Geology Reference
In-Depth Information
for enhanced flood flows in the early years of
forest growth (because of the reduced intercep-
tion and transpiration). More importantly, the
increased interception and transpiration of the
later years reduced streamflows to levels that
could be detrimental to aquatic habitats and
threatened drinking water abstraction. Soil ero-
sion was correspondingly enhanced in the early
years but reduced in the later years of the planta-
tion cycle (Fig. 14.7).
Another form of exploration was carried out
by Birkinshaw and Bathurst (2006) to investigate
the scaling relationship between specific sedi-
ment yield and basin area, as a function of
sediment source, land use and rainfall distribu-
tion. Two basins which had already been mod-
elled by SHETRAN were used in the study.
However, the simulations were not representa-
tions of the real-world basins. They simply used
the topographies and river networks as a frame-
work within which to investigate systematically
the scaling relationship for a number of specified
conditions. The study showed that the relation-
ship could be inverse or direct, depending on the
conditions.
14.7.4 Use of model for hypothetical
exploration in the absence of data
A criticism frequently levelled at physically-
based, spatially-distributed models is that data
provision does not match the models' require-
ments for parameterization and calibration (e.g.
Merritt
et al
., 2003). However, this criticism
misses two important points. The first is that
decisions have to be made, whatever the state of
the data. In such cases physically-based models
make full use of whatever data are available and
can be used to quantify the uncertainty associ-
ated with making a decision on the basis of those
data. The second is that, even if there are no data
for the basin, a physically-based model can still
be parameterized using estimates and data from
the literature, an option not available to other
types of model. Of course, the uncertainty asso-
ciated with such a parameterization is large but
the model can provide an important first step in
exploring a potential problem area. The extent to
which the uncertainty is acceptable has to be
weighed against the cost of its reduction through
the collection of more detailed spatially and
temporally varying data. If the exploration sug-
gests a potential for an undesirable or unaccept-
able basin response, then it is clear that the more
detailed application is required. In that case the
model can be used to indicate which data are
most needed to reduce the uncertainty and can
thus guide the field measurements. An example
is presented by Bathurst
et al
. (1998c) to demon-
strate the potential impact of forest plantation
on runoff and soil erosion in southern Chile. A
SHETRAN model of a hypothetical basin was
created using basin data and meteorological data
from other applications in similar climatic envi-
ronments. Simulations were carried out for a
period of 20 years, a typical plantation cycle.
First a simulation was run for native forest, cali-
brated to produce equilibrium conditions. Then
a simulation was carried out for a plantation for-
est of exotic species, growing from a bare soil
condition through to full cover and mature
height. Comparison of the simulations showed
the plantation conditions to produce a potential
14.8
Future Research Needs
Challenging but exciting research needs concern
model parameterization, design and use. Some
examples of parameterization developments rel-
evant to erosion modelling include:
●
use of airborne LIDAR (LIght Detection And
Ranging), which has the capability of providing
vertical resolutions of a few centimetres and hor-
izontal resolutions of a few decimetres to metres
(e.g. McKean & Roering, 2004), to produce accu-
rate microtopography maps and long-term ero-
sion and deposition measurements;
●
use of additional remote-sensing techniques to
characterize soil surface conditions and spatial
variability in the factors controlling erosion and
transport (e.g. King
et al
., 2005);
●
a means of quantifying the soil erodibility coef-
ficients from measurable soil properties, to reduce
the current reliance on calibration;
