Geology Reference
In-Depth Information
the equations for the erosion processes on to a
hydrological model, which was used to generate
runoff and transport the resulting water flow over
the landscape. Examples include AGNPS (Young
et al
., 1989), ANSWERS (Beasley
et al
., 1980) and
GAMES (Rudra
et al
., 1986). Some models, such
as CREAMS (Knisel, 1980), allowed a choice
between daily simulations based on the USCS
Curve Number, a coefficient which describes the
soil, slope and land cover characteristics, and
storm simulations in which runoff is calculated
as the excess of the rainfall intensity over the
infiltration rate of the soil. This generation of ero-
sion models was process-based as regards the
simulation of runoff and sediment, but relied on
the factors of the USLE to describe soil erodibility
(
K
), slope length (
L
), cropping (
C
) and manage-
ment (
P
) effects.
Current research on erosion modelling is
concerned with replacing the coefficients related
to soil, slope and land cover with parameters
that measure their properties directly and which
can therefore take account of variations in both
time and space. Instead of a single value to
express
K
, soils are described by properties such
as cohesion, shear strength and surface rough-
ness, and land cover by architectural properties
of the vegetation such as height, percentage
cover, stem size and stem density. This means
that soil, for example, can be modelled dynami-
cally allowing for changes in cohesion as the
surface crusts or seals under raindrop impact
(Moussa
et al
., 2003) or human or animal tram-
pling, or as surface roughness changes as a result
of different tillage practices. Similarly, plant
cover effects can be altered in relation to sea-
sonal plant growth and decay. The effects of soil
and plant cover are sometimes described sepa-
rately for each of the four processes of erosion,
namely detachment of soil particles by raindrops
and runoff, and the transport of the detached
material by rainfall and runoff. The outcome is
that erosion models have become more complex,
since they now incorporate many submodels to
describe the behaviour of soil and vegetation.
Future models may well account for the move-
ment of soil over the landscape by tillage using
methodology developed by Govers
et al
. (1994)
and van Muysen
et al
. (2002). WATEM (van Oost
et al
., 2000; Verstraeten
et al
., 2002) combines
these equations with a modification of the
Revised Universal Soil Loss Equation (RUSLE)
(Renard
et al
., 1991) to simulate the transport
of sediment by runoff and tillage on a mean
annual basis. Improved descriptions of the effect
of soil will take account of its aggregate struc-
ture rather than the size distribution of the
primary particles of clay, silt and sand, by using
parameters based on aggregate stability (Issa
et al
., 2006).
With greater concern environmentally about
the fate of eroded sediment has come the recogni-
tion that the way in which many erosion models
simulate the deposition of sediment is too sim-
plistic. Just comparing the amount of material
available for transport with the transport capac-
ity, and dumping the sediment which cannot be
transported, results in patterns of deposition over
the landscape which are unrealistic. Too much
material is deposited too quickly. In models such
as WEPP and EUROSEM, an attempt is made to
control the rate of deposition by taking account of
the settling velocity of the soil particles in the
flow and a coefficient expressing the efficiency of
the deposition process. Although this approach
produces better results, it is analogous to the use
of coefficients to describe the effects of plant cover
on erosion rather than simulating the physical
process. Future models are likely to model deposi-
tion explicitly taking account not only of particle
settling velocity, but also the velocity of the run-
off and the depth of flow. The approaches devel-
oped to predict sediment deposition in filter strips
(Tollner
et al
., 1976; Rose
et al
., 2003) and farm
ponds (Verstraeten & Poesen, 2001) are likely to
be adapted to describe deposition from runoff.
Such developments will lead to even greater com-
plexity in erosion models as they attempt to
describe the erosion processes more fully. For
example, the four erosion processes identified
by Meyer and Wischmeier (1969) and the process
of deposition will need to be simulated sepa-
rately for each soil particle size. Alternatively, ero-
sion, transport and deposition can be modelled
