Calibration of Erosion Models - Erosion Modelling - page 45

values and disregarding more complex hydro-
graph information, we are increasing uncertainty
unnecessarily.
When modelling larger areas with low spatial
resolution, sources and sinks are not explicitly
modelled because they occur within one spatial
element (grid cell) and sediment delivery ratios
are used. There is a lack of calibration, which is
understandable in view of the lack of erosion/soil
loss data on this scale. More effort should perhaps
be made here because the model results at this
scale are often used by governments and form the
basis for national and international policies, per-
haps more than the model results at more detailed
scales. While in the scientific literature USLE
variations are used in proper ways with adapta-
tions to produce sediment delivery estimates (see
examples above), the original USLE is still popu-
lar as a consultancy tool (see Chapter 8), although
sometimes used completely outside its original
purpose.
In terms of model quality and calibration, the
earlier conclusions can be repeated here, and are
since confirmed by many authors: calibration
should always be done if possible, and the range
of data on which the model is calibrated should
be a large as possible. Different calibration data-
sets or strategies may be needed for large events
and small events, and generally small events or
totals are less well predicted. When comparing
the performance for lumped results (e.g. annual
soil loss), the more empirical models perform just
as well as physically-based models that attempt
to simulate the hydrological and erosion proc-
esses in detail. They do not perform any better,
but are more attractive as they generally have
lower data requirements. Under a given set of cir-
cumstances, some models perform better than
others, but generalizations on model performance
cannot be made. The same model is shown to
perform well by one author and only moderately
well by others.
Assuming that the calibrations discussed above
used a best-possible parameterization of the
model, it seems that specifying patterns of input
parameters based on land use, landscape and/or
geostatistics is not sufficient to obtain more than

moderate results. Several authors argue that a
more distributed approach is needed to provide
spatially distributed predictions of soil erosion
and sediment transport within a catchment.
Different sources of spatial information are avail-
able: agricultural information, radioactive and
chemical elements that can be used as tracers,
and earth observation data of increasing temporal
and spatial scales. Thus the necessary informa-
tion seems to be available and accessible and
should be used in the future (see Chapter 17).

References

Arhonditsis, G., Koulouri, M., Giourga, C. & Loumou,
A. (2002) Quantitative assessment of agricultural
runoff and soil erosion using mathematical model-
ling: applications in the Mediterranean region.
Environment Management 30 : 434-53.
Bathurst, J.C. & Lukey, B. (1998) Modelling badlands ero-
sion with SHETRAN at Draix, southeast France. In
Modelling Soil Erosion, Sediment Transport and
Closely Related Hydrological Processes , Proceedings of
the International Association of Hydrological Sciences
Vienna Symposium. IAHS Publication 249 : 129-36.
Bayramin, I., Erdogan, E.H. & Erpul, G. (2007) Use of
USLE/GIS Methodology for predicting soil loss in a
semiarid agricultural watershed. Environmental
Monitoring and Assessment 131 : 153-61.
Beasley, D.B., Huggins, L.F. & Monke, E.J. (1980)
ANSWERS: a model for watershed planning.
Transactions of the American Society of Agricultural
Engineers 23 : 939-44.
Beven, K.J. (1997) Distributed Modelling in Hydrology:
Applications of TOPMODEL . John Wiley & Sons,
Chichester.
Beven, K. (2002) Towards an alternative blueprint for a
physically based digitally simulated hydrologic
response modelling system. Hydrological Processes
16 : 189-206.
Beven, K. & Freer, J. (2001) Equifinality, data assimila-
tion, and uncertainty estimation in mechanistic
modelling of complex environmental systems using
the GLUE methodology. Journal of Hydrology 249 :
11-29.
Boardman, J. & Favis-Mortlock, D. (1998) Modelling
soil erosion by water . NATO ASI Series 1: Global
Environmental Change, Volume 55. Springer-Verlag,
Berlin: p. 531.