Geoscience Reference
In-Depth Information
10
1
cm, Gillette et al. (
1982
) measured
U
t
from 20 to 100 cm s
1
over loose
or disturbed surfaces with a portable wind tunnel. A similar approach was used
by Arya (
1975
) to determine the wind stress on Arctic pack ice. Marticorena
and Bergametti (
1995
) extended this concept to aeolian studies. As suggested by
Arya (
1975
), they assumed that an internal boundary layer (IBL) develops behind
the roughness elements, similarly to the development of the IBL occurring after
a sudden change in roughness. At the intersection of the IBL and the boundary
layer produced by the roughness elements, the ratio of local to total wind shear
stress (or the ratio of
U
*
s) can be expressed as a function of the roughness length
of the smooth erodible surface and of the roughness length characterising the
overall surface, including the effect of the roughness elements. This drag partition
scheme predicts that the efficiency for eroding particles is greatly reduced for
Z
0
of about 0.1 cm, above which one would not expect dust emissions under most
natural conditions (Marticorena and Bergametti
1995
). Compared to the wind-
tunnel dataset of Marshall (
1971
), this parameterisation gives a comparable level
of agreement to that of Raupach et al. (
1993
). It also reproduces the erosion
threshold measured with a portable wind tunnel (Gillette et al.
1980
) over natural
non-vegetated surface across a large range of
Z
0
(Marticorena et al.
1997a
):
ln .Z
0
=
z
0s
/=ln
a.X=
z
0s
/
b
f
R
.Z
0
;
z
0s
/
D
1
(5.7)
where
z
0s
is the aerodynamic roughness length of the erodible surface, and
a
and
b
are empirical coefficient describing the evolution of the IBL as a function of the
distance (X) estimated, respectively, to be
a
D
0.35,
b
D
0.8 and
X
D
10 (with
Z
,
Z
0
and
z
0s
in cm).
These two drag partition schemes are now widely used and have allowed
significant progress in the prediction of erosion threshold over natural rough
surfaces. As an example, Laurent et al. (
2005
) used such a partition scheme and
Z
0
derived from satellite to estimate erosion thresholds over East Asian deserts.
The estimated values, expressed as wind velocity at a height of 10 m, vary from
7ms
1
in smooth sandy deserts such as the Taklamakan to 15-18 m s
1
in the stony
Gobi Desert, in agreement with field observations (Natsagdorj et al.
2003
;Wang
et al.
2003
). The spatial distribution of these thresholds is in remarkable agreement
with wind velocities associated with synoptic records of dust storms (Kurosaki and
Mikami
2007
).
The partition schemes discussed above have been developed for relatively low
roughness densities and assume an isotropic arrangement of roughness elements.
An extensive comparison performed by Darmenova et al. (
2009
) shows that when
using consistent input data, the two models give similar results for loose soils
and for rough surfaces with sparse roughness elements, that is, the most easily
erodible surfaces. In contrast, large discrepancies are found in the case of dense
vegetation or solid elements, such as stony surfaces (regs or gobi deserts) and
grasslands. More generally, these schemes do not reproduce the erosion thresholds
over complex surfaces, such as surfaces with dense vegetation cover and/or specific
Search WWH ::
Custom Search