Geology Reference
In-Depth Information
Table 14.4
Values of model erodibility coefficients applied in field and basin simulations.
Raindrop impact
erodibility coefficient (J
−1
)
Overland flow erodibility
coefficient (mg m
−2
s
−1
)
Location
Area (km
2
)
Reference
UK fields
0.034/0.047
1.3-11.8
0.65-5.9
Norouzi Banis
et al
. (2004)
Iowa fields
0.051/0.064
28-82
0.14-0.33
Wicks and Bathurst (1996)
Laval, France
0.86
0.1-10
1-20
Lukey
et al
. (2000)
Rimbaud, France
1.46
0.01-5
0.1-10
Lukey
et al
. (1995)
Ijuez, Spain
45
0.05-0.2
0.5-2
Bathurst
et al
. (2007)
Valsassina, Italy
180
0.05-0.2
0.5-2
Bathurst
et al
. (2005)
Llobregat, Spain
505
0.05-0.2
0.5-2
Bathurst
et al
. (2006)
Cobres, Portugal
701
0.13/2
1.3/20
Bathurst
et al
. (1996)
Agri, Italy
1532
0.1-10
1-20
Bathurst
et al
. (2002)
been determined using caesium-137 measure-
ments. These rates were successfully contained
within the range of simulated values that repre-
sented the uncertainty in model output derived
from uncertainty in the model parameterization.
In addition, the spatial variability in the long-
term erosion and deposition rates was reproduced
excellently at one site and partially at the other.
This suggests that the model does have the poten-
tial to represent spatial variability correctly, given
the appropriate data. It also shows how such field
data can improve the severity of the validation
procedure, accounting for internal as well as out-
let conditions.
(2000) simulated two of the Draix research basins
administered by the French agency CEMAGREF
in southeast France. A model was first con-
structed for the 86 hectare Laval basin which is
severely affected by badlands erosion. The model
was then altered to represent the Laval basin as
if it were equivalent to the neighbouring 108 ha
Brusquet basin, which was successfully rescued
from badlands erosion by reforestation in the
early 1900s. Changes in the model parameters
included specification of forest vegetation (e.g.
with appropriate leaf drip parameters), an increase
in overland flow resistance and elimination of
erosion by overland flow. Uncertainty bounds
were developed both for the simulation results
and for the measured data. The results showed a
good ability to simulate the observed two orders
of magnitude reduction in sediment yield from
the Laval (127 t ha
−1
y
−1
) to the Brusquet (0.03-
2.75 t ha
−1
y
−1
) basin, with the simulated changes
exceeding the output uncertainty. Lukey
et al
.
(1995) tested the ability of SHETRAN to simulate
sediment yield in the 1.46 km
2
Rimbaud basin
(administered by CEMAGREF in southeast
France) following a fire in August 1990 that
destroyed approximately 85% of the original
cover of maquis and chestnut plantation. Output
uncertainty bounds were derived as a function of
specified uncertainty in the model erodibility
coefficients. The results showed a good ability to
contain measured sediment discharge values
within the simulation bounds for individual
14.7.3
Land use and climate change studies
The SHETRAN erosion model has been used to
simulate the impacts of river basin change in two
main ways. The first is testing the ability to repre-
sent change using measured data, either from paired
basins with different land covers or from a single
basin that has undergone a change; generally these
concern the effect of forest logging or planting. The
second is calibration for the current basin condi-
tions and then application to scenarios for possible
future land uses or climates. A third approach has
also been proposed in which the model is used in a
completely hypothetical manner to allow the ini-
tial exploration of a potential problem.
The tests using measured data are, almost by
necessity, limited to small basins. Lukey
et al
.
