Identifying Sources of Aeolian Mineral Dust: Present and Past - Mineral Dust: A Key Player in the Earth System - page 58

(Ghienne et al. 2002 ; Drake and Bristow 2006 ). The northern portion of this
former lakebed, south of the Tibesti Mountains, is a topographic low, the Bodélé
Depression, which is estimated to be the largest single source of dust on the Earth
at present based on TOMS and other data (Middleton and Goudie 2001 ;Prospero
et al. 2002 ; Washington et al. 2003 , 2009 ). Recent field studies in the basin itself,
where there are abundant erosional remnant yardangs, composed of diatomaceous
sediments, confirm the dramatic amount of sediment removal by wind (Warren et al.
2007 ; Bristow et al. 2009 ).
Earlier, we alluded to the study of Ginoux et al. ( 2012 ) identifying the Great
Plains region of North America as an important source of dust. Much of the
area where these dust plumes originate is covered by a Quaternary-age aeolian
sediment known as the Blackwater Draw Formation (Holliday 1989 )(Fig. 3.3 d)
and is disturbed by extensive agriculture (Lee et al. 2012 ). Alluvial sources in
desert regions have also been identified in MODIS imagery, such as dust seen on
the western Saudi Arabian coast, moving southwestward over the Red Sea and
toward Sudan (Fig. 3.3 b). The outermost part of the Saudi Arabian coast is an
emergent coral reef terrace, with silt-rich sabkhah deposits, just landward of the
terrace. However, it appears that the main dust sources are alluvial deposits that are
landward of the coast. Drainage development in the mountains that parallel the Red
Sea coast of Saudi Arabia has generated not only wadis that are filled with alluvium
but also older alluvial terraces and pediment surfaces (Brown et al. 1989 ), all of
which are potential dust sources.

3.3.4 Back-Trajectory Analyses to Identify Dust Sources

The analysis of the atmospheric trajectories taken by parcels of air has also
assisted in the investigation of dust sources. In this approach, models are used to
reconstruct the route that an air mass takes to, or from, a given point in time and
space. Trajectories can therefore be examined both “backward” and “forward,” and
assessments in both directions have provided considerable insight into dust source
location and behavior.
Trajectory models typically use three-dimensional input datasets of remodeled
gridded wind fields, available in varying spatial and temporal resolutions; thus,
performance of the analysis is dependent on the quality of input data (Gebhart
et al. 2005 ). The model outcomes also depend on the specified initial conditions,
with key variables including the starting height above the surface for the air parcel,
because this affects mean speed of the advected parcel. Studies commonly run
trajectories at more than one height to show the potential variation associated
with the start height or where atmospheric structure is known to be important for
dust transport routes (Engelstaedter et al. 2009 ; Alonso-Perez et al. 2012 ). The
length of time for which the air parcel is simulated is another important issue in
the reliability of the trajectories produced. With increased time, the accuracy of
the modeled route depreciates, but the distance between dust source and sampling