Geoscience Reference
In-Depth Information
applications of the theory have been undertaken and those
that have been carried out involve a very limited range of
vegetation types (e.g. Lancaster and Baas, 1998; Craw-
ley and Nickling, 2003); far more experimental data are
required to develop the model further. Furthermore, appli-
cation of the Raupach, Gillette and Leys (1993) model at a
regional or global scale may be hampered by the relatively
complex input parameters required.
Okin (2008) has recently presented a new model for
determining the effect of vegetation on wind erosion that
partly resolves this problem. In recognition of the ob-
servation that vegetation in drylands is rarely regularly
patterned, Okin has devised a model that focuses on the
size distribution of
gaps
between nonerodible roughness
elements as being the driving force controlling the amount
of potential erosion (rather than the structural parameters
of the roughness elements themselves). Initial testing of
the model has shown good results but a significant ad-
vantage is that it requires relatively straightforward input
data that could be acquired from fieldwork or large-scale
image analysis.
where
n
=
number of roughness elements
b
=
width of roughness elements
h
=
height of roughness elements
s
=
surface area
However, such approaches have been seen to be unsatis-
factory when describing complex roughness distributions
and three-dimensional objects (Minvielle
et al.
, 2003;
Musick, Trujillo and Truman, 1996); problems may arise
in the case of vegetation where effective height may vary
in response to changing wind speeds and plant pliability,
and the drag provided by vegetation may vary with stem
porosity.
However, there are two models for determining the
impact of shear stress partitioning in the presence of
nonerodible roughness elements that have become widely
recognised in the research literature. While both mod-
els compare the ratio of stress required for erosion on a
bare surface to the stress required for erosion on a surface
including roughness elements, they each have a distinct
approach to the problem (see King, Nickling and Gillies,
2005, for a full comparison). The first is the model of
Marticorena and Bergametti (1995), which defines the ra-
tio between the aerodynamic roughness of the surface be-
tween roughness elements (
z
0s
) and the total aerodynamic
roughness (
z
0
). This model often forms the basis for the
emission schemes in global-scale dust models (Zender,
Bian and Newman, 2003; see Chapter 20). The other ap-
proach is that of Raupach (1992) and Raupach, Gillette
and Leys (1993), in which the wake development behind
individual roughness elements is modelled to generate
a ratio between the erosion threshold on a bare surface
and that incorporating nonerodible roughness elements.
The input parameters to the Raupach, Gillette and Leys
(1993) model are more complex than for the Marticorena
and Bergametti (1995) model and include the vegetation
roughness density (
18.3
Sediment in air
18.3.1
Grain entrainment
Sediment is entrained into the airflow when forces acting
to move a stationary particle overcome the forces resisting
sediment movement. The relevant forces for the entrain-
ment of dry, bare sand are shown diagrammatically in
Figure 18.10.
Particles are subjected to three forces of movement:
lift, surface drag and form drag. Lift is a result of the
air flowing directly over the particle forming a region
of low pressure (in contrast to relatively high pressure
beneath the particle); hence there is a tendency for the
particle to be 'sucked' into the airflow. Surface drag is
the shear stress on the particle provided by the velocity
profile and the form drag is also related to upwind and
downwind pressure differences around the particle. When
these forces overcome the forces of particle cohesion,
packing and weight, the particle tends to shake in place
and then lift off, spinning into the airstream.
It has been shown that aerodynamic entrainment is pri-
marily a function of the mean grain size of the particles
involved combined with the erosivity of the wind, with
u
∗
the preferred measure (Williams, Butterfield and Clark,
1990). Bagnold (1941) studied these relationships and
derived values of
critical threshold shear velocity
(
u
∗
ct
)
) combined with terms describing the
vegetation aspect ratio and erosion threshold delineated as
a function of the maximum shear stress.
The Raupach, Gillette and Leys (1993) model shows
good general agreement with field and wind tunnel data
(Figure 18.9) and experimental testing of the procedure
has confirmed its robustness (Brown, Nickling and Gillies,
2008; King, Nickling and Gillies, 2005, 2006; Gillette,
Herrick and Herbert, 2006; Gillies
et al.
, 2000, 2010;
Gillies, Nickling and King, 2007). However, very few
λ
