Transcript Document

Dryland River Modelling of Water and Sediment Fluxes using a
Representative River Stretch Approach
Eva N. Mueller (1), Ramon J. Batalla (2,3), Axel Bronstert (1)
(1) Institute of Geoecology, University of Potsdam, Postfach 60 15 53, 14415 Potsdam, Germany (2) Department of Environmental and Soil Sciences, University of Lleida, 25918 Lleida, Spain(3) Forest
Technology Centre of Catalonia, Pujada del Seminari, 25280 Solsona, Spain ([email protected])
Modelling Approach
Result III. Composite Modelling of the River
The study investigates process-based modelling of sediment
transport of a Mediterranean mountainous dryland river
within the meso-scale Isabena watershed (445 km2) of the
Pre-Pyrenean region in NE Spain. The modelling study is
carried out to enable the quantification of sediment fluxes
that erode mainly from local badland areas during highintensity rainstorm events, resulting in high-density sediment Badland Area
fluxes in the river system and severe sedimentation of a
downstream reservoir thus threatening future water supply
(see map). The transport of sediments from the badland
areas in the river’s main stem was modelled using a
composition of five representative river stretches. The
proposed modelling framework enabled a detailed spatial
and temporal examination of complex deposition and river
bed degradation patterns as well as an insight into the
temporary sediment storage behaviour of the riverbed and of
its floodplain along the entire river flow path. Central
research questions are: (1) how to parameterise
heterogeneous river sections at the meso-scale, (2) how
sensitive are the model parameters of the sediment routine to
high or low flow conditions, (3) how to model the complex
processes of temporary storage and re-degradation of fine
sediments in the riverbed?
The water routing is based on the kinematic wave approximation after Muskingum. Flow rate,
velocity and flow depth are calculated for each river stretch and an hourly time step using
Manning’s flow equation. Sediment transport is modelled using the transport capacity concept
based on a power function of the stream velocity as implemented e.g. in the SWAT model. The
maximum concentration of suspended sediment that can be transported by the water is given by:
The Isabena River downstream of the badland area was modelled as a composite of characteristic
stretches along a 32.9 km stem using measured time series of water discharge from the year 2000
with an hourly time step as input data. As no data for sediment discharge was available, a
hypothetical scenario was assumed with a constant sediment input of 50 g/l occurring when the
daily rainfall amount exceeds 20 mm. The graphics below show the water and sediment fluxes for
no degradation on the left size and full re-degradation of deposited sediments on the right side.
WITHOUT DEGRADATION
WITH DEGRADATION
1
150
4
8
16
32
64
128
256
4
150
200
250
300
-50
0
50
River width [m]
100
6
5
4
3
0
50
100
150
200
250
16
32
64
128
256
2
512 1,024 2,048 4,096
0
-150
100
5
4
3
2
50
100
150
200
250
8
16
32
64
128
256
Flow velocity [m/s]
Depth [m]
4
4
2
512 1,024 2,048 4,096
6
5
4
3
2
0
-100
-50
0
50
River width [m]
100
150
5
50
100
150
200
250
4
8
16
32
64
128
256
512 1,024 2,048 4,096
2
0
-150
5
4
3
2
1
-100
-50
0
50
River width [m]
100
150
Deposition [tons/h]
5
60
40
20
10
25
50
50
100
150
200
Discharge [m3/s]
250
300
25
50
100
120,000
25,000,000
100,000
20,000,000
80,000
15,000,000
60,000
10,000,000
40,000
20,000
5
10
15
20
River length [km]
Sediment outflow
25
Deposition
30
120,000
25,000,000
100,000
20,000,000
80,000
15,000,000
60,000
10,000,000
40,000
5,000,000
20,000
0
0
35
0
0
Degradation
5
Water outflow
10
15
20
River length [km]
Sediment outflow
25
Deposition
30
35
Degradation
Time [hours]
0.0
0.0
1.4
2.7
3.3
1.4
2.7
3.3
7.0
7.0
10.1
10.1
11.7
11.7
14.3
15.0
14.3
15.0
21.8
22.5
21.8
22.5
24.3
24.3
5,000
10,000
5,000
5
10 25
50 100 150 300
1
5
10
25
50
5
10
25
50
10,000
5,000
1
Discharge [m3/s]
5
10
25
50
100
Discharge [m3/s]
20,000
100
80
60
40
20
15,000
10,000
5,000
0
5
10 25
50 100 150 300
15,000
10,000
5,000
0
1
5
10
25
50
100
1
Discharge [m3/s]
5
10
25
50
100
Discharge [m3/s]
20,000
100
80
60
40
20
5
10 25
50 100 150 300
Discharge [m3/s]
• The sensitivity analysis showed that the uncertainty in regard to the model parameters of the
transport capacity equation is enormous. It thus appears unfeasible to use the modelling approach
for ungauged rivers. It is furthermore questionable to which extent measured data from a specific
type of river section can be used for the parameterisation of another river section type.
• Temporary storage of sediments in the riverbed appears to play an important role for the
sediment export out of the river system. For the model scenario without degradation, a large
amount of sediments is deposited within the entire river stem and does not reach the outlet. For the
scenario with degradation, a large amount of the sediment is temporarily stored preferably in the
shallow river stretches with large floodplains, and remobilised during small floods several days
after the main flood event. The rather crude model approach thus enabled the reproduction of
complex, highly non-linear transport processes along the river.
15,000
10,000
5,000
0
1
32.9
• The river data of the field study led to the derivation of five distinct types of river sections that
enabled the identification of key model parameters which adequately describe the intrinsic
heterogeneity of the mountainous dryland at the meso-scale.
15,000
100
29.4
30.3
31.3
100
0
1
29.4
30.3
31.3
26.9
27.4
Discussion
Discharge [m3/s]
15,000
26.9
27.4
32.9
10,000
100
0
0
0
10
15,000
Discharge [m3/s]
120
6
5
0
1
Sediment outflow [tons/h]
2
5,000
50 100 150 300
80
1
Max. Sediment Conc. [g/l]
Flow velocity [m/s]
Depth [m]
1
4
1
Discharge [m3/s]
10,000
Discharge [m3/s]
7
6
10 25
100
300
8
8
100
20,000
Discharge [m3/s]
Type E
shallow, narrow
Slope: 0.75 %
Manning: 0.04
50
0
0
0
10
25
120
1
0
-150
20
1
8
6
10
15,000
Discharge [m3/s]
7
2
40
Discharge [m3/s]
8
1
60
300
5,000
Discharge [m3/s]
0
0
10
Type D
shallow, wide
Slope: 0.75 %
Manning: 0.02
6
150
5
120
0
-50
0
50
River width [m]
1
Discharge [m3/s]
1
-100
80
1
Max. Sediment Conc. [g/l]
8
4
50 100 150 300
100
300
8
Flow velocity [m/s]
Depth [m]
4
140,000
10,000
20,000
Discharge [m3/s]
6
10 25
0
7
2
5
120
7
150
8
1
1
1
-100
30,000,000
15,000
0
Discharge [m3/s]
2
10
Type C
medium, regular
Slope: 1.3 %
Manning: 0.03
100
8
6
0
-150
50
5,000
0
Discharge [m3/s]
2
512 1,024 2,048 4,096
0
0
Flow velocity [m/s]
Depth [m]
2
20
Deposition [tons/h]
100
40
10,000
Deposition [tons/h]
-50
0
50
River width [m]
60
15,000
Deposition [tons/h]
-100
8
1
3
2
10
Type B
steep, deep, wide
Slope: 2 %
Manning: 0.02
4
80
Sediment outflow [tons/h]
0
-150
DEPOSITION
Deposition [tons/h]
2
1,024 2,048 4,096
5
Sediment outflow [tons/h]
4
6
100
Sediment outflow [tons/h]
512
140,000
Time [hours]
20,000
Sediment outflow [tons/h]
256
6
SEDIMENT OUTFLOW
120
Max. Sediment Conc. [g/l]
8
Max. Sediment Conc. [g/l]
128
30,000,000
River length [km]
MAX. SEDIMENT CONC.
Max. Sediment Conc. [g/l]
64
160,000
Water outflow
River length [km]
FLOW VELOCITY
Flow velocity [m/s]
Depth [m]
32
35,000,000
The uncertainties associated with the right choice of the a and b parameters are considerable, more
pronounced for the steep, upslope stretches, and increase over proportionally for larger water
discharge.
7
16
35,000,000
0
Critical model input parameters are: the Manning‘s roughness factor, slope, shape of cross-section,
channel cover factor, the a and b parameters of the power function and the erodibility factor of the
riverbed. The first four parameters were derived from the field data for the five representative river
stretch types. A sensitivity analysis was carried out to investigate the influence of the power function
parameters with a having a literature reference range of 0.01 – 0.0001 and b between 1.0-1.5. The
first two columns below show the flow velocity and the corresponding maximum sediment
concentrations at maximal transport capacity as a function of discharge for the five characteristic
river types. The other two columns depict the corresponding sediment discharge out of and the
deposition inside a reach for a model scenario with a hypothetical incoming sediment concentration
of 50 g/l. The circles O show the model results for a parameter set of (a, b) = (1.7, 0.016), that were
derived from a limited data set on water and sediment discharge data for the Isabena River.
8
8
180,000
0
Result II: Sensitivity Analysis
10
4
40,000,000
160,000
5,000,000
where Conccurrent is the sediment concentration at the current time step (ton/m3), V is the volume of
water in the reach segment (m3), C is the channel cover factor (-) and E is the channel erodibility
factor (-).
The Isabena River is characterised by a very heterogeneous spatial distribution of river forms and
properties, which makes the parameterisation of state-of-the-art river models a difficult task.
Steep, narrow, deep incised mountain torrents with rocky, gravely riverbeds in the upper parts of
the catchment alternate with shallow, plain and very wide riverbeds and large floodplains with silty
riverbed materials in the lower catchment area, with parts of the river system having an ephemeral
flow regime. To enable model parameterisation, five representative river stretches were derived
from the results of a field campaign that investigated central model parameters such as the crosssectional profile, slope, roughness and the gradation of the riverbed material.
180,000
Total water volume [m3]
SedDegradation = (Concmax – Conccurrent) · V · C · E
Result I: Field Studies
Type A
steep, narrow
Slope: 3 %
Manning: 0.02
Total water volume [m3]
where v is the flow velocity in (m/s), Concmax is the maximum sediment concentration for each river
stretch in (ton/m3), and a and b are user-defined coefficients. Riverbed degradation is calculated by:
40,000,000
Total sediment mass [tons]
Concmax = a · v b
Total sediment mass [tons]
Introduction
15,000
Acknowledgements
10,000
5,000
This research was carried out within the SESAM (Sediment Export from Semi-Arid Catchments:
Measurement and Modelling) project and was funded by the Deutsche Forschungsgemeinschaft.
0
1
5
10
25
50
Discharge [m3/s]
100
1
5
10
25
50
Discharge [m3/s]
100