Geology Reference
In-Depth Information
Table 13.3
Additional criteria used for model selection.
Model evaluation
Criterion
WEPP
EUROSEM
MMF
Physically-based
***
***
***
Easy to use
*
*
**
Describes erosion, runoff and deposition
*
*
***
Accounts for effects of vegetation explicitly
**
***
***
Proven applicability to UK conditions
***
***
***
Describes soil particle sizes
***
*
***
Annual/seasonal model
***
*
***
Available or readily attainable data requirements
**
**
***
Enables routing of water and sediment over multiple elements
***
***
***
from field to farm and small catchment scales
Can be combined with land cover and land management options
***
***
***
Reasonable simulation time requirements
***
***
***
Has been applied to vegetated buffer design
**
**
*
Has been applied to vegetated buffer placement
**
**
*
Low cost to obtain and use
***
***
***
Suitability scoring: *, low; **, moderate; ***, high.
flow, and previously trapped sediment does not
affect future depositional capacity (Wenger, 1999).
These assumptions are likely to be particularly
limiting where, as in the second field site, con-
centrated flow dominates.
Deletic (2000) showed that the methods
currently used in models for simulating the
transport of sediment through a grass buffer are
very poor. This is attributed to previous work
focusing on the overall performance of the buffers
rather than the processes involved. For example,
it is questionable whether any of the models
adequately define the erosive potential of the
runoff once it leaves a grass strip. EUROSEM is
based on the assumption that, if transport
capacity has not been reached, then the flow is
still erosive. If all or much of the sediment is
deposited within or upslope of a buffer, then a
flow's erosion potential should be increased, as
demonstrated in laboratory experiments by
Emama Ligdi and Morgan (1995). Styzcen and
Morgan (1995) suggested that as sediment builds
up within the buffer, the foreslope of the
resultant sediment wedge becomes steeper than
the ground slope, as simulated in the GRASSF
model. Therefore, once the foreslope has mig-
rated to the downslope edge of the strip, the
velocity of the flow leaving the buffer will be
increased.
Despite the need to select a model that simu-
lates correctly the processes operating within
the buffer, there is also a need to choose a model
that matches the availability of the input data.
Generally, model users require simple models
that are easy to operate and understand rather
than parameter-intensive models with large
data requirements. Merritt
et al
. (2003) also noted
that model accuracy does not always increase
with model complexity because there can be an
accumulation of error through inaccuracies in
the large number of parameters required. Thus
models like REMM are restricted to trained
users as well as requiring large amounts of
data.
With the exception of TRAVA and EUROSEM,
most of the models listed in Table 13.2 were
developed in the US and, of these, only WEPP
has been widely used in Europe. Combining the
