Geology Reference
In-Depth Information
sediment fan at the base of the gully at site 2.
Within the buffer, the model shows that the
transport capacity is greater than detachment for
sand, but not for clay or silt. This explains the
deposition of some clay and silt particles within
the buffer as the sediment load reduces to equal
the transport capacity. As expected, the model
indicates that sediment movement through the
buffer is limited by the transport capacity.
The failure of the model to predict the delivery
of sand to the buffer was somewhat surprising
because when applied to field sites in Bedfordshire
and Cambridgeshire, the eroded material leaving
the base of the fields simulated on sandy loam
soils was predicted to contain as much as 79-83%
sand (Morgan & Duzant, 2008). The difference
may well reflect the greater slope of the latter
sites which were 4-9° compared with only 2-3° at
the sites in this study.
Having shown that the chosen model can be set
up to give acceptable predictions of sediment depo-
sition within a vegetated buffer, as compared to
field measurement, the model was then run to
determine whether the predicted mean annual soil
loss from the field at Site 2 fell within acceptable
limits. Using meteorological data from Taunton,
the following input data were used: mean annual
rainfall, 1075.5 mm; mean annual temperature,
9.86 °C; and the mean annual number of rain days,
84. The mean annual soil loss from the field's con-
tributing area (60 m long and 33 m wide) was pre-
dicted as 18.2 t ha
−1
, which is far higher than the 1 t
ha
−1
proposed as the upper limit to control down-
stream pollution arising from runoff from agricul-
tural land (Moldenhauer & Onstad, 1975; Morgan,
2005). As already shown, much of this soil is
trapped within the downslope buffer. Below the
buffer, the mean annual soil loss from the field was
predicted at 0.8 t ha
−1
, indicating that the buffer is
very effective. However, this calculation assumes
that the sediment passes uniformly through the
whole of the buffer width. If, given the field obser-
vations above, the model is set up to simulate the
concentration of the runoff and sediment into only
8 m of the buffer width, the effective area becomes
2044 m
2
(1980 m
2
for the field and 64 m
2
for the
buffer) and the predicted mean annual soil loss
below the buffer then increases slightly to 0.9 t ha
−1
,
which is still acceptable. This result indicates that
the present buffer feature should be capable of
reducing sediment delivery to the adjacent water-
course to a level which will also control pollution
from phosphorus transported by surface runoff in
particulate form. In reality, however, the recent
installation of the grass waterway in the contribut-
ing area upslope of the buffer (Plate 12) has proba-
bly made the buffer redundant. The grass waterway
controls the detachment and transport of soil par-
ticles by runoff closer to source rather than relying
on the vegetated buffer to create a sediment sink.
13.5 Conclusions
The best results with the model are obtained when
both the effective contributing area and the effec-
tive buffer area are defined from field observations
and measurements. Only where the slope is planar
and there are no concentrations of flow is the
model likely to give reasonable predictions using
gross areas of the individual slope elements. When
operated with field observations for the input data,
and set up to reflect the patterns of the flow paths
and the spatial distribution of erosion and sedi-
mentation observed in the field, the model gives
good predictions of the amount of sedimentation
within the vegetated buffers. The model thus deals
reasonably well with concentrated flow in the
contributing upslope area and the delivery of run-
off and sediment to the buffer at a limited number
of points rather than uniformly across the buffer
width. At present the model does not predict the
particle size of the deposited material particularly
well because it simulates the deposition of most of
the sand particles close to their point of detach-
ment. Further development of the model will be
needed to represent better the detachment, trans-
port and deposition of sand. One approach would
be to use lower fall velocities than those conven-
tionally recognized for sand.
Despite its limitations, the model does simulate
the effectiveness of vegetated buffer strips using
explicit input parameters to describe the vegeta-
tion. It is therefore appropriate to use in the design
