Geology Reference
In-Depth Information
processes and by indicating cumulative, basin-
scale effects. They also provide the data with
which to create and test models. However, the
sheer complexity and natural variation of sedi-
ment yield controls make it difficult to transfer
experimental results between even similar basins
with more than order-of-magnitude accuracy.
Furthermore, field experiments cannot quantify
the impacts of a land-use change on basin-scale
sediment yield in a period of less than several
years. Modelling, by contrast, is able to investi-
gate the integrated effects of the various proc-
esses, including changes in the relative importance
of the processes. Modelling can also deliver rapid,
systematic investigation of a range of land man-
agement strategies, allowing selection of the
strategy that minimizes the accelerated erosion
before there is any intervention on the ground. A
difficulty, though, with models is that there are a
number of sources of uncertainty which must be
considered carefully when interpreting simula-
tion results.
This chapter considers the use of physically-
based models for predicting the impacts of land
use and climate change on sediment yield at spa-
tial scales ranging from the hillslope to river
basins of up to 1000 km
2
. It reviews the require-
ments of such models, discusses their capabilities
and limitations (including uncertainty) and dem-
onstrates their application, focusing on the
SHETRAN basin modelling system in particular.
The overall aim is to provide a sufficient back-
ground for a user to understand the issues behind
model selection, parameterization and applica-
tion and the interpretation of model results.
supply of sediment into the river system, with its
potentially detrimental consequences for salmon-
spawning and sport fishing;
●
the extent to which soil loss can be reduced
through the introduction of sympathetic agricul-
tural techniques such as seed ploughing and
buffer strips;
●
the effect of agricultural fertilizer application
on non-point source pollution, involving the
transport of phosphorus adsorbed to sediment
particles;
●
the impact on reservoir sedimentation of the
replacement of native forest by cropland in the
upstream basin; and
●
the impact of future changes in rainfall regime
on erosion and sediment yield.
Tackling such problems involves repeat model-
ling of the focus area with different management
and climate scenarios and the comparison of out-
put data in such forms as annual sediment yields,
peak sediment concentrations in rivers and
spatial variation in soil erosion. A number of spe-
cific modelling requirements then need to be
satisfied.
(1) Capability to predict the impacts of future
change
Of the available modelling approaches
(e.g. empirical, conceptual, stochastic, physical-
ly-based), only the physically-based can truly be
said to have this capability (e.g. Abbott
et al
.,
1986a). These models are based on the fundamen-
tal equations of physics, such as the equations of
mass, momentum and energy conservation, and
theoretically and experimentally derived rela-
tionships (such as the factor-of-safety equation
for slope stability). Their parameters have a phys-
ical meaning (such as soil permeability, vegeta-
tion root cohesion and percentage of bare soil)
and can therefore be changed to represent speci-
fied future conditions on the basis of physical rea-
soning, available databases and expert judgment.
It is not possible so to change the parameters
of, for example, an empirically-derived regression
equation as they are a product of the regression
procedure and have no physical meaning. This
chapter is therefore concerned solely with physi-
cally-based modelling.
14.2 Model Requirements
The overall requirement is to be able to predict
the impacts of possible future changes in land
use, vegetation cover and climate on the erosion,
sediment transport and sediment yield of a basin.
Typical sediment-related management problems
which need to be addressed include:
●
the effect of forest logging in upland areas on
the incidence of shallow landslides and thus the
