Geology Reference
In-Depth Information
different storms, and by the apparent inability of
LISEM to simulate erosion patterns correctly
(Takken
et al
., 1999; Hessel
et al
., 2003a; Jetten
et
al
., 2003). The first problem might be less signifi-
cant if the same calibrated storm is used for all
scenarios. The implications of the second prob-
lem should be investigated further, and are likely
to differ depending on the aim of the scenario
simulation.
The second assumption is our use of multipli-
cation factors. It is important to realize that the
effect that is predicted for soil and water conser-
vation (SWC) measures is determined by the val-
ues of the multiplication factors that were
assumed (Table 12.2). More quantitative data on
the effects of SWC measures should be gathered
to be able to select values with a higher degree of
certainty, especially for sensitive parameters such
as saturated conductivity and Manning's
n
.
A third assumption is that the selected storm
is representative. In reality, it seems likely that
the effect of soil and water conservation meas-
ures will depend on the intensity and size of a
storm. Gong and Jiang (1979), for example,
reported that reforestation and planting grasses
had similar effects for low-intensity rain, but that
planting grasses was less effective for heavy rain.
Farmers in our study area also indicated that for
really large storms it does not matter what con-
servation measures you have, because there will
be severe erosion anyway. In LISEM, the decreas-
ing effectiveness of conservation measures in
large storms is only because rainfall intensity is
higher, but saturated conductivity remains the
same. LISEM cannot, however, simulate other
reasons for their decreasing effectiveness in larger
storms, such as exceedance of storage capacities
(e.g. of ditches), and destruction of measures.
Despite the fact that LISEM only simulates part
of the effect of changed storm size, Hessel (2002)
showed that if a storm of half the size of the origi-
nal storm was used, only 1% of the runoff was
produced, while a storm of double the size showed
a 6.6-fold increase in predicted soil loss.
Thus, more research is needed before we can
say to what degree the simulation results of
LISEM reflect reality. LISEM is a state-of-the-art
erosion model, and other erosion models will suf-
fer from similar limitations. Therefore, care must
be taken not to read too much into scenario simu-
lation results. Scenario simulations with erosion
models give us useful insights into what might
happen, but they do not tell us what will happen.
It should also be realised that this chapter only
discussed the physical effectiveness of measures,
while for adoption, the socio-economy should
also be considered. Measures can be very effec-
tive, but if they are not acceptable to stakehold-
ers, they cannot be implemented.
12.8 Conclusions
LISEM was adapted to conditions on the Chinese
Loess Plateau, such as steep slopes, high sedi-
ment concentrations and the occurrence of gul-
lies. These adaptations improved the performance
of LISEM slightly after calibration.
Calibration was done both for the catchment
outlet, and for spatial patterns inside the catch-
ment. It was found that LISEM can be calibrated
well for large events at the outlet, but that small
events cannot be simulated properly. The main rea-
son for this seems to be that small discharge events
in the catchment are caused by localized heavy
rainfall. Observed and simulated erosion patterns
showed general similarities, but were quite differ-
ent in detail. These results indicate that LISEM is
more reliable for larger events, and that simulated
erosion patterns should be regarded with caution.
Therefore, a calibrated large event was used to
evaluate the effects of different land use scenarios,
and observed patterns were analysed in relative
terms only. Although such simulations cannot
tell us what will happen after land use change,
they do indicate that a major change in land use
would be far more effective than a change in man-
agement practices.
Acknowledgements
The research described in this chapter was part
of the INCO-DEV Erochina project (Contract
IC18-CT97-0158). We would like to thank all
