Geology Reference
In-Depth Information
they were air-dried at room temperature for a min-
imum of 48 hours. The sediment was then
carefully removed and the dry weight obtained.
The particle size distribution (<1 mm) was deter-
mined by mechanical sieving down to 63
Table 13.2
List of models considered.
Model
Reference
CREAMS
Flanagan
et al
. (1989)
m and
then by sedigraph analysis. The cumulative parti-
cle size distribution was determined, from which
the percentages of sand (>63
μ
WEPP
Nearing
et al
. (1989)
RUSLE
Renard
et al
. (1997)
VFSMOD
Muñoz-Carpena & Parsons (2004)
μ
m), silt (2-63
μ
m)
MMF
Morgan & Duzant (2008)
and clay (<2
m) were obtained. Table 13.1 shows
the amount of sediment collected in the buffers
in the two fields. This information provided
the measured data for comparison with model
predictions.
μ
EUROSEM
Morgan
et al
. (1998)
TRAVA
Deletic (2001)
SEDIMOT
Wilson
et al
. (1981)
GRASSF
Barfield
et al
. (1979)
AGNPS
Tim & Jolly (1994)
REMM
Lowrance
et al
. (2000)
13.3 Model Selection
The usefulness of a model for evaluating vege-
tated buffers as a method of erosion control depends
on its ability to simulate the main functions
performed by the buffer, namely enhancing
infiltration, increasing sedimentation and increa-
sing the filtration of material suspended in the
runoff. Particular attention is therefore given to
determining which processes a model simulates
and how the role of vegetation is described. The
following criteria were established as design
requirements that a model should satisfy for use
in the field study area:
(1)
it must be physically-based in order to offer
greater potential than a conceptual or empirical
model in predicting the spatial distribution of
runoff and sediment;
(2)
it must be simple and easily understood
without requiring a trained user;
(3)
it must estimate infiltration, runoff, erosion
and sediment deposition in order to allow
evaluation of sediment and water retained by,
and passing through, the buffer; and
(4)
it must take vegetative cover into account
explicitly so that guidance can be given on the
design of buffers to improve their performance.
Table 13.2 lists a number of models in which
an attempt is made to represent the relevant
filtration and sedimentation processes. Many
of these have not been fully tested using inde-
pendently collected field data, and none has
been tested using standard datasets, so compar-
ison between them is difficult. Based on an
understanding of the bioengineering role of veg-
etation (Coppin & Richards, 1990), it is clearly
important that at least the effects of ground
cover and the density of the plant leaves and
stems are modelled explicitly. Of the models
listed, only EUROSEM and the modified MMF
model use these parameters directly. The major-
ity of the models express plant cover effects
through some form of coefficient, such as the
C
factor of the Universal Soil Loss Equation,
which, as indicated by Styczen and Morgan
(1995), is not a process-based approach. While
these models may represent the conditions for
which the
C
factor values have been experi-
mentally derived, they cannot be used to pre-
dict the effect of the same or different vegetation
in other climatic and soil conditions, and
C
fac-
tors are not designed to account for deposition
processes.
Flanagan
et al
. (1989) demonstrated that, as
long as a number of assumptions are met, a
simplified version of the CREAMS model could
effectively simulate the processes of sediment
deposition within a vegetated buffer. These
assumptions are that the flow is shallow and
uniformly distributed along the upslope edge of
the buffer, concentrated flow effects are minimal,
the grass is not submerged or flattened by the
