Dust Production Mechanisms - Mineral Dust: A Key Player in the Earth System - page 113

Fig. 5.6 Probability density function of the 10 m wind velocity fields from the ECMWF
meteorological model over the Western Sahara for March and December 1991 (Adapted from
Marticorena and Bergametti 1996 )
particles in a soil is often encountered in salty or silty depressions, such as dry lakes,
which explain their high potential for dust emissions (Callot et al. 2000 ; Bullard
et al. 2011 ; see also Chap. 3 ) .
However, a high potential for dust emission is not sufficient to make a dust
source; sufficiently, high winds are needed. The activity of a dust source depends on
the fraction of surface winds exceeding the erosion threshold defined by the local
surface properties and its intensity depends on the local wind velocity and erosion
thresholds and on the soil properties. The distribution of the surface wind velocities
(at 10 m height) from the analysis of the European Centre for Medium-Range
Weather Forecast over the Western Sahara is illustrated in Fig. 5.6 for the month
of March and December 1991. A small fraction of the surface winds corresponds
to velocities higher than the minimum erosion threshold (29 % > 6ms 1 in March
and 15 % in December). The fraction of high winds, responsible for the most intense
dust emissions, is even smaller (4 % > 10 m s 1 in March and 2 % in December).
Representing the surface winds in dust source regions is thus a challenging issue
(see Chap. 6 ) . Finally, a proper description of both the surface characteristics and
the surface wind distribution is required to estimate whether a source region is a
“preferential source” or not, either in terms of frequency or in terms of intensity.
The knowledge of dust emission processes has allowed the establishment of
conceptual and numerical models that describe the dependence of the emission
fluxes on U * and surface and soil parameters. When all input parameters are
measured or fixed in reasonable and realistic ranges, dust mass fluxes can be
properly estimated by physical dust emission models as a function of U * (Shao et al.
2011 ). The major weakness in the present description of dust emission processes