Environmental Engineering Reference
In-Depth Information
successfully for the high-gradient experiments than for
the low-gradient ones. A similar result was seen for most
of the other six experiments. Note, in Figure 4.8, that
the rain-impacted texture of the surface of the soil is not
reproduced by the model.
Thus, with microtopography as the only spatially
explicit input, the CA model RillGrow 2 is able to predict
the pattern of rills that will be formed. It does this by
considering rill networks to be emergent, whole-system
responses to interactions between flow, detachment and
deposition at a scale of millimetres. RillGrow continues
to be developed, and a detailed description of the lat-
est version of the model (RillGrow 5) is forthcoming
(Favis-Mortlock, in preparation).
Table 4.2
some characteristics of the Leicester flume
experiments used to evaluate RillGrow 2.
Experiment
Slope angle
Rainfall
(mm h
−
1
)
no.
(degrees)
X09
5
108
X10
10
129
X11
15
125
X12
5
117
X13
10
127
X14
15
126
X15
5
120
X16
10
126
X17
15
131
X18
5
121
X19
10
123
X20
15
125
4.5 Summary and conclusions
Source:
Reproduced with permission from Favis-Mortlock
et al
.
This chapter started by considering the apparently ever-
increasing complexity of models which results from
the desire to improve them. Following this overview,
recent advances in understanding non-linear systems
were described, starting with deterministic chaos and
fractals, and followed by self-organization and emergence.
Examples were drawn from a number of scientific areas.
Next came a description of the CA modelling approach,
which builds upon these concepts, and a more detailed
look at one CA model. Where has this discussion led us?
between splash and flow is most troublesome when flow
detachment is low.
A more difficult test is for the model to reproduce the
observed patterns of erosion at the end of the experiments.
Results for six experiments are shown in Figure 4.7, with
a 3D view in Figure 4.8.
For both experiments, the model was able to reproduce
the
main
elements
of
the
rill
pattern;
again
more
50
600
500
40
400
30
300
20
200
10
100
0
0
0
100 200 300
Measured total discharge (I)
400
500
600
0
10 20 30
Measured total sediment loss (kg)
40
50
(a)
(b)
Figure 4.6
Total discharge (a) and sediment delivery (b) at the end of the Leicester flume experiments, measured values and as
simulated by RillGrow 2. The black square indicates the value for X11; this was used to calibrate the model (Reproduced with
permission from Favis-Mortlock
et al
.).
Search WWH ::
Custom Search