See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/274859620

SEDIMENT YIELD AND TRANSPORTATION ANALYSIS

Article · January 2008

DOI: 10.2208/prohe.52.157

CITATIONS READS

4 102

2 authors, including:

Tetsuya Sumi
Kyoto University
121 PUBLICATIONS   545 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Integrated Reservoir Operation System for Mitigation of Flood Damage due to Super Typhoon View project

The impacts of hydropower dams uptream on the Vietnamese Mekong Delta View project

All content following this page was uploaded by Tetsuya Sumi on 15 June 2015.

The user has requested enhancement of the downloaded file.

 Annual Journal of Hydraulic Engineering, JSCE, Vol.52, 2008, February
Annual Journal of Hydraulic Engineering, JSCE, Vol.52, 2008, February

SEDIMENT YIELD AND TRANSPORTATION

ANALYSIS: CASE STUDY ON MANAGAWA
RIVER BASIN

Chadin CHUTACHINDAKATE1 and Tetsuya SUMI2

1Member of JSCE, Doctoral Student, Dept. of Civil and Earth Resources Engineering, Kyoto University (Katsura
Campus, Nishikyo-ku, Kyoto 615-8540, Japan)
2Member of JSCE, D. Eng. , Associate Professor, Dept. of Civil and Earth Resources Engineering, Kyoto University
(Katsura Campus, Nishikyo-ku, Kyoto 615-8540, Japan)

Sediment system in watershed is not only sediment yield but also including sediment transportation
along the rivers. In this study, the Geographic Information System (GIS) combined with sediment yield
model can be enhancing the evaluation of soil erosion estimation. Surface erosion on Managawa river
basin is computed with the Modified Universal Soil Loss Equation (MUSLE) and it is verified to reflect
the hydrological processes be able to estimate soil losses. In the sediment transport routing module, total
load equation is applied to carry sediment from soil surface erosion to deposit in Managawa dam.
According to annual accumulation sediment volume data in Managawa reservoir during 1981 – 2003, the
establish model and simulation results are satisfy. The efficiency of the Modified Universal Equation with
sediment routing in rivers is more than the simple Modified Universal Equation.

Key Words : Sediment Yield, Sediment Transportation, Soil Erosion, MUSLE, Managawa Dam

1. INTRODUCTION objective of this study is computing annual

depositing sediment volume in Managawa dam by
Managawa Dam is located on the high rainfall using soil erosion model and sediment
density and steep slope area caused sediment transportation model. Erosion on sub-basin caused
problems. Every year the bottom levels of reservoir by rainfall and surface runoff is computed with
are observed then we can know annual sediment Modified Universal Soil Loss Equation (MUSLE),
depositing volume. There exist many kinds of soil which time interval is day continuously from 1983
erosion models, both physical models and empirical to 2004. After outlets of each sub-basin, sediment
models and also there are many useful numerical will be transported by main channel. This study
formulas to predict annual sediment yield. The most shows the results of these combined systems.
popular soil erosion equation is Universal Soil Loss
Equation (USLE). Simple empirical methods such 2. STUDY AREA
as Universal Soil Loss Equation (USLE) (Musgrave,
1947; Wishchmeier and Smith, 1965), Revised Managawa Dam constructed during 1965-1977
Universal Soil Loss Equation (RUSLE) (Renard et in Fukui prefecture is located at latitude 35º 55' 50"
al., 1991) or Modified Universal Soil Loss Equation N and longitude 136º 32' 31" E. Managawa Dam is
(MUSLE) (Williams, 1975) are frequently used for a concrete arch dam with 127.5 m height, 357 m
estimation of surface erosion and sediment yield width and 115 MCM capacity designed for
from catchment areas. In Modified Universal Soil irrigation, water supply and power generation where
Loss Equation (MUSLE), the rainfall energy factor Managawa river is a tributary of Kuzurui river.
is replaced with a runoff factor and optimizes Catchment area above the dam is about 223.7 km2
hydrologic process of sediment yield thus these that the mean elevation is 830 m above mean sea
improve the sediment yield prediction. The main level and land slope is about 0.45. Since there are

- 157 -
 Fig.1 Managawa river basin Fig.2 Land use 1997, 2 = Agricultures, 5= Forest, 6=Waste
Land, 7=Building, A= Small Building, B=Water Body
Kumokawa Dam and Sasougawa Dam situated on
up stream of Managawa Dam as shown in Fig.1, most of land use of the study area is forest and it
sediment will be captured by those dams but does not so much change by time as shown in Fig.2.
Kumokawa Dam has been filled full by sediment. Calculating soil erosion, Modified Universal
Therefore sediment supply to Managawa Dam is Soil Loss Equation (MUSLE), in this study apply
also generated from watershed above Kumokawa with the Soil and Water Assessment Tool (SWAT)
Dam. During the study period, 1981-2004, the that the watershed modeling framework is
average annual rainfall is 2391 mm. The area is delineated starting from the digital description of the
covered by forest where accounts for 94% area of landscape as Digital Elevation Map (DEM), land
total watershed. The major soil types in the study use and soil data sets using ArcView interface,
area are sandstone, mudstone and conglomerate Spatial Analyst, with geomorphologic assessment
(Managawa Dam office, 2005) procedures to obtain soil erosion from each
sub-basin. Managawa river basin was divided in to 9
3. METHODOLOGY sub-basins which one sub-basin on upper
Sasougawa Dam is neglected because that dam will
capture all sediment from upper part. After that the
In this study, the computing an annual sediment calculated daily sediment supply from each
volume depositing in dam by mathematical model is
sub-basin will be taken to input as lateral sediment
necessary to input the models; soil erosion model, inflow to Managawa River model which
hydraulic model and sediment transport model, with hydrodynamic (HD) and sediment transport (ST)
hydrology data, hydrodynamic data, sediment data,
models are calculated by MIKE 11 developed by
geographic data and topographic data. DHI Water and Environment. Flow chart of this
study is shown in Fig.3.
(1) Data collection and data analysis
Hourly rainfall data from 1981-2004 were Soil type
Rainfall DEM Land use
collected by 8 rain gauges located on Managawa
river basin and the results from plotting double mass
Input
curve of each rainfall stations are reliable.
Distribution of rainfall could be affected by MUSLE (SWAT model)
topographical data such as elevation and so on, the Output
Thiessen method was used to estimate rainfall
within the entire catchment. Soil erosion
Discharge and water level data are available in
hourly to input as boundary condition in hydraulic Discharge X-section Sediment
model which outflows from Kumokawa Dam and River line Property
Sasougawa Dam are the upper boundaries and water
level at Managawa Dam is the lower boundary. Input
Geographic Information System (GIS) data used Hydrodynamic and Sediment Transport model
for finding out parameters in soil erosion model are
Digital Elevation Model (DEM) 50m x 50m, land Output
use and soil type. For land use information, there are Sediment volume
available in 1976, 1987, 1991 and 1997 however the
Fig.3 Scope of this study

- 158 -
(2) Modified Universal Soil Loss Equation
(MUSLE)
The Soil and Water Assessment Tool (SWAT) is
a long term distributed parameter model, designed
to predict the impact of land management practice in
a watershed (Arnold et al., 1998). In this study, the
SWAT ArcView interface (DiLuzio et al., 2001)
was used to write SWAT input files from GIS data
layers. SWAT model calculates soil erosion caused
by rainfall-runoff process using MUSLE. The model
is a modified form of the USLE. The difference
between the two approaches that in MUSLE rainfall
energy factor is replaced with a runoff factor which
represents energy used in detaching and transporting
sediment. SWAT model requires a Digital Elevation Fig.4 Sub-basins in this study
Model (DEM) from which it determines the No.6
drainage network and divides the basin into No.1 No.4 No.7 No.8
sub-basins defined by grid cells, spatially related No.2 No.5
one to another, that each has geographic position in No.3
the watershed defined by surface topography.
Sub-basin No.
This study applies SWAT model only to find out
the soil erosion of each sub-basin at each outlet
point. The MUSLE is used in this study which is 8.2 km
given as Eq.(1). From MUSLE, the shortest time
10.6 km
interval of output is daily and this study need daily
sediment yield data to input in sediment transport 14.4 km
model. Fig.5 Location that sediment yield of each sub-basin from
MUSLE supplied to the main river system

Y = 11.8(Qs q p Aarea ) 0.56 K ⋅ C ⋅ P ⋅ LS ⋅ CFRG (1) (3) Hydrodynamic and Sediment Transport
models
Where Y is the sediment yield on a given day The hydrodynamic module (HD) contains an
(ton), Qs is the surface runoff (mm), qp is the peak implicit, finite difference computation of unsteady
runoff rate (m3/s), Aarea is area (km2), K is the USLE flows in rivers. The non-linear equations of open
soil erodibility factor, C is the USLE cover and channel flow (Saint Venant Equation) can be solved
management factor, P is the USLE support practice numerical between all grid points at specified time
factor, LS is the USLE topographic factor and intervals for given boundary conditions.
CFRG is the coarse fragment factor. SWAT The basic equation is governed by the continuity
estimates the surface runoff (Qs) with the SCS curve and momentum equations:
number method and the peak runoff rate is
calculated with the rational method: ∂Q ∂A
+ =q (3)
∂x ∂t
c ⋅ i ⋅ Aarea ⎛ Q2 ⎞
qp = (2) ∂⎜⎜ α ⎟⎟
3 .6 ∂Q ⎝ A ⎠ ⎛ ∂h ⎞ n gA Q Q
2

+ + gA⎜ + S 0 ⎟ + = 0 (4)
∂t ∂x ⎝ ∂x ⎠
4
A2 R 3
Where qp is the peak runoff rate (m3/s), Aarea is
area (km2), c is the runoff coefficient, i is the rainfall
intensity (mm/hr) and 3.6 is a unit conversion factor. Where Q is discharge (m3/s), A is cross section
There are 8 sub-basins which sub-basin no.1, 2 area (m2), q is lateral inflow (m2/s), α is
and 3 are on the dam area so those sub-basins will momentum distribution coefficient, h is water level
directly supply sediment into dam. The soil erosion (m), S0 is bed slope, n is Manning’s roughness
of those areas from MUSLE will be directly sum up coefficient (s/m1/3) and R is hydraulic radius (m)
to total sediment volume as shown in Fig.4 and The sediment transport computations are made
Fig.5 and the sub-basin No.9 is not supplying in parallel with the hydrodynamic computations.
sediment to Managawa Dam. The sediment transport equations are solved in time
and space as an implicit function of the

- 159 -
corresponding values of the hydraulic parameters. In 80

Changing in sediment volume (%)

this study, total load model Acker and White (1973)
60
presented a semi-empirical sediment transport mode
is used for computation. 40
a) Boundary conditions, calibration and 20
verification for Hydrodynamic model
Outflows from Kumokawa Dam and Sasougawa 0
Dam are set to upstream hydrodynamic boundaries 0 0.002 0.004 0.006 0.008 0.01
-20
and water level at Managawa Dam is set to down
stream boundary as shown in Fig.6. -40
Grain size daimeter (m)
To calibrate Hydrodynamic model by changing
Fig.8 Sensitivity analysis of grain size diameter effect to
roughness coefficient which this study select year
accumulation sediment volume in 1995
1995 and 2004 for calibration and year 1998, 2002
and 2003 are for verification. The efficiency index
For reference, sensitivity analysis of grain size
(EI) for calibration is 94 % and for verification is
diameter is shown in Fig.8. When grain size
88% while roughness coefficients, Manning M, are
becomes finer, sediment accumulation volume
20, 20 and 15 for branch 1, 2 and 3 in sequence.
becomes larger.
b) Sensitivity analysis of sediment transport
model
Daily sediment yield from MUSLE computation
4. RESULTS
will be supplied to the main river system at outlet
points of each sub-basin. Acker and White equation (1) Topographic results
need to input grain size diameter (D50). The This study has generated a watershed to 8
sub-basins as shown in Fig.4 excluding Sasougawa
distributed grain size diagram of 4 sediment
dam basin. From 50m x 50m grid, we can obtain
samplings in Managawa Dam in 1998 is shown in
Fig.7. Sample 1, 2 and 3 were taken near dam topographic information as area, average slope and
average elevation in Table 1.
body and sample 4 was taken around upstream delta
deposits. Based on these data, the sediment
transportation model will apply the approximate (2) Soil surface erosion results
Average annual sediment yields (t/ha) for each
value for D50 as 0.005 m in this study.
sub-basin during 1981 to 2004 were computed by
MUSLE with SWAT model as shown in Table 2.
Managawa Dam
D/S BD Table 1 Topographic information of each sub-basin
Branch3
Nagashima Sub Area Slope Elevation Manning’s
station basin (km2) (%) (m) n
Kumokawa Dam 1 9.39 0.44 546 0.035
Sasougawa Dam 2 12.76 0.52 363 0.035
U/S BD
Branch1 Branch2 U/S BD 3 11.54 0.52 546 0.035
4 9.84 0.45 704 0.035
5 27.35 0.45 557 0.035
Fig.6 Hydrodynamic boundary conditions
6 8.01 0.49 1089 0.035
7 17.64 0.42 705 0.035
100
8 53.81 0.44 629 0.035
80
Table 2 Annual average soil surface erosion from MUSLE
60
Percent

Sub-basin Average Erosion (t/ha/year)

Sample1
40 1 9.66
Sample2
Sample3 2 12.04
20 Sample4
Model
3 12.08
4 9.86
0
0.00001 0.0001 0.001 0.01 0.1 1 5 10.39
Size (m) 6 10.94
Fig.7 Sediment distribution curve of 4 samplings in 7 9.26
Managawa Dam, 1998 8 10.58

- 160 -
 Sub-basin No. 2 and 3 show high erosion rate 5. DISCUSSION
because these slopes are so steep. The average
sediment yield from the whole watershed calculated The computed sediment volumes in this study
from Eq.(5) is about 60,161 m3/year. show large differences with the observed data
because the observed data have some errors in some
∑ Ya ⋅ Area ⋅ 100 year. Therefore, total accumulated sediment
Sy = (5) volumes were compared between them. Observed
σ
and computed total sediment volumes until year
Where Sy is the average sediment yield (m3), Ya
2003 were 1,747,197 m3 and 1,560,381 m3
is the sediment yield (t/ha), Area is sub-basin area
respectively, which error is about 10%. In some
(km2) and σ is soil density about 2.65 t/m3.
year, the accumulative data are not much difference
as shown in Fig.9. Therefore, the sediment yield and
(3) Sediment volume in dam
sediment transport model in this study can be used
The volume of sediment depositing in the large
to estimate the sediment accumulation volume in
dams are measured every year in Japan. Sediment
Managawa Dam.
volume data of Managawa Dam is also available
In order to determine efficiencies of the MUSLE
from starting operation until recently year but some
model, a logarithmic form of Nash-Sutchliffe model
data was disappear because of technical error as
is adopted, which is given by:
shown in Table 3.
After sediment yield from each sub-basin was
n
computed by MUSLE, those data were used to input
in total load transport model, Acker and White ∑ (ln( A
e =1
eo ) − ln( Aep )) 2
(1973). And then sediment will be routed along the Z ln = 1 − n (6)
river and deposited in the dam reservoir. Computed
results are shown with observed ones in Table 3.
∑ (ln( A
e =1
eo ) − Aln m ) 2

Observed data in 2004 was extremely large mainly

because of the Fukui heavy rainfall in July 18, 2004.
Where Aeo is the observed soil loss for event
Table 3 Observed and computed sediment volume, 1981-2004, e, Aep is the predicted soil loss and Aln m is the
in Managawa Dam
mean value of ln( Aeo ) for all the events selected.
Year 3
Observed Volume (m ) 3
Computed Volume (m ) The Nash-Sutchliffe efficiency for this
1981 - 42,304 combined MUSLE and sediment transport model is
1982 - 75,108 about 0.30 where a model efficiency of 1.0
1983 78,664 77,730 represents a perfect fit of the model to observed
1984 15,768 66,523 values. Negative values indicate that use of the
1985 166,907 53,337 average is a better predictor than the model
1986 16,726 73,660 (Nash-Sutchliffe, 1970). The Nash-Sutchliffe
1987 23,477 75,525 efficiency for the simple MUSLE is -0.84 and then
1988 59,272 54,324 the present model of MUSLE with the river routing
1989 79,652 93,670 system is increasing the efficiency.
1990 353,591 106,494
1991 - 58,157 Rainfall
0 0 0 0 0 0 0 0 0 00 0 0
5 0 1 0 1 5 20 2 5 3 0 3 5

1992 - 41,525 Computed

Sediment Volume (cu.m)

Observed
6

1993 175,980 66,901

Annual Rainfall (mm)
+0 E+0 E+0 E+0 E+0 E+0
3

1994 - 52,652
6

1995 106,271 60,281

3
6

1996 - 52,339
2
6

1997 4,812 42,950

2
6

1998 249,132 121,724

1
5

1999 41,822 66,315

5E
19 0

2000 - 50,654
81
83
85
87
89
91
93
95
97
99
01
03
19
19
19
19
19
19
19
19
19
20
20

2001 101,227 81,551

Year
2002 237,574 83,042
Fig.9 Annual rainfall, observed and computed accumulation
2003 36,322 63,617
sediment volume
2004 1,078,341 145,913
(-) data is not available

- 161 -
 3500
reservoir sedimentation volumes and its tendency if
basic characteristics of each catchment such as
Annual Rainfall (mm)
3000
2500 topographical, geological, meteorological conditions
2000 may change.
1500
1000
500 6. CONCLUSION
0
0 1 2 3 4 5
Annual Disharge (MCM)
This study is an attempt to estimate the yearly
volume of sediment deposition in Managawa Dam
Fig.10 Relationship between Annual Discharge and Annual using the Modified Universal Soil Loss Equation
Rainfall, 1981-2004 (MUSLE) with sediment transport model. The
Computed Sediment Volume

400 MUSLE was developed for simulating the impact of

350 land management practices on eco-hydrologic
300
system (Arnold et al., 1995). The results of annual
1000x(cu.m)

250
200 sediment accumulating volume for this study area
150 show large differences with the observed data but
100 total volumes almost coincided very well. In order
50
0
to improve the model, other modifications may be
0 1 2 3 4 5 needed to compute sediment yield and transport
Annual Discharge (MCM) process in high flood periods. However, this model
can be used to estimate reservoir sedimentation
Fig.11 Relationship between Annual Discharge and Computed
volumes and its tendency if catchment’s conditions
Sediment Volume
may change.
Observed Sediment Volume

400
350
300 ACKNOWLEDGMENT: The authors would like
1000x(cu.m)

250
200
to express our sincere gratitude to Managawa Dam
150 Office and NEWJEC for their willingness to supply
100 the necessary data for analysis in this research.
50
0
0 1 2 3 4 5
REFERENCES
Annual Discharge (MCM) 1) Eileen Chen and D. Scott Mackay: Effects of
distribution-based parameter aggregation on a spatially
Fig.12 Relationship between Annual Discharge and Observed distributed agricultural non point source pollution model,
Sediment Volume Science Direct, Journal of Hydrology, 2004
2) Fumitoshi IMAIZUMI and Roy C. SIDLE: Relationship
In order to improve the model, the relationship between Sediment Supply and Transport Process in
among annual rainfall, annual discharge, and Miyagawa Dam Catchment, Annuals of Disas. Prev. Res.
observed and computed sediment volumes are Inst., Kyoto University, 2005
analyzed. The annual runoff flowed into Managawa 3) S.L. Neitsch, J.G. Arnold, J.R. Kiniry, J.R. Williams, Soil
Dam has good correlation with annual rainfall as and Water Assessment Tool Theoretical Documentation,
plotted in Fig.10. The annual sediment volume in Agriculture Reserch Service and Experiment Station, 2005
dam is not only depended on the soil surface 4) Chjeng-Lun Shieh and Sen-Yaun Lee, Sediment yield model
properties and rainfall but depended on discharge for a watershed: a case study of Choushui River, Taiwan,
also. Fig.11 and Fig.12 show that the computed Eropean Geosciences Union, 2005
sediment volume from the model has moderately 5) Rabin Bhattarai and Dushmata Dutta, Estimation of Soil
Erosion and Sediment Yeild Using GIS at Catchment Scale,
correlation with annual discharge but the observed
Springer Science, 2006
one is varying more widely. R-square of computed
6) Guoqiang WANG, Qing FU, Kuniyoshi TAKEUCHI and
sediment volume and annual discharge is 0.34,
Hiroshi ISHIDAIRA, Estimation of River Sediment
whereas the one of observed sediment volume and Concentrations during Hydrologic Event, Annual Journal of
annual discharge is 0.16. Correlation may be Hydraulic Engineering, JSCE, 2007
dropped by the large rainfall events and other 7) Seyed Hamidreza Sadeghi, Takahisa Mizuyama and Babak
modifications are needed to compute sediment yield Ghaderi Vangh, Conformity of MUSLE Estimates and
and transport process in such high flood periods. Erosion Plot Data for Storm-Wise Sediment Yield
However, this study shows that this MUSLE with Estimation, Terr. Atmos. Ocean Sci., 2007
sediment transport model can be used to estimate
(Received September 30, 2007)

- 162 -

View publication stats