Geology Reference
In-Depth Information
were corrected for particular events using data
from a portable Delta-T theta probe. These data
were collected close to the date of the event.
Rainfall was measured using six tipping-
bucket rain gauges (1998-2000) and four simple
rain gauges that measure total rainfall only
(1999-2000). The rain gauges were distributed
throughout the catchment. Thus, the number of
rain gauges used in LISEM was between six and
ten. Discharge and sediment concentration were
measured at a V-shaped weir built in 1998 (posi-
tion indicated in Plate 9). The area upstream of
the weir is slightly over 2 km 2 , but the total area
of the catchment is 3.5 km 2 . In the three-year
study period only six events could be measured,
one of which was not used for calibration because
no sediment concentration data were available.
erodible croplands of the Danangou catch-
ment, flow velocity was independent of slope
angle. The Manning equation, however, pre-
dicts an increase of velocity with slope angle.
The most pragmatic solution to this problem
was to allow Manning's n to vary for cropland
as a function of slope, instead of taking a value
that is constant for each particular land use.
For other land uses, no slope dependency was
found, and a constant value for Manning's n
was used.
Use of alternative transport equations. So
far, the LISEM model has only used the stream-
power based equation developed by Govers
(1990) to predict transport capacity. A num-
ber of other equations were tested for the
Danangou catchment (Hessel & Jetten, 2007).
These equations were selected on their abil-
ity to deal with steep slopes and with high
sediment concentrations. Concentration-
dependent fluid density and viscosity were
used in these equations.
(ii) High sediment concentrations
Introduction of a sediment concentration-
dependent particle fall velocity. At high sedi-
ment concentrations, settling velocity is
significantly lower than settling velocity in
clear water. The Chien and Wan (1983; as
reported in Shen & Julien, 1993) equation
was implemented because it was developed
for Chinese conditions.
(iii) Gullies
Introduction of a map with loose material
present in gullies. Observations in the
Danangou catchment showed that loose
material accumulates on gully floors in
between runoff events due to soil falls from
the gully walls. A daily-based gully model
was developed by Hessel and van Asch (2003)
to model the amount of loose material avail-
able on gully floors. During the LISEM run
the only factor determining whether or not
the material is removed is the availability of
transport capacity. The remaining available
material is recalculated during each timestep
and erosion will stop when there is no mate-
rial remaining.
12.5.2
Adaptations to LISEM
Because of steep slopes, high sediment concen-
trations and permanent gullies in the area, the
following changes were made to LISEM.
(i) Steep slopes
Correction for overland flow distance.
Previously, LISEM used the distance between
pixel centres as flow distance. The grid is,
however, essentially a horizontal grid. For
steep slopes the overland flow distance is not
equal to the distance between pixel centres.
For example, if the slope is 45° and the dis-
tance according to the grid is 10 m, the actual
distance over the surface is 14.4 m. To correct
for this, a map showing the overland flow dis-
tance was calculated from the slope map.
Use of sine instead of tangent, both in the
Manning equation, and in equations for stream
power. This is theoretically better since the
slope in these equations is the energy slope.
The sine of the slope angle gives the actual dis-
tance over which friction is exerted on the
flow. For steep slopes, the tangent is much
larger than the sine. Therefore, flow velocity
and stream power will be smaller for steep
slopes when sine is used instead of tangent.
Use of a slope-dependent Manning's n .
Hessel et al . (2003b) found that for the steep
Search WWH ::




Custom Search