Geoscience Reference
In-Depth Information
or simply a reactivation of pre-existing dunes and sand-plains that had lost their
stabilising plant cover as a result of increased aridity or other reasons, including fire
or the impact upon vegetation of changes in atmospheric carbon dioxide concentration
(Hesse et al.,
2005
; Hesse,
2010
). A fundamental issue that needs to be resolved is the
unproven assumption that the presence of now vegetated and stable dunes in presently
semi-arid areas does indeed reflect a former reduction in precipitation. As we shall
see in the following sections, dune formation reflects a variety of controlling factors,
each of which may obscure or outweigh the influence of local precipitation.
8.3 Sand transport by wind
Dune development is controlled by three main independent variables: wind speed,
sand supply and vegetation cover. The early observations by Bagnold in the Libyan
Desert, supplemented by his sand flume studies, demonstrated that the movement
of individual sand grains by saltation is a function of wind velocity and grain size,
with the threshold value for sand mobilisation and transport ranging from 4 m/sec
(Bagnold,
1941
, p. 70) to 6 m/sec, depending on sand particle size (Wasson et al.,
1983
, p. 126). As wind velocity increases, there is an exponential increase in sand
movement (
Figure 8.5
). Analyses of changes in sand particle size in sediment cores
off the west coast of the Sahara show that the Trade Winds were stronger during
the last glacial, probably as a result of enhanced anticyclonic circulation over the
tropical deserts linked to steeper temperature and pressure gradients between trop-
ical and equatorial latitudes (Parkin and Shackleton,
1973
; Parkin,
1974
; Sarnthein
et al.,
1981
).
Several quite different models of sand transport have been invoked to account
for the formation and movement of linear dunes. One model favoured by Mabbutt
(
1977
) and by Hollands et al. (
2006
) is the 'wind-rift' model in which turbulent wind
vortices remove sand particles from the sand-floored swales between the linear dunes
and deposit the sand on the flanks of the dunes. The parent sand in the swales can
be alluvial, as in the case of the late Quaternary alluvial sands in the north-western
Simpson Desert of Australia dated by Hollands et al. (
2006
), or reworked eolian sand
formed during an earlier phase of dune degradation, as is presently occurring to linear
dunes in the Kimberley region of north-west Australia (Goudie et al.,
1993
).
A frequently invoked model is one where there is movement by lateral accretion of
sand in the downwind direction by sand avalanching at the proximal slip face, with the
linear dune ultimately advancing several hundred kilometres or more from its initial
point of origin (Twidale,
1972
; Pye and Tsoar,
1990
; Lancaster,
1995
).
A third model involves vertical accretion (Telfer and Thomas,
2007
; Stone and
Thomas,
2008
; Cohen et al.,
2010a
) and is supported by the fact that OSL ages become
progressively older with depth within the same linear dune. This model seems to
apply especially to linear dunes formed from transverse source-bordering dunes whose
