed depression.