847

THE BASIN-WIDE TWO-DIMENSIONAL MODEL OF SEDIMENT

PRODUCTION AND TRANSPORTATION FOR ESTUARY SEDIMENT
SIMULATION

Qiang LUO
Numerical Simulation Department, INA Corporation, 1-44-10 Sekiguchi,
Bunkyou-ku, Tokyo 112-8668, Japan E-mail: q-luo@ina-eng.co.jp

Kodama TOSHIO
Numerical Simulation Department, INA Corporation, 1-44-10 Sekiguchi,
Bunkyou-ku, Tokyo 112-8668, Japan

Takagi TOSHIMITSU
Numerical Simulation Department, INA Corporation, 1-44-10 Sekiguchi,
Bunkyou-ku, Tokyo 112-8668, Japan


Abstract: Researches and observation in Japan shows that seashore profile change around an
estuary is apparently related to the volume of rivers sediment input, and a model to generate this
sediment input becomes necessary for estuary sediment transportation simulation. In this
research, the physics-based, two-dimensional model for non-uniform sediment production is
presented. In this model, the governing equations of diffusive wave model are employed to
route the surface water flow. Meanwhile, sediment is produced from each grid relating to
rainfall intensity, velocity, steepest slope, land use, etc. After that, the concentration continuity
equation is utilized to depict the sediment transportation across the basin. Then, the sediment
gradually reaches and accumulates in the channel, in which it falls into and is computed in two
transportation patterns: bed load and suspended load, according to the flow capacity. One-
dimensional equations are adopted to simulate the sediment movements in the river and riverbed
change, and finally output the sediment volume and hydrograph that can be utilized as input or
boundary conditions for estuary simulation. The model is verified in a miniature test basin and a
trial run is carried out in a real watershed.

Key words: Estuary, Sediment, Sediment control

1. INTRODUCTION AND LITERATURE REVIEW
In Japan, there are plenty of studies and observations that related seashore profile change
and erosion around the estuary to sediment input of the river. An estuary, as the terminal of
river sediment transportation, is the important source of beach sediment transportation
(Yoshida, 2003). Reservoir construction boomed during 1960s in Japan, and this resulted in
abrupt cut of sediment supply to seashores around the estuary and in turn caused beach
erosion (Sato and Aita, 1975). Kajimura, et al. (2001) and Sato (2003) had undertaken
separately long-term investigations in the beach change of the Nakoso Coast, located at
Fukushima Prefecture alongside the Pacific Ocean, from 1960 to 2000. In this period, two
dams were constructed, in 1962 and 1984 respectively, on the Samekawa River flowing into
the Ocean along the beach. There were data shown that, after the second dams construction,
annual average deposition before the two dams is 1.310
5
m
3
/year, which is roughly equal to
the decrease of seashore sediment transportation volume, 1.010
5
m
3
/year. This correlation
International Conference on Estuaries and Coasts
November 9-11, 2003, Hangzhou, China
848
reveals that beach profile change around the estuary depends largely on the sediment supply
from the river. Another research shows that drainage channels also affect beach profiles
(Iwaya, 1998). A survey of the worlds seashores points out that this phenomenon (rivers
sediment supply change accounts for part of the beach profile variation) is not unique on
Japans seashores (Marine Construction Technological Committee, 2001).
In order to reach accurate model output of beach profile change simulation around an
estuary, it is apparent that data of sediment input from the river are necessary. This research
presents a physics-based, two-dimensional model for producing sediment time series
(sediment discharge hydrograph) during a rain event and/or a period without rain, applying to
rivers with and without dams.
In the late of 20th century, the deposition issues in reservoirs became more and more
severe and began to bother Japanese Engineers and researches. Among these issues, the most
difficult one is determination of sediment volume produced on the slopes of the basin. In that
age, distributed hydrological models were not so popular as later years just before the
dawning of the 21st century when more powerful PCs have been developed. Many lumped
models for sediment production from a basin had been developed in Japan, as well as in the
other parts of the world. Takebayashi, et al. (1992) gave a summary to these models applied
in Japan, and a report by Deposition Flushing Division of Dam Technology Committee
(2001) also included two models employed in the United States.
All the above models developed linear and/or non-linear relations of sediment production
with all or some of the following factors: watershed area; slope; area of landslide; land use;
precipitation; flood runoff; reservoir storage; etc. However, dots of observation data and
model outputs in correlation charts scattered up and suggest very weak correlation between
them (Takabayashi, 1992). This phenomenon may be caused by the complexity of a real
basin. Realizing this point, researchers began to turn to the distributed models, which
decompose a large basin into smaller grids and therefore reduced the complexity of each grid.
In recent years, some distributed models were available in Japan. However, many these
models routed the overland flow with kinematic wave model equations and sediment
transport dose not involve the continuity equation. These models are not physics-based
distributed models, but just lumped models applied to grids.
Johnson et al. (2000) developed a physics-based, two-dimensional upland erosion model
(CASC2D-SED) and Julien and Rojas (2002) improved the same model by introducing
sediment transportation in the channel. In this model, the overland flow is routed with
diffusive wave model equations, and an empirical function is employed to account slope
sediment production or upland erosion as the authors terming. In the function, sediment
discharge is related to water flow discharge, slope, soil and vegetation properties, and
conservation practices. However, the sediment discharges in x and y directions are directly
related to those water flow components and the concentration continuity equation is not
involved. One merit of this model is that sediment is split into three classes (sand, silt and
clay), and can be taken as a non-uniform sediment model. And in the channel, sediment
deposition is possible but erosion is not allowed.
In this research, a fully physics-based, two-dimensional watershed sediment production
and transportation model is developed. In the model, two-dimensional diffusive wave model
equations are employed to simulate the water flows (both overland flow and channel flow) in
the watershed, concentration continuity equation is adopted to govern the sediment
transportation on the overland grids while two new lateral terms are introduce to the equation
to account sediment production on the slope. In the channel, both bed load and suspended
load are computed and the suspended load is accounted with the concentration continuity
equation. In addition, this model is a fully multi-particle-size model and sediment can be
simulated in a particle distribution at will. Finally, this model not only can produce long-
849
term and/or short-term sediment discharge hydrograph at the basin outlet or any spot along
the stream system, but also can simulate riverbed change due to deposition and/or erosion, in
channels with or without dams.
2. AN INTRODUCTION TO LUO-TAMAI MODEL FOR SURFACE WATER FLOW
The LUO-TAMI model (Luo, 2000) has been developed for water balance simulation in
large-scale complex watersheds (LCW), which include not only mountainous areas but also a
large fraction of flatland. It has also a complex stream system with reservoirs, lakes,
conjunctions, divergences, loop channels, sources, etc, as well as human impacts. The main
features of this model are: 1) both overland grids and channel grids are placed in the same
physical frame and governed by the same set of equations; 2) channel grids are not the
boundary conditions of the overland grids but grids with the same properties as the overland
grids, and therefore, there are not only mass exchanges but also momentum exchanges
between them; 3) it resolves easily the difficulties in the routing of surface flows in a flat
basin, in channels with conjunctions, divergences, loop channels, lakes and reservoirs.
Together, this model runs with the evapotranspiration model, infiltration model and
groundwater models (Luo, et al., 2000). The model can output discharge hydrograph at any
point on the stream system, and of course the outlet of the watershed. In this section, the
LUO-TAMAI model is introduced briefly.
The two-dimensional diffusive wave model Saint Venant equations are utilized as the
governing differential equations, which can be derived from the two-dimensional free surface
flow equations after some simplifications. They are written as below:
( ) ( )

= +

= +

0
0
fy
fx
S
y
z
S
x
z
q
t
h
y
vh
x
uh
(1)
where u and v are the x and y components of velocity respectively, h is the water depth, z is
the elevation of the water surface and
0
z h z + = , z
0
is the land surface elevation, q is the lateral
flow in the vertical dimension, S
fx
and S
fy
is the friction slopes in x and y directions
respectively, which can be obtained from the following Manning equations:
( ) ( )
( ) ( )

=
=
3 / 4 2
3 / 4 2
h v v n S
h u u n S
y fy
x fx
(2)
in which n
x
and n
y
are the Manning coefficients in x and y directions respectively. From
equation (2), one can see that the friction slope S
fx
and S
fy
have the same directions with the
velocities of u and v respectively.
In order to route the basins flat areas correctly, the staggered scheme (Fig. 1) is adopted to
discretize the governing equations. Once a direction is assigned to the flow (positive for
flowing out of the grid, and negative for in flow), flows in Fig. 1 (a) can be denoted in a
simplified way as Fig. 1 (b). Considering the possible flow direction of diagonal for channel
grids (Fig. 2), the governing equations become:
( )

= =

=
=

+
=
9 2 1
1
3 2
9
1
,..., , , m z B A z
z L n
R A
A u
q
y x
A
t
h
A u
y x
m m m m
m m
m m
m m
g
m
m m
(3)
850
i
i-1
i+1
j-1 j j+1
y
ij
u
y
ij
Q
+ y
ij
u
+ y
ij
Q
x
ij
u
x
ij
Q
+ x
ij
u
+ x
ij
Q
z
ij
z
i+1j
z
i-1j
z
i-1j+1
z
i-1j-1
z
i+1j-1
z
i+1j+1
z
ij+1
z
ij-1
ch
ij
u
ch
ij
Q
+ ch
ij
u
+ ch
ij
Q
i
i-1
i+1
j-1 j j+1
y
ij
u
y
ij
Q
+ y
ij
u
+ y
ij
Q
x
ij
u
x
ij
Q
+ x
ij
u
+ x
ij
Q
z
ij
z
i+1j
z
i-1j
z
i-1j+1
z
i-1j-1
z
i+1j-1
z
i+1j+1
z
ij+1
z
ij-1
ch
ij
u
ch
ij
Q
+ ch
ij
u
+ ch
ij
Q
j-1 j j+1
m
ij
u
m
ij
Q
z
ij
m=1
m=5
m=3
m=6 m=2 m=7
m=4
m=8
m
ij
z
j-1 j j+1
m
ij
u
m
ij
Q
z
ij
m=1
m=5
m=3
m=6 m=2 m=7
m=4
m=8
m
ij
z

(a) The staggered grid (b) Simplified denotation
Fig. 1 The staggered grid scheme in this research

x
y
River grids
Q
y
Q
x
Overland grids
Q
x
Q
y
Q
ch
x
y
River grids
Q
y
Q
x
Overland grids
Q
x
Q
y
Q
ch

(a) (b)
Fig. 2 Possible river flow directions and discharges

in which all subscripts of ij are omitted, and:
( )

=
=
m m m m
m
ij ij m
z L n R B
z z z
3 2
(4)
For a more effective convergence of numerical solutions, the SIMPLE algorithm is
adopted to solve equations (3).
3. SEDIMENT PRODUCTION AND TRANSPORTATION ON OVERLAND GRIDS
On overland grids, the modified two-dimensional concentration continuity equation is
employed to govern the sediment transportation of sediment:
( ) ( ) ( )
l fl Issl
l l l
c w q
y
vh c
x
uh c
t
h c
=

(5)
where
l
c is the average concentration of a kind of particles, and subscript l is the index of
particles with the representative diameter d
l
;
Issl
q is the sediment production term, which is
the volume of sediment of unit area and unit time, eroded from the grid, and has the velocity
dimension: (m/s); and
fl
w (m/s) is the falling velocity for particles of index l. Together, the
right-hand term: ( ) c w q
fl Issl
serves as the source term accounting for the sediment exchange
between the flow and the land surface.
After intensive investigation into worlds large watersheds, Morris and Fan (1998)
attributed basin sediment yield to the combination of the following factors: geology, slope,
851
climate, drainage density, and patterns of human disturbance. And, they also find that land
use and precipitation are two dominant factors. Based on these findings and those from the
study of references in the Introduction section and the researches of the authors, the sediment
production term in equation (5) is written as:
2 2
0
v u P s q
l Issl
+ = (6)
in which,
l
is the sediment production coefficient, a dimensionless coefficient (different
for different particle index l ) relating to land use, geological properties of soil, and human
impact and other factors, and it is subject to calibration;
0
s is the steepest slope of the grid;
and, P is the normalized precipitation with no unit.
The falling velocity is given by the following function:
l l
l
s
fl
d d
gd w

6 36
1
3
2
2
2
+
|
|
.
|

\
|
= (7)
where
s
is the density of sand; is the density of water; g is the gravitational
acceleration (=9.8 m/s
2
);
l
d is the representative diameter of particles of index l; and, is
the kinematic viscosity of water.
Equation (5) is discretized according to the staggered scheme and solved iteratively with
the implicit method.
4. SEDIMENT TRANSPORTATION IN A CHANNEL
Sediment transporting in a channel or bed material comprises two parts according to their
different flow properties. One is suspended load and the other bed load. These two parts of
sediment transportation are computed separately and then unified in the riverbed change
simulation.
4.1 SUSPENDED LOAD
Suspended load is depicted with the similar equation to that for sediment transportation in
the overland flow:
( ) ( ) ( )
bl fl sul
l l l
c w q
y
vh c
x
uh c
t
h c
=

(8)
in which,
sul
q is sediment pick-up rate from the riverbed, with the dimension of velocity;
and
bl
c is the concentration near the riverbed. According to Hydraulics Committee (2000),
sul
q is a function of applied sheer stress (
l *
) or friction velocity (
*
u ), and falling velocity;
and
bl
c is related to the average concentration
l
c , and therefore, equation (8) becomes:
( ) ( ) ( )
( ) ( )
l fl l fl sul
l l l
c w u w u q
y
vh c
x
uh c
t
h c
, ,
* *
=

(9)
Equation (9) has only one variable (
l
c ) and therefore can be solved. In this research, a
stream is taken as a one-dimensional domain. However, the two-dimensional equation is
employed and the excessive dimension is utilized as the linkage with overland grids to obtain
source input of sediment from them.
4.2 BED LOAD
The dimensionless bed load for sediment of non-uniform particles is calculated with
Ashida and Michiue formulas (Hydraulics Committee, 1999). These formulas related the
dimensionless bed load to Reynolds Number (
*
R ), applied sheer stress (
l *
) and critical
sheer stress (
cl *
), representative diameters (
l
d ) and the average diameter (
M
d ) of particles,
852
and some empirical coefficients. In the calculation of dimensionless bed load for non-
uniform sediment, the critical sheer stress for each particle index l is adopted. It can be
obtained from the average critical sheer stress after some rectification.
The volume of bed load of unit width can be obtained by multiplying the dimensionless
bed load by the normalization factor, shown as below:
( )
bl l s bl
q gd q
*
3
1 = (10)
in which
bl
q
*
is the dimensionless bed load, and
bl
q is bed load volume of unit width
(m
2
/s).
4.3 RIVERBED CHANGE SIMULATION
When sediment transports along with water flow in a channel, it deposits onto the riverbed
and erodes the riverbed as well. This phenomenon must be simulated, otherwise the model
will give incorrect sediment yield at the basin outlet or estuary. Riverbed change includes
variations of both elevation and particle size distribution (Hydraulics Committee, 2000).
The equation for elevation variation is given as below:
( ) ( )

|
.
|

\
|

l
sul bl fl
l
bl
b
q c w
x
q B
B t
z 1
1 (11)
in which, is the porosity of the riverbed soil,
b
z is riverbed elevation, and, B is river
width. Riverbed sediment particle distribution change is accounted with the following
equations:
( ) ( )
( )

=
=

1
1
1 1
*
l
l
sul bl fl
bl b
l
l
p
q c w
x
B q
B t
z
p
t
p

(12)
where is the thickness of riverbed sediment exchange layer,
l
p is the present percentage
of particles of index l,
*
l
p is the percentage of riverbed sediment given by the following
equations:

<

=
0 ,
0 ,
0
*
t z if p
t z if p
p
b l
b l
l
(13)
in which
l
p
0
is the percentage of original distribution of the riverbed particles. The
riverbed sediment exchange layer is a thin layer with a constant thickness but variable
upper and lower boundaries changing with time (Sediment Control Association, 2000).
Equations (11) and (12) can be solved explicitly.
5. MODEL CALIBRATION AND VALIDATION IN A TEST BASIN
Sediment gauging in the river flow is very difficult and there exist very few measured data.
However, Santos et al. (1998) provided a miniature basin (100m100m, grid size=2m, max
elevation difference=10m, two streams), and relatively detailed data of precipitation,
discharges and sediment quantities during several rainfall events.
Rainfall event 4 is used to calibrate the model. First of all, the water flow discharge is
calibrated. The precipitation data are applied evenly over all grids of the miniature basin and
the model outputs water flow discharge hydrograph (Fig. 3). Then, sediment yield is
calibrated. No sediment discharge hydrograph but the total amount of sediment yield is
available, and the sediment yield in rainfall event 4 is 1,200kg. The calibrated model gives a
sediment yield of 1,113kg with the proper sediment production coefficient
l
.
853
0.000
0.010
0.020
0.030
0.040
0.050
0 20 40 60 80 100
Time (min)
Q

(
m
3
/
s
)
0
100
200
300
400
500
P
r
e
c
i
p
i
t
a
t
i
o
n

(
Observed
Computed
Precipitation

Fig. 3 Hydrographs for calibration (rainfall event 4)

Rainfall event 13 is utilized to validate the model with the calibrated
l
. A comparison of
water flow hydrograph is shown in Fig. 4. The observed sediment yield is 4,000kg and the
simulated sediment yield is 4,110kg. These results can be taken as good agreements.

0.000
0.050
0.100
0.150
0 20 40 60 80 100
Time (min)
Q

(
m
3
/
s
)
0
100
200
300
400
500
P
r
e
c
i
p
i
t
a
t
i
o
n

(
Observed
Computed
Precipitation

Fig. 4 Hydrographs for validation (rainfall event 13)
6. A TRAIL RUN IN A REAL BASIN
Since model accuracy may affected model outputs, especially in a small domain of
hydrological modeling. The above miniature basin is an artificial basin, and both the spatial
and temporal scales are too small to demonstrate the models ability. A small watershed with
an area of 300km
2
, Kusaki, located at a tributary of the Tone River in Kanto Region of Japan,
is selected to take a trail run (Fig. 5). In this test simulation, time step varies with rainfall
intensity from several seconds to one hour, and time step for data output is one hour.
In the digital basin (Fig. 5 b), the black grids are basin grids, and the gray grids are stream
grids. The grid size is 1km, and distributed radar precipitation data in the same grid size are
available across the basin. Other data, such as DEM, land use, and outlet water discharge
hydrograph during the rain event, are also available. But there are no data of sediment
discharge. Fig. 6 shows water flow calibration during a rainfall event.
854

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2
3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 3
4 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 0 0 1 1 1 0 0 0 0 0 0 0 0 4
5 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 5
6 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 6
7 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 7
8 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 8
9 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 8 1 1 1 9 1 1 1 0 0 0 0 0 0 0 9
10 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 8 1 1 9 1 1 1 0 0 0 0 0 0 0 0 10
11 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 8 1 9 9 1 1 0 0 0 0 0 0 0 0 0 11
12 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 9 1 1 1 1 0 0 0 0 0 0 0 0 0 12
13 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 9 1 1 1 1 0 0 0 0 0 0 0 0 0 13
14 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 9 1 1 1 1 0 0 0 0 0 0 0 0 0 14
15 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 9 1 1 1 1 1 0 0 0 0 0 0 0 0 0 15
16 0 0 0 0 0 0 1 1 1 1 1 1 1 1 9 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 16
17 0 0 0 0 0 0 0 0 1 1 1 1 9 9 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 17
18 0 0 0 0 0 0 0 0 1 1 1 1 9 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 18
19 0 0 0 0 0 0 0 0 1 1 1 1 9 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 19
20 0 0 0 0 0 0 0 0 1 1 1 1 9 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 20
21 0 0 0 0 0 0 0 0 1 1 1 9 9 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 21
22 0 0 0 0 0 0 0 0 1 1 9 9 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 22
23 0 0 0 0 0 0 0 0 1 1 9 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 23
24 0 0 0 0 0 0 0 0 1 1 9 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 24
25 0 0 0 0 0 0 0 0 1 1 9 1 1 1 1 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 25
26 0 0 0 0 0 0 0 0 0 9 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 26
27 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 27
28 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 28
29 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 29
30 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
0: Grids outside
the basin
1: Basin grids
8, 9: streams

(a) Bird eye view of the basin (b) Digitalized basin
Fig. 5 Kusaki basin for the trial run

0
100
200
300
400
500
600
700
800
900
1000
1100
1200
0 12 24 36 48 60 72
Time (h)
Q

(
m
3
/
s
)
0
40
80
120
160
200
P
r
e
c
i
p
i
t
a
t
i
o
n
Observed
Computed
Precipitation

Fig. 6 Discharge comparison in a rainfall event

In this mountainous watershed, there are only two types of land use: forest and water body.
For water body, there is no sediment production and 0 =
l
. For forest, six different values
of
l
are selected to test the model, and the simulated sediment discharges at the basin outlet
are shown in Fig. 7. This figure demonstrates that the sediment production coefficient
affected greatly sediment output at the basin outlet, or sediment output is very sensitive to the
sediment production coefficient. Fig. 8 shows sediment discharges response to water flow
discharge or precipitation.
As mentioned in section 1, this model not only can produce sediment discharge
hydrographs at any point of the stream system, but also can simulate riverbed changes of both
elevation evolution and particle size distribution. Fig. 9 shows the riverbed change related to
the original riverbed elevation in a time period of 72 hours, and Fig. 10 is the average particle
size evolution in the same time periods. From this figure, one can find an interesting
phenomenon that the eroded sections (with negative riverbed elevation change) have their
average particle size increased, and vice versa for the sections of deposition (with positive
riverbed elevation change).
855
0.0
10.0
20.0
30.0
40.0
0 12 24 36 48 60 72
Time (h)
S
e
d
i
m
e
n
t

D
i
s
c
h
a
r
g
e

3
/
s
)
Gamma=0.05 Gamma=0.10 Gamma=0.25
Gamma=0.50 Gamma=0.75 Gamma=1.00

Fig.7 Sediment discharge sensibility to sediment production coefficient
l


0.01
0.1
1
10
100
1000
10000
0 12 24 36 48 60 72
Time (h)
D
i
s
c
h
a
r
g
e

a
n
d

S
e
d
i
m
e
n
t

3
/
s
)
0
60
120
180
240
300
P
r
e
c
i
p
i
t
a
t
i
o
n

Computed
Sediment
Precipitation
Discharge

Fig. 8 Sediment discharge response to water flow discharge ( 00 . 1 =
l
)

-0.030
-0.025
-0.020
-0.015
-0.010
-0.005
0.000
0.005
0.010
0 5 10 15 20 25
Distance from outlet (km)
R
i
v
e
r
b
e
d

e
l
e
v
a
t
i
o
n

c
h
T=12(h) T=24(h) T=36(h) T=48(h) T=60(h) T=72(h)

Fig. 9 Riverbed elevation change related to original riverbed elevation ( 00 . 1 =
l
)
856
43.5
44.0
44.5
45.0
45.5
46.0
46.5
0 5 10 15 20 25
Distance from outlet (km)
A
v
e
r
a
g
e

p
a
r
t
i
c
l
e

s
i
Original T=12(h) T=24(h) T=36(h) T=48(h)
T=60(h) T=72(h)

Fig. 10 Riverbed average particle size evolution ( 00 . 1 =
l
)
7. CONCLUSIONS
This research presents a physics-based, two-dimensional model for watershed sediment
production and transportation. After the above analysis and study, it can be concluded that:
1) the physics-based two-dimensional model for watershed sediment production is corrected
and effective for sediment output; 2) the introduction of sediment production term and
sediment production coefficient for overland grids is successful; 3) the models capability is
not limited to generation of sediment discharge hydrographs, it is also capable to simulate
riverbed change.
However, due to lack of observation date of sediment in a real basin, the model still needs
further calibration and validation before applied to a real basin.
REFERENCES
Deposition Flushing Division of Dam Technology Committee, 2001. Reservoir deposition flushing
strategy: present situation and problems, Oodamu (Great Dam), No. 176 (2001-7), pp. 1-49. (In
Japanese)
Hydraulics Committee, 1999. A Collection of Hydraulic Formulas, JSCE. (In Japanese)
Hydraulics Committee, 2000. A Collection of Hydraulic Formulas in CD-ROM, JSCE. (In Japanese)
Iwaya, T., 1998. Study on the floodway in the seaside dune blockade area, J. Hydrau., Coast. Environ.
Eng., JSCE, No. 586/II-42, pp. 61-75. (In Japanese)
Johnson, B.E., et al, 2000. The two-dimensional upland erosion model CASC2D-SED, Journal of the
American Water Resources Association, Vol. 36, No. 1, pp. 31-42.
Julien, P., and R. Rojas, 2002. Upland erosion modeling with CASC2D-SED, International Journal of
Sediment Research, Vol. 17, No. 4, pp. 265-274.
Kajmura, T., et al., 2001. Sand movement and long-term beach evolution in an alluvial system
composed of the same river and the Nakoso coast, J. Hydrau., Coast. Environ. Eng., JSCE, No.
691/II-57, pp. 121-132. (In Japanese)
LUO, Qiang, 2000. A distributed water balance model in large-scale complex watersheds (LCW) and its
application to the Kanto Region, Ph.D. dissertation of the University of Tokyo, Tokyo, Japan.
LUO, Qiang, et al., 2000. Non-optimization direct method of the inverse problem for groundwater
modeling (NODMIP) in a large scale basin, Journal of Japan Society of Hydrology and Water
Resources, Vol. 13, No. 2, pp. 148-155.
Marine Construction Technological Committee, 2001. Marine Construction Technologies of the World,
Sankaidou Press, Tokyo, Japan. (In Japanese)
Morris, G.L., and J. Fan, 1997. Reservoir sedimentation handbook, McGraw-Hill.
857
Santos, C.A.G., et al., 1998. Sediment yield due to heavy rainfall from a test field in Brazil and its
analysis by a runoff-erosion model, J. Hydrau., Coast. Environ. Eng., JSCE, No. 586/II-42, pp. 117-
126.
Sato, S., 2003. Coastal conservation based on comprehensive sediment control in the fluvial system,
Kasen (River), No. 680, Japan River Association, pp. 15-19.
Sato, S., and Aita, 1975. Civil Engineering Handbook for Easy Understandings, Shokoku Press, Tokyo,
Japan. (In Japanses)
Sediment Control Association, 2000. Numerical Simulation Methods for Riverbed Evolution in
Mountainous Rivers, Sankaido. (In Japanese)
Takebayashi, S., et al., 1992. A method to estimate amount of sedimentation in dam-reservoirs
sedimentation as stochastic process, Dam Engineering, No. 8, pp. 6-20. (In Japanese)
Yoshita, T., 2003. Relation between seacoasts and estuaries, Kasen (River), No. 680, Japan River
Association, pp. 6-8. (In Japanese)