Geology Reference
In-Depth Information
D
f
=
y
· (
TC
−
C
) ·
Q
(12.2a)
between summer storms and winter storms. De
Roo and Jetten (1999) identified saturated conduc-
tivity, suction at the wetting front, and initial mois-
ture content as sensitive parameters.
More recently, several authors have pointed to
the necessity of calibrating process-based, distrib-
uted models in a spatial way (e.g. Jetten
et al
., 1996,
2003; Takken
et al
. 1999; Beven, 2002; see also
Section 3.5). Such a calibration is a logical step since
the main advantage of distributed models over
lumped models should be their ability to predict
spatial patterns. Also, there are circumstances
where the location of erosion or deposition in a
catchment is more important than the precise
amount of water/sediment passing the outlet: for
example, to design effective anti-erosion measures.
D
p
=
(
TC
−
C
)
· w ·
W
· DX
(12.2b)
where
D
f/p
is the net rate of sediment detach-
ment/deposition by flow (kg s
−1
),
y
is the
efficiency coefficient (value between 0 and 1),
TC
is the sediment transport capacity (kg m
−3
),
C
is the sediment concentration (kg m
−3
), Q is the
discharge (m
3
s
−1
),
w
is the particle settling veloc-
ity (m s
−1
),
w
is the flow width (m) and
DX
is the
pixel length (m).
To avoid mass balance problems, the detach-
ment in a timestep cannot exceed the remaining
transport capacity (
TC
−
C
), while the deposition
cannot be more than there is sediment in the
flow. Current versions of LISEM (2.5 and above)
have a possibility of including a stationary base-
flow in the simulations.
12.4
LISEM in relation to the Loess Plateau
Some of the main algorithms of LISEM are only
valid for a given range of circumstances. Since the
circumstances on the Loess Plateau are very differ-
ent from the circumstances for which LISEM was
developed, it is possible that application limits
will be encountered when applying LISEM to the
Loess Plateau. At present LISEM does not specifi-
cally take into account the effects of steep slopes,
high sediment concentrations and the presence of
gullies.
Soil erosion models have not been applied to
steep slopes very often. Since their focus has been
on predicting erosion from arable land in Europe
and the US, which is generally on gentle to moder-
ate slopes, not much attention has been paid to
slope steepness. Slopes of 10% are usually consid-
ered 'steep', while in many other areas of the
world, including China, cropland occurs on much
steeper slopes. The velocity equations and the
sediment transport equations that are commonly
used in erosion models have not been developed or
tested for such steep slopes as those of the Chinese
Loess Plateau. LISEM calculates flow velocity
with Manning's equation and subsequently calcu-
lates stream power and transport capacity. Since
unit stream power is the product of the velocity
and the (energy) slope, the slope angle influences
12.3.2
LISEM calibration and validation
Although, theoretically, fully physically-based
models should not have to be calibrated, reality is
different (see Chapter 3 for a more general discus-
sion of this issue). Models are never fully physi-
cally-based, and many authors have demonstrated
the need to calibrate process-based erosion mod-
els to obtain an acceptable predictive quality (e.g.
Jetten
et al
., 1999). In the case of hydrological/
erosion models, calibration has mostly been done
using measured data at the outlet of the plot or
catchment.
LISEM was developed for the province of
Limburg, The Netherlands, and has been previously
calibrated and validated for several catchments in
the Loess region of northwestern Europe. De Roo
et al
. (1996b) found that the most sensitive variable
in the prediction of runoff is saturated hydraulic
conductivity. For the prediction of soil erosion,
LISEM was most sensitive to changes in Manning's
n
and transport capacity. De Roo
et al
. (1996b) found
that LISEM gave reasonable results for 60% of the
storms that were modelled. They attributed the dis-
crepancy for the other 40% to spatial and temporal
variability in saturated hydraulic conductivity
and initial moisture content, and to differences
