Earth Surf. Process. Landforms35, 1666–1673 (2010) Copyright © 2010 John Wiley & Sons, Ltd.
Published online 28 June 2010 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/esp.2006
Effect of reservoir construction on suspended
sediment load in a large river system: thresholds
and complex response
Jiongxin Xu and Yunxia Yan*
Institute of Geographical Sciences and Natural Resources Research, Key Laboratory for Water Cycle and Related Land Surface Processes, Chinese Academy of Sciences, Beijing, China
Received 9 September 2009; Revised 25 January 2010; Accepted 2 February 2010
* Correspondence to: Yunxia Yan, Institute of Geographical Sciences and Natural Resources Research, Key Laboratory for Water Cycle and Related Land Surface Processes, Chinese Academy of Sciences, Beijing, China. E-mail: [email protected]
ABSTRACT: Dam construction greatly alters the channel boundary of rivers, making the dammed river system a human-controlled system. Based on hydrometric data in the upper Changjiang River basin, the change in behaviour of sediment transport of some dammed rivers was studied. As a result, some phenomena of threshold and complex response were found. When the coeffi cient (Cr,a) of actual runoff regulation by reservoirs, defi ned as the ratio of total capacity of reservoirs to annual runoff input, is smaller
than 10%, suspended sediment load at Yichang station, the control station of the Changjiang River, shows a mild decreasing trend. When this coeffi cient becomes larger than 10%, suspended sediment load decreases sharply. The coeffi cient of 10% can be regarded as a threshold. The Cr,a of 10% is also a threshold, when the variation of suspended sediment concentration (SSC)
with Cr,a at Yichang station is considered. The impacts of reservoir construction can be divided into several stages, including road
construction, dam building and closure, water storage and sediment trapping. During these stages, some complex response was identifi ed. At the station below the dam, SSC increases and reaches a maximum, and then declines sharply. This phenomenon was found on the main-stem and several major tributaries of the upper Changjiang River. In the Minjiang River, where a series of dams were built successively, the response of SSC is more complicated. Copyright © 2010 John Wiley & Sons, Ltd.
KEYWORDS: reservoir; sediment transport; threshold; complex response; upper Changjiang River
Introduction
The impact of human activity on sediment yield and transport in rivers is an important subject both in theory and in engi-neering practice. Reservoir construction greatly changes sedi-ment transport and deposition of rivers and many research results have been published on this topic (Ackermann et al., 1973; Petts, 1984; Williams and Wolman, 1984; Church, 1995; Fergus, 1997; Brandt, 2000; Graf, 2005; Francis and Nislow, 2005; Chien, 1985; Chien et al., 1987; Han, 2003;
Phillips et al., 2005). The sediment load of the upper
Changjiang River, and its major tributaries, has been altered since a large number of reservoirs were built, including the Three Gorges Dam (Zhang and Wen, 2002; Xu et al., 2004; Yang et al., 2005; Fu et al., 2006; Dai et al., 2007; Zhang and Chen, 2007; Xu, 2007). Over the past 50 years, reservoir construction in the upper Changjiang River basin has under-gone several stages. In the fi rst stage, only some small-sized reservoirs were built, on small tributaries. In the second stage, reservoirs were built on major tributaries. In the third, large reservoirs were built on the main stem of the river, such as Gezhouba Dam and the Three Gorge Dam. Since 2000, hydro-electric development has entered a new stage, and
many cascade dams have been planned on the Jinshajiang, Minjiang, Jialingjiang and Wujiang Rivers. During large-scale hydropower development, sediment trapping by reservoirs greatly increases and becomes the most important impact on sediment reduction of the upper Changjiang River. The closure and water storage of the Three Gorges Dam has dramatically changed the sediment load of the upper Changjiang River. In the near future, four cascade dams will be built on the Jinshaliang River, i.e. the Xiangjiaba Dam with a capacity of 51·85 × 108 m3, the Xiluodu Dam with a capacity of 126·7 ×
108
m3
, the Baihetan Dam with a capacity of 193·8 × 108
m3
, and the Dongwude Dam with a capacity of 39·4 × 108 m3
(Chen and Huang, 2005). These reservoirs will further change the hydrological regime of the upper Changjiang River and trap large quantities of sediment, further reducing sediment transport. The construction of the Xiangjiaba and Xiluodu Dams are underway. At the same time, construction of cascade dams on Jialingjiang, Minjiang and Wujiang Rivers also occurred. To predict the changes of sediment load of the upper Changjiang River in the future, it is important to study the impact of reservoirs on sediment transport.
Dams may be regarded as a human-made boundary condi-tion of a river, and after construccondi-tion of reservoirs, the river
network and dams constitute a complex system, i.e. a human-controlled natural system, which is more complicated than a natural river system. Thresholds, complex response and scale effect are important characteristics of the behaviour of a fl uvial system (Schumm, 1973; Becker et al., 1999; Blöschl, 1999). Much research has explored the complex response of river channels to reservoir construction (Petts, 1979; Xu, 1990a, 1990b, 1991; Sherrard and Erskine, 1991; Benn and Erskine, 1994). Many results have been reported on changes of sedi-ment transport in the Changjiang River (Lu and Higgitt, 1998, 1999, 2001; Yang et al., 2002; Zhang and Wen, 2002; Xu et al., 2004; Yang et al., 2005; Fu et al., 2006; Dai et al., 2005; Zhang and Chen, 2007), and some researchers considered the effect of cascade dams on hydrological regime of rivers (Chen and Huang, 2005). However, little attention has been paid to thresholds and complex response phenomena related with the effect of reservoir construction on suspended sediment of the upper Changjiang River. This will be dealt with in the present study, to get a deeper insight to the complex behaviour of rivers under human impacts.
Study Area, Data Source and Method
The drainage basin above Yichang, draining 100·5 million km2, is the upper Changjiang basin (Figure 1), which covers
the Qinghai-Tibet Plateau, the Yunnan-Guizhou Plateau, the Sichuan Basin and the Qinling and Dabashan Mountains. According to the data from 1950 to 2000, mean annual runoff and suspended sediment load at Yichang station were 4382 × 108
m3
and 5·01 × 108
t, respectively. The major tributaries
include the Jinshajiang, Minjiang, Jialingjiang and Wujiang Rivers, and their runoff and sediment loads are shown in Table I. The runoff from Jinshajiang and Jialiangjiang Rivers accounts for 32·6% and 15·3% of that at Yichang station, respectively, but the sediment load accounts for 50·9% and 20·4% of that at Yichang, respectively. Thus, the two rivers are major sedi-ment source of the upper Chanjiang River, and the annual suspended sediment concentrations (SSCs) of the Jinshajiang and Jialiangjiang Rivers are 1·54 and 1·57 times that at Yichang station, respectively (Xu et al., 2004).
Since the 1950s, reservoirs have been built in the upper Changjiang River basin. The reservoir data used in the present study, including the date of reservoir completion and the capacity of the reservoir, are obtained from various sources, mainly from Yang et al. (2005) and Hydrological Bureau of Changjiang River Water Conservancy Commission (2008). The data on runoff, suspended sediment load and concentra-tion are provided by relevant hydrometric staconcentra-tions, including Yichang station on the Changjiang River and some other sta-tions on its tributaries. All the measurements of water stage, discharge and SSC, and load at hydrometric stations are con-ducted following national standards issued by the Chinese Ministry of Water Conservancy and Electric Power (e.g. 1962, 1975). The procedures used for hydrological survey, sampling and laboratory analyses at hydrometric stations in China are basically the same as used internationally. Strict checks and error analyses were carried out following the standards, to ensure the accuracy of the data. A detailed introduction is provided in Yan (1984) and the Sedimentation Commission of Chinese Society of Hydraulic Engineering (1992). As data on bedload transport are not available, the present study deals
Figure 1. A map showing the Changjiang River basin. (1)Three Gorges Dam; (2) Gezhouba Dam; (3) Ertan Dam; (4) Wujiangdu Dam; (5) Gongzui Dam; (6) Tongjiezi Dam; (7) Bikou Dam; (8) Baozhusi Dam.
Table I. Hydrometric characteristics for the rivers studied
River Station
Drainage area (km2)
Annual suspended load Annual runoff
Annual suspended concentration Period 108 t Percentage of that at Yichang 108 m3 Percentage of that at Yichang (kg/m3)
Jingshajiang River Pingshan 485099 2·5544 32·6 1426·0 50·9 1·76 154·4 1950–2000
Minjiang River Gaochang 135378 0·1210 19·8 865·7 9·5 0·62 54·4 1956–2000
Jialingjiang River Beibei 156142 1·2011 15·3 665·4 24·0 1·80 157·9 1956–2000
Wujiang River Wulong 83035 0·2803 11·4 497·3 5·6 0·55 48·2 1956–2000
Changjiang River Yichang 1005501 5·0094 100 4381·0 100 1·14 100 1950–2000
with suspended sediment transport only. Suspended sediment median size (D50) at Yichang station was used. The annual D50
is calculated as the weighted average of monthly D50, with
monthly sediment load as the weight. Based on the earlier data, time series analysis and statistical analysis are con-ducted, to determine the thresholds and understand the nature of system response.
Results and Explanation
Regulation of reservoir and impacts on annual
river fl ow
To describe the impact of the reservoir on annual river fl ow, three indices are introduced: (1) total capacity (Ct) of all
res-ervoirs in a river basin above a hydrometric station; (2) coef-fi cient of reservoir regulation on annual mean river fl ow (Cr,m),
defi ned as the ratio of total capacity of reservoirs in a certain year to annual mean river fl ow at the hydrometric station over a period of years; (3) coeffi cient of reservoir regulation on annual river fl ow (Cr,a), defi ned as the ratio of total capacity
of reservoirs in a certain year to annual river fl ow of this year. The Cr,m index expresses the general effect of regulation of
reservoirs on river fl ow over a period of time, and the Cr,a
index expresses the actual effect of regulation of reservoirs on river fl ow for each year. Therefore, the Cr,m may be named
‘coeffi cient of reservoir capacity’ and the Cr,a ‘coeffi cient of
actual runoff regulation by reservoirs’.
The temporal variation of total capacity of reservoirs (Ct) is
shown in Figure 2, which indicates a clear increasing trend. Before 1995, Ct increased slowly. After this year, Ct increased
sharply, due to the completion of a series of large-sized res-ervoirs including Ertan and Three Gorge Resres-ervoirs. The tem-poral variations of Cr,m and Cr,a are plotted in Figure 3, showing
similar increasing trend. Because of the inter-annual fl uctua-tion in river fl ow, the Cr,a has someinter-annual fl uctuation.
Note that, due to the abnormal draught in the summer of 2006, annual runof at Yichang station in 2006 was 2848 × 108
m3
, only 65·8% of the mean average. Hence, the coeffi -cient of actual runoff regulation (Cr,a) by reservoirs is rather
high.
Impact of reservoirs on annual sediment transport
of rivers
Construction of the reservoir raises local base level of the river at the dam, and channel gradient above the dam decreases signifi cantly. As a result, sediment deposition occurs, and the sediment input to the reservoir is larger than the sediment
output from the reservoir. Thus, sediment transport below a reservoir decreases. Due to sediment trapping by reservoirs, sediment yield decreases. Obviously, the larger the total capacity of reservoirs above a hydrometric station is, the larger the reduction of sediment yield is. Since sediment trapping effi ciency is controlled by the ratio of reservoir capacity to runoff input to the reservoir (Brune, 1953), sediment transport below the dam may be related to the coeffi cient (Cr,a) of actual
runoff regulation by reservoirs. Figure 4(a) shows the relation-ship between annual suspended sediment load at Yichang station and the Cr,a index for the upper Changjiang River basin
above Yichang, based on yearly data from 1954 to 2003. Close negative correlation can be seen with R2=
0·78, which is signifi cant at a level of p < 0·001. It is notable that the relationship is non-linear and can be fi tted by two straight lines with different slopes. When Cr,a issmaller than 10%, a
weak decreasing trend of annual suspended load (Qs,Yichang) at
Yichang station with the increased Cr,a can be seen, and the
scatter of points is largely due to inter-annual variation of rainfall, especially that of rainstorms. However, when Cr,a
becomes larger than 10%, points become concentrated, and Qs,Yichang decreases rapidly with the increased Cr,a.Thus, the
percentage of 10% may be regarded as a threshold, where the response of Qs,Yichang to the increased Cr,a changes. When the
Cr,a is very small, the amount of sedimentation in reservoirs is
also small and only accounts for a minor proportion of Qs,Yichang,
and the variation in reservoir sedimentation is smaller than the inter-annual fl uctuation in Qs,Yichang. Thus, the decreasing trend
in Qs,Yichang is not clear. Only if the Cr,a becomes suffi ciently
large does the reservoir sedimentation become clear, and a signifi cant decrease in Qs,Yichang can be observed. In Figure 4,
regression lines are fi tted for the points with Cr,a> 10% and
800 700 600 500 400 300 200 100 1950 1960 1970 1980 1990 2000 2010 0 Year T o
tal capacity of reser
v o irs (10 8m 3) 0 0.05 0.1 0.15 0.2 0.25 0.3 1950 1960 1970 1980 1990 2000 2010 Year
Coefficient of reservoir capacity
Coefficent of actual runoff regulation b
y
reservoirs
C. of reservoir capacity C. of actual runoff regulation
For Cr,a<0.1: 1950-1y = 54473e-3.4156x R2 = 0.1137 For Cr,a>0.1: y = 313284e-22.473x R2 = 0.9779
For all dada: y = 72160e-12.547x R2 = 0.7777 100 1000 10000 100000 0 0.05 0.1 0.15 0.2 0.25 0.3
Coefficient (Cr,a) of actual runoff regulation by reservoirs
Annual suspended load (10
8t/yr)
Figure 2. Temporal variation in total capacity of reservoirs in the upper Changjiang River basin.
Figure 3. Temporal variation in the coeffi cient of reservoir capacity and the coeffi cient of actual runoff regulation by reservoirs in the upper Changjiang River basin. This fi gure is available in colour online at wileyonlinelibrary.com
Figure 4. Relationship between annual suspended sediment load at Yichang station and the coeffi cient of actual runoff regulation by reservoirs in the upper Changjiang River basin.
<10%, respectively, and regression equations are given. The slope of the two regression lines is quite different, it is −3·42 for the former and −22·5 for the latter, indicating that the rate at which Qs,Yichang decreases with Cr,a is 6·6 times larger when
Cr,a is >10% than that when Cr,a is <10%.
The relationship between annual suspended sediment con-centration (SSCYichang) at Yichang station (SSCYichang) and the
coeffi cient of actual runoff regulation by reservoirs in the upper Changjiang River basin (Figure 5) is similar to Figure 4. The squared correlation coeffi cient is R2=
0·8068, signifi cant at a level of p < 0·001. This is larger than the R2 for the
Qs,Yichang–Cr,a relationship. The relationship in Figure 5 is also
non-linear. When Cr,a is smaller than 10%, the rate at which
SSCYichang decreases with the increased Cr,a is very low, but
when Cr,a becomes larger than 10%, it becomes very high.
Thus, Cr,a = 10% is also a threshold. In Figure 5, regression
lines are fi tted for the points with Cr,a > 10% and <10%,
respectively, and regression equations are given. The slope of the regression line for the former is −3·60, and that for the latter is −20·4, meaning that the rate at which SSCYichang
decreases with Cr,a is 5·7 times higher when Cr,a is > 10% than
that when Cr,a is < 10%.
Apart from the discussion earlier, an alternative explanation can be given for the breaks in Figures 4 and 5. Brune (1953) showed that the value of the ratio of reservoir capability to runoff was correlated with reservoir sediment trapping effi -ciency for the particular group of reservoirs he studied. That is, the ratio explains a statistically signifi cant portion of the differences in sediment trapping between reservoirs. The Brune’s relationship is valid only if the characteristics of the reservoirs constructed and their operations did not change over time. The situation for reservoirs involved in Figures 4 and 5 is somehow different. In the Upper Changjiang River basin, the reservoir construction showed some difference for the periods before and after 1995. Before 1995, small and moderated sized reservoirs were dominant, but afterwards, a number of large sized reservoirs were built. It can be expect that a large number of small reservoirs would have a very different sediment trapping effi ciency than one large hydro-power reservoir, even though the value of the capacity to runoff ratio was identical. The completion of a series of large-sized reservoirs after 1995 coincidentally increased the Cr,a
value from approximately 0·08 to greater than 0·20. Thus, the points from the period before 1995 and those after 1995 should be regarded as two sample populations for the use of the Brune’s relationship. Hence, different slopes appear for the two sample populations. The break between the two straight lines may be regarded as a threshold, which refl ects the fact that large sized reservoirs have higher sediment trapping
effi ciency than small and moderate sized reservoirs, when the Cr,a is equal.
Xu (2007) has studied temporal variation of suspended sedi-ment median grain size (D50) of the Upper Changjiang River
and its major tributaries and found that over the past 40 years, the D50 of these rivers showed a clear fi ning trend. The
decrease in D50 can be related to human activity such as soil
conservation measures and reservoir construction. Regulation of runoff by reservoirs causes sediment deposition in the res-ervoir. In general, coarser sediment has higher probability to deposit, and thus, the sediment released from the reservoir becomes fi ner than that before reservoir construction. This is evidenced by the relationship between annual SSC at Yichang station and the coeffi cient (Cr,a) of actual runoff regulation by
reservoirs in the Upper Changjiang River basin (Figure 6), which shows a negative correlation, signifi cant at a level of p< 0·01.
Complex response of SSC during
reservoir construction
River systems are complex. After a series of reservoirs are built on a main-stem and in tributaries, channel boundary condi-tions could change dramatically, and the river system would become more complicated, during and after the construction of reservoirs.
The change in river sediment transport can be caused by both changing human factors (soil conservation, land-use change, reservoir construction, etc.) and climate factors (pre-cipitation, etc.). In general, when precipitation increases, both river fl ow and sediment load of a river may increase. Then the increase in sediment concentration, which may be expressed as the ratio of sediment load to river fl ow, may be smaller than the increase in sediment load alone. River fl ow is closely correlated with precipitation. Thus, when SSC is used, the infl uence of changed precipitation may be offset partly. Hence, we use SSC as an index to discuss the sediment transport behaviour of dammed rivers in response to reservoir construction.
Before the dam building, roads are usually built to carry heavy machineries and building materials. During road build-ing, large quantities of rocks and soils are dumped to the river. When the dam is built, more wasted materials are derived for slope excavation, which are also dumped to the river. This will increase sediment transport or SSC in the affected river reaches. After the closure of the dam, sediment trapping starts, and sediment transport or SSC below the dam will decrease. When the project is completed and the reservoir starts to store For Cr,a<0.1:
y = 1.2638e-3.5987x
R2 = 0.2121 For Cr,a>0.1: y = 5.8306e-20.442x
R2 = 0.9534
For all data: y = 1.6237e-11.769x R2 = 0.807 0.01 0.1 1 10 0 0.05 0.1 0.15 0.2 0.25 0.3
Coefficient (Cr,a) of actual runoff regulation by reservoirs
Annual
SSC
(kg/m
3)
Figure 5. Relationship between annual suspended sediment concen-tration at Yichang station and the coeffi cient of actual runoff regula-tion by reservoirs in the upper Changjiang River basin.
y = 0.0031x-0.3306 R2 = 0.6663 0.001 0.01 0.1 0.001 0.01 0.1 1
Coefficient of actual runoff regulation by reservoirs
Annual suspended sediment
D50
(mm)
Figure 6. Relationship between annual suspended sediment concen-tration at Yichang station and the coeffi cient of runoff regulation by reservoirs in the Upper Changjiang River basin.
water, more sediment is trapped and sediment transport or SSC will further decline. During the whole course, SSC shows a cyclic variation, namely, SSC increases to a maximum, fol-lowed by a decrease. This may by regarded as a complex response induced by reservoir construction, which is well evidenced by the data from a number of dams on the main stem of the Changjiang River and its tributaries.
The Bailongjiang River, the major tributary of the Jialingjiang River, is the most important sediment source of the Jialingjiang River. On the Bailongjiang River, two large-sized reservoirs were built, i.e. the Bikou Reservoir and the Baozhusi Reservoir. The Bikou Dam is 101 m high, and the total capacity of the reservoir is 5·21 × 108
m3
. The drainage area controlled by the dam is 26,000 km2, the mean annual runoff input is 86·72
× 108
m3
, the annual sediment load input is 2460 × 104
t, and the mean annual SSC is 2·64 kg/m3. Since the data of SSC are
not available, Figure 7(a) shows the temporal variation in annual suspended sediment load at Bikou station before and after the construction of the Bikou Reservoir. During the con-struction of the reservoir, sediment load at the Bikou station immediately below the dam increased to a maximum. When the dam was closed in 1971, sediment load started to decline. After the reservoir was completed and used for water storage and electricity generation in 1975, sediment load further decreased to a level much lower than that before the reservoir construction.
The Baozhusi Dam is located downstream from the Bikou Dam, controlling a drainage area of 28,425 km2. The total
capacity is 25·5 × 108
m3
. The dam construction was started in 1989, the electricity generation was started in 1996, and the project was completed in 1998. The temporal variation of SSC at Sanleiba station below the Baozhusi Reservoir is shown in Figure 7(b). When the dam construction was started, annual SSC at Sanleiba station increased. Before the dam was closed, SSC increased to a maximum, and then decreased. After the reservoir was used for water storage and hydro-power genera-tion, SSC further declined to a level much lower than that before the reservoir construction.
The Wujiangdu Reservoir was built on the middle reach of the Wujiang River, controlling a drainage area of 27,800 km2.
The height of the dam is 165 m, and the total capacity is 23·0
× 108 m3. The dam construction was started in 1973, and the
project was completed in 1983. Figure 7(c) shows the tempo-ral variation in annual SSC at Wulong station downstream from the reservoir before and after the construction of the Wujiangdu Reservoir. It can be seen that during the construc-tion of the reservoir, SSC increased. After the dam was closed, SSC decreased. When the project was completed and the reservoir was used for water storage and electricity generation, SSC further declined to a level much lower than that before the reservoir construction.
The Ertan Reservoir was built on the lower reach of the Yalongjiang River, the largest tributary of the Jinshajiang River. The dam controls the drainage area of 116,400 km2. The total
capacity is 58·0 × 108
m3
. The road construction was started in 1987, the dam construction was started in 1991, the elec-tricity generation was started in 1998 and the project was fi nally completed in 2000. The Ertan Hydropower station is the largest one of those built in the twentieth centaury in China. Figure 7(d) shows the temporal variation in annual SSC at Xiaodeshi station below the reservoir before and after the construction of the dam. It can be seen that during the con-struction of the reservoir, SSC increased. After the dam was closed, SSC decreased. When the project was completed and the reservoir was used for water storage and electricity genera-tion, SSC further declined to a level much lower than that before the reservoir construction.
0 5 10 15 20 25 30 35 40 1955 1960 1965 1970 1975 1980 1985 Year
Annual suspended load at Bikou
(10 6t/yr) 1971, dam closure 1975,Water storage starting Construction starting (a) 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1988 1990 1992 1994 1996 1998 2000 2002 Year Annual SSC at Sanleiba (kg/m 3) 1989, Construction starting 1996, Electric generation starting 1998, Projetc completion (b) 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1970 1975 1980 1985 1990 1995 Year Annual SSC at Wolong (kg/m 3) 1973, starting
of construction1983, water storage starting
(c) 0 0.5 1 1.5 2 2.5 1988 1990 1992 1994 1996 1998 2000 2002 Year Annual SSC at Xiaodeshi (kg/m 3) 1991, dam construction starting 2000, project completion 1998,electricity generation starting (d) 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 Year Annual SSC at Yichang (kg/m 3) 1998 large flood
Initial water storage
Cutting of the river for dam construction Construction starting
(e)
Figure 7. Temporal variation in annual suspended sediment load or concentration at various stations before and after the construction of the reservoir. (a) Bikou station before and after the construction of the Bikou Reservoir. (b) Sanleiba station before and after the construction of the Baozhusi Reservoir. (c) Wulong station before and after the construction of the Wujingdu Reservoir. (d) Xiaodeshi station before and after the construction of the Ertan Reservoir. (e) Yichang station before and after the construction of the Three Gorges Reservoir.
The example of the Three Gorge Project is similar. The total capacity of the reservoir is 393 × 108 m3, controlling a
drain-age area of 100 million km2
, accounting for 56% of the total of the Changjiang River. The project was started in 1993. The fi rst phase of construction was 1993–1997, and the river damming occurred in 1997. The second phase was from 1997 to 2003. In 2003 the dam was closed and initial water storage and electricity generation were started. The third phase is 2003–2009, and the project will be fi nally completed in 2009. Figure 7(e) shows the temporal variation in annual SSC at Yichang station downstream from the dam during the con-struction of the Three Gorges Reservoir. When the construc-tion of the project was started, large quantities of wasted rocks and soils were dumped to the river, which resulted in an increased in annual SSC. SSC further increased to a maximum in 1998, one year after the damming of the river. The abnor-mal large fl ood in 1998 made the maximum SSC one year later. Afterwards, SSC decreased. When the initial water storage and electricity generation were started in 2003, SSC was further decreased to a very low level.
Wolman and Schick (1967) studied effects of construction on fl uvial sediment in urban and suburban areas of Maryland and found a cyclic variation, namely, fl uvial sediment increased during the construction, but decreased after the construction was completed. When the construction is under-way, the land surface was disturbed and therefore stronger erosion occurs. In the meantime, large quantities of soil and rocks are disposed to the river directly. Then, sediment load of the river increases. When the urban construction is com-pleted, a considerable proportion of land surface is covered by impermeable materials, namely, cement etc., which are erosion-resistant, and no more construction wastes are dis-posed to the river. Hence, sediment load of the river may signifi cantly decrease. The situation of construction of large sized reservoirs is similar to the situation of urban construc-tion. The disturbed land surface, especially steep hillslopes, becomes easier to be eroded. In the meantime, huge quantities of soil and rocks are delivered to the river. Thus, sediment load or SSC increases. With the completion of the dam con-struction, less soil and fewer rocks are disposed to the river. The disturbed land surface and hillslopes are protected by engineering measurements or gradually established vegeta-tion, and as a result, erosion is greatly reduced. Therefore, the complex response of fl uvial sediment to reservoir construction occurs in a similar way to urban construction.
If several reservoirs were built on a river, the situation may be more complicated. In general, different reservoirs were built in different times, and thus, the sediment trapping induced by different reservoirs has different phases. Hence, the superposed effects of these reservoirs may be more com-plicated than the effect caused by a single reservoir. This was observed from the Minjiang River.
In 1967, the Gongzui Reservoir was completed and used for water storage and electricity generation. With a total capacity of 3·57 × 108
m3
, the reservoir is located in the middle reach of Daduhe River, the largest tributary of the Minjiang River, controlling drainage area of 76,400 km2
, which is 54·4% of the total drainage area of the Minjiang River. In 1967, the river was cut off for dam building. Water storage and electricity generation were started in 1972, and the project was fi nally completed in 1978. Serious siltation occurred above the dam, and up to 1981, 2·2 × 108 m3 of
sediment was trapped by the reservoir. After the river was cut for dam building, SSC decreased. When water storage was started, SSC further decreased. When the reservoir was silted up, sediment can be released from the dam, and SSC increased. In 1985, the construction of Tongjiezi Dam located 37 km
downstream from the Gongzui Reservoir was started, the total capacity of which is 2·0 × 108 m3. In 1992, the dam was
completed and water storage and electricity generation were started. The sediment trapping by the Tongjiezi reservoir made SSC of the Minjiang River decrease again. After 2000, a series of cascade hydropower stations were started to build on the Minjiang River main stem and its tributaries. Huge quantities of excavated soils and rocks were dumped to the river, causing the SSC of the Minjiang River to start to increase. It can be expected that, with the completion and water storage of these cascade dams, the SSC at Gaochang station would decrease again. The temporal variation of SSC of the Minjiang River shows a picture of complex response of river sediment regime to a series of reservoirs that are successively built on a river.
Figure 8 shows the temporal variation in annual SSC at Gaochang station on the Minjiang River, and it is complicated. From 1956 to the early 1960s, because of deforestation in the river basin, SSC increased. Afterward, a series of reservoirs were built successively. The impact of these reservoirs on SSC can be generalized as fi ve stages: (1) a reservoir was con-structed and the downstream SSC decreased; (2) this reservoir was silted up, sediment was released and SSC increased; (3) another reservoir was constructed and trapped sediment, then SSC decreased; (4) construction of a series of dams were started, large quantities of rocks and soil were dumped to the river, SSC increased; (5) these dams are completed and trap sediment, then SSC decreases.
Distal effects of sediment reduction
by reservoir construction
Once a reservoir is built, due to sediment trapping, sediment load below the dam will decrease signifi cantly. However, two factors may increase the downstream sediment load. Firstly, downstream from the dam, with the increase in drainage area, sediment supply from the below-dam drainage area also increases. Secondly, since the water fl ow released from the dam has lower SSC than before, the river bed will be scoured. The sediment reduction by sediment trapping of the reservoir may be offset by the sediment increase due to the above two factors, and thus, the effect of sediment reduction may become smaller and smaller in the downstream direction. This sedi-ment reduction effect further downstream from a reservoir can be seen from the example of the Bikou Reservoir.
As pointed out earlier, the Bikou Reservoir was built on the Bailongjiang River, a tributary of the Jialingjiang River. The reservoir was fi nally completed and used for water storage and electricity generation in 1975. The hydrometric station imme-diately below the dam is Bikou station, near the confl uence
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1950 1960 1970 1980 1990 2000 2010 Year Annual SSC at Gaochang (kg/m 3) Sediment trapping by Gongzui reservoir Sediment releasing when Gongzui Reservoir silted up Sediment trapping by Tongjiezi reservoir Sediment increase during cascade dams construction Forest destruction Tongjiezi reservoir started
Cascade dams started
Figure 8. Temporal variation in annual suspended sediment concen-tration at Gaochang station on the Minjiang River. The regression line is fi tted by a 5° polynomial.
of the Bailongjiang to the Jialingjiang. Downstream from the confl uence, there are four hydrometric stations on the Jialingjiang River, namely, Sanleiba, Tingzikou, Wusheng and Beibei stations. From 1975 to 1989, no large dams were built on the Jialingjiang River basin. The data of sediment load on these stations can be used to assess the downstream variation of the sediment reduction effect induced by the Bikou Reservoir. The period from 1954 to 1975 is considered as pre-dam period, and the period from 1976 to 1989 as post-dam period. The annual suspended sediment load for the pre- and post-dam periods are Qs,pre and Qs,post, respectively.
Thus, the rate (RSR) of sediment reduction induced by the
Bikou Reservoir is defi ned as follows:
RSR=
(
Qs,pre−Qs,post)
Qs,pre (1)The ratio (Rda)of drainage area at Bikou station to that at
Sanleiba, Tingzikou, Wusheng and Beibei stations is also cal-culated, which is used to express the downstream variation of the control of the Bikou Reservoir on sediment load at these stations. Table II shows that with the downstream decrease in Rda, RSR also decreases. Figure 9 shows the relationship
between RSR and Rda, based on data from the fi ve stations. The
regression equation is as follows:
R e R R n
SR= ⋅0 0472 2 4572⋅ da
(
2= ⋅0 987, =5)
This formula may be used to assess the downstream decrease in sediment reduction by the construction of the Bikou Reservoir. Since the trend in Figure 9 shows how the sediment reduction effect becomes smaller with the increase of uncon-trolled drainage area by the reservoir, it may be regarded as a scale effect related to reservoir impact.
Conclusions
Reservoir construction greatly alters river channel boundary conditions, and the river system becomes a man-controlled
Table II. Downstream variation of sediment reduction effect by the Bikou Reservoir
Stations
Drainage area (km2)
Annual sediment load (1954–1975) (104 t)
Annual sediment load (1976–1989) (104 t)
Ratio of drainage area controlled by the reservoir to total area
Rate of sediment reduction by the Bikou Reservoir
Bikou 26086 2081 906 1 0·565 Sanleiba 29247 2000 1145 0·892 0·428 Tingzikou 62182 6505 5781 0·420 0·111 Wusheng 80381 7599 6721 0·325 0·116 Beibei 157900 13975 12944 0·165 0·074 y = 0.0472e2.4572x R2 = 0.9872 0.01 0.1 1 0 0.2 0.4 0.6 0.8 1 1.2
Ratio of reservoir controlled drainage area to the total drainage area
Ratio of sediment reduction by
reservoir Beibei Wusheng
Tingzikou
Sanleiba Bikou
Figure 9. Relationship between ratio of sediment reduction and the ratio of controlled drainage area to the total drainage area of various downstream stations.
system. Based on hydrometric data and information of reser-voir construction in the upper Changjiang River basin, some thresholds and complex response caused by reservoir con-struction have been found.
When the coeffi cient of actual runoff regulation by reser-voirs is smaller than 10%, the sediment load at Yichang station on the Changjiang River shows a mild decreasing trend, but when this coeffi cient becomes larger than 10%, sediment load decreases sharply. The coeffi cient of 10% can be regarded as a threshold. The coeffi cient of 10% is also a threshold when sediment concentration is considered.
During the whole course of reservoir construction including the stages of road constriction, dam building and closure, water storage and sediment trapping, complex response may occur in SSC at stations below the dam. SSC increases fi rstly, reaches a maximum, and then declines sharply. This has been documented by analysis of data from a number of reservoirs on the main-stem of upper Changjiang River and its tributaries.
The successive construction of a series of dams on the Minjiang River resulted in more complicated response: (1) a reservoir was constructed and the downstream SSC decreased; (2) this reservoir was silted up, sediment was released and SSC increased; (3) another reservoir was constructed and trapped sediment, then SSC decreased; (4) construction of a series of dams were started, large quantities of rocks and soil were delivered to the river, SSC increased; (5) these dams are com-pleted and trap sediment, then SSC decreases.
When a reservoir is built and used for water storage, the effect of sediment reduction may attenuate in the downstream direction due to the increase of the uncontrolled drainage area. This scale effect can be described by an exponential equation.
Acknowledgements—The fi nancial supports from Chinese Ministry of Water Resources (2007SHZ090134) and the National Natural Science Foundation of China (40788001) is gratefully acknowledged. All hydrometrical data used in this study come from the Changjiang River Water Conservancy Commission. The author is grateful to two anony-mous reviewers, whose comments and suggestion were valuable for the improvement of the manuscript, to Professor S. N. Lane, who polished up the English in the manuscript.
References
Ackermann WC, White GF, Worthington EB (eds). 1973. Man-made Lakes: Their Problems and Environmental Effects, American Geo-physical Union Monograph 17. American GeoGeo-physical Union: Washington, DC.
Becker A, Blöschl G, Hall A. 1999. Land surface heterogeneity and scaling in hydrology. Journal of Hydrology217(3–4): 1–174. Benn PC, Erskine WD. 1994. Complex channel response to fl ow
regu-lation: Cudgegong River below Windamere Dam, Australia. Applied Geography14(2): 153–168.
Blöschl G. 1999. Scaling issues in snow hydrology. Hydrological Processes13(14–15): 2149–2175.
Brandt SA. 2000. Classifi cation of geomorphological effects down-stream of dams. Catena40(4): 375–401.
Brune GM. 1953. Trap effi ciency of reservoirs. Transactions American Geophysical Union34(3): 407–418.
Chen J, Huang W. 2005. Preliminary study on the change of water fl ow and sediment process in the Changjiang River. Resources and Environment in the Yangtze Basin. 14(6): 786–792. (in Chinese). Chien N. 1985. Changes in river regime after the construction of
upstream reservoirs. Earth Surface Processes and Landforms 10: 143–159.
Chien N, Zhang R, Zhou ZD. 1987. River Channel Processes. Science Press: Beijing (in Chinese).
Church M. 1995. Geomorphic response to river fl ow regulation: case studies and time scales. Regulated Rivers11: 3–22.
Dai HC, Wang LL, Jiang DG. 2007. Recent variation of water fl ow and silt concentration of Yangtze River before and after impounding of Three Gorges Reservoir. Journal of Hydraulic Engineering Sup-plement: 226–231 (in Chinese).
Fergus T. 1997. Geomorphological response of a river regulated for hydropower: River Fortun, Norway. Regulated Rivers 13: 449–462.
Francis JM, Nislow KH. 2005. Changes in hydrologic regime by dams. Geomorphology71(1–2): 61–78.
Fu RS, Qi ML, Fang HW, Chen XQ. 2006.Incoming runoff and sedi-ment runoff to Yichang hydrometric station affected by the projects on the upstream of Yangtze River. Journal of Hydroelectric Engi-neering, 25(2): 103–111 (in Chinese)
Graf WL. 2005. Downstrea m hydrologic and geomorphic effects of large dams on American rivers. Geomorphology79(3–4): 336–360. Han QW. 2003. Sedimentation in Reservoirs. Science Press: Beijing
(in Chinese).
Hydrological Bureau of Changjiang River Water Conservancy Com-mission. 2008. A study of sediment trapping by reservoirs in the upper Changjiang river basin during 1990s. In A Study of Sediment Issues of the Three Gorges Project, Experts Group on Sediment Issues of the Three Gorges Project (eds). Press of Intellectual Prop-erty Right: Beijing; vol. 1, 250–300 (in Chinese).
Lu XX, Higgitt DL. 1998. Recent changes of sediment yield in the Upper Yangtze, China. Environmental Management22: 697–709. Lu XX, Higgitt DL. 1999. Sediment yields variability in the Upper Yangtze.
Earth Surface Processes and Landforms24(11): 1077–1093.
Lu XX, Higgitt DL. 2001. Sediment delivery to the Three Gorges (2): local response. Geomorphology 41: 157–169.
Ministry of Water Conservancy and Electric Power. 1962. National Standards for Hydrological Survey. China Industry Press: Beijing; vo1s 1–7 (in Chinese).
Ministry of Water Conservancy and Electric Power. 1975. Handbook for Hydrological Survey. Water Conservancy and Electric Power Press: Beijing; vols 1–3 (in Chinese).
Petts GE. 1979. Complex response of river channel morphology sub-sequent to reservoir construction. Progress in Physical Geography 3: 329–362.
Petts GE. 1984. Impounded Rivers. John Wiley & Sons: Chichester. Phillips JD, Slattery MC, Musselman Z. 2005. Channel adjustments
of the lower Trinity River, Texas, downstream of Livingston Dam. Earth Surface Processes and Landforms 30(11): 1419– 1439.
Schumm SA. 1973. Geomorphic thresholds and complex response of drainage systems. In Fluvial Geomorphology, Morisawa M (ed.). New York State University Publications in Geomorphology: Bing-hamton, NY; 299–309.
Sedimentation Commission of Chinese Society of Hydraulic Engineer-ing. 1992. A Handbook for Sediment Engineering. China Publishing House for Environment Science: Beijing (in Chinese).
Sherrard JJ, Erskine WD. 1991. Complex response of a sand-bed stream to upstream impoundment. Regulated Rivers 6: 53– 70.
Williams GP, Wolman MG. 1984. Downstream Effects of Dams on Alluvial Rivers, US Geological Survey Professional Paper No. 1286. US Geological Survey: Reston, VA.
Wolman MG, Schick AP. 1967. Effects of construction on fl uvial sedi-ment: urban and suburban areas of Maryland. Water Resources Research3: 451–462.
Xu JX. 1990a. An experiment study of complex response in river channel adjustment downstream from a reservoir. Earth Surface Processes and Landforms15: 43–53.
Xu JX. 1990b. Complex response in adjustment of the Weihe River channel to the construction of the Sanmenxia Reservoir. Zeitschrift fur Geomorphologie34(2): 233–245.
Xu JX. 1991. Complex response in channel adjustment downstream from the Danjiangkou Reservoir. Chinese Science Bulletin36(10): 838–842.
Xu JX. 2007. Trends in grain size of suspended sediment in upper Yangtze River and its tributaries, as infl uenced by human activities. Hydrological Sciences Journal52(4): 777–791.
Xu QX, Shi GY, Chen ZF. 2004. Analysis of recent changing charac-teristics and tendency runoff and sediment transport in the upper reach of Yangtze River. Advances in Water Science15(4): 420–426 (in Chinese).
Yan YS. 1984. An Introduction to Hydrometry. Water Conservancy and Electric Power Press: Beijing (in Chinese).
Yang SL, Zhao QY, Belkin IM. 2002. Temporal variation in the sedi-ment load of the Yangtze River and the infl uences of human activi-ties. Journal of Hydrology263(1–4): 56–71.
Yang SL, Zhang J, Zhu J, Smith JP, Dai SB, Gao A. 2005. Impact of dams on Yangtze River sediment supply to the sea and delta wetland response. Journal of Geophysical Research110: F03006. DOI. 10.1029/2004JF000271
Zhang L, Chen J. 2007. Analysis on change of water fl ow and sedi-ment in upstream of Yangtze River. Journal of Yangtze River Scien-tifi c Research Institute24(6): 34–37 (in Chinese).
Zhang XB, Wen AB. 2002. Variations of sediment in upper stream of Yangtze River and its tributary. Journal of Hydraulic Engineering4: 56–59 (in Chinese).
