Geoscience Reference
In-Depth Information
An alternative to comparing field-measured time se-
ries of wind and sand transport data involves the direct
determination of entrainment thresholds from small-scale
field wind tunnels. Such experiments reduce wind and
surface variability and so tend to give better correlations
between sand transport and wind velocity. While tradi-
tional portable wind tunnels have provided useful data on
entrainment thresholds (Belnap and Gillette, 1998) they
are bulky to manoeuvre and so there are limitations on
the surfaces that can be sampled. A recent development
has been that of the Portable In-Situ Wind Erosion Lab-
oratory (PI-SWERL) (Etyemezian
et al.
, 2007), which is
a relatively small circular wind tunnel. The portability
of the PI-SWERL offers the ability to obtain numerous
replicates of the critical wind speed required for erosion
over a broad range of surfaces that have previously been
impractical to measure in the field.
100
80
No STEs
STEs
0
Low
u
t
High
Wind speed
Figure 18.13
An idealised example of the modified time frac-
tion equivalence method for determining the threshold of sed-
iment entrainment using a cumulative frequency plot of wind
velocity and a calculated saltation intermittency value (
γ
)of
20 %. The threshold corresponds to the 80th percentile value
of wind velocity (from Wiggs, Atherton and Baird, 2004).
18.5
Surface modifications to
entrainment thresholds and transport flux
While the relationships shown in Figure 18.11 have been
found to be satisfactory for loose, dry, flat and homo-
geneous surfaces (Williams, Butterfield and Clark, 1994),
the critical thresholds of motion and grain transport fluxes
on natural sediment beds are also influenced by variations
in factors such as sediment mixtures, surface crusting,
surface slope, moisture and vegetation. The relationships
between these parameters are frequently complex and are
not yet fully understood. The effects of vegetation are
dealt with above. Three other especially important mod-
ifications are provided by surface crusting, surface slope
and moisture content.
transport events above threshold. Wiggs, Atherton and
Baird (2004) noted that in their data set such problems re-
sulted in
60 % of their sand transport events accounted
for by the calculated threshold.
Improvements to the explanatory power of thresholds
calculated in this way can be achieved if a lag time is
assumed between the changing wind velocity and the re-
sponse of the sand transport mechanism. Many investiga-
tors have found that the response time of the sand transport
system to wind variability is of the order of
<
1 second
(Wiggs, Atherton and Baird, 2004; Davidson-Arnott and
Bauer, 2009; Weaver, 2008) and so lagging sand transport
behind velocity by this amount results in a much closer
association between these two data series.
While the identification of a single entrainment thresh-
old remains an appealing prospect, it is clear that variabil-
ity inherent in the natural system (in terms of paraemeters
such as grain size, microtopography, packing, moisture)
makes it an unachievable goal. Such complexity in the
system means that entrainment thresholds for a particular
surface are perhaps more appropriately represented by a
range of possible thresholds (Wiggs, Atherton and Baird,
2004; Davidson-Arnott, MacQuarrie and Aagaard, 2005;
Davidson-Arnott
et al.
, 2008) whereby the probability of
erosion may be modelled by investigation of the cumu-
lative probability density functions of time series data of
wind velocity and sand transport (Davidson-Arnott and
∼
18.5.1
Surface crusting
A surface crust lying above loose, erodible sediment sig-
nificantly inhibits the entrainment of sediment into the
wind. In wind tunnel studies Zobeck (1991) found that
soils without surface crusts were 40-70 times more erodi-
ble than crusted soils and similar impacts on erodibility
have been found by investigations in wind tunnels (Rice,
Mullins and McEwan, 1997; Rice and McEwan, 2001;
McKenna-Neuman and Maxwell, 2002) and the field
(Rajot
et al.
, 2003; Houser and Nickling, 2001a, 2001b;
Hupy, 2004). However, the effect of crust characteristics
on aeolian processes is still poorly understood and so its
impact is often disregarded in aeolian transport models
