Geomorphology 91 (2007) 65 79

www.elsevier.com/locate/geomorph

Fluvial processes and morphological response in the Yellow and

Weihe Rivers to closure and operation of Sanmenxia Dam
ZhaoYin Wang , Baosheng Wu, Guangqian Wang
State Key Laboratory of Hydroscience and Engineering, Tsinghua University, Beijing 100084, China
Received 4 July 2006; received in revised form 27 January 2007; accepted 27 January 2007
Available online 7 February 2007

Abstract
The fluvial and morphological processes induced by impoundment of the Sanmenxia Reservoir and relevant human activities
on the Yellow River and its tributaries are complex. The longterm annual sediment load of the Yellow River was 1.6 billion tons,
ranking first of all the world's rivers. In 1960, Sanmenxia Dam began filling. Sediment transport in the river then was greatly
disturbed and a new cycle of the fluvial processes was induced. First, the dam caused not only anticipated sedimentation in the
reservoir, but also serious sedimentation in the largest tributary of the river (the Weihe River). The response of fluvial process to the
dam closure varies in space and time. Second, the downstream reaches of the dam experienced erosion and resiltation, changes of
river pattern, and development of meanders. Moreover, the downstream reaches of the dam have experienced more and more water
diversion, which has induced readjustment of the longitudinal profile of the river. The study reveals that sedimentation in the
Sanmenxia Reservoir enhanced the bed elevation at Tongguan, where the Weihe River flows into the Yellow River. The rising
Tongguan's elevation caused retrogressive siltation waves in the Weihe River, which propagated at a speed of about 10 km/yr. An
equilibrium sedimentation model is proposed, which agrees well with the data of sedimentation in the Weihe River. In the reaches
below the dam the river changes from braided to wandering, or from wanderingbraided to wanderingmeandering. The discharge
released to the downstream reaches has been regulated by the reservoir and it decreases along the course because the quantity of
water diversions is more than the inflow from tributaries. The reduction in discharge causes readjustment of the longitudinal bed
profile. By using the minimum stream power theory, we prove that the riverbed profile is developing toward an Sshape profile.
2007 Elsevier B.V. All rights reserved.
Keywords: Fluvial processes; Sanmenxia Reservoir; River pattern; Bed profile; Water diversion

1. Introduction
According to the International Commission on Large
Dams, the world's rivers are now obstructed by more
than 40,000 large dams. From 1949 to 1990, the number
of large dams in China increased from only eight to
more than 19,000. These large dams have provided
extensive benefits during the past century and have
Corresponding author.
E-mail address: zywang@tsinghua.edu.cn (Z.Y. Wang).
0169-555X/$ - see front matter 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.geomorph.2007.01.022

fueled the economy by providing cheap power, irrigation, and municipal water supplies. Dam construction
and reservoir operations are great disturbance to fluvial
processes of rivers. Aggradation in the reservoir and
upstream reaches and degradation in the downstream
reaches have taken place in many of the world's rivers
(Leopold, 1973; Gregory and Park, 1974; Petts, 1979;
Mahmood, 1987; Collier et al., 1996; Kondolf, 1997).
Many studies have investigated changes in river patterns
responding to dam closure, showing rivers with braided
or braidedmeandering transitional characteristics

66

Z.Y. Wang et al. / Geomorphology 91 (2007) 6579

(Neill, 1973; Church, 1983; Ferguson and Werritty,

1983; Knighton and Nanson, 1993). Lateral adjustment
of river channels due to reservoir operation is more
irregular. The channel width may respond to dam
closure with trends of widening, narrowing, and no
change (Williams and Wolman, 1984). Any changes
imposed on a fluvial system tend to be absorbed by the
system through a series of channel adjustments
(Schumm, 1973). The description of channel adjustment
and evolution by nonlinear decay functions are well
documented (Hey, 1979; Robins and Simon, 1983) Graf
(1977) used exponential functions to describe the
relaxation time necessary to achieve equilibrium
following a disturbance. Simon (1989) found that both
exponential and power equations were initially fitted to
the observed data.
In the upstream reaches, the primary consequence of
impoundment of rivers is sedimentation. However,
sedimentation issues are not confined solely to the
reservoir. The backwater reach of the reservoir can
extend hundreds of kilometers upstream, as in the case
of the Three Gorges Project on the Yangtze River in
China. The current velocity and sediment carrying
capacity of the flow are reduced by reduction in energy
slope; thence sedimentation occurs in the backwater
region. The aggradation, in turn, raises the local water
surface elevation, creating additional backwater and
deposition even farther upstream and in tributaries, as in
the case of the Sanmenxia reservoir on the Yellow River.
This feedback mechanism allows the depositional
environment to propagate much farther upstream than
the initial hydraulic backwater curve might suggest
(Goodwin et al., 2001).
Qian et al. (1987) have defined the wandering river as
a river pattern exhibiting very unstable channel and high
migration rate, which usually occurred in rivers carrying
high sediment load like the Yellow River in China. The
Danjiangkou Reservoir on the Hanjiang River in China
has changed the river from a braided river to a
wanderingbraided river, which is caused by strong
erosion of the riverbank initiated by operation of the
reservoir. Large quantities of sediment were supplied to
the channel by bank erosion and deposited at many
midchannel bars during floods. Hence a wandering
braided channel pattern with many unstable mid
channel bars has developed. Strikingly, while the river
was developing from a braided river into a wandering
braided river, the sediment quantity measured at the
upstream station was approximately equal to that
measured at the downstream station (Xu, 1996). This
implies that a huge amount of sediment on the bed and
banks was removed, while only a small amount of

sediment was transported through the channel. Church

(1983) also reported changes of river patterns from
braided or multithread to single thread. The construction
of the Black Butte Dam on Stony Creek in the U.S. in
1963 caused the braided pattern of the reaches downstream of the dam to change to a singlethread, incised
meandering pattern by 1967.
Although many case studies on fluvial processes
induced by dams have been reported, general laws on
the processes do not exist. The aim of the study is to
comprehend the complicated fluvial processes and
morphological responses in the upstream and downstream of the Sanmenxia Reservoir. In the upstream
sedimentation in Sanmenxia Reservoir has changed
the lower boundary of the Weihe River (the bed
elevation at the confluence), which has induced
continuous sedimentation and a new morphological
process in the tributary. The laws of sedimentation
responding to varying lower boundary and equilibrium
sedimentation volume in the Weihe River for given
lower boundary conditions are studied. In the downstream reaches the water and sediment load have
reduced greatly since the impoundment of the
reservoir, partly because the reservoir trapped sediment and much water and sediment load were diverted
from the river along the course. The river has adjusted
itself to match the changing flow conditions. The
lower Yellow River is a wandering river. The fluvial
processes and morphological responses to the closure
and operation of the Sanmenxia Reservoir and water
diversion projects along the river are complex. The
speed of channel migration did not reduce after the
impoundment as predicted, although the sand bars in
the channel had reduced, which is different from the
general morphological response to dam closure. This
paper analyzes the stability of the channel and
development trend of the bed profiles under the
changing water and sediment conditions. The research
results reported here may shed light on the management of heavily sedimentladen rivers under changing
conditions.
2. Background and data collection
As shown in Fig. 1, the Yellow River has a drainage
area of 795,000 km2 and a length of 5464 km making it
the second longest river in China. The longterm
annual sediment load at the Sanmenxia Station was
1.6 billion tons before 1980, ranking first of all the
world's rivers (Qian and Dai, 1980), although the
sediment load has reduced greatly in the past 20 yr.
The huge amount of sediment is mainly from the loess

Z.Y. Wang et al. / Geomorphology 91 (2007) 6579

67

Fig. 1. The Yellow River and its tributaries (of which the Weihe River is the largest), the locations of the Sanmenxia and Xiaolangdi Dams, the
hydrological stations: Hekouzhen, Longmen, Tongguan, Huayuankou, Aishan and Lijin, and measurement crosssections: Tiexie and Gubaizui,
along the river course.

plateau in the middle reaches of the river. In 1960, the

first dam on the river, i.e., the Sanmenxia Dam, began
filling. Sediment transport in the river then was greatly
disturbed and a new cycle of fluvial processes was
induced. Nevertheless, the dam has caused unforeseen
impacts on the fluvial processes and river morphology
both in upstream and downstream reaches, which were
not fully comprehended during the project planning
process. This is not surprising since the fluvial processes
and morphological responses to the dam closure and
reservoir operation at the basin scale are immensely
complex.
First, the dam caused not only anticipated sedimentation in the reservoir, but also serious sedimentation in
the largest tributary the Weihe River, which was not
predicted. Second, the reaches downstream of the dam
experienced erosion and resiltation, channel migration,
and development of meanders. Moreover, the downstream reaches have experienced more and more water
diversion the past decades, which has induced readjustment of the longitudinal profile of the river. The
river flow and sediment carrying capacity have been
greatly changed, which caused a new morphological
development.
Sedimentation in the reservoir and the Weihe River
has caused severe flooding. A recent flood disaster
occurred in the lower Weihe River in the 2003 fall, which
was extremely exacerbated by sedimentation in the
Sanmenxia Reservoir on the Yellow River (Wang et al.,
2004a,b). The flood caused great economic loss and

affected million people, and has rekindled the argument

on decommission of the dam. The main cause for the
disaster was the continuously enhancing riverbed and
flood plain due to sedimentation. If there were no
Sanmenxia Dam the riverbed would be much lower and
the flood would cause no such a great disaster. Nevertheless, only 3 yr before the event, when the Yellow
River Conservancy Commission celebrated the 40 yr
anniversary of Sanmenxia Reservoir, many people spoke
highly on the reservoir and awarded Sanmenxia Reservoir a great achievement in the hydroconstruction in
China. It is a great practice in the training of heavily
sedimentladen rivers (rivers carrying sediment load
often higher than 100 kg/m3). The half century safety of
the lower Yellow River reaches and development of
the river basin were attributed to the operation of the
reservoir, which played an important role in flood
control, icejam flood control, power generation, irrigation and water supply. The contrary opinions against
Sanmenxia Reservoir were due to a lack of understanding of the fluvial processes and morphological responses of the Weihe and Yellow Rivers to the reservoir
operations and insufficiency of effective management
strategies.
Sediment measurement has been performed for a long
period of time on the Yellow River; especially a
systematic and regular measurement began in 1950. In
the period from 19501980 the Yellow and Weihe Rivers
experienced a high sediment load period, and, from 1985
the sediment load in the two rivers has been remarkably

68

Z.Y. Wang et al. / Geomorphology 91 (2007) 6579

reduced due partly to the climate change and mainly to

human activities. Sanmenxia Reservoir began to impound
just in the period of highest sediment load and experienced
an extremely high rate of reservoir sedimentation.
The measurement for calculation of reservoir sedimentation and erosion has been conducted since 1960
after the completion of the reservoir. Usually high
sediment concentration and intensive sedimentation and
erosion occur in flood season from July to September
and range line surveys (measurement of river bed cross
sections at fixed places) have been performed in a
unified way twice a year prior to and after the flood
period. Additional measurement has been made only
after large flood events. There are 32 measurement cross
sections for the 134km river reach from Longmen to
Tongguan, 41 measurement cross sections for the 113
km river reach from Tongguan to the Sanmenxia Dam,
and 107 cross sections for the entire lower Yellow River
from the dam to the river mouth (increased to 145 in
1998). Range surveys have also been conducted for the
lower Weihe River with 28 measurement cross sections
for the 128km river reach from Lintong to the
confluence with the Yellow River.
Traditionally, a manual sounding pole was commonly used for depth measurement, and the positioning
of sounding points was located by intersections
determined by transits from two points. In recent
years, the Differential Global Positioning System
(DGPS) interfaced with a underwater depth sounding
system has been used in river range line surveys. A
volumetric method is used for computing the amount of
deposition or erosion. The volume enclosed by the
crosssectional areas of two adjacent cross sections
underneath a datum plan is computed, and the difference
of the volume between two consecutive surveys gives
the amount of deposition or erosion.
The majority of the transported sediment in the middle
and lower Yellow River is suspended load. Bedload is
only a small portion of the total load, varying from 0.3% to
0.7% on average at all the hydrologic stations (Long and
Zhang, 2002). Therefore, suspended load is measured on a
regular basis at hydrologic stations along the Yellow
River, while bedload is only measured occasionally for
special purposes. Suspended sediment discharge is
usually measured by dividing the crosssectional areas
into a number of subsections and sampling. Sediment
discharge passing through each subsection is first
obtained by taking measurements along the vertical
within the portion of the subsection it represents.
Sediment concentration in a vertical is usually obtained
by a multipoint method, in which samples are taken from
three to five points. Unit samples are also taken at

hydrometric stations during floods as a supplement to the

conventional method. The sediment concentration is
obtained by separating sediment from water in the
samples and weighing. The relative error in the
concentration measurement is less than 1%. Analysis
indicated that there were systematic errors in the
measurement of total sediment discharge at hydrometric
stations, mainly resulting from the unmeasured suspended
sediment and the bed load. The reported suspended
sediment load at some hydrometric stations was smaller
than the true values by about 37% (Li and Long, 1994).
The study on the fluvial processes in the Yellow and
Weihe Rivers is performed based on the data of range line
surveys and measurements of sediment transport at the
hydrological stations. There is no formula for estimation
of the total error in the sedimentation volume calculation.
The accuracy of the computed amount of sediment
deposition depends mainly on the density of measurement
crosssections and errors in measurements. The accuracy
and relative error depends also on the amount of
deposition in the studied reach. The larger is the
sedimentation volume the smaller is the relative error.
The accuracy and maximum error have been studied in
two ways: 1) comparison of the result of the range line
surveys and sediment load variation measured at two
hydrological stations or using the sediment continuity
equation; 2) adding measurement crosssections up to a
density of about 89 crosssections per 10 km. The results
show that the calculated sedimentation with the measured
crosssections generally agrees with the sediment load
variation at the hydrological stations. There is no
accumulative error in the measurement crosssections.
By analysis of the data observed with high density cross
sections in an experimental reach, the accuracy of
deposition computation was about 7085% for a reach
with an appropriate amount of deposition (Long et al.,
2002). The relative error would be less and less along with
an increase in the amount of deposition. The cumulative
sedimentation volume in the reservoir is huge and the
relative error is smaller than 30%.
3. Sedimentation of the Weihe River induced by the
Sanmenxia Reservoir
As shown in Fig. 1 the middle reaches of the river
extend from the Hekouzhen hydrological station in the
north of the basin to Xiaolangdi. The Sanmenxia Dam is
located in the lower part of the middle reach of the river.
The river flows south from Hekouzhen to Tongguan,
then makes a 90 turn and flows east. The Weihe River,
which is the largest tributary, flows into the Yellow
River at Tongguan. The construction of the Sanmenxia

Z.Y. Wang et al. / Geomorphology 91 (2007) 6579

Dam was initiated in 1957, and water impoundment

commenced in September 1960. The crest elevation of
the dam is 353 m, and the original capacity of the
reservoir was 9.705 billion m3 with a pool level of
335 m. The reservoir area extends upstream a distance of
246 km to Longmen. The lower Weihe River is affected
by the reservoir as well.
The Weihe River is 818 km long and has a
drainage area of 134,800 km 2 with more than
23 million people dwelling in the river basin. The
river basin was known as the 800 li (1 li = 0.5 km)
fertile Qin Valley. The most serious adverse effect of
the Sanmenxia Dam is the unanticipated sedimentation
in the lower Weihe River and, consequently, the high
flooding risk to the lower Weihe Basin and Xi'an, an
ancient capital of China. Sedimentation in the Weihe
River has changed the valley into a swamp with a
high groundwater table. Local people complained and
some officials and scientists suggested decommissioning the dam. The Weihe River has been experiencing a
striking change in fluvial processes since the
impoundment of the Sanmenxia Reservoir. The river
channel has been changing from meandering with a
sinuosity of 1.65 to straight with a sinuosity of only
1.06 and slightly meandering with a sinuosity about
1.3.
The longterm average annual runoff of the Weihe
River is 8.06 billion m3, and annual sediment load is
386.6 million tons, which compose about onefifth of
the annual runoff and onethird of the annual sediment
load of the Yellow River at Sanmenxia. In the past
decades, the water and sediment load in the Weihe
River and the Yellow River have been reducing
mainly because of human activities. Table 1 shows
water and sediment load in the rivers in the periods
19602001 and 19862001. Water and sediment load
in the two periods are less than the average values
before 1980, but the ratios of water and sediment load
from the Weihe River to the Yellow River remain

69

unchanged. The majority of the sediment load consists

of silt with a median diameter of about 0.03 mm.
Before the impoundment of Sanmenxia Dam, the
Weihe River carried 386.6 million tons of sediment
into the Yellow River annually and the Weihe River
itself remained a relatively stable longitudinal bed
profile.
The elevation of Tongguan or Tongguan's elevation
is defined as a flood stage corresponding to a discharge
of 1000 m3/s at the Tongguan Hydrological Station on
the Yellow River, which acts as the base level of the bed
profile of the Weihe River. Before the Sanmenxia Dam
Tongguan's elevation was about 323.5 m. Since
impoundment of the Sanmenxia Reservoir, sediment
has been depositing in the reservoir, which causes
Tongguan's elevation to increase. The energy slope and
sediment carrying capacity of the flow in the Weihe
River have been reduced. The sediment load could not
be transported into the Yellow River and sedimentation
occurred in the lower Weihe River. In other words, the
rising Tongguan's elevation has changed the lower
boundary of the Weihe River, hence inducing a new
cycle of fluvial processes.
The filling of Sanmenxia Reservoir began in
September 1960 and the pool level reached the highest
pool level of 332.58 m on Feb. 9, 1961. The reservoir
functioned as a storage basin until March 1962. Severe
sedimentation problem became evident immediately
after impoundment. During the first 18 months, 93% of
the incoming sediment load was trapped in the reservoir
and caused 17% of capacity loss below an elevation of
335 m or 26% of capacity loss below an elevation of
330 m. Tongguan's Elevation had risen about 5.5 m in
Oct. 1961. A flood with a discharge of 2700 m3/s from
the Weihe River was blocked, which flooded 17,000 ha
of farmland in the lower Weihe Plain and caused a great
economic lose. To mitigate the sedimentation, the
operation scheme was changed to detain only flood
water in flood seasons. However, the floodreleasing

Table 1
Water and sediment load of the Yellow and Weihe Rivers in the past decades
River/hydrologic
station

Distance to the
Yellow River
mouth L (km)

Annual runoff
(19602001)
(bil. m3)

Annual sediment
load (19602001)
(mil. tons)

Average sediment
concentration
(19602001) (kg/m3)

Annual runoff
(19862001)
(bil. m3)

Annual sediment
load (19862001)
(mil. tons)

Weihe/Huaxian
Yellow/Tongguan
Yellow/Sanmenxia
Yellow/Huayuankou
Yellow/Aishan
Yellow/Lijin

1177
1092
996
734
374
100

6.79
34.61
34.69
37.44
33.07
28.56

312
1043
1009
910
770
700

46.04
30.13
29.09
24.20
25.00
36.80

4.66
25.16
24.62
25.88
19.16
13.56

248
722
712
610
440
350

70

Z.Y. Wang et al. / Geomorphology 91 (2007) 6579

capacity of the outlet structures was limited. In order to

increase the capacity to discharge sediment from the
reservoir a project of reconstruction of outlet structures
was carried out in two stages during the period 1964
1973. Ten bottom outlets were reopened, thus, sediment
deposit can be sluiced out off the reservoir through these
outlets with the lowest elevation at 280 m.
The operation scheme of the Sanmenxia Reservoir
has been substantially changed to achieve a balance
between sediment inflow and outflow in the following
three reservoir operation modes (Wu and Wang, 2004):
(i) storage from September 1960 to March 1962, the
reservoir was operated at a high storage level the whole
year round; (ii) detaining flood water and sluicing
sediment from March 1962 to October 1973, the
reservoir was operated at a low storage level throughout
the year, detaining floods only during flood seasons and
sluicing sediment with the largest possible discharges;
and (iii) storing clear water and releasing turbid water
from November 1973 to the present, the reservoir has
been operated at a high level (315320 m) to store
relatively clear water in nonflood seasons (November
June) and at a low level (302305 m) to release high
sediment concentrations in flood seasons (JulyOctober), as shown in Fig. 2. Moreover, the bottom outlets
have also been used to discharge high concentration
density currents. As the concentration is higher than
200 kg/m3, density current occurs in the reservoir. The
density current may be directly released from the
reservoir though the bottom outlets even if the pool
level remains high.
The longitudinal profile in the reservoir has varied
with the changes of operation modes as shown in Fig. 3.
During the storage operation period the reservoir was
severely silted. The measured sedimentation volume in
1964 was about 1.95 billion tons, representing 70% of
the incoming sediment load in the first 4 yr. Changing
the operation modes has reduced the sedimentation

volume, and the bed profiles have been relatively stable

since the 1970s. The crosssection CS 41 is at
Tongguan, at which the Weihe River flows into the
reservoir.
Fig. 4 shows the variations in Tongguan's elevation
over time from 1960 to 2001. Three ascending periods
are denoted by I, II, and III; and two descending periods
are denoted by 1 and 2. The abrupt rise and fall in 1960
and 1962 were caused by the impoundment in 1960 and
change of the operation mode from storage to flood
detention. The time of high elevation (329 m in Fig. 3)
was short and its influence on the Weihe River
sedimentation was temporary, although it caused an
obvious flood stage rise in 1961. Therefore, the period
of 19601962 is not separated from the ascending
period I.
The ascent and descent of Tongguan's elevation were
results of reservoir sedimentation and erosion, which in
turn were caused by variations in the pool level of the
reservoir. Generally speaking, sedimentation in the
lower Weihe River occurred during the periods when
Tongguan's Elevation rose, and erosion occurred during
the periods when it fell. The total volume of sediment

Fig. 2. Variation of pool level of Sanmenxia Reservoir in different

operation modes.

Fig. 4. Variation of Tongguan's elevation (watersurface elevation at

Tongguan for a flow of 1000 m3/s).

Fig. 3. Longitudinal profiles at Sanmenxia Reservoir during different

periods of operation (in which CS12CS48 are the measurement cross
sections on the reservoir reach of the Yellow River).

Z.Y. Wang et al. / Geomorphology 91 (2007) 6579

Fig. 5. Aggradation of the lower Weihe River measured at crosssections WY2 (21 km from Tongguan) and WY7 (59 km from
Tongguan) from 1960 to 2001 (Wang and Li, 2003).

deposited in the lower Weihe River up to the year 2001

was about 1.3 billion m3. The sedimentation was
distributed mainly in a 100kmlong reach from the
confluence. The accumulated deposition volume per
unit length was high near the confluence, reduced
upstream, and to nearly zero near Xi'an. Fig. 5 shows
the transect of the profiles of the channel bed and
floodplain in the lower Weihe River measured in 1960
and 2001 at the crosssections WY2 and WY7, which
are 21 and 59 km from Tongguan, respectively. The
floodplain elevation had risen by 3 to 5 m from
sedimentation, and the main channel had shrunken and
become more unstable. The flood discharge capacity of
the channel was hence reduced and the flood stage at the
same discharge was substantially enhanced.
Simon (1989) and Simon and Thorne (1996) studied
channel response in disturbed alluvial channels and
found that the changes imposed on a fluvial system tend
to be absorbed by the system through several stages of
channel adjustment and following exponential decay
equations. The response of the Weihe River to the
Sanmenxia Dam closure is more complex because the
raised Tongguan's elevation is not stable and the effect
has transmitted from the confluence to Xianyang Station
(180 km upstream from Tongguan. Erosion and
sedimentation caused by the ascending and descending
of Tongguan's elevation propagated upstream in retrogressive waves. Fig. 6ac shows the distribution of the
deposition rate per unit river length in the periods 1960
1969, 19691973, and 19731980, respectively, in
which the horizontal axis is the number of the
measurement crosssections on the Weihe River; the
average distance between the neighboring crosssections

71

is about 6 km. In the period from 1960 to 1969,

Tongguan's Elevation rose abruptly from 323.5 to
328.5 m (see Fig. 4). As a result, sedimentation occurred
in the reach around Huaxian at a rate of up to
2.5 million tons/km/yr (Fig. 6a). The mark I indicates
that the sedimentation corresponding to the first ascending period of Tongguan's elevation. In the period from
1969 to 1973, the sedimentation wave moved upward to
the reach between Huaxian and Lintong, but the rate of
sedimentation decreased to about 0.75 million tons per
km per year (Fig. 6b). In the meantime, the first erosion
wave occurred near the river mouth, which corresponded
to the first descending period of Tongguan's elevation,
indicated by the mark 1. In 19731980, the first
sedimentation wave had moved upstream to Lintong, the
first erosion wave had moved to Huaxian and the peaks
had obviously decreased too. During this period, the
second sedimentation wave occurred in the reach
between the river mouth and Huaxian indicated by the
mark II. This wave of sedimentation was associated

Fig. 6. Erosion () and sedimentation (+) per unit length per year
showing retrogressive waves in the lower Weihe River as a result of
ascending and descending of Tongguan's elevation. (The cross
sections are numbered from the river mouth. Huaxian, Lintong, and
Xianyang are hydrological stations by the river and are about 50 km,
128 km and 180 km upstream from Tongguan. The distance between
neighboring crosssections is about 6 km). (Wang and Li, 2003).

72

Z.Y. Wang et al. / Geomorphology 91 (2007) 6579

with the second ascending period of Tongguan's

elevation. The ascending and descending of Tongguan's
Elevation generated erosion and sedimentation waves,
which propagated retrogressively along the Weihe River,
at a speed of about 10 km/yr.
4. Equilibrium sedimentation model
Two questions to be answered about the fluvial
processes in the Weihe River induced by the Sanmenxia
Dam are: Is there any equilibrium of sedimentation in
the Weihe River? And whether the sedimentation has
reached the equilibrium? We propose a simple model to
answer the questions (Wang et al., 2004a,b). Assume
there is an equilibrium sedimentation volume, Ve for a
given increment of Tongguan's elevation. If the real
sedimentation volume, V, is much less than Ve the rate
of sedimentation in the river is high. The rate of
sedimentation is proportional to the difference between
the equilibrium and real sedimentation volume:
dV
K V e  V
dt

in which K is a constant with dimension of [1/T]. The

solution of the equation is
V ekt

R

KVe eKt dt const

The equilibrium sedimentation volume Ve is proportional to the enhancement of Tongguan's elevation Zt,
which is given by Zt = Zt 323.5, in which Zt is
Tongguan's elevation at time t and 323.5 m is the
Tongguan's elevation before the dam. Simply, the
equilibrium sedimentation volume can be imagined to
have a shape like a cone, then it may be assumed
Ve ADZt =2

unchanged at 5 m (Zt, = 328.5 323.5 = 5), which

shows that the equilibrium sedimentation volume is
around 1.3 billion m3.
As shown in Fig. 7, the model agrees well with the data
of sedimentation, which proves that for a given Zt, there is
indeed an equilibrium sedimentation volume. If the increment in Tongguan's elevation remains unchanged, the
sedimentation of the lower Weihe River may reach
equilibrium in about 25 yr. At present, the sedimentation
of the lower Weihe River is approaching to the equilibrium
volume, and there will be no great volume of accumulated
sedimentation if Tongguan's elevation stops rising.
Nevertheless, the equilibrium sedimentation volume is
dynamic and increases with rising lower boundary. If
Tongguan's Elevation continues to rise the equilibrium
sedimentation volume will be greater than 1.3 billion m3
and longer time is needed to reach the equilibrium.
Sanmenxia Dam not only caused retrogressive sedimentation and erosion in the lower Weihe River, but also
changed the river patterns. Before the reservoir began to
be used, the lower Weihe River was a meandering river,
with a value of sinuosity of about 1.65, in which sinuosity
is defined as the ratio of the length of the channel to the
length of the river valley. The closure of the dam reduced
the sinuosity to 1.06 in 1968, as shown in Fig. 8a. Very
quick sedimentation in this period buried the meandering
channel. In the meantime a straight channel developed
which was affected mainly by the reservoir operation. In
the period from 1970 to 1975 the Weihe River experienced erosion and the channel developed gradually from
straight to meandering. The sinuosity had gradually
increased to 1.2. In the following period more and more
meanders have developed and the lower Weihe River has
been developing toward meandering with a sinuosity
about 1.3.
Moreover, the river channel has become quite unstable since the closure of the dam. Fig. 8b shows the

in which A is a representative area of riverbed and

floodplain on which sedimentation occurs. Substituting
Eq. (3) into (2) yields
1
V AKeKt
2

Z


DZt e dt  DZt
Kt

in which t is the time from 1960 when Sanmenxia Dam

begin to fill and Tongguan's elevation began to rise. The
parameters in the equation are determined from data as
A = 5.30 108(m2) and K = 0.15/yr. Fig. 7 shows the
calculation result of the sedimentation volume (solid
curve) in comparison with the real sedimentation
volume (pyramids). The dashed curve in the figure is
the calculation result with the value of Zt remaining

Fig. 7. Calculated cumulative sedimentation volume with Eq. (4) (solid

curve) in comparison with the real sedimentation volume (pyramids).
The dashed curve in the figure is the calculation result with the value of
Zt remaining unchanged at 5 m (Wang and Li, 2003).

Z.Y. Wang et al. / Geomorphology 91 (2007) 6579

Fig. 8. (a) Variation of sinuosity of the lower Weihe River; (b)

Migration distances of the stream channel measured at crosssections
WY5WY35 during the first ascending and first descending periods of
Tongguan's elevation (Wang and Li, 2003).

migration distances of the stream channel measured at

crosssections WY535 during the first ascending and
descending periods of Tongguan's elevation. The
migration distance was up to 1.8 km at the cross
sections near Huaxian (WY11). The dam had less effect
in the reaches farther upstream, and the migration
distance was < 1 km at crosssections WY1835.
5. Erosion and resiltation below Sanmenxia Dam
Sanmenxia Reservoir has trapped about 7.1 billion
m3 of sediment in 45 yr, including the sedimentation
volume in the Weihe River. The sediment load and
runoff in the downstream reaches was then greatly
reduced, especially in the first 4 yr, which induced a
complex morphological process downstream of the
dam. The water and sediment released from Sanmenxia
Dam varied with the operation modes of the reservoir.
Consequently, erosion and resiltation occurred in the
reaches downstream from the dam. The process of bed
erosion and resiltation was very fast because the
sediment load was high and the bed material was
erodible. The erosion and resiltation occurred mainly in
the reach about 180600 km downstream from the dam.
About 2.31 billion tons of sediment had been eroded
from the riverbed in the first 4 yr since the closure of the
dam. In the following 9 yr, however, the reservoir
changed its operation mode from storage to detaining
flood water and sluicing sediment; and the downstream
channel was resilted at a high rate, with a total volume of

73

sediment deposition of about 3.95 billion tons (Yang

et al., 1994). Erosion and resiltation occurred both in the
stream channel and on the floodplain, with roughly 60%
in the channel and 40% on the floodplain.
The lower Yellow River was a wandering river,
although it has been confined within the strong grand
levees that are on both sides of the river 525 km apart
from each other. The migration rate of the channel was
quite high. The closure of Sanmenxia Dam did not
change this situation. Fig. 9 shows the wandering of the
lower Yellow River channel within the grand levees
during the storage operation (19601964) and late
operation periods (19801984) of Sanmenxia Dam. The
river migrated at high speed with a maximum value of
more than 5 km/yr. Even during the period immediately
following the closure of the dam, when clear water was
released into the reach, the channel migrated more than
3 km/yr.
Generally, dams tend to cause a reduction in
migration rates in the downstream reaches. For instance
the closure of the Danjiangkou Dam on the braided
Hanjiang River caused an initial reduction in bank
erosion intensity from about 25 m/yr during 19551960
to about 7.0 m/yr during a period of 17 yr immediately
after the dam closure (Xu, 1997). The lower Yellow
River did not respond the dam closure with reduced
channel migration because of the specific features of
reservoir operation.
The migration of river channel is a result of erosion
and sedimentation, which are caused by varying
discharge and sediment load. Wang and Wu (2001)
studied the channel migration in the lower Yellow River
and found that the speed of channel migration is
proportional to the square root of the fluctuation
intensity of discharge, Qrms, which is defined as follows:
 Z T
1=2
1
2
Qrms
Q  Qm dt
5
T 0
where Qm is the mean flow discharge averaged over the
period of T, Q is the discharge measured at time t
(0 < t < T). Generally, the time period of measurement, T,
is more than 5 yr. For the lower Yellow River, the
fluctuation intensity of discharge is on the same order or
even larger than the mean flow discharge. Sanmenxia
Reservoir has been operating at low pool level in flood
season since the second year after the closure and the
fluctuation intensity of flow discharge has not been
remarkably changed by the reservoir. For instance, both
the fluctuation intensity of discharge and the mean
discharge in the period of 19811985 were in the range
of 13001800 m3/s, almost the same as that before the
dam closure. Therefore, the migration of the channel

74

Z.Y. Wang et al. / Geomorphology 91 (2007) 6579

Fig. 9. The lower Yellow River channel wandered within the grand levees during the storage operation (19601964) and late operation periods
(19801984) of Sanmenxia Dam. (The curves are the thalweg of the channels) (Wang et al., 2004a,b).

remains unchanged, which is different from the

morphological responses to dam closure on other rivers.
To control the migration of the channel, local people
have constructed many spur dykes by the channel. The
spur dykes have, to a certain degree, fixed the channel
and concentrated the flow and cause sediment deposition between the spur dykes. The channel, therefore, was
deepened and relatively stabilized. A channelization
degree is defined as the ratio of the total length of the
spur dykes to the length of the channel, or the length of
spur dykes per channel length. Fig. 10 shows the
distribution of channelization degree along the river
downstream from Sanmenxia Reservoir. From the 1970s

to 2002, the degree has increased from 0.20.8 to 0.8

1.35. Nevertheless, the natural fluvial processes tend to
break the constraint of the spur dykes, and the flow
scours the dykes and causes them to collapse.
Fig. 11 shows the probability of collapse of each
dyke as a function of the channelization degree. The
probability is calculated with the total times of collapse
per year over the number of spur dykes. The probability
is low as if the channelization degree is < 0.8. For a
channelization degree >0.8, however, the probability of
dyke collapse abruptly increases from 10% to 30%. The
high probability of dyke collapse is due to the conflict
between the natural fluvial processes and the constraint

Fig. 10. Distribution of channelization degree (the ratio of the length of

the spur dykes to the length of the channel) along the lower Yellow
River (Wang et al., 2004a,b).

Fig. 11. Probability of dyke collapse as a function of the channelization

degree (Wang et al., 2004a,b).

Z.Y. Wang et al. / Geomorphology 91 (2007) 6579

Fig. 12. Probability of dyke failure as a function of bankfull discharge

(Wang et al., 2004a,b).

of channelization. In fact, the strongest conflict occurs

for channelization degrees in the range of 0.81.0,
therefore, dyke failure has a corresponding high
probability. Nevertheless, if the channelization degree
approaches to 2 (both sides of the channel are
completely controlled with spur dykes), the channel
motion will change from lateral to vertical. The channel
will be deepened, resulting in an increase in the bankfull
discharge. Fig. 12 shows the probability of dyke failure
against the bankfull discharge. Following an increase in
bankfull discharge, the probability of dyke failure
decreases.
Sanmenxia Reservoir has caused the lower Yellow
River to change from a wanderingbraided into a
wanderingsingle thread channel. Fig. 13 shows the
channel morphology of the TiexiePeiyu reach, which

75

is about 157189 km downstream from Sanmenxia

Dam, before and after the construction of the dam (Yang
et al., 1994). There were many sand bars before closure
of the dam; the number of bars had decreased 3 yr after
the dam was used for impoundment. The river had
become a singlethread channel by 1964. In the
meantime, the sinuosity of the river increased. The
number of meanders in a 300kmlong river reach (150
450 km downstream from the dam) had increased from
16 to 22. This reach developed from a wandering
braided channel to a wanderingmeandering channel.
Meanders have generally developed after the Sanmenxia Dam. The reach from the dam to Tiexie (0 to
157 km directly below the dam) is constrained by
mountains and no meanders develop within it. Statistics
are made for a 400kmlong reach, from 150 to 550 km
below the dam, which was an active fluvial reach. Before
the impoundment of the dam, only 16 meanders were
located in the 400kmlong reach, and more meanders
have generally developed after the impoundment. Fig.
14 shows the numbers of meanders with different
wavelength in the reach in the 1970s, 1980s, and 1990s.
The meander wavelength is defined as the distance
from one turning point of the channel on one side of the
valley to the next turning point on the same side. As
shown in Figs. 14, 17 small meanders exist in the reach
in the 1970s. Some meanders were separated by straight
sections and some other meanders connected with each
other and form small meandering sections. Between two

Fig. 13. Channel morphology of the TiexiePeiyu reach (157189 km from Sanmenxia) pre and postSanmenxia Dam.

76

Z.Y. Wang et al. / Geomorphology 91 (2007) 6579

process, manmade spur dykes affected, more or less, the

development of meanders.
6. Impacts of discharge reduction on fluvial
processes

Fig. 14. Numbers of meanders with different wavelengths in a 400km

long reach downstream of Sanmenxia Reservoir in the 1970s, 1980s,
and 1990s (Wang et al., 2004a,b).

small meandering sections was a section with straight

channel. Lately the reservoir operation became stable,
more meanders developed and the meandering sections
became longer. In the 1980s, however, 22 meanders
with a wavelength from 3 to 30 km had developed in the
reach. In the 1990s, the number of meanders continued
increasing and the meanders became regular; 31 of them
have meander wavelength within the range of 615 km.
The river became more and more meandering. In the

Sanmenxia Reservoir regulates, more or less, the

flow discharge for the downstream reaches, and the
peak discharge of extreme events was cut down.
Moreover, the annual runoff released to the downstream reaches has been reduced. The average
precipitation of the Yellow River basin is 476 mm,
but the pan evaporation is 10003000 mm/yr. The
total surface runoff of the watershed is 58 billion m3 ,
about 2% of the total of China. The downstream
channel is a perched river, with its riverbed 10 m
higher than the surrounding land. This poses a
flooding risk but also provides the potential for
water diversion to farmland and numerous cities and
towns within and outside of the Yellow River basin.
For instance, residents in Tianjin and Qingdao cities,
which are several hundred kilometers from the Yellow
River and outside of the river basin, are drinking water
from the Yellow River. Currently more than 4500

Fig. 15. Variation of annual runoff and sediment load in the period from 1960 to 1997 at Xiaolangdi (130 km from Sanmenxia Dam) and Lijin
(900 km from Sanmenxia Dam). The differences between the two stations are due to the inflow from tributaries and water diversions (Wang et al.,
2004a,b).

Z.Y. Wang et al. / Geomorphology 91 (2007) 6579

Fig. 16. Bankfull discharge along the lower Yellow River course
during different periods (Wang et al., 2004a,b).

water diversion projects and 29,000 pumping stations

have been completed for irrigation and water supply
since the dam construction. The irrigation area has
increased from 0.8 million ha in 1950 to 7 million ha
in 1995.
Water diversion inevitably affects the fluvial processes. Water diversion may even change a section of a
perennial stream to an ephemeral river section (Fogg
and Muller, 1999). For the first time in its history, the
Yellow River failed to reach the ocean in 1972. In 1997,
the Yellow River failed to reach the sea twothirds of the
year because most of the river water was diverted. While
water diversion projects have become a popular and
important strategy to meet increasing water demand, the
stream flow, sediment transport, and fluvial processes of
rivers are increasingly affected.
Fig. 15 (a) and (b) shows the variation of the annual water and sediment load from 1960 to 1997 at the
Xiaolangdi (130 km downstream from Sanmenxia) and
Lijin (900 km downstream from Sanmenxia) hydrologic
stations, in which the horizontal lines represent the
average runoff and sediment load. The differences
between the figures at the two stations are due to the
inflow from tributaries and outflow by water diversions
along the course from Xiaolangdi to Lijin. From 1960
1969, more water flowed through Lijin than Xiaolangdi
because water diversion was less than the inflow from
tributaries. From 19701985, the annual runoff at Lijin
was equal to or slightly less than at Xiaolangdi because
more water had been diverted. From 1986 to the present,
however, the total volume of water diverted was much
more than the inflow from tributaries, and the water
runoff decreased along the course. The annual runoff was
about 11 billion m3 less at Lijin than at Xiaolangdi. The
reduction in runoff over a long stretch of the river elicited
a sharp reduction in the flow's sedimentcarrying
capacity. Therefore, the annual load was much less at
Lijin than at Xiaolandi from 1986 to the present.
Table 1 also shows the decrease of water and
sediment load along the course. From 1986, water and

77

sediment loads increased along the course and reached

their maximum values at Huayuankou and then
decreased farther downstream from diversion. The
sediment load at Lijin is less than that at Sanmenxia
by more than 300 million tons, which must have been
deposited in the reach between Sanmenxia and Lijin
and consequently changed the morphology of the
river.
One of the impacts of the runoff reduction on the
fluvial processes was the shrinkage of the channel.
Fig. 16 shows the bankfull discharge of the lower Yellow River during different periods. Water diversion has
reduced the discharge and sedimentcarrying capacity,
and sediment has been deposited in the channel, which
has made the channel shallow and unstable. As a result,
the bankfull discharge has decreased steadily. The
bankfull discharge was about 9000 m3/s in 1958 and
1964; it decreased to about 6000 m3/s in 1985, and to
only 3000 m3/s in 1999. The shallow channel cannot
accommodate floodwater as before; so the flood stage
has become extremely high and the phenomenon known
as the little flood with high flooding disasters has
occurred in the river basin.
The second important impact of water diversions is
the adjustment of the riverbed profile. Field evidence
from natural streams shows that variations in successive
processes and forms result from a system's tendency to
minimize the rate of energy dissipation with time
(Simon, 1992). According to the minimum stream
power theory, the morphology of fluvial rivers develops
to reach the minimum stream power (Yang, 1996). This
can be described by the following equation:


dP
d
ds
dQ
csQ c Q s
0
dx dx
dx
dx

in which P is the stream power in ton/s, is the specific

weight of water in ton/m3, s is the riverbed slope, x is

Fig. 17. Longitudinal bed profiles of the lower Yellow River in 1977
and 1997 (Wang et al., 2004a,b).

78

Z.Y. Wang et al. / Geomorphology 91 (2007) 6579

the distance along the river course in km, and Q is the

discharge in m3 /s. For most rivers, the discharge
increases along the course from the inflow from
tributaries; thus, the term sdQ / dx is positive. According
to Eq. (6), the term Qds / dx must be negative, or the
slope of the riverbed decreases along the course, so that
these rivers exhibit concave riverbed profiles. Eq. (6)
indicates the direction of morphological processes and
equilibrium state of longitudinal river profile. Sediment
load plays an important role in the speed of morphological process but does not change the direction and the
final equilibrium of the profile. The higher the sediment
load the faster is the morphological process. For a low
sediment load river the riverbed profile often does not
meet Eq. (6) because it takes a very long time to reach
the minimum stream power profile.
The Yellow River carries heavy sediment load and the
morphological processes are fast. The large quantity of
water diverted along the course of the Yellow River
makes the term sdQ / dx negative. For instance, since
1986, the average discharge has decreased along the
Yellow River course in the reach downstream of
Huayuankou, i.e., dQ / dx < 0. According to Eq. (6) the
term Qds / dx must be positive. In this case, the riverbed
profiles will develop toward a convex shape, which is
different from the normal concave curve. Fig. 17 shows
the bed profiles of the lower Yellow River for 1977 and
1997. The mean bed elevation is the average bed
elevation of the channel with a crosssection of wet area
about 500 m2. The figure shows that the lower section of
the river is developing toward a convex profile. Because
the profile of the upper section is concave, the river
shows an Sshape longitudinal bed profile. The trend
will continue, and the turning point in the profile will
move upstream because the water diversion is continuing
and more water will be diverted in the foreseeable future.

In the reaches downstream from the reservoir the

channel responses to the dam closure with erosion and
resiltation. The channel migrates several kilometers per
year within the levees after the closure, remains in the
same range of migration distance. The main reason for
the wandering nature is the high fluctuation intensity of
discharge. To stabilize the channel, many spur dykes
have been constructed within the 1020km wide
valley defined by the Grand Levees. Channelization
degree is defined as the ratio of the total length of the
spur dykes to the total length of the banks. The highest
probability of dyke failure occurs if the channelization
degree is between 0.8 and 1 because of the conflict
between the natural fluvial processes and the constraint
of channelization. The braided channel has changed
into a singlethread channel after the dam closure.
Meanders have gradually developed, and the river
changes from braided to wandering or from wanderingbraided to wanderingmeandering. The flow
discharge decreases along the course because of the
operation of the reservoir and water diversion, which
induces morphological responses. According to the
minimum stream power theory the riverbed profiles are
developing toward an Sshape longitudinal bed profile
with a concave profile in the upper reach and a convex
profile in the lower reach because the quantity of water
diversion along the course is more than the inflow from
its tributaries.
Acknowledgements
This work was sponsored by the Science Fund for
Creative Research Groups of the National Natural
Science Foundation of China (No. 502219033) and
the SinoDutch Project (2004CB720402).
References

7. Conclusions
Sedimentation in Sanmenxia Reservoir raised the
Tongguan's elevation, which acts as the base level of
the bed profile of the Weihe River. The ascent and descent
of Tongguan's elevation have induced sedimentation and
erosion in the lower Weihe River in the form of retrogressive waves. The waves propagated toward upstream
along the Weihe River, at a speed of about 10 km/yr. For a
given increment in Tongguan's elevation, sedimentation
volume in the Weihe River is in equilibrium. The model
proposed in the paper is based on the assumption that the
sedimentation rate is proportional to the difference
between the real sedimentation volume and equilibrium
sedimentation volume agrees well with the data.

Church, M., 1983. Pattern of instability in a wandering gravel bed

channel. In: Collinson, J.D., Lewin, J. (Eds.), Modern and Ancient
Fluvial Systems, Special Publication No. 6. IAS University Press,
Cambridge, UK, pp. 169180.
Collier, M., Webb, R.H., Schmidt, J.C., 1996. Dams and rivers. A
Primer on the downstream effects of dams. U.S. Geological Survey
Circular, vol. 1126 (94 pp.).
Ferguson, R.I., Werritty, A., 1983. Bar development and channel
changes in the gravelly River Feshie, Scotland. In: Collinson,
J.D., Lewin, J. (Eds.), Modern and Ancient Fluvial Systems,
Special Publication no. 6. IAS University Press, Cambridge, UK,
pp. 181193.
Fogg, J.L., Muller, D.P., 1999. Resource values, instream flows and
ground water dependence of an oasis stream in the Mojave Desert.
Annual Summer Specialty Conference Proceedings, Science into
Policy: Water in the Public Realm/Wildland Hydrology, June 30
July 2, Bozeman, Montana.

Z.Y. Wang et al. / Geomorphology 91 (2007) 6579

Goodwin, P., Falte, M., Betts, A.D.K., 2001. Managing for unforeseen consequences of large dam operations. Working Paper of the
World Commission on Dams, Vlaeberg, Cape Town, South
Africa.
Graf, W.L., 1977. The rate law in fluvial geomorphology. American
Journal of Science 277, 178191.
Gregory, K.J., Park, C., 1974. Adjustment of river channel capacity
downstream from a reservoir. Water Resources Research 10,
870873.
Hey, R.D., 1979. Dynamic processresponse model of river channel
development. Earth surface Process and Landforms 4, 5972.
Knighton, A.D., Nanson, G.C., 1993. Anastomosis and continuum of
channel pattern. Earth Surface Processes and Landforms 18,
613625.
Kondolf, M.G., 1997. Hungry water: effects of dams and gravel
mining on river channels. Environmental Management 21 (4),
533551.
Leopold, L.B., 1973. River channel change with time: an example.
Geological Society of America Bulletin 84, 18451860.
Li, S.H., Long, Y.Q., 1994. Application of modified method for
sediment discharge in the Lower Yellow River. Journal of
Sediment Research 2, 4048 (in Chinese).
Long, Y.Q., Zhang, Y.F., 2002. Study on the Yellow River sediment by
the viewpoint of total sediment. Yellow River 24 (8), 2829 (in
Chinese).
Long, Y.Q., Liang, G.T., Zhang, Y.F., Shen, G.Q., Zhang, L.Z., Cheng,
L.Y., 2002. Range survey of deposition in the Lower Yellow River.
Journal of Sediment Research 17 (2), 212.
Mahmood, K., 1987. Reservoir sedimentation: Impact, extent,
mitigation. World Bank Tech. Rep., vol. 71. Washington D.C.
Neill, C.R., 1973. Hydraulic and morphologic characteristics of
Athabasca River near Fort Assinboine. Alberta Research Council,
Edmonton, Hway River Engng. Div. Rep. REH/73/3. 23 pp.
Petts, G.E., 1979. Complex response of river channel morphology
subsequent to reservoir construction. Progress in Physical
Geography 3, 329362.
Qian Ning (Ning Chien), Dai Dingzhong, 1980. The problems of river
sedimentation and present status of its research in China.
Proceedings of International Symposium on River Sedimentation.
Guanghua Press, China.
Qian Ning (Ning Chien), Zhang Ren, Zhou Zhide, 1987. Fluvial
processes. China Science Press. (in Chinese).
Robins, C.H., Simon, A., 1983. Maninduced channel adjustment in
Tennessee streams. United States Geological Survey, Water
Research Investigations report, 824098, p. 129.

79

Schumm, S.A., 1973. Geomorphic thresholds and the complex

response of drainage system. In: Morisawa, M. (Ed.), Fluvial
Geomorphology. State University of New York, Binghamton,
pp. 229310.
Simon, A., 1989. A model of channel response in disturbed alluvial
channels. Earth Surface Process and Landforms 14, 1126.
Simon, A., 1992. Energy, time, and channel evolution in catastrophically disturbed fluvial systems. Geomorphology 5, 345372.
Simon, A., Thorne, C.R., 1996. Channel adjustment of an unstable
coarsegrained stream: opposing trends of boundary and critical
shear stress, and the applicability of extremal hypotheses. Earth
Surface Process and Landforms 21, 155180.
Wang, Z.Y., Li, C.Z., 2003. A study on the effect of Tongguan's
elevation on sedimentation in the Weihe River. Research Report,
Institute of River and Coastal Engineering, Tsinghua University (in
Chinese).
Wang, Z.Y., Wu, B.S., 2001. Sedimentremoving capacity and river
motion dynamics. International Journal of Sediment Research 16
(2), 105115.
Wang, ZhaoYin, Cheng, Dongsheng, Li, Changzhi, Zhou, Jing,
2004a. Analysis on fluvial processes and failure of spur dykes in
the Lower Yellow River. Research Report, Institute of River and
Coastal Engineering, Tsinghua University (in Chinese).
Wang, ZhaoYin, Wu, Baosheng, Li, Changzhi, 2004b. Has the
sedimentation in the Lower Weihe River reached equilibrium?
People's Yellow River, No. 5 (in Chinese).
Williams, G.P., Wolman, M.G., 1984. Downstream effects of dams on
alluvial rivers. U.S. Geological Survey Professional Paper 1286.
83 pp.
Wu, Baosheng, Wang, ZhaoYin, 2004. Impacts of Sanmenxia Dam
and management strategies, keynote lecture. Proceedings of the
International Conference on Hydraulics of Dams and River
Structures, Tehran, Iran.
Xu, Jiongxin, 1996. Wandering braided river channel pattern
developed under quasiequilibrium: an example from the Hanjiang
River, China. Journal of Hydrology 181, 85103.
Xu, Jiongxin, 1997. Evolution of midchannel bars in a braided river
and complex response to reservoir construction: an example from
the middle Hanjiang River, China. Earth Surface Processes and
Landforms 22 953965.
Yang, C. Ted, 1996. Sediment TransportTheory and Practice. The
McGrawHill Companies, Inc., New York.
Yang, Qingan, Long, Yuqian, Miao, Fengju, 1994. Operational
Studies of Sanmenxia Project on the Yellow River. Henan People's
Press. (in Chinese).